The Newsletter 75 Autumn 2016

Hierarchy and contact: re-evaluating the Burmese dialects

Patrick McCormick

<p>We think of Burma as a country of great linguistic and ethnic diversity. The government today classifies the population into 135 ‘national races’, most of which they place under eight larger overarching categories. The <a href=\"\"><em>Ethnologue </em>website</a>, which provides information about languages across the globe, says that there are 117 living languages in the country. This discrepancy points to both differences in classification schemes, but also perhaps highlights an expectation—common at least in the English-speaking world—that language and ethnicity should be nearly co-equal. Since the colonial era, indigenous ways of making sense of identity and community, which may or may not have been based on language, have been overtaken by a practice that the British introduced: equating language first with race, and later, as the idea of race evolved, with ethnic group.</p>

I explore this intersection between language, identity, and the equation between language and ethnicity by looking at three examples of variation within the Burmese language itself: among the Rakhaing/Marma, Tavoyan and Intha dialects. Speakers of Rakhaing, which varies comparatively less from Standard Burmese than Tavoyan does, have a separate sense of identity, while many speakers of Tavoyan generally consider themselves to be ‘Burman’. Such identifications come out of complex, incompletely understood historical and political contingencies.

Image removed.

Looking at the larger overall picture, in whatever way Burmese dialect speakers think of themselves, their communities occupy a position near the top of local sociolinguistic hierarchies. Much scholarship has focused on the hierarchical interpersonal relations that are prominent features of lowland, court-based Southeast Asian societies. The idea of hierarchy is also useful for understanding how entire languages and their societies stand in relation to each other.

Ethnicity in Burma

In Burma, ethnicity is central to social organization, culture and politics. When the British arrived in the country in 1824, they tried to make sense of the diversity they found. The categories of caste and religion, which they had used in India, were not useful for categorizing the people. The British created a new technology of governance that would allow them to create subjects and, for example, recruit suitable people into the army: they equated language with ‘race’ to form racial categories. Outside of Burma, this earlier concept of race has gradually evolved into the concept of ‘ethnicity’ or ‘ethnic groups’. Two profound consequences of this first racial (later ethnic) thinking have been that groups of people in Burma have defined and redefined themselves into ethnic identities, where language plays a key, if equivocal role. These identities have been both a cause and consequence of long-standing conflict.

Scholars also disagree about ‘ethnicity’; many resist a historicization of ethnicity, that is, a discussion of how ethnicity as a way of understanding and categorizing difference has arisen in a particular context. Some take such arguments as a claim that the British somehow created difference. Others choose to downplay the role of ethnicity, treating it as a sort of political neurosis, a mania for a particular way of organizing the world, of the English-speaking world. I argue that difference among human populations has always existed, and people have always been aware of it. The practices of organizing and understanding difference through ethnicity—solid, singular, non-porous, identities, which are projected into the past—is a recent development stemming from European romantic nationalism.

Image removed.

In Burma, practices and ideologies of race began evolving during the colonial period (1824-1948). Over time, ethnicity has encroached on the earlier practices of identification from the pre-colonial Burmese empires, but which compete with ethnicity. Family and kinship connections, networks of patronage and loyalty, religious and cultural practices have all shaped how people understood themselves and wider communities. Identifications can vary depending on the context. In the pre-colonial period, not everyone who spoke the same language considered themselves to be part of the same community or to have the same identity.

Following the logic of ethnicity, language is central to identifying and recognizing difference. Linguistic scholarship, however, validate or falsify claims of ethnic unity or separation. It can provide insights into how languages and their varieties are connected, how long ago they separated, and how languages function in society. Political and historical processes have created ethnic communities, not language alone. For example, at the level of every day speech, Hindi and Urdu are the same language sharing a common origin, yet politics have divided them into separate languages. On the other hand, as Kojima (in this Focus) discusses the Palaung, who speak widely divergent varieties (perhaps even languages), nevertheless consider themselves to form one community.

The Burmese language and its varieties in Burma and beyond

Compared with the other national languages of Southeast Asia, Burmese varies little regionally. Out of a population of approximately fifty-two million, thirty-two million speak Burmese as a native language and another ten million speak it as a second language.1 Numbers throughout are from the Ethnologue website,  This relative lack of variation perhaps reflects the fact that the language has spread from Upper Burma, centered in Pagan and Mandalay, only in the past few centuries. Speakers are aware of the minor differences between the speech of Mandalay, Yangon, and Mawlamyaing, but they tend not to attach much meaning to this variation. Rather, they tend to be more attuned to the accents of second-language speakers, which along with visual cues such as physical appearance and clothing, can form the basis of stereotyped evaluations of those second-language speakers.

Dialects differing markedly from Standard Burmese have, however, arisen in geographical areas peripheral to the earlier, historical extent of the language, even into areas that are now other countries. To the far west of Burma, the Rakhaing Yoma mountain range cuts off Rakhaing State from the rest of central Burma. The state covers most of the historical region of Arakan, kingdoms whose power reached well into what is now Bangladesh. Rakhaing is the largest and best-known dialect, with about 800,000 native speakers in Rakhaing State and another 200,000 in Bangladesh. Another million speak Rakhaing as a second language, including speakers of Chin languages, Chakma (Daingnet), and Bengali dialects. Marma is a dialect of Rakhaing spoken by 150,000 people in Bangladesh and another 30,000 in Mizoram and Tripura in Northeast India. The Marma are the only group of Burmese dialect speakers who live entirely outside of Burma, and represent a population who fled Arakan after wars with central Burma in the eighteenth century.

Image removed.

On the other side of the country along the Tenasserim (Taninthāyi) coast, around 400,000 people speak Tavoyan in and around the city of Tavoy (Dawe), control of which has passed between Burmese and Thai courts over the centuries. The Tavoyans possibly came from further north for trade. The sounds of Tavoyan suggest that it broke off from the main body of speakers perhaps as early as the Pagan period (approximately 9th-13th centuries AD), just before Burmese was first written. To this day, the Tavoyan word for Burmans from further north of Tavoy is ganthà, literally ‘Pagan’ + ‘child’, the first syllable coming from a reduced form of the name of the ancient kingdom, Pagan. Finally, further north, about 90,000 speakers of the Intha dialect live on and around Inle Lake in southern Shan State. Historically, the Shan, who speak a Tai-Kadai language related to Thai, surrounded the Intha. Burmese speakers may have moved to the region as part of a military outpost. There are other Burmese dialects, some of which have been only sporadically described, including the Yaw and Beik (Mergui) varieties, which appear to be fairly close to Standard Burmese, and Taung’yo and Danu, both spoken in Shan State and which appear to more divergent.2 In August 2016, I interviewed Danu and Taung’yo speakers near Pindaya in Shan State. Both groups seem to have a distinct sense of identity.

Variation and intelligibility

Deciding on the division between ‘language’ and ‘dialect’ is fraught with difficulty, and has as much to do with ideology, history and politics, as with any kind of ‘objective’ assessment of how close or far two speech varieties are. For example, the fact that speakers of two varieties can understand each other does not guarantee that the one group will automatically reject or accept the other as members of their same group. Furthermore, what it means for two varieties to be ‘mutually intelligible’ also depends on the extent to which the various speakers involved have been exposed to language variation. When dealing with dialects, the influence and prestige of the standard language cannot be discounted – ‘standard’ forms can often displace dialectal forms. When trying to understand the development of the Burmese language and the dialects, linguists rely on how the Burmese language is written, or ‘Written Burmese’. To simplify a little, it preserves the sounds of the language from an earlier period. Written forms, for example, indicate consonant sounds before many of the sweeping sound changes that all spoken forms have undergone in recent centuries.

To understand something of how varieties of the spoken language diverge, we can consider the word for chicken, written <KRAK>.3 Following convention, forms between brackets <> represent transliterations from another alphabet (here Burmese), while forms between slashes // represent sounds.  which represents an earlier pronunciation. Through a series of sound changes, the modern Burmese pronunciation is /cɛʔ/, but in Rakhaing /kraʔ/. In this instance, Rakhaing seems to be more conservative than Standard Burmese. Judging by British sources, the change of the sound of /r/ to /y/ appears to have been continuing into the nineteenth century, and is the origin of such forms as ‘Rangoon’, which in current pronunciation is Yangoun. On the other hand, the word for temple or school, written /KYOṄḤ/, is /càuŋ/ in Standard Burmese and Rakhaing, but /klɔ̀ŋ/ in Tavoyan. The presence of the ‘l’ sound in Tavoyan suggests that it preserves a sound that existed in the oldest stages of Burmese, which was first written in the eleventh century AD.

Tavoyan Speaker, July 2016. Photo courtesy of the author.

Imagine this kind of change throughout the sound systems, together with differences in vocabulary and a few in syntax (or ‘grammar’). Differences in vocabulary can reflect borrowings from other languages. The Standard Burmese word loungyi [sarong] comes from Bengali lungi, while the Rakhaing word is təyɔ̀ and the Tavoyan is θəʃɔ̀ŋ, itself from Malay sarong.4 Okell, J. 1995. “Three Burmese dialects”, in D. Bradley (ed.) 1995. Papers in Southeast Asian Linguistics no. 13: Studies in Burmese Languages, Canberra: Department of Linguistics, Australian National University, pp.1-138.  Other words are simply made up of different native components: ‘boy’ in Standard Burmese is kauŋ.gəlè, literally ‘animal + little’, compared with Tavoyan pʰá.sú, literally ‘male + little’. Tavoyan now occurs only as an element in girls’ names in Standard Burmese. The calque is another mode of more subtle borrowing in which native elements replicate a model from another language. ‘Chili’ in Burmese is ŋəyouʔ.θì, literally ‘pepper + fruit’, but Intha saʔ.θì, literally ‘spicy + fruit’, replicates Shan maak pʰit, ‘fruit + spicy’.

Differences in syntax are less numerous. They may include the example of how questions are asked in Intha. For example, pʰa loʊʔ ne (ha) for ‘what are you doing?’ In Burmese, this would be ba louʔ ne (ðə) lɛ̀. The final lɛ̀ here indicates a relative question (that is, involving who, what, where, when, who), and may be dropped off in very casual speech in some circumstances. In Intha, however, it appears to be normal to leave it off or even replace it with the ha, apparently from Shan. The other dialects follow Standard Burmese in this regard.

Changing names, changing identifications?

Dealing with the names of languages, peoples, and what we call ‘ethnonyms’ today is in general highly problematic. The exact content of a term, or the people and situations in which a term is used, can change radically over time. People themselves change names and reinterpret them. One such example in English is the term ‘Dutch’, which refers to the language of the Netherlands but is cognate with the name Germans have for themselves and their language, Deutsch.

When dealing with Burmese dialect speakers, we see similar shifts and changes. In Burmese English, today the official name of the country, language and people is Myanmar, written <MRANMĀ>. The form <BAMĀ>, the origin of such terms as Burma and Burman in English and Bamā or Bamar in Burmese English, is from the same word, through a particular sound change. According to interviews I conducted in 2016, Rakhaings call themselves rɔʔkʰaiŋ and their language rɔʔkʰaiŋ.zəgɔ̀, literally ‘Rakhaing’ + ‘language’. Marmas call themselves mərəma and their language mərəma.ɟəgɔ̀. We have already seen how the sound change from r to y has affected Standard Burmese, so the connection between Marma and Myanma is immediate and intriguing. In the nineteenth century, British officials such as Hamilton found that terms like Rakhaing and Yanbye referred only to local parts of Arakan, and that the people of Arakan used Maramā to refer to themselves. They also used the terms Maramā-grī and Mranmā-grī. Today, however, this term refers only to the Baruas, the Bengali-speaking Buddhist population of what is now Bangladesh.

Tavoyans tend to call themselves Bamā, the same as the Standard Burmese name of the ethnic majority, and their language bəmagà, although sometimes (particularly for ideological reasons), some call themselves dəwɛ and their language dəwɛgà. Inthas call themselves ʔɛ̀nsà and their language ʔɛ̀n.səkà. British sources also use ‘Dawe’ for Intha. Whether this was a local appellation or some kind of classificatory confusion remains to be investigated. British scholars tried to draw connections between the various dialect speakers.

Such variation in names seem to index as much emerging categories as historical and political developments (the rise of nationalism, the political benefits of being an ethnic group), and emerging schemes of knowledge. These last include developing linguistic classifications, and government technologies of governance, starting with British practices and evolving over time, through to the most recent government census of 2014. A final element is no doubt people’s own lived experiences, in which earlier pre-colonial practices, such as having multiple or shifting identities, was possible and normal. One way to understand the shift in the meaning of Mranmā-grī is to note that, in what is now Bangladesh, Rakhaings are a Buddhist community with close ties to the highly institutionalized and prestigious Burmese Buddhism. If Buddhist practices were less organized or institutionalized among the Barua, they would seek training and ordination in Rakhaing institutions, where they would become exposed to the language and thus in a sense ‘become’ a kind of Rakhaing.

Positioning Burmese and its dialects hierarchically

In whatever way the speakers of these Burmese dialects evaluate themselves, they appear to be at or near the top of the local language hierarchy. Whatever the names for themselves and their languages, dialect speakers tend to have the same status as Burmans in the language ecology. My observation comes out of a body of scholarship on multilingualism and language contact. In situations where there is multilingualism—people regularly speaking many languages—the overall trend is that who speaks what language, or who does and does not learn a specific language, reflects where speech communities fall on a hierarchy. A corollary of ‘learning up’ is that speakers of lower-placed languages can replicate both matter (usually thought of as ‘borrowings’) and patterns (or ‘syntax’) by reanalyzing native words or forms and using them in the same way as the model language.

Women on Tractor, Inle Lake. Image reproduced under a creative commons license courtesy of fabulousfabs on flickr.

Said another way, people tend to learn the languages higher up on the hierarchy and not the other way. Inside Burma, the general trend is that the more languages one speaks, the lower one is on the hierarchy. Slightly confusingly, higher status languages live in lower altitudes—the lowlands and in mountain valleys—and lower status language speakers live in higher altitudes—the uplands, highlands, and mountains. An example from the Kachin world would be a Maru or Lhaovo speaker, who will also know Jinghpaw and Burmese, and possibly Shan or Chinese. Jinghpaw and Shan speakers would not learn Lhaovo unless they had family connections. Burmese rarely learn minority languages, learning instead English and such economically useful languages as Korean or Japanese. Parallels outside of Burma are many, such as eastern Europeans who learn many western European languages for work and education opportunities, but not vice-versa.

This trend of speaking and replicating up is a general pattern, and there are important exceptions. The kind of multilingualism meant here is long-term and stable, and a not rapid language shift. Certain populations will learn languages lower on the hierarchy, such as Chinese and other traders who learn many languages to facilitate business. Absolute numbers can make a crucial difference: in some parts of lower Burma, where the Mons are the local majority, some Burmese do in fact learn Mon. Based purely on linguistic observation, in Bangladesh, a Muslim-majority country, the Bengali language stands above Marma and Rakhaing, yet within local Buddhist communities, Rakhaing and Marma are at the apex. The situation in Bangladesh highlights the importance of Burmese Buddhism, with its great prestige and institutional complexity. The Burmese government in fact indirectly supports Buddhism in Bangladesh, by facilitating monks from there coming to Burma to study.

Absolute numbers may also help explain the phonological situation in some of the Burmese dialects. Historically, apparently Shans learned Intha but not the other way around, and speakers of Karen languages around Tavoyan learn Tavoyan but not the other way around. When large numbers of speakers speak a higher-status language, they may take with them certain speech habits. If the higher-status language is spoken by a fairly small number of people, the surrounding speech habits may work their way into the speech of monolinguals. Following this observation, Intha lacks the sound /θ/, having instead /s/, which represents the influence of Shan.

If we look at Tavoyan ‘on the page’, at how each word in a Tavoyan sentence corresponds with Standard Burmese, the two are quite close once we understand how older sounds in each variety have shifted down to the present. But when we listen to Tavoyan, it is at first strikingly, incomprehensibly different for most speakers of Standard Burmese. The prosody or intonation is quite different, lacking the drawn-out sounds that Burmese uses for emphasis. There is the presence of the /l/ as in /klɔ̀ŋ/ [school], and the nasal vowels of Standard Burmese are often not nasalized. Finally, the sounds /b/ and /d/ often sound ‘imploded’ as /ɓ/ /ɗ/ as in Khmer or Vietnamese (see Jenny’s essay in this Focus). These exact points of difference may have parallels in the Karen languages. In linguistically assimilating to Tavoyan, they may have carried these features over into Tavoyan speech. I caution, however, that much more work would have to be done to thoroughly describe all aspects of the sound systems of the dialects and of the language(s) they have come into contact with before we can make such assumptions certain.

In conclusion, a sense of difference, social distance, ideas of allegiance and patronage may no doubt be quite old, but how language feeds into the creation and maintenance of difference is equivocal. Linguistic scholarship alone cannot make or break an ethnic identification, but it can provide useful insights into how speech communities relate to each other. Studying sound changes is a big part of what the study of historical linguistics is. The question of how the sound system of Old Burmese developed into what we now hear in Burmese and the dialects may seem to be the arcane concern of latter-day philologists being paid cushy money from European universities. But sound change can offer direct insights into historical and social processes. Linguistic data thus is a form of historical evidence.

Patrick McCormick, École française d'Extrême-Orient (Yangon) and researcher in the Department of Comparative Linguistics, University of Zürich (

First photo: Marma Nuns, Yangon, 2016. Courtesy of the author.