A short introduction to Belhare and its speakers

Balthasar Bickel

University of California at Berkeley and University of Zürich

NOTE: For a more elaborate (and more recent) summary on Belhare, download the g'zipped PDF draft of my contribution to Thurgood & LaPolla's The Sino-Tibetan languages,to be published by Curzon Press.


Belhare (in Nepali Belhaareor, in sanskritizing newspaper language, Belhaariiya)is spoken by about two thousand people living on the Belhara hill, one of the small but steep southern foothills of the Himalayas situated in theDhankuta (Nep. DhankuTaa)district of the Koshi zone in Eastern Nepal (87° 18' E and 26° 57' N; ca. 1150 m above sea-level; click here for a map).* The language has also been recorded as Athpare (Nep. AaThpaare or AaThpahariiya).As shown by Vikal & Rai (2051), this designation has much cultural but hardly any linguistic justification. 'Athpare' refers to the ethnic unit that is constituted by the close cultural ties between the Belhare and the autochthonous inhabitants of neighbouring Dhankuta bazaar. The languages of the Athpare from Dhankuta and from Belhara are very similar but beyond mutual intelligibility (see Ebert 1994 and 1997 on the Athpare from Dhankuta). Although calling themselves both Athpare, people recognise this linguistic difference. If clarification is required, the two sub-groups are distinguished by calling the Dhankuta people Noupagari (Nep. Nau-pagaDii,literally, 'nine-turbans') and the Belhare people Athpagari (Nep. AaTh-pagaDii, literally, 'eight-turbans'), an ethnonymy reflecting dominant socio-cultural categories (see op. cit.).Occasionally and somewhat misleadingly, the Belhare are also identified as Khalsali (e.g., Dahal 1985), i.e., as the people of Khalsa,the region including Belhara but stretching out in the west into a linguistically distinct area. On the other hand, the Athpare from Dhankuta are sometimes set off from the Belhare by the native term SanaNguchi,which is perceived unpleasant. (I failed to record a (folk) etymology, though.) To avoid any confusion, the Linguistic Survey of Nepal has adopted the toponymically derived term Belhareas the unique designation of the language spoken by the original inhabitants of the Belhara hill (Hanßon 1991).

The Belhare people, or, for that matter, the Athpare as a whole, are classified as Raaii.This term reflects the ethnic categorisation that arose in the aftermath of the forced unification and foundation of modern Nepal under king Prthvii Naraayan Shaah (imperabat1742 - 1775). Originally an administrative title, the term was more and more perceived as the name of a social and political unit. Although indeed accepted by many groups in Eastern Nepal, the term is frequently, but not unanimously, rejected by Athpare people. Apart from being of dubious ethnological significance, the term Raaiidoes not correspond to a linguistically valid distinction because it does not extend to the neighbouring Limbu language, which belongs to the same genetic sub-branch as Belhare and Athpare, viz. to the Kiranti (Nep. Kiraa~tii) family. The Kiranti family includes most languages of the Eastern area (Nep. Puurbaañcal)of Nepal and extends also to adjacent areas in the west (Janakpur zone) and east (Sikkim). Click here to look at the position of Belhare in an abbreviated Stammbaum,drawn on the basis of Hanßon's (1991) tentative classification.

The precise position of the Kiranti group within Sino-Tibetan is a matter of dispute and any statement seems premature before the reconstruction of Proto-Kiranti has received enough elaboration and sophistication. Most authors classify the Kiranti family as Bodic, if such an intermediate node (including languages like Tibetan, Magar, Lepcha etc.) within Tibeto-Burman is assumed at all. On the basis of verbal agreement morphology, the Kiranti family has been compared to the Qiangic and Nungish groups spoken, respectively, in Sichuan and the China-Burma border area (Thurgood 1984), but the level of genetic relation is far from clear. Expanding on recurrent speculations in the literature, van Driem (1993) discusses morphological evidence for a close genetic relationship of Kiranti with Newar (Nepaal bhaashaa), constituting a major division ('Mahaakiraa~tii') within the Bodic family.

Virtually all speakers of Belhare are bilingual in Nepali, the national Indo-Aryan lingua franca that has had wider currency at least since the unification of Nepal (see van Driem 1991, Pradhan 1991). Although this bilingualism has resulted in frequent code mixing and a large amount of Nepali loan-words, the grammar of Belhare has maintained its distinct Kiranti characteristics (cf. Bickel 1995b, 1997a) and continues to exert a strong influence on the Nepali spoken by Belhare people. This confirms Turner's (1931: xiv) early observation that Nepali is deeply affected by Tibeto-Burman grammar but is strongly resistant against such influences in its vocabulary.

Unlike many other Sino-Tibetan languages, but in line with the Kiranti group in general, the grammar of Belhare is characterized by an elaborate morphology in both the nominal and verbal domain. Nominals are inflected for number and case. The number category on nouns generally distinguishes unmarked forms from non-singular (-chi) but is further differentiated into dual (-chi) and plural (-i, -kha) values in first and second person pronouns (-chi 'd' vs. -i 'p'), demonstratives (-khachi 'd' vs. -kha 'p') and the suffixal article (-na 's', -khachi 'd', -kha 'p'). Dual and plural are also distinguished in the verbal agreement morphology (with -chi for dual in all persons and -i for plural in first and second person only). Pronouns and verbal agreeement incorporate an exclusive [-addressee] vs. inclusive [+addressee] opposition in the first person, e.g. Nkechi 'we two incl.' vs. NkeNchiN 'we two excl.'. The case suffix system includes the categories absolutive (with zero marking), oblique (in -Na or -a; encompassing ergative, instrumental, and causal functions), genitive (in -Nahak or -hak) comitative (-lok), ablative (-huN or -nahuN), directive (-leN), mediative (-lamma 'via, through the medium of'), general locative (-et) and the 'environmental space' locatives that differentiate 'upward' (-ttaN), 'downward' (-pmu) and 'across' (-?ya) locations (see Bickel 1997b). Oblique marking with actor noun phrases ('ergative') is compulsory, except with the first person singular pronoun, which never appears in this case. Syntactically, Belhare constructions monitor partly an accusative, partly an ergative style pivot, but in terms of frequency, accusative syntax is more prominent (Bickel, in press). Both clause combining and simple sentence constructions crucially involve topic marking (with -na) (Bickel 1993). Together with a rich system of focus and illocutive marking devices this feature makes the information structure of a Belhare text highly explicit.

This is not to say that word order is not also exploited to signal information structure. Indeed, the order of clause and sentence constituents varies considerably in both dialogues and narratives and seems to largely follow discourse principles. Word order is not relevant for syntax. There is no syntactic rule or constraint sensitive to constituent order or clausal configurationality, and there is no evidence whatsoever for a VP (Bickel, in press). In contrast with clause and sentence structure, nominal constituents are configurationally regimented in a rather rigid way. There is a complex noun phrase syntax and elements of a noun phrase are subject to extraction constraints.

Belhare belongs to what is known to Tibeto-Burman linguists since Hodgson (1856) as a 'complex pronominalized language'. The characterisation refers to the elaborate agreement morphology in the verb which reduces the use of independent pronouns to contrastive purposes and allows the typical Belhare clause to consist of a verb only. The Belhare verb cross-references both the actor (A) and undergoer (U) argument of a transitive clause. In intransitives, the agentive and patientive character of the single argument is neutralised and the verb agrees with the subject (S) irrespective of its role. The morphology expressing person, number and role values is analysed in Bickel (1995a), where it is shown that there is no systematic alignment of the inflectional system in either an ergative or accusative style. The Belhare verb also distinguishes two tenses (past in -t or -yuk and nonpast in -he or -att), five marked moods (subjunctive in -a, definitive in -yuk , inconsequential in -kone, imperative in -a or -an, and optative in ak-), four aspect markers (temporary in -hett, spatially distributed temporary in -kon, imperfective in -yakt ~ ya- ~ yau-, inceptive in -kett), and two types of perfect (a simple perfect in -Na...-khak or -sa...-khak, and a resultative perfect in -Ne or -se). See Bickel (1996) for a monograph on this system. The tense, mood, and aspect markers are affixed to verb stems which can be expanded by series of Aktionsart modifiers ranging from categories like telicity over dimittive to reflexive.

The elaborate verb agreement morphology notwithstanding, Belhare is not a concentric language in the sense of Milewski (1950) but rather belongs to the intermediate group of 'double-marking' languages in Nichols' (1986) typology. This is due to the fact that oblique case morphology (in -Na or -a) is usually obligatory if there is an actor noun phrase in the transitive clause, i.e., the ergative function is always marked. Likewise, both attribute and head of a noun phrase are usually marked morphologically. The head noun optionally cross-references the attribute by a possessive prefix. The attribute is marked by a genitive suffix if it is a common noun. Verbal attributes and demonstratives are marked either by a general nominaliser (-khak)or, if the noun phrase has specific reference, by the suffixed article. Colour terms, which constitute a distinct lexical category, are suffixed by -ma if the reference is specific and by the standard nominaliser if specificity is not at issue. Numerals are marked by a classifier.

Belhare phonology distinguishes voicing and aspiration with stops and employs a five-vowel system, including nasalized /i/ and /u/, as well as a series of falling (closing) diphthongs. Whereas the prosodic and the morphological word often do not align, Belhare is subject to principles of prosodic morpheme integrity which trigger various processes of gemination, resyllabification, and destressing (see Bickel 1998). Together with a trochaic stress pattern, this gives the language a highly characteristic rhythm.


Bickel, Balthasar, 1993. Belhare subordination and the theory of topic. In: Karen H. Ebert [ed.], Studies in Clause Linkage, pp. 23 - 55. Zürich: ASAS-Verlag.

Bickel, Balthasar, 1995a. In the vestibule of meaning: Transitivity inversion as a morphological phenomenon. Studies in Language 19, 73 - 127.

Bickel, Balthasar, 1995b. Relatives à antécédent interne, nominalisation et focalisation: entre syntaxe et morphologie en bélharien. Bulletin de la Sociétéde Linguistique de Paris 90, 391 - 427.

Bickel, Balthasar, 1996. Aspect, mood, and time in Belhare. Studies in the semantics - pragmatics interface of a Himalayan language.Zürich: ASAS-Verlag.

Bickel, Balthasar, 1997a. The possessive of experience in Belhare. In: David Bradley [ed.], Tibeto-Burman Languages of the Himalayas,pp. 129 - 149. Canberra: Pacific Linguistics.

Bickel, Balthasar, 1997b. Spatial operations in deixis, cognition, and culture: Where to orient oneself in Belhare. In: Jan Nuyts & Eric Pederson [eds.], Language and Conceptualization, pp. 46 - 83. Cambridge: Cambridge University Press.

Bickel, Balthasar 1998. Rhythm and feet in Belhare morphology. Ms., University of California at Berkeley [available from the Rutgers Optimality Archive, ROA No. 287]

Bickel, Balthasar, in press. Hidden syntax in Belhare. In: George van Driem [ed.], Himalayan Linguistics,Berlin: Mouton de Gruyter. [Ms. downloadable]

Dahal, Dilli Ram, 1985. An ethnographic study of social change among the Athpahariya Rais of Dhankuta.Kathmandu: Centre for Nepal and Asian Studies.

Driem, George van, 1991. Taal en identiteit: indo-arisch expansionisme in oostelijk Nepal. Bijdragen tot de taal-, land- en volkenkunde147, 61 - 73.

Driem, George van, 1993. The Newar verb in Tibeto-Burman perspective. Acta Linguistica Hafniensia 26, 23 - 43.

Ebert, Karen H., 1994. The structure of Kiranti languages: comparative grammar and texts.Zürich: ASAS-Verlag.

Ebert, Karen H., 1997. Athpare.München: LINCOM-EUROPA.

Hanßon, Gerd, 1991. The Rai of Eastern Nepal: Ethnic and linguistic grouping. Kirtipur, Kathmandu: Linguistic Survey of Nepal and Centre for Nepal and Asia Studies.

Hodgson, Brian Houghton, 1856. Aborigines of the Nilgiris, with remarks on their affinities. Journal of the Asiatic Society of Bengal 25, 498 - 522.

Milewski, Tadeusz, 1950. La structure de la phrase dans les langues indigènes de l'Amérique du Nord. In: Tadeusz Milewski [ed.], Études typologiques sur les langues indigènes de l'Amérique, pp. 70 - 101. Kraków: Polska Akademia Nauk 1967.

Nichols, Johanna, 1986. Head-marking and dependent-marking grammar. Language62, 56 - 119.

Pradhan, Kumar, 1991. The Gorkha Conquests.The process and consequences of the unification of Nepal, with particular reference to Eastern Nepal. Calcutta: Oxford University Press.

Thurgood, Graham, 1984. The Rung languages: a major new Tibeto-Burman subgroup. Proceedings of the 10th Annual Meeting of the Berkeley Linguistics Society 338 - 49.

Turner, Ralph Lilley, 1931. A Comparative and Etymological Dictionary of the Nepali Language. Reprint New Delhi: Allied Publishers Ltd. 1980.

Vikal, Bal Bahaadur (= Balthasar Bickel) & Lekh Bahaadur Rai, 2051. AaThpahariiya jaati ke ho? ('What is the Athpahariya caste?'). Janavishvaas Saaptaahik 2/23, 3 - 4.


Last updated 11/3/98/BB.