More than seven thousand languages are spoken across the world today, yet their distribution is significantly unequal. About ninety-seven percent of the world’s people speak only about four percent of these languages. Behind these numbers are communities whose histories, ways of understanding the world, and identities are tied to languages spoken by only a few thousand people. UNESCO estimates that nearly half of the world’s languages are endangered and that one language disappears approximately every two weeks. Language loss therefore represents one of the most pressing cultural crises of the twenty-first century.

Bangladesh is not exempt from this global reality. Our national identity is historically intertwined with language politics. The Language Movement of 1952 established Bangla as a central symbol of sovereignty and dignity, and it inspired UNESCO to recognise 21 February as International Mother Language Day. Yet this powerful national narrative has often obscured the country’s internal linguistic diversity. Hyow is one such language, part of the linguistic and cultural diversity of Bangladesh, and one that I have been researching for the last twenty years.

Hyow (its endonym), officially known as Khiang (an exonym), has a population of 4,826 according to the Population and Housing Census 2022. The word ‘Hyow’ refers to the Chin ethnic community, which belongs to the South Central (formerly Kuki-Chin) branch of the Tibeto-Burman language family. It is spoken in around thirty-one villages across three districts of the Chittagong Hill Tracts and along the border areas of Chattogram and Bandarban.

Hyow has two varieties, Laitu and Kongtu, spoken in lowland and highland areas respectively. They show only minor linguistic differences. Hyow forms part of a broader linguistic continuum that stretches from the Chittagong Hill Tracts into the Chin and Rakhine states of Myanmar. A language called Laitu is spoken in around twenty-two villages along the Yangon–Sittwe Asia Highway, and another language, Le(i)tu, is found along the Lemyo (or Lemro) River in Rakhine. Although their names are similar and they share linguistic features, they are not fully mutually intelligible with Hyow. Some external classifications treat Hyow as a dialect of Asho, spoken in southern Rakhine. However, linguistic evidence shows that Hyow is a distinct language and not mutually intelligible with Asho.

Becoming a linguist was not what I had planned as an undergraduate student at the University of Dhaka. I began working as an interior designer in my second year as an undergraduate and was preparing to pursue it as a full-time profession. Everything changed in 2006 when I attended a workshop on language documentation conducted by David Peterson, who had been working on South Central languages in the Chittagong Hill Tracts since 1998.

After a month of training, we travelled to Bandarban for fieldwork. I was assigned to work on Hyow alongside another student. Our language consultant, Hlakray Prue, later became a pioneering figure in building educational and professional opportunities within the Hyow community. That one week of fieldwork—collecting wordlists, analysing basic linguistic constructions, and interacting directly with speakers—left a lasting impression on me. It was the first time I realised how much knowledge a language holds and how fragile that knowledge can be.

Language documentation emerged in the mid-1990s as a response to global language loss. Therefore, its introduction as a methodology in Bangladesh was not too late. The workshop on language documentation that I attended was its second edition; another cohort had participated in a similar workshop in 2004. In both 2004 and 2006, the groups of participants included students as well as teachers. Unfortunately, there was no institutional capacity building for language documentation even after these two workshops. As it was not easy for Dr Peterson to travel to the Chittagong Hill Tracts regularly due to his teaching commitments at Dartmouth College, he employed me as his research assistant to transcribe and translate audio recordings of Hyow oral stories.

From 2007 to 2009, I conducted intermittent fieldwork on Hyow. During this period, I was also part of Dr Peterson’s expedition to track Rengmitca speakers in the remote areas of Alikadam in Bandarban. Even after these opportunities, I was unable to develop my skills in documentary and theoretical linguistics fully, partly due to a lack of institutional resources. I completed my MA in Linguistics in 2009 and was then diverted from my research on Hyow by my new job as a teacher at Scholastica.

In my experience, Hyow is a relatively stable language, but it is not without the risk of losing its distinctiveness. I already hear an overwhelming proportion of Bangla words when young Hyow speakers communicate with one another. With intense and long-standing contact, there is a strong possibility of structural change in Hyow.

It was not until 2013 that I was able to prepare myself to pursue a PhD programme and finish what I had started. Out of the forty-eight months of my PhD programme at Nanyang Technological University in Singapore, I spent almost fourteen months in Bandarban recording, transcribing, translating, and analysing different genres of texts in Hyow. In the early stages of my PhD, I concentrated on the sound system of Hyow. It took me almost six months to conduct acoustic analysis and measure the physical properties of different vowels, consonants, and tones in the language. Generally, researchers use three to four hours of annotated texts to write a grammar. Writing a grammar as part of a PhD dissertation, however, is a mammoth task. As a result, researchers also rely on elicitation, a technique that was used extensively in earlier periods to translate sentences for investigating specific grammatical features of a language. However, since direct translations from English or other major languages are problematic, modern language documentation methodologies place emphasis on natural texts. As I already had several hours of Hyow texts from my fieldwork between 2007 and 2009, the corpus size available for writing the grammar of Hyow was very strong. In total, I had ten hours of annotated natural texts on which to base the grammar of Hyow.

For typologists who compare structural patterns across languages, such as word order or agreement systems, Hyow and its closely related languages are particularly important. They exhibit structural features that are rarely documented elsewhere, making them valuable not only for the community itself but also for a broader understanding of human language.

Language shift among younger generations is increasingly evident. Migration to urban centres, Bangla-dominant education systems, and the consumption of digital media are accelerating the erosion of intergenerational transmission. At the same time, digital technologies are increasingly mediating access to education, public services, economic opportunities, and cultural participation. In my experience, Hyow is a relatively stable language, but it is not without the risk of losing its distinctiveness. I already hear an overwhelming proportion of Bangla words when young Hyow speakers communicate with one another. With intense and long-standing contact, there is a strong possibility of structural change in Hyow. This is true not only for Hyow but also for other Indigenous languages in the Chittagong Hill Tracts, and it is not merely an impressionistic observation.

These gradual changes may seem subtle, but they raise an important question: what must be done before a language begins to lose not only its words but also its linguistic structure and identity? In recent years, there have been small-scale initiatives to introduce education in the mother tongue. We have also seen large-scale initiatives aimed at developing digital tools and digitising Indigenous languages. We can expect new initiatives focused on developing AI-based digital tools to emerge very soon because of their global appeal. If we assess the outcomes of the first two sets of initiatives, we may gain insights into what to expect from the latter. From my conversations with local scholars involved in mother-tongue education initiatives, I detect frustration in their voices. This is largely due to administrative weaknesses. There appears to be a lack of coordinated effort to make these projects successful. As a linguist, I also believe that such initiatives must be linguistically well informed. Large-scale initiatives to develop digital tools for Indigenous languages, such as keyboards, likewise need to be grounded in sound linguistic analysis. Without this, there is a real risk of errors in these tools and a waste of public funds. We already have strong linguistic research on Cak, Chakma, Hyow, Khumi, Koda, Marma, and Pangkhua. Researchers working on these and other languages should be involved in any initiatives of this kind. Anticipated initiatives to develop AI-based tools for languages in Bangladesh will require even greater care, as their quality depends directly on the quality of the data on which they are trained. Moreover, such projects will require the development of scholars within local communities. In the case of Hyow, a robust corpus already exists, and some community members possess strong metalinguistic knowledge due to their involvement in earlier research.

The development of digital tools and artificial intelligence may offer new possibilities, but technology alone cannot safeguard a language. Without careful linguistic analysis, high-quality data, and sustained community involvement, even the most advanced systems will fail.

Before a language falls silent, what it needs most is commitment from scholars, institutions, and the community itself. Ekushey teaches us that language is dignity, and languages survive when people decide that they matter.

Muhammad Zakaria is a Swiss National Science Foundation Advanced Postdoctoral Fellow in Linguistics at the University of Bern.

Send your articles for Slow Reads to [email protected]. Check out our submission guidelines for details.