Languages of India, 1931

The 1931 census and its language return

The census of 1931 was the last in British India to be fully tabulated. It covered the provinces under British administration and the princely states and agencies, and it included Burma, which was separated from India in 1937. Enumeration took place on the night of 26 February 1931. Table I puts the population at 352,837,778. Table XV, like the other tables built from individual returns, covers 350,529,557 people. The missing 2,308,221 were estimated, not enumerated: 2,212,837 in the agencies and tribal areas of the North-West Frontier; 75,735 in Ahmedabad, where a Congress boycott disrupted the count; and 19,649 in East Manglun in the Shan States. The tables were published in 1933 in Volume I, Part II (Imperial Tables), under J. H. Hutton as Census Commissioner.

Every household schedule had a mother-tongue column. For the first time there was also a column for any other language habitually spoken “in daily or domestic life”. Infants and deaf-mutes were recorded with their mother’s language. Part I of Table XV, mapped here, gives the mother-tongue figures. Part II gives the subsidiary-language figures, but it is a sample, not a count (see Cautions). Hutton’s report counts 223 languages under Grierson’s classification, against 222 in 1921 (para. 150). Some later summaries give 225.

Hutton’s Chapter X explains why the census asked about languages beyond pure linguistics. The first reason was the “not unreasonable desire of many Indians for a redistribution of provinces on a linguistic basis”. The second was that the franchise might be extended to large tribal populations. The third was the wish to start primary education “in the real language of the child” (para. 147). The Orissa Boundary Committee asked for the returns from the Oriya–Telugu borderland before they had even been extracted. Orissa became a separate province in 1936, and those figures were among its foundations. The same questions came back after independence, in the States Reorganisation Commission of 1953–55, and 1931 was the last full set of language figures from before Partition.

The classification (Family › Sub-family › Branch › Group › Language) follows Sir George Grierson’s Linguistic Survey of India, whose introductory volume came out in 1927. On this site you can map any level of it, from “Dravidian Family” down to single entries such as “Kurukh or Oraon”. Where the printed table gives no total for a group, the site adds up its members and says so.

This site was prompted by India State Stories’ essay “Legacies of the 1931 Linguistic Census”, which discusses these tables and their later use.

Cautions: read these before trusting a number

1. “Western Hindi”, “Eastern Hindi” and “Bihari” are a clerk’s division, not what people said. Nobody told an enumerator they spoke “Western Hindi”. The census used the single term Hindustani for both Hindi and Urdu as spoken languages. In the United Provinces Hindustani was the only entry permitted, to keep out the Hindi–Urdu controversy. The India table then reassigned Hindustani to Grierson’s languages by locality. In the United Provinces all 48 million Hindustani speakers went under “Western Hindi”, including the Awadhi-speaking east and the Bhojpuri-speaking districts that Grierson placed under Eastern Hindi and Bihari. That is why the United Provinces shows 100% Western Hindi and almost no Eastern Hindi. Hutton says that even for his linguistic map the Eastern/Western line was “taken arbitrarily as the boundary (more or less) of Agra and Oudh”, and that telling Lahnda from Panjabi was “more than the census enumerators could grasp” (para. 152). The sharp edges these languages show on the map are therefore artefacts of this reclassification. The census recorded no split between Urdu and Hindi at all.
2. Some returns were politically coloured. In the Oriya–Telugu and Oriya–Bengali borderlands there were rival campaigns over the boundaries of a new Orissa province. Hutton says they led people to “plump” for one language and to hide their bilingualism (para. 148). In Bhopal (in the Central India Agency), Urdu was returned “in deliberate disregard of the actual facts” for Hindus as well as Muslims. Elsewhere some Muslims returned Urdu (Hindustani) as their mother tongue and the local vernacular as a second language, “though the precise opposite was the real fact”. In Bastar, some Gonds returned Halbi instead of Gondi because Halbi carried more status (para. 149).
3. Tribal languages are probably undercounted. Speakers who were bilingual in a regional language were often recorded under that language. For instance, the Central Provinces superintendent, quoted by Hutton, found that of 84,971 Sawaras “not a single one returned the tribal language”. The new subsidiary-language column helped, but only in some places.
4. Units are administrative, not linguistic. The map shades whole provinces and agencies. Bengal, Madras or Hyderabad each contained several large language areas. For district detail you need the provincial volumes (see Further reading).
5. Frontier areas were only partly enumerated. Most of the North-West Frontier’s agencies and tribal areas (an estimated 2.2 million people) were not enumerated individually and are not in Table XV. The “NWFP Agencies & Tribal Areas” row covers only 46,451 people, 96% of them men, in effect the garrisons and posts. Its hover box and info pane say so, so that its shading is not read as describing the tribes. The census also left out or estimated parts of the Assam hill tracts and the Burmese frontier. The north-eastern tracts beyond the administered border are grey on the map and have no figures.
6. Burma and Aden were part of “India” in 1931. Both are counted in the India totals. Burma accounts for most of the Tibeto-Chinese and Karen figures. Aden sits inside Bombay (row 7), and its figures are shown in Bombay’s panel.
7. Migrants show up everywhere. A few thousand speakers of a northern language in a southern province usually means soldiers, traders, labourers or officials, not a settled community. The “Females/1,000 males” measure helps spot these groups: migrant groups tend to be mostly male.
8. The boundaries drawn here are approximate. They are good enough to show which unit is which, but they are not survey-accurate. Data explains how they were built.
9. Transcription. Every figure was copied by hand from the printed scan and checked arithmetically (see Data). Where the printed table does not add up, the printed figure has been kept and flagged, not silently corrected. For anything you intend to publish, check the figure against the page image. Every language panel links to its page.

Where the data comes from

The figures

The source is Census of India, 1931. Vol. I – India. Part II – Imperial Tables (Delhi: Manager of Publications, 1933), Table XV, “Language”, Part I, “Mother Tongue”. This site uses the scan on the Internet Archive, india.history.resource.92539, scan leaves 477–514 (printed pages 472–509). The same volume is also held by the Gokhale Institute (GIPE) DSpace and Tamil Digital Library.

The table has 958 numbered columns: Persons, Males and Females for each language and each group total. It has 44 rows: India, the Provinces and States subtotals, 35 provinces, states and agencies, and sub-rows for Aden, Cochin, Travancore and the other Madras States. People read the whole table from page images, without OCR, and then checked it three ways:

• persons = males + females in every cell;
• the provincial and state rows add up to the printed Provinces, States and India totals in every column;
• the members of each group add up to the group total printed in the table.

A separate reader then re-read a random sample of 250 cells without seeing the transcription. 247 agreed with it. Re-checking the scan showed that the other 3 were the second reader’s own slips (details in source/verification).

The results of these checks are listed below.

Every figure can be searched, filtered and checked against its page scan in the Province tables. Downloads: table15.csv (every figure, long format), languages.csv (the classification with all-India totals), and the page-by-page transcription files in source/transcribed/.

District figures

The provincial and state volumes print Table XV by district. The district map draws those figures on present-day district boundaries. Each 1931 area is linked to the districts made from it through the district lineage of Jolad, Kalra and Singh (CC BY 4.0), and the few that cannot be separated are grouped together. How the 24 volumes were made comparable is explained on that page and, rule by rule, in docs/regional.md.

The map

The base boundaries are the 1931 province/state layer from W. Tennant’s plague_india repository (LGPL-3.0). Its latitudes have been rescaled to WGS84. That layer merges several census units, so they were separated here using modern district boundaries from geoBoundaries, which run close to the old lines at this scale:

• Sylhet moved from Bengal to Assam, where it belonged in 1931;
• Bahawalpur goes to the Punjab States Agency;
• Khairpur goes to the Bombay States;
• Kalat, Las Bela and Kharan become the Baluchistan States, and Gwadar, then Muscat’s, is shown as foreign;
• the Malakand states, the agencies and the tribal belt are separated from the NWFP districts;
• Kathiawar, Cutch and Palanpur form the Western India States Agency;
• the Simla Hill States are separated from the Punjab States Agency.

Some 1931 boundaries match modern districts closely, and for these the modern districts replace plague_india’s hand trace. They are Coorg (Kodagu), the NWFP districts and tribal agencies, Tehri Garhwal and Rampur, Manipur, Tripura and Cooch Behar, the Simla Hill States, Bahawalpur, Sind, and Burma (modern Myanmar). The same goes for the lines that became today’s international borders. The Rann of Kutch (from OpenStreetMap, ODbL) is drawn as salt waste, since Table I counts it in no unit’s area. The north-east frontier tracts, Portuguese and French India, and Muscat’s Gwadar are grey.

Every modern state and province then has a list of the 1931 units that could lie inside it. Any other unit found there must be an overshoot of the base map, and goes to the neighbouring unit it belongs with. The list keeps deliberate exceptions for real historical outliers: Madras’s Munagala and Bhadrachalam in Telangana, Tonk’s Sironj in Madhya Pradesh, Khariar in Odisha, Seraikela and Kharsawan in Jharkhand, and Datia and Samthar in Uttar Pradesh. After that, thin strips where two sources’ lines disagree, and leftover fragments, go to the unit they sit against. All shapes are simplified together, so neighbours share one line with no gaps or overlaps. The scripts are tools/build_geo.py, tools/geo_audit.py and tools/area_check.py.

To check the boundaries, each polygon’s area was compared with the area printed for its unit in Table I. The map total comes within 1.4% of the printed all-India figure (1,808,679 sq. mi), and every unit is within about ±10%, except the Andaman and Nicobar Islands, whose coastlines are modern. The full comparison is below.

Coastlines and neighbouring countries come from Natural Earth (public domain). The projection is Mercator, following the approach of the Japanese Empire map this site is modelled on. Scattered enclaves (dozens of small states inside Gujarat, Central India and the Deccan) are merged into their agency. Small French and Portuguese territories other than Goa are not drawn. The build scripts are in tools/.

Further reading and sources

This is a reading guide, not a full bibliography. Each entry says why it matters for these tables. The list is still thin on Burma, on Adivasi perspectives and on the northeast. Suggestions are welcome.

Read the census

• Census of India, 1931. Vol. I – India. Part II – Imperial Tables (1933). The source of every figure on this site: Table XV Part I (mother tongue) and Part II (bilingualism). Internet Archive · GIPE · Census of India digital library.
• J. H. Hutton, Census of India, 1931. Vol. I – India. Part I – Report (1933), ch. X “Language”. The Census Commissioner’s own account of what went wrong: Hindustani, Orissa, Bhopal, the Gonds. Internet Archive.
• The provincial and state volumes (Vols. II–XXVIII), whose Table XV gives districts, tahsils and states, often with district bilingualism. Examples: Assam; Burma (district language and bilingualism tables); Western India States; Jammu and Kashmir.
• G. A. Grierson, Linguistic Survey of India (1903–1928), especially Vol. I Part I (1927). The classification the 1931 tables follow. Searchable at the Digital South Asia Library.
• Later censuses, for comparison. The categories are not directly comparable across years. The 1961 census was the first to publish full mother-tongue and bilingualism tables for the Indian Union (e.g. Part II-C(ii), Uttar Pradesh). The 2011 table C-17, “Population by bilingualism and trilingualism” is the modern successor to Part II. The Census’s Language Census Data portal explains how language returns are rationalised.

How colonial categories were made

• Bernard S. Cohn, “The Census, Social Structure and Objectification in South Asia”, in An Anthropologist among the Historians and Other Essays (Oxford University Press, 1987). The classic argument that enumeration helped make the categories it counted.
• David Lelyveld, “Colonial Knowledge and the Fate of Hindustani”, Comparative Studies in Society and History 35:4 (1993), 665–682. doi:10.1017/S0010417500018661. Essential background to the site’s Hindustani / Western Hindi problem.
• Asha Sarangi, “Enumeration and the Linguistic Identity Formation in Colonial North India”, Studies in History 25:2 (2009), 197–227. SAGE. How census operations standardised linguistic diversity in the north.
• Thomas R. Trautmann, Languages and Nations: The Dravidian Proof in Colonial Madras (University of California Press, 2006). The intellectual history behind the table’s “Dravidian Family”.
• Javed Majeed, Colonialism and Knowledge in Grierson’s Linguistic Survey of India (Routledge, 2018). How the survey behind this classification was made.
• Javed Majeed, Nation and Region in Grierson’s Linguistic Survey of India (Routledge, 2018). The companion volume: how the survey mapped languages onto regions and nations.
• Richard Steadman-Jones, Colonialism and Grammatical Representation: John Gilchrist and the Analysis of the ‘Hindustani’ Language in the Late Eighteenth and Early Nineteenth Centuries (Philological Society, 2007). Where the British category “Hindustani” came from.
• Pritipuspa Mishra, “The Mortality of Hindustani”, Parallax 18 (2012), 71–83; Alok Rai, “The Persistence of Hindustani”, Annual of Urdu Studies 20 (2005), 135–44. Two views of the category that the 1931 provincial tables print but Vol. I breaks up.
• Walter N. Hakala, Negotiating Languages: Urdu, Hindi, and the Definition of Modern South Asia (Columbia University Press, 2016). How dictionaries and lexicographers drew the line between Hindi and Urdu.
• Farina Mir, “Imperial Policy, Provincial Practices: Colonial Language Policy in Nineteenth-Century India”, Indian Economic and Social History Review 43:4 (2006), 395–427. Why provinces made different language choices under one imperial policy.
• Judith T. Irvine, “Language Fields: Robert Needham Cust’s Language Map of South Asia, 1878”, in Cynthia Talbot (ed.), Knowing India: Colonial and Modern Constructions of the Past (Yoda Press, 2011). How language maps like this site’s ancestors were made.
• Leslie F. Taylor, The Ethnographical and Linguistic Survey of Burma: An Account of Its Institution and Progress (Heffer, 1948); Carol A. Boshier, “Leslie Fernandes Taylor and the ‘Lost’ Linguistic and Ethnographical Survey of Burma”, SOAS Journal of Postgraduate Research 6 (2014), 49–76. Burma’s own, unfinished language survey.
• Christopher A. Bayly, Empire and Information: Intelligence Gathering and Social Communication in India, 1780–1870 (Cambridge University Press, 1996). The information order that the census grew out of.
• Colin P. Masica, The Indo-Aryan Languages (Cambridge University Press, 1991). The standard linguistic reference for the groups behind Western and Eastern Hindi, Bihari, Rajasthani and Lahnda.
• Christopher R. King, One Language, Two Scripts: The Hindi Movement in Nineteenth Century North India (Oxford University Press, 1994). The script politics that led to the 1931 return of “Hindustani”.

Regional language movements

• Pritipuspa Mishra, Language and the Making of Modern India: Nationalism and the Vernacular in Colonial Odisha, 1803–1956 (Cambridge University Press, 2020), open access. The making of Orissa, the first linguistic province, including how Adivasi languages were pushed aside.
• Sumathi Ramaswamy, Passions of the Tongue: Language Devotion in Tamil India, 1891–1970 (University of California Press, 1997), free online. Tamil devotion and politics.
• Lisa Mitchell, Language, Emotion, and Politics in South India: The Making of a Mother Tongue (Indiana University Press, 2009). How “mother tongue”, the very census category, became an object of attachment in Telugu.
• Veena Naregal, Language Politics, Elites, and the Public Sphere: Western India under Colonialism (Permanent Black, 2001). Marathi print, bilingual elites and linguistic hierarchy in western India.
• Mithilesh Kumar Jha, Language Politics and Public Sphere in North India: Making of the Maithili Movement (Oxford University Press, 2018), Oxford Academic. How Maithili became “minor” within the Hindi sphere, including through census classification and mapping. Maithili is folded into “Bihari” in these tables.
• Richard Burghart, “A Quarrel in the Language Family: Agency and Representations of Speech in Mithila”, Modern Asian Studies 27:4 (1993), 761–804. Maithili against Hindi, and what “Bihari” in the tables covers.
• Paul R. Brass, Language, Religion and Politics in North India (Cambridge University Press, 1974); Alok Rai, Hindi Nationalism (Orient Longman, 2001). Hindi, Urdu and Punjabi politics around and after these returns.
• Farina Mir, The Social Space of Language: Vernacular Culture in British Colonial Punjab (University of California Press, 2010). Punjabi literary culture under a colonial state that favoured Urdu.
• Christopher Shackle, From Wuch to Southern Lahnda: A Century of Siraiki Studies in English (Bazm-e Saqafat, 1983). The history of the “Lahnda” category and what became Siraiki.
• Chitralekha Zutshi, Languages of Belonging: Islam, Regional Identity, and the Making of Kashmir (Hurst, 2004). Language and identity in Jammu and Kashmir.
• Jayeeta Sharma, Empire’s Garden: Assam and the Making of Modern India (Duke University Press, 2011). Tea, migrant labour and Assamese identity: why Mundari, Oriya and Santali appear in Assam’s tables.
• S. C. Patra, Formation of the Province of Orissa: The Success of the First Linguistic Movement in India (Punthi Pustak, 1979). The Orissa province that the 1931 returns helped create.
• Report of the Orissa Committee (1932). The boundary committee that asked for the 1931 language returns.

Adivasi and minoritized languages

• UNESCO and IGNCA, Tribal and Indigenous Languages of India (2022), UNESCO. A review of what official records and surveys do and do not say about tribal languages, and a route into endangerment and documentation.
• G. N. Devy (chief ed.), People’s Linguistic Survey of India (Orient BlackSwan, 2012–), volumes. A speaker-centred counterpoint to official classification, recording some 780 languages.
• Anvita Abbi, “Vanishing Diversities and Submerging Identities”, in Asha Sarangi (ed.), Language and Politics in India (below). Tribal languages under the pressure of regional and national ones, which is the process Hutton’s “survival of tribal languages” section describes.
• Prathama Banerjee, Politics of Time: “Primitive” and History-Writing in a Colonial Society (Oxford University Press, 2006). How the Santals were written into and out of history.
• Felix Padel, The Sacrifice of Human Being: British Rule and the Khonds of Orissa (Oxford University Press, 2001). Colonial rule and the Kond (Kui) speakers counted in these tables.

After Partition

• Asha Sarangi (ed.), Language and Politics in India (Oxford University Press, 2009), OUP. A bridge from colonial classification to the constitution, linguistic states (Schwartzberg), minorities and education.
• Tariq Rahman, Language and Politics in Pakistan (Oxford University Press, Karachi, 1996). What became of Urdu, Bengali, Sindhi, Pashto, Siraiki, Balochi and Punjabi in the western half of these tables, and of the Bengali movement.
• Sanjib Baruah, India Against Itself: Assam and the Politics of Nationality (University of Pennsylvania Press, 1999), Penn Press. Assam and the northeast, where the 1931 Assamese–Bengali figures later became explosive.
• Sudipta Kaviraj, “Writing, Speaking, Being: Language and the Historical Formation of Identities in India”, in D. Hellmann-Rajanayagam and D. Rothermund (eds.), Nationalstaat und Sprachkonflikte in Süd- und Südostasien (1992). An influential essay on how languages became identities.
• Jyotirindra Das Gupta, Language Conflict and National Development: Group Politics and National Language Policy in India (University of California Press, 1970); Robert D. King, Nehru and the Language Politics of India (Oxford University Press, 1997). The fight over a national language after 1947.
• B. R. Ambedkar, Thoughts on Linguistic States (1955), and the Report of the States Reorganisation Commission (Government of India, 1955). The redrawing of states along language lines that these tables anticipated.

Maps

• The linguistic map that accompanies Vol. I Part I of the 1931 report, and the language maps in the Imperial Gazetteer of India, Atlas (1909). DSAL Imperial Gazetteer maps.
• India State Stories, “Legacies of the 1931 Linguistic Census” and “Districts in Colonial Era”. The essay that prompted this site, and district maps from 1872 to 1941.
• The Japanese Empire map, the SVG map this site’s design follows.