14 | Findings on Language, Religion and Caste and Tribal status in the 1891 Census Report

Baines opened the religion chapter with a deceptively simple question: what constituted a distinct religion in India, and what was merely a sect or local variation? For census purposes, he distinguished Brahmanism (his preferred term for Hinduism) from its offshoots, such as Sikhism and Jainism, and from foreign religions like Islam, Christianity, Zoroastrianism, and Judaism. Tribal belief systems were grouped under the category of Animism, which he defined, drawing on C. P. Tiele, as belief in spirits and souls, often expressed in India through tribal religions centred on magic, spirit worship, and exorcism.

Yet Baines recognised that these categories were far from clear-cut. In many forest and hill regions, the boundary between Animism and Brahmanism was highly arbitrary, with census enumerators often classifying marginal groups outside the Brahmanical fold. 

He preferred the term Brahmanism because it highlighted what he saw as its defining features: hereditary social status and priestly authority. Although over 72 percent of Indians were classified under this category,  Baines cautioned that it (Brahmanism) encompassed an immense diversity of beliefs and communities. Its enduring strength, he argued (drawing on the French scholar Barth), lay in its remarkable capacity for absorption and adaptation, enabling it to survive challenges such as Buddhism and incorporate local tribal deities and practices into its fold.

Table: Religious Distribution of India (1891) with Variation since 1881

Baines provided a concise history of Buddhism in India, tracing its rise as a reaction against Brahmanic sacerdotalism, its spread under royal patronage, and its eventual decline. The key to Buddhism’s early success, he argued, was its rejection of the Brahmanic monopoly over salvation. Its remnants in the 1891 census, 7.1 million persons (mostly in Burma) were almost entirely of Thibetan or Burmese extraction.

Muslims constituted nearly 20 percent of India’s population, 57.3 million persons, or a large majority of the world’s Muslims at the time. Baines distinguished between foreign Muslims (descendants of invaders and immigrants) and Indian converts. The former are concentrated in Sindh (77 percent), Kashmir (70 percent), and the North-West Frontier; the latter are found across the subcontinent, particularly in eastern Bengal.

The conversion of India to Islam, Baines noted, was not primarily a matter of doctrinal persuasion but of social and political pressure. What emerged was a distinctly Indian Islam, in which “the simplicity of the Kurán made but a feeble appeal to the mind of those accustomed to the sensuous rites of the new Brahmanism.”

In eastern Bengal, however, Islam has had a more transformative effect. The provincial census superintendent attributed the higher growth rate of the Muslim population to three factors: rise in social status, greater variety of diet, and, crucially, the absence of prohibition on widow marriage and the later age of marriage for girls.

Baines noted Christianity, like Islam, appealed to “a class of the population whose position is hereditarily and permanently degraded by their own religion.” On the Malabar coast, early Portuguese missionaries found it necessary to maintain caste distinctions within the church – separating upper from lower castes, to attract converts.

The Jewish population (17,194) was also concentrated in Bombay, Aden, Cochin, and Calcutta. The Census 1891 report noted the ‘White Jews’ of Cochin claimed to have kept aloof from intermarriage; while the ‘Black Jews’ did not. Both communities, Baines noted, “speak a northern dialect of Arabic” in Aden and Marathi or Malayalam in India.

A detailed table below shows the proportion of each religion to 10,000 of the population in each province and state.

A few immediate patterns stand out from this chart:

• Burma (Upper and Lower) is overwhelmingly Buddhist (>85%).

• Sindh, Kashmir, Aden, and Punjab show Muslim majorities.

• Punjab States has a substantial Sikh share, unlike most other regions.

• Central Provinces States, Assam, and Central India show notable Animistic populations, reflecting the census category used for many tribal communities.

• Madras States stands out for its unusually large Christian population.

• Most princely states of western and southern India (Mysore, Baroda, Hyderabad, Rajputana) remain predominantly classified as Brahmanic/Hindu.

Defining Caste

The third section of Chapter V, which is dedicated to caste, is the longest and most complex. Baines tackles the question of caste and its relationship to race, occupation, and religion.

He began by rejecting any simple equation of caste with Hinduism, and notes “Caste is a development of the special tendency to which the social atmosphere of India is abnormally favourable, and is not, therefore, the peculiar attribute of the Brahmanic form of religion”.

Baines suggested a distinctly racial origin of caste, noting that “the distinction between the Aryan and the non-Aryan was one of colour.” The Brahmans, as a hierarchy, “could easily prevail on the Arya laity to exclude the subject races from participation in what seemed to be the peculiar privilege of the superior.” He further contended that this

The census classification of castes, tribes, and races was based primarily on function, with a secondary grouping by race for communities of foreign origin. The scheme comprised 60 groups, later collapsed into 21 classes for the purposes of the Census 1891 report. The following table shows the proportion of each major class to the total population.

Major Caste Groups

Baines provided detailed commentary on the principal castes within each group. Military and Dominant Castes (1,025 per 10,000) included Rajputs (10.4 million), Jats (6.7 million), Gujars (2.2 million), Marathas (3.3 million), and Nairs (980,000). Other Cultivators (1,670 per 10,000) formed the largest single group. The Kunbi or Kurmi (10.5 million) was noted as “the ordinary term for a tiller of the soil” in the Deccan and Gangetic valley. 

Forest Tribes (551 per 10,000) were the most heterogeneous group. Baines subdivided them into nine geographical groups, from the Kolarian tribes of the north-east (Santhal, Munda, Ho) to the Bhil and Gond of central India to the Mongoloid tribes of Assam (Khasi, Kachari, Garo) and Burma (Kuki, Lushai, Khyin). 

Leather Workers and Lower Village Menials (1,073 per 10,000) included the Chamar (leather workers, 11.3 million) and a range of watchman and scavenger castes (Dosadh, Pasi, Mahar, Dhed, Paraiya, Dom). Baines observed that the leather workers were traditionally “members of the village community, and were always in their present degraded rank.” But the export trade in hides and skins, and the establishment of tanneries using European capital, raised the position of this class above the ordinary village menial.

One of the intriguing passages in the general report concerns the origin of European gipsies. Baines reviewed the evidence connecting the Indian Dom and Jat castes with the Rom of Europe. The gypsies, he suggested, “are really the descendants of the 12,000 or so of musicians and earthworkers… transported to Persia, probably from Sindh, in the time of Behrám Gaur, about 400 A.D.”

Their language, customs, and caste-exclusiveness all point to an Indus Valley origin. Baines is cautious about the specific connection between Dom and Rom, noting that “the interchange of R and a palatal D” at the beginning of a word was phonetically problematic. But he concluded that the overall case for an Indian origin was strong.

1881 vs 1891

The caste enumeration in 1891 was far more systematic than in 1881. The separation of religion, sect, caste, and sub-caste into four distinct columns made it easier for the enumerators to fill the schedules. The instructions also prohibited vague terms such as “Hindustani, Marwadi, Panjabi” in the caste column, forcing a more precise response.

Nevertheless, Baines acknowledged some inconsistencies still prevailed. Two large provinces, Bengal and the North-West Provinces, submitted their caste returns six months after all others, making reclassification difficult.

We can see the improvement in enumeration is more clearly in the increase of female returns in certain castes and the more accurate recording of tribal affiliations. But Baines concluded with a note of caution that  the caste system is fluid, with communities constantly “leaping over or creeping under” its barriers, that any static classification can only be approximate.

The 1891 census introduced a fundamental change in how occupation was recorded. In 1881, enumerators had counted only those who actually worked, excluding dependents, women engaged in domestic labour, and children. As a result, Baine’s noted that “no less than 53 percent of the population got excluded.” Women, in particular, were poorly enumerated making their figures difficult for further examinations (Baines, 1893).

Although the new methods employed improved the enumerations, Baines acknowledged the difficulties with the process involved. “In the most favourable circumstances,” he wrote, “the completeness and accuracy of the enumeration is a matter of very considerable uncertainty” (Baines, 1893). People often returned caste names instead of occupations. They had multiple livelihoods and could not specify which was the principal one. Vague terms like ‘service’, ‘shopkeeping’, and ‘labour’ abounded in the census returns.

Despite these imperfections, the results were unambiguous. Baines isolates just twenty ‘primitive’ occupations that together account for nearly 85 percent of India’s population.  Agriculture dominated without a doubt. Landholders and tenants alone constitute 52.98 percent of the total population. Agricultural labourers add another 6.50 percent. When general labourers (8.87 percent), pastoralists (1.45 percent), and village servants (1.07 percent) are included, the proportion of the population whose lives are tied directly to the land approaches three-quarters.

(Miller with 2 oxen, watercolour painting, c.1850s. Source: New York Public Library Digital Collections)

The remaining ‘primitive’ occupations included cotton workers (3.07 percent), goldsmiths (0.62 percent), blacksmiths (0.55 percent), brass and coppersmiths (0.14 percent), carpenters (1.06 percent), masons (0.36 percent), barbers (0.90 percent), washermen (0.72 percent), fishermen (0.95 percent), oil pressers (0.69 percent), potters (0.82 percent), leather workers (1.11 percent), scavengers (0.40 percent), priests (0.60 percent), and mendicants (1.95 percent). Baines concludes, “the 20 village occupations… comprise no less than 88 percent of the rural population, and nearly 85 percent of the whole of the inhabitants of India.”

Based on Baine's explanations in the general report, we can observe that the village was not merely a residential cluster but a self-sufficient economic microcosm. Within this traditional system, the carpenter made the plough; the blacksmith forged the ploughshare; the potter supplied the water jars; the oil-presser provided cooking oil and lamp fuel; the barber performed ritual and surgical functions; the washerman, the tanner, the scavenger, each had a hereditary place. 

One of Baines’s most important observations concerns the combination of agriculture with other occupations. Using supplementary returns from provinces that tabulated partly agricultural status, he shows that: 

Urban Contrasts

Although only 9.48 percent of India’s population lived in towns classified as urban, the occupational structure of cities varied dramatically. Baines provides a comparative table for five major cities: Bombay, Calcutta, Madras, Rangoon, and Haidrabad (Hyderabad). The contrasts are revealing.

Table: Top 10 Categories of Occupation

The census narrative suggests five distinct urban economies:

• Bombay was a textile-commercial city, with the highest concentration of textile workers and traders.

• Calcutta was an administrative-port city, characterized by domestic service, transport, and commerce.

• Madras was a professional and service centre, leading in professions and food supply occupations.

• Rangoon was a port-frontier economy, specializing in food provisioning, transport, timber, and migrant labour.

• Hyderabad was a princely capital, dominated by administration, defence, retainers, domestic service, and dependent classes.

In Haidrabad (Hyderabad), the capital of the Nizam’s dominions, the administrative element constituted 14.12 percent of the population, nearly three times the urban average. Defence personnel added another 9.75 percent.

Bombay, by contrast, showed the marks of industrial capitalism. Textile workers formed 15.82 percent of its population, and commerce supported 11.28 percent. The city, Baines adds, bred only a quarter of its own population, the rest were immigrants, mostly male, drawn by the mills and docks.

The literacy inquiry in 1891 asked each person whether they were ‘Learning’ (under instruction), ‘Literate’ (able to read and write but no longer a pupil), or ‘Illiterate’. A supplementary column recorded knowledge of English among the literate. The instructions explicitly excluded those who could only sign their names or who attended rote schools where only oral recitation was taught.

The geographical coverage for literacy was incomplete. The wild tracts of the eastern frontier, Rajputana, Central India, and Kashmir were omitted from the literacy return, leaving a population of just under 262 million for analysis. Baines notes that in 1881, similar omissions had been made, but the 1891 census added Upper Burma, Quettah, Aden, and the Andamans to the enumerated area.

Baines breaks these figures down by age. Among males under 15, 6 percent were literate or under instruction; between 15 and 25, the proportion rose to 15 percent; after 25, it fell back to 13.8 percent. 

TABLE B: Proportional Abstract of Literacy

The bulge in the 15–25 age group reflects the cumulative impact of education, but also the confusion between ‘pupils’ and ‘literate’ noted by several provincial superintendents. The Mysore superintendent, himself a Brahman, observed that students “returned themselves in a vast number of cases as learned, though they were still learning, probably in a spirit of bravado.”

Baines writes, “Thus the stimulus to female education was imparted little earlier than the beginning of the present generation.”

Baines attributed such low level of literacy to three main factors: 

• First, the occupational bias of the population. He noted that Indian agriculturalists are likely not significantly far behind their counterparts in European nations, however, the difference lies in the substantially larger proportion of their population. “The raiyat (cultivator), sends his son to school”, Baines observed, “but the necessities of the farm generally lead to the lad’s removal at a rudimentary stage.”

• Second, he observed that the long existence of a hereditary literate class such as Brahman and Kayasth, whose “object it has been to maintain their own monopoly of all book-learning as the chief buttress of their social supremacy. Sacerdotalism knows that it can reign over none but an ignorant populace”

• Third, and most fundamentally, is the position of women. “It is by the position of women in it that we judge to a great extent of the relative progress made by a community towards the complete life that we call civilization,” He also blames the irony of social reformers who preach progressive values in public but lead orthodox lives in private.

One of the most contentious parts concerns the comparison between the census return of pupils and the annual returns of the State Department of Public Instruction. The departmental return exceeded the census by 25 percent for males and by 87 percent for females. Baines attributed this discrepancy to three causes.

• First, the departmental returns include “rote schools” (Kuran and Puran schools) where students learn by oral repetition only and cannot read or write. The census explicitly excluded these. In Panjab alone, 57,397 such pupils were returned departmentally.

• Second, fraudulent practices in the departmental returns were well documented. In Bengal, fabrication of returns in two municipalities were discovered. One schoolmaster in Chittagong transferred his entire school across a district boundary to claim a second return.

• Third, the census’s stricter definition of literacy (ability to both read and write) excluded many who appeared in departmental figures as ‘under instruction’. Baines concluded that “it is not necessary to assume that the difference between the returns and those of the State department betokens any defect in the former.”

Literacy in colonial India in 1891 was highest amongst Parsi, Jewish and Jain communities. Other groups such as Hindus, Christians, Muslims, Buddhists and others exhibit varying levels of literacy, with significant disparities in female literacy. However, these religious aggregates obscure the real concentration of literacy.

These groups are: priests (Brahmans), temple servants, writers (Kayasth and Karan), herbalists and astrologers, traders, Burmese, Parsis, Europeans, Eurasians, Native Christians, and Goanese Christians.

TABLE: Literacy by Caste (proportion)

English and the Universities

Baines notes that with an average annual attendance of 337,000 studying English over the preceding five years, one might expect 700,000 to 800,000 English-knowing literates. The discrepancy, he suggests, reflects the rudimentary stage at which English study ends for most.

Of the 15,200 who presented themselves annually for matriculation, only about 34 percent passed. Of those who passed, only a fraction proceeded to a degree. The university examinations, Baines observed, produce relatively infinitesimal numbers.

Baines concludes the literacy chapter with a brief survey of Indian literature and journalism. In 1890–91, 5,595 books were published across all Indian languages. Most were textbooks or ephemeral works; “a few works on Sanskrit texts, with an occasional drama on a historical occurrence or a subject of the day are all that are likely to survive the year of their birth.”