Endangered Languages of Nepal: UNESCO Danger Levels and Speakers

Nepal is one of the most linguistically diverse countries on earth relative to its size. The 2021 National Population and Housing Census (BS 2078) recorded 124 distinct mother tongues, drawn from four language families, Indo-Aryan and Tibeto-Burman (Sino-Tibetan) alongside smaller Austroasiatic (Munda) and Dravidian communities, plus one language isolate, Kusunda, that belongs to no known family. Nepali, written in Devanagari, is the mother tongue of about 44.9 percent of the population, roughly 13.09 million people, and serves as the official language of the federation.

The UNESCO World Atlas of Languages, formerly the Atlas of the World's Languages in Danger (third edition, 2010), classifies a large share of Nepal's smaller languages as endangered. A widely used compilation of the Atlas counts 73 Nepali languages across four active danger levels: roughly 18 vulnerable, about 38 definitely endangered, around a dozen severely endangered and a handful critically endangered, plus languages that linguists now treat as effectively extinct. As Himalayan linguist George van Driem summarised, the region has more than forty vulnerable or definitely endangered languages and more than a dozen severely or critically endangered ones.

Endangerment is heavily concentrated among the Kiranti (Rai) languages of the eastern hills, the small Tibeto-Burman tongues of the central hills such as Dura and Baram, and tiny forest or foraging communities like the Kusunda and Bankariya. By contrast, the languages considered safe are Nepali and the large Indo-Aryan languages of the Tarai, Maithili, Bhojpuri and Awadhi, together with widely spoken groups such as Tharu and Tamang. In short, Nepal's linguistic diversity is real but fragile: a few big languages are thriving while scores of small ones are fading within a generation or two.

UNESCO ranks languages on a five-step scale based mainly on intergenerational transmission, that is, whether children are still learning the language at home. Understanding these labels is essential to reading the directory below, because the danger badge measures the health of transmission, not simply how many people once claimed the language.

The single most important caveat for Nepal is that census mother-tongue figures usually record ethnic identity rather than fluency. Many respondents name their ancestral language out of pride even when they can no longer speak it, so census totals routinely overstate the number of living speakers, sometimes dramatically. Dumi, for example, was reported by 8,638 people in 2021 but has only a few hundred genuinely fluent elderly speakers; Dura was reported by nearly 2,000 people despite having no remaining native speakers at all; and Kusunda's 23 census respondents include semi-speakers, with just one person able to use the language fluently.

The reverse also happens. Some languages show census increases, Kagate rose from 99 speakers in 2011 to 611 in 2021 and Tilung from 1,424 to 1,969, that reflect ethnic revival movements and better enumeration rather than a real rise in everyday use. Read the directory with this gap in mind: the UNESCO danger level reflects whether the language is still being passed to children, while the census columns reflect self-identification.

The list below is ordered roughly from most critical to less critical, showing each language's UNESCO danger level and its census mother-tongue trend from 2011 to 2021. Where UNESCO and the census disagree, both are given, because the disagreement is itself the story: a shrinking pool of real speakers hidden behind a larger number of people who claim the language as heritage.

Almost every entry shares one pattern, the count of fluent daily speakers is far smaller than either the census figure or the ethnic population. A community of several thousand people may carry a language kept alive by only a few dozen elders. That is why documentation and school programmes, discussed later, are racing against the clock.

Kusunda is the emblem of language endangerment in Nepal. It is a language isolate, meaning it has no proven relationship to any other language in the world, which makes each remaining fragment linguistically priceless. The Kusunda were historically a small forest-dwelling, hunting community of the central-western hills; today the wider Kusunda ethnic group numbers only around 250 people according to the 2021 census.

For years two sisters were the last fluent speakers. Gyani Maiya Sen-Kusunda, widely reported as the last fully fluent speaker and a tireless campaigner for the language, died on 25 January 2020 at the age of 85. Her younger sister, Kamala Sen-Khatri, is now generally regarded as the sole remaining fluent speaker. The 2021 census recorded 23 people naming Kusunda as their mother tongue, but linguists note that most are semi-speakers with limited proficiency, mainly elderly.

There is, however, an active rescue effort. Under Nepal's Language Commission, Kamala Khatri has taught Kusunda classes to young community members in Dang district, and scholars including David Watters and Uday Raj Aaley have produced a grammar and dictionary of the language. Whether these efforts can rebuild an actual community of speakers, rather than only an archive, remains the central question for Kusunda's survival.

A distinct category within Nepal's endangered languages is the group that is effectively dormant, languages with an identifiable ethnic community but essentially no fluent native speakers left. Dura, of Lamjung district in central Nepal, is the clearest example. Media reports in the 2000s named an elderly woman as the last speaker, but subsequent linguistic study concluded she was speaking a Nepali-based argot rather than Dura itself. Scholars now treat Dura as extinct as a spoken mother tongue, even though nearly 2,000 people reported it in the 2021 census on the basis of ethnic identity.

Baram (also called Bramu), of Gorkha district, is close behind. The ethnic Baram community numbers several thousand, yet only a few dozen elderly speakers remain, and the language is classified critically endangered. Tilung and Sam, both eastern Kiranti languages, and Surel, a small Dolakha community, sit in the same fragile band, spoken chiefly by older generations with little transmission to children.

For these languages, revival is a project of reconstruction rather than maintenance. Where recordings, wordlists and older grammars survive, communities and linguists can attempt to relearn the language from documentation. In the Dura case, a Nepali-speaking enthusiast has reportedly learned to speak reconstructed Dura as a second language, a fragile but genuine sign that dormant languages need not be permanently lost if the documentation is good enough.

The pressures on Nepal's minority languages are structural and reinforcing. Nepali is the language of government, courts, higher education, media and inter-ethnic communication, so families often switch to it to give children better prospects. Once parents stop speaking the ancestral language at home, the chain of transmission breaks and the language slides from definitely endangered to critically endangered within a generation.

Migration and urbanisation scatter small speech communities, labour migration abroad and movement to towns dissolve the dense village networks in which a minority language is used daily. Historic stigma against indigenous languages, and decades in which schooling was conducted only in Nepali, taught many speakers to see their mother tongue as a barrier rather than an asset. Communities with only a few hundred members, such as the Kusunda, Bankariya, Surel and Baram, are especially exposed because they lack the numbers to sustain everyday use.

Nepal's 2015 Constitution (BS 2072) gave minority languages their strongest legal footing to date. Article 6 declares that all mother tongues spoken in Nepal are languages of the nation, Article 7 keeps Nepali in Devanagari as the official language while allowing provinces to adopt additional official languages, and Article 32 guarantees every community the right to use and preserve its language and culture. To turn these principles into policy, Article 287 established the Language Commission (Bhasha Aayog), constituted in 2016 (BS 2073), which studies the status of languages and recommends measures for their protection, promotion and use in official life.

On the ground, revitalisation takes several forms: community-run language classes, mother-tongue and multilingual education in early grades, documentation projects that produce dictionaries and grammars, and cultural programmes that raise the prestige of speaking indigenous languages. The Kusunda classes in Dang, the Dura reconstruction effort in Lamjung, and academic surveys of Baram, Tilung and the Kiranti languages are examples of this work in practice.

The realistic outlook is mixed. For languages still spoken by children, such as several vulnerable Kiranti tongues, well-designed school programmes can slow or reverse decline. For critically endangered and dormant languages like Kusunda and Dura, the immediate goal is thorough documentation so that the knowledge is preserved, with revival dependent on a small, committed group of learners. Either way, the window is narrow, and most of the crucial knowledge sits with elderly speakers whose numbers fall every year.