4 min readMar 4, 2026 06:20 AM IST
First published on: Mar 4, 2026 at 06:20 AM IST
How many caste communities does India have and how did they emerge? While there are several anthropological and historical theories pointing to possible answers, we do not have a definite answer. It is the same with the question: How many “Other Backward” communities does India have? Much less attended is the question: How many Denotified and Nomadic communities does India have? The DNTs started getting “notified” as “criminal tribes” with the infamous Criminal Tribes Act (CTA) of 1871. They were “denotified” in 1952, and placed variously in the SC, ST or General lists. India has no exact official count of the DNTs; we know neither the number of DNT communities, nor the exact population. As a historical irony, the CTA was passed in the year in which the first-ever Census was carried out.
It is surprising that we do not know the precise composition of Indian society. We also don’t have an accurate list of languages spoken in India and their scripts. The idea of carrying out a proper survey of languages in the Subcontinent was proposed by George Abraham Grierson in 1886. Eight years later, the colonial government initiated the survey. After an elaborate study for three decades, Grierson reported 179 languages and 544 dialects. The map of India surveyed by him was different from the present one. The Constitution defines India as “a Union of States”, and the states are organised primarily as “linguistic states”. Yet since Independence, there has been no official Linguistic Survey — the country depends on data gathered during the Census.
The 1961 Census reported 1,652 “mother tongues”. The 1971 Census had a list of 108 “languages”, plus another classification called “others”. The 2011 Census had listed 1,369 “mother tongues” and 121 “languages”. No one ever asked in Parliament why 283 “mother tongues” disappeared during the half century between 1961 and 2011.
The Census that should have been carried out in 2021 is now on the horizon. It will be completed in 2027. The questions to be asked do not include any on OBCs and DNTs. With regard to language-related question(s), there is a strange rider attached: If a language community does not have more than 10,000 speakers, its name will not be published in the Census reports. This has no scientific basis. No linguist has ever argued or even remotely suggested that for a language to be a language it must have a minimum number of speakers. Besides, the distinction between “mother tongue” and “language” that the Census uses defies logic. It is unfortunate that in the past these criteria were applied while processing the language-related data in the Census. India has an age-long tradition of language diversity, and no government should fear the magnitude of such diversity. Added to the Census practice in the past, the current government has initiated another absurd project in relation to language diversity.
Since the late 18th century, when William Jones identified what came to be known as the Indo-European family of languages, the languages of the world have been classified into various linguistic “families”. The more important language “families” in India are the Indo-Aryan, Dravidian, Sino-Tibetan and Austro-Asiatic families. The Bharatiya Bhasha Samiti, inspired by the historiography held dear by the RSS, is now busy promoting the idea that all of these are essentially a single “Bharat bhasha parivar” of a common origin, pre-dating the Indus Valley civilisation and with Sanskrit as its primary “janani”. This view has mythical trappings but no scientific validity. Neither the reductive Census, being carried out primarily to facilitate delimitation of Assembly and Parliament seats, nor wishful linguistics can tell us who and how many we are and what the extent of our social and cultural diversity is.
Devy is a writer and cultural activist
