Counting the uncountable: linguistic diversity in Nepal

The 21st of February marked International Mother Language day, an annual UNESCO Heritage Day that celebrates linguistic diversity and multilingualism around the world. This year’s International Mother Language Day has particular importance, as 2019 has been marked as the United Nations International Year of Indigenous Languages.

Most estimates place the total number of languages in the world at around 7,000. Calculating the number of languages in a given country or region is important for both linguists and policy makers (as well as a host of other professions) for many reasons. By knowing the numbers of speakers of a particular language, policy makers can more effectively plan for linguistically-inclusive public communication and administration, consider the role of different languages as the medium or subject of instruction in education systems, and direct efforts for language preservation and revitalisation.

However, while empirical data on other social characteristics is usually readily available or easy to collect, data on language has often proved more difficult to enumerate. This post will explore how the question, the context, and the response all provide room for subjective interpretation and can lead to vastly different figures for the number of languages in a single country: Nepal.

Nepal

The small South Asian nation of Nepal boasts huge linguistic diversity relative to its geographic, economic, and population size. Inhabitants of modern-day Nepal were historically made up of hundreds of distinct groups of people with different cultures, languages, and leaders, who were only united under a single ruler in 1768. Topographical barriers like the great mountains of the Himalayas and the sweeping plains of the lowland Terai region meant that many languages developed in relative isolation.

Nepal has collected language data through regular decennial censuses for the last 60 years. Yet within this relatively short period, various censuses and other independent linguistic surveys have returned different tallies for the number of languages present in Nepal. Estimates have ranged from as few as 17 in the 1971 census, to 123 languages in the most recent census of 2011, to an estimate of over 140 in a 2005 linguistic survey, as shown in the table:

While it’s true that language, and language use, shifts naturally over time due to factors such as generational change, revitalisation efforts, or migration, the radical differences between estimates in Nepal over just sixty years suggests something more complicated is at play.

While many census or survey questions can be answered objectively and impartially, others, such as questions on mother tongue, language use, ethnicity, or religion, require the respondent to make a subjective judgement. In Nepal, as elsewhere, the wording of the question, the wider social and political context within which the question is asked, and the respondents outlook and ideology towards language, have all influenced efforts to calculate the country’s languages.

The Question

The first way in which respondents may be influenced in their reporting of language is the wording of the question itself. Subtle differences in the terminology, phrasing, or layout can change the way a question is interpreted and answered.

The 2011 Census in Nepal included one question on language (Yadava 2014, p. 52):

Q. 10. What are the mother tongue and the second language of …………… (a given respondent)? 1. Mother tongue: …………… 2. Second language: ……………

Looking closely at this question, there are several possible ways respondents could have been influenced in the responses they provided.

Firstly, the terminology itself may have induced certain responses. Nepal’s Central Bureau of Statistics opted to use the term “mother tongue” (मातृभाषा /mātr̥bhāṣā/ in Nepali) in the 2011 Census, over alternatives such as “main language”, “usual language”, “home language”, etc.

For some people, these terms could be considered synonyms as they elicit the same response. For example, an Australian, living in Australia, of British-heritage parents, would answer “English” in all three situations. For others, the response may be different based on the subtly different slant of each term. For example, a child of a Vietnamese immigrant mother in Montreal who first learned Vietnamese (the “mother tongue”), but uses mostly French when interacting with the Canadian father and local friends (the “home language”), and speaks English to colleagues in a multinational workplace (the “usual language”).

In his chapter in the ground-breaking book Census and Identity: The Politics of Race, Ethnicity, and Language in National Censuses’, Arel (2002) refers to this as different “language situations”. Canada is one of the few countries that attempts to collect data on all three possible language situations in its surveys – most countries, including Australia and Nepal, only ask about one or two potential language situations. By focusing on the “mother tongue” of the respondent, ignoring other potential terms such as “usual” or “home” language, the 2011 Census elicited a certain response from respondents.

Though the 2011 Census in Nepal did also request respondents to provide another language as their “second language”, no further definition or guidance was given to clarify what was meant by “second language”. Given that the majority of respondents (56%) who chose to answer this question listed the national language, Nepali, as their second language, it is likely that many took this as an opportunity to assert their knowledge of the language of social mobility.

Another potential challenge is the layout of the census form itself. In 2011, a single space was provided for each part of the question, forgoing the possibility of more than one language being spoken in favour of simpler enumeration, indicating a bias towards mono- or bilingualism in a country where multilingualism is common, if not the norm.

In 2011, the response format was open-ended, with no list of languages to choose from. This led to confusion between duplicate, indistinct, or unknown languages that was ultimately resolved by grouping some 21,000 responses into a single ‘other’ category, and another 47,000 as ‘non-responses’ (Yadava, 2014) – potentially missing smaller, lesser known or documented languages.

Conversely, in 1981, the census question around language provided a list of the five largest, most dominant languages and a catch-all ‘other’ category, reflecting the wider socio-political context at the time, as we will see soon. It is likely that the inherent bias contained in this question influenced the way people responded.

The Context

Quantifying the number of languages in a country or region can also be influenced by the prevailing political, social, and cultural climate. As Sebba (2017) aptly puts it when describing the inclusion of a language question in the 2011 British Census: “inevitably, questions about language are asked within a social and historical context which both constrains the possible answers and motivates respondents to select certain answers rather than others from those available, in accordance with prevailing ideologies about (among others) nation, ethnicity and language. The act of census-taking (…) is always politically and ideologically charged.”

The relatively low number of languages reported in Nepal’s early censuses were undoubtedly influenced by the assimilation policies in place at the time, and the generally higher levels of social exclusion. From 1962 to 1990, under a political system known as “Panchayat” and controlled by an authoritarian monarchy, the state viewed linguistic, gender, ethnic, and spiritual diversity as barriers to be overcome in the pursuit of a ‘unified’, ‘modern’ Nepal. Cultural ‘unity’ was projected as essential to nation-building and the maintenance of independence. The relatively low number of languages reported in the 1962, 1971, 1981, and 1991 Census reflected the widespread restrictions on cultural and linguistic expression that the Nepali population was experiencing during these years.

Through the 1990s and 2000s, the Maoist ‘People’s War’ raged against the monarchy, promising greater representation to minority caste, ethnic, linguistic, and gender groups. The decade-long civil war was linked to the populations’ growing awareness of social and cultural inequalities that had persisted for generations. The substantial increase in the number of languages reported in the 1996 Census, and all censuses thereafter, was therefore a manifestation of the wider phenomenon of ethno-linguistic awakening.

Since the conclusion of the civil war, the Government of Nepal (which today includes a majority of Maoist and Communist party members) has for the most part promoted linguistic diversity and harmony – for example, the 2015 Constitution declares that ‘all languages spoken in Nepal’ are to be considered national languages, and has opened the doors for individual provinces to declare their own official languages as part of the new federal structure – which is reflected in the 2011 Census results.

The Response

Arel (2002, p. 106) argues that in responding to questions around language, respondents can take a ‘forward looking’ or ‘backwards looking’ stance in providing their response. Arel points to the Belgium census of 1947 as an example of this, when many Flemish citizens provided “forward looking” responses by identifying French, as the language they wished their children would use in order to move up in social status.

Many respondents in Nepal appear to be providing ‘backwards looking’ responses in the 2011 Census. A ‘backwards looking’ response is one that reflects the language of one’s parents or ancestors, regardless of the individual’s actual knowledge or regular use of the language. ‘Backwards looking’ responses may be politically or ideologically driven – the lack of knowledge of one’s ancestors’ language being seen as a temporary state brought about by authoritarian state policies – or simply a nostalgic and sentimental nod to historical or cultural roots.

Nepal’s 2011 Census data reported that there was 1,424 people who speak Tilung as their mother tongue. The Language Commission of Nepal, through their own local-level surveys and consultations, found there to be only two fluent speakers of Tilung. Other members of the wider ethnic group reported Tilung as their mother tongue despite not speaking more than a handful of isolated words, thus displaying a ‘backwards looking’ approach in their responses. The Language Commission has received many anecdotal reports of other languages similarly being over-represented in Census results due to ‘backwards looking’ reporting of cultural heritage versus the language most often or most fluently spoken. And while there are still living speakers of a language, as with Tilung, correcting this practice would not necessarily change the total number of languages present; but knowing the precise number of speakers of a language allows government to better target language documentation and preservation efforts, particularly in a resource-strained context like Nepal. A language with 1,424 speakers might be considered only ‘threatened’, but with two elderly speakers it is ‘almost extinct’ (see Expanded Graded Intergenerational Disruption Scale).

Looking forward

Regardless of the precise number of languages reported, Nepal is and always has been a multilingual and multicultural country. Nepal is already planning for the 2021 Census, and by further refining its Census protocol and considering the various ways the socio-political context and personal ideologies may influence responses, the Government of Nepal will be able to better plan and implement linguistically-inclusive policies for its citizens.

References

Arel, D. (2002). Language categories in censuses: backward- or forward-looking? In D. I. Kertzer & D. Arel (Eds.), Census and Identity: The Politics of Race, Ethnicity, and Language in National Censuses (p. 97). Cambridge, UK ; New York: Cambridge University Press.

Bennett, L. (2005, December). Gender, Caste and Ethnic Exclusion in Nepal: Following the Policy Process from Analysis to Action. (p. 7) Paper presented at the New Frontiers of Social Policy Conference, Arusha, Tanzania.

Sebba, M. (2017). Awkward questions: language issues in the 2011 census in England. Journal of Multilingual and Multicultural Development. https://doi.org/10.1080/01434632.2017.1342651

Yadava, Y. (2014). Language Use in Nepal. In Central Bureau of Statistics. Population Monograph of Nepal Volume II: Social Demography. Central Bureau of Statistics.