Limbu (Yakthung) Grammar — Cross-check & Dialect Reference

A structured grammar reference for Limbu, built primarily from a full descriptive grammar of the Chhatthare dialect, with an explicit, separately-labelled section noting where Chhatthare differs from the standard Phedappe / Panthare ("van Driem") Limbu. The dialect-difference list exists so that a translator or TTS/ASR developer does NOT accidentally mix inventories across dialects.

Sources

1. Tumbahang, Govinda Bahadur (2007). A Descriptive Grammar of Chhatthare Limbu. A Dissertation Report, Central Department of English, Tribhuvan University, Kirtipur. — The substantive grammar; nearly every paradigm and page citation below is from this work. Page citations are to the printed footer page numbers of the PDF.
2. Sighak, Bishnu Maya (2024). "Limbu News Bulletins in Radio Broadcasting: A Study of Present Status and Future Adaptation." Educational Journal 3(2), Research Management Cell, TU Sanothimi. — A sociolinguistic / media study, NOT a grammar. It contributes background only (dialect inventory, speaker numbers, the fact that Radio Nepal standardises on Panchthare). It has no grammatical paradigms; do not expect morphology from it.

Caveat on source #2. The Sighak article is a journalism/media paper. Its only relevance here is corroborating (a) the four-dialect division — Panchthare, Phedappe, Tamarkhole, Chhathare — and (b) that the broadcast "standard" is the Panchthare dialect, which non-Panchthare speakers find hard to follow (Sighak 2024, p. 38–39). Everything grammatical is from Tumbahang (2007).

Cross-check status. Tumbahang himself ran a 3-day comparative workshop (Dharan) with Chhatthare, Phedappe, Panthare and Taplejungge speakers, and repeatedly contrasts his Chhatthare data against van Driem 1987 (Phedappe), Wiedert & Subba 1985 (Panthare/"standard"), and Mikhailovsky 2003 (Mewakhole/Taplejungge). So the dialect-difference claims below are the author's own explicit comparisons, not my inference. Where I add a comparison not stated in the text I mark it [inferred].

Transcription key (as in the source, normalised to IPA where useful)

The PDF's text layer renders the open vowel ɔ (and occasionally a schwa) as a box. It is written /ɔ/ below. Other source conventions, kept as-is in cited forms:

Gloss abbreviations follow the source: 1/2/3 person, s/d/p singular/dual/plural, ns non-singular, i/e inclusive/exclusive, S subject, A agent, O object, pO/dO/sO plural/dual/singular object, pA/dA/sA plural/dual/singular agent, PT past, NPT non-past, NEG, REFL, RECIP, ERG, ABS, GEN, LOC, COM, DEF, NML nominaliser, SEQ sequential, Prg progressive, CONV converb, 1→2 first-acts-on-second portmanteau, etc.

1. Typological overview

• Eastern Kiranti, Tibeto-Burman; Bodic > East Himalayish (Kiranti). (Tumbahang 2007, p. 40, Fig. 1)
• Verb-final (SOV). "Limbu is a verb final language" (p. 319). Almost fully head-final / modifier-before-head, with a small set of head-after-modifier exceptions in kinship phrases (§7).
• Morphologically ergative but with split ergativity (§6); the verb is biactantial / polypersonal — it agrees with both agent and object.
• Mixed morphological type: agglutinative in nominal morphology, fusional / portmanteau-heavy in verb morphology (p. 42–44). Many verb affixes are portmanteaux.
• Phonology: 20 consonants, 7 vowels, no vowel-length contrast in Chhatthare (p. 52). (This is itself a dialect difference — see §9.)

2. Nouns: number, case, gender, classifiers

2.1 Number — three-way (singular / dual / plural)

Number is marked by nominal suffixes. The diachronic order is singular → plural → dual (dual is "a special case of non-singular", developed later in Kiranti). (p. 122)

• pu 'bird' → pu-gʰa 'birds' → pu-gʰa-cʰi 'two birds' (p. 122)
• -si marks non-singularity as an identity operator (e.g. appHaN-si 'they (pl.) are my uncles', anni-si 'they (du.) are my aunts'). (p. 123)
• When a numeral modifies the noun, the noun is NOT marked for number: nEccʰi pi? 'two cows', sumsi pHak 'three pigs'. (p. 124)

2.2 Case — twelve cases

Nouns inflect for case; pronouns do NOT take ergative or absolutive (§3). The full declension of napmi 'man' (Table 14, p. 140–141):

Notes on the key cases:

• Absolutive ∅ — zero-marked; transitive object and intransitive/reflexive subject take absolutive (p. 126). A definite absolutive -iŋ appears on definite/identifiable nouns: napmi-iŋ laŋgʰEk 'the man walks' (p. 127). The definite -iŋ assimilates: -miŋ after bilabials, -niŋ after dentals, -ŋiŋ after velars. It never occurs on personal pronouns, but does occur on demonstratives (p. 128).
• Ergative -Na ~ -na ~ -ma (oblique marker) — agent of a transitive verb. Allomorphy by place of articulation of the preceding consonant: -Na after velars, -na after dentals, -ma after bilabials (p. 129). A definite ergative -i combines with it (koco-ŋa-i napmi har-u 'the dog bit the man'). The ergative never occurs on personal pronouns (only on demonstratives / third-person referents) (p. 130). This single suffix -Na/-na/-ma is the multifunctional oblique marking ergative, instrumental and causative (p. 333).
• Genitive -NaN/-naN/-maN — possessor; the possessed noun then takes a possessive prefix (napmi-ŋaN ku-baŋ 'a man's house'). The genitive is collapsing into the bare nominative synchronically (lahaŋ ku-baŋ for lahaŋ-ŋaN ku-baŋ 'Lahang's house') (p. 135).
• Locative -o (with homorganic consonant insertion: -mo/-no/-No), also marks direction depending on the verb (p. 136).
• Comitative -nuŋ 'with'; negative counterpart deprivative -ma?E 'without' (p. 136).
• Mediative & Ablative share -lam and -nuŋ ('through' / 'from / path') (p. 137–138).
• Directive -naN ~ -lEkkʰaŋ 'towards' (p. 139).
• Allative -dʰarik 'as far as' (p. 139).
• Comparative -aŋ, or periphrastically comitative -nuŋ + subordinator -E → -nuŋNE; synchronically often Nepali bʰanda 'than'. Superlative = kErEk 'all' before the comparative (p. 140, p. 336–337).
• Vocative -E ~ -o — used only with kinship nouns (p. 134, p. 337).

2.3 Gender — minimal, non-productive

Only a few kinship terms, ethnonyms and a handful of common nouns inflect: masculine -pa/-ba, feminine -ma (p. 117–118). E.g. nak-pa 'nephew' / nak-ma 'niece'; yakthuŋ-ba 'Limbu (m)' / yakthuŋ-ma 'Limbu (f)'. Adjectives agree (cuk-pa 'small (m)' / cuk-ma 'small (f)'). Most kinship terms do NOT inflect for gender; sex pairs like yEmba 'husband' / me? 'wife' are separate lexemes, not grammatical gender (p. 119).

2.4 Numerals & human classifiers

Native counting in normal speech only goes up to sumsi 'three'; higher numerals exist but Nepali is used in practice (p. 119–121).

lɔtʰik 1, nEccʰi 2, sumsi 3, lisi 4, nasi 5, tuksi 6, nusi 7, yEtcʰi 8, pʰaŋsi 9, thiboŋ 10, kiptʰik 100.

Human classifiers (used when counting people):

• -pa (tHippa) — singular human (one person).
• -pHu (nEppHu, sumbHu) — non-singular human (two/three persons). (p. 121) tHippa 'one person', nEppHu 'two people', sumbHu 'three people' — these also function as nominals and inflect for case.

2.5 Pronominalisation on nouns

Possession is by prefix: a- 1sPOSS, ka- 2sPOSS, ku- 3sPOSS (ku-miN 'his name', ka-mik 'your eye', a-huk 'my hand') (p. 124). Nouns + pronominal suffixes form verbless (copula-less) sentences: napmi-Na 'I am a man', napmi-na 'you are a man', napmi-si-Na 'we are people' (p. 125).

3. Pronouns

3.1 Personal pronouns (Table 15, p. 145)

First & second person distinguish singular / dual / plural; first person dual & plural distinguish inclusive / exclusive. Third person distinguishes only singular / non-singular.

Morphemic make-up (p. 142): 1s underlying an (surfaces as a word-finally), dual -cʰi, plural -i, exclusive -ŋa (inclusive = unmarked/a-). Third person -cʰi covers both dual and plural (a non-singular marker).

Key structural fact for translators: personal pronouns do NOT take ergative or absolutive case markers and stay unchanged as subject / agent / object (p. 143). Only demonstratives and other 3rd-person referents inflect for ERG/ABS. This is the basis of the split-ergativity in §6.

Personal pronouns DO inflect for the oblique cases — locative, comitative, mediative, ablative, elative, allative, comparative (Tables 16–18, p. 145–146). E.g. 1s: a-o LOC, a-nuŋ COM, a-lam MED/ABL, a-dʰarik ALL, a-aN COMPR.

3.2 Possessive pronouns / prefixes

Bound possessive prefixes: a- 1s, ka- 2s, ku- 3s; non-singular forms use the full pronoun-derived prefixes (ancʰi- 1di, ani- 1pi, kʰuncʰi- 3ns, etc.) (p. 146). Independent ("objective", like English mine) possessives carry a clear genitive -Niŋ/-iŋ: ba paŋ-Niŋ aŋ 'this house is mine', ba paŋ-Niŋ kʰEnEŋ 'this house is yours' (p. 147). Note the ambiguity: ku- 3sPOSS is not co-referential with any noun in the clause, so kukku = 'our uncle' or 'his uncle' (homophony of duplicated ku- 'our' and ku- 'his'), disambiguated only by context (p. 148).

3.3 Demonstrative pronouns

Two-way proximate vs. remote, distance reckoned from the speaker (p. 149):

• Proximate ba ~ kumba 'this' (Table 19, p. 150)
• Remote hamba 'that' (Table 20, p. 150)

Demonstratives — unlike personal pronouns — DO take ergative and absolutive (kumba-ŋa 'this-ERG', hamba-iŋ 'that-ABS'), can be used adnominally (ba napmi 'this man'), and are now functioning as the de-facto 3rd-person pronouns (p. 143–144, p. 149). Their synthetic dual/plural: hambagʰacʰi, hambagʰa.

3.4 Interrogative pronouns

• Human: sa 'who' — inflects for all cases like a noun (Table 21, p. 150).
• General/non-human: hwiN 'which', hE 'what' — decline like nominals (Tables 22–23, p. 151).

4. The biactantial (polypersonal) verb

Chhatthare finite verbs mark person, number, case, reflexivity, tense, inclusivity and exclusivity by affixes. The verb agrees with both agent and object ("biactantial"). Each affix occupies a slot; many are portmanteaux. There are 3 prefix slots and 10 suffix slots (p. 240–241). Animacy hierarchy (1&2 > 3) drives the ordering, not raw grammatical role (p. 237).

4.1 Affix slots

Prefix slots (Table 51, p. 241):

Speech-act participants (1 & 2) share Pf1 (equal animacy); 3rd person fills Pf2; negation fills Pf3.

Suffix slots (Table 52, p. 241):

4.2 Person markers

Number suffixes: dual -cʰi, plural -i (S/O), plural agent -m, non-singular object -si. (p. 247–248)

Worked forms (transitive lɔm(ma) 'to beat', NPT) — note double-marking of 1st-person object in 3→1:

• Ø-lɔps-u-Ø 3sA-beat-3O-sO 'He beats him.' (p. 249)
• ka-lɔps-Ø-u-Ø 2-beat-sA-3O-sO 'You beat him.'
• lɔps-Ø-Ø-u-ŋ-Ø beat-NPT-sA-3O-1e-sO 'I beat him.'
• a-Ø-lɔm-Ø-Ø-ma 1-3sA-beat-NPT-sO-1e 'He beats me.' (1s object double-marked by prefix a- AND suffix -ma, p. 243)
• lɔm-na-Ø-Ø beat-1→2-sO-sA 'I beat you.' (the 1→2 portmanteau -na, p. 248)
• mu-lɔps-u-Ø 3pA-beat-3O-sO 'They beat him.' (p. 246)

4.3 Inclusive vs. exclusive

The original 1st-person pronoun aŋga split into inclusive a- and exclusive -ŋa (Fig. 14, p. 268). Inclusivity of subject/object/agent = a-; exclusivity = -ŋa. 1st-person singular is inherently exclusive (it always excludes the addressee); inclusive exists only in non-singular (you can't include the hearer in a singular). (p. 268–269)

• a-lok-Ø-cʰi 1i-run-NPT-dS 'we (incl.) run' vs. lok-Ø-cʰi-Na run-NPT-dS-1e 'we (excl.) run' (Table 56, p. 268)

4.4 Reflexive / reciprocal

One morpheme -cʰin (REFL), with allomorph -nE in non-singular (reciprocal), in Sf1 (p. 264–267):

• Ø-lɔm-cʰin-Ø 3sS-beat-REFL-NPT 'He beats himself.' (p. 264)
• Ø-lɔm-nE-Ø-cʰi 3-beat-RECIP-NPT-nsS 'They beat each other.' (p. 265)

Reflexive verbs are derived from transitive verbs and conjugate like intransitive/middle verbs; many are lexicalised (yuŋ-cʰiŋ 'sits', originally reflexive of yuks 'put') (p. 266).

4.5 Tense & aspect

• Tense: non-past ∅, past -a (in Sf2). Past is neutralised before the vocalic suffixes -u (3O) or -i (pl.) — past and non-past then look identical (p. 266, p. 275).
• Progressive: suffix -ro ~ -lo on the stem + auxiliary wa 'be' (te-ro wa 'he is going'; lɔps-u-ro wa 'he is beating him'). Negation is not permitted in progressive (p. 277–278).
• Perfect: sequential subordinator -aŋ + auxiliary wa 'be' (teg-a-aŋ wa 'he has gone'; past aux wama → pluperfect teg-a-aŋ wah-a 'he had gone') (p. 282).

4.6 Mood

Indicative is the bare finite verb. Other moods by particles/suffixes (p. 281, p. 342): imperative; hortative (drop the 1st-person -a suffix); irrealis particle mEn ('would have'); possibility laye ('perhaps'); topic/focus gɔ (= Nepali cahĩ); contrary-to-expectation ri.

4.7 Non-finite verbs

• Infinitive -ma ~ -na ~ -Na (e.g. pap-ma 'to speak'). (p. 318)
• Purposive -na ~ -ma ~ -Na; converb -?E (marked only under negation).
• Active participle ka- … -pa; passive participle -na-ba. (p. 318)

5. Negation

Negation is a discontinuous (circumfixal) morpheme, basic morph ma(n)- … -nEn(n) (NEG), surrounding the verb stem; the two parts cannot negate in isolation (p. 269). The first part appears as man- / ma- / m- / n- / N- (place-assimilating prefix in Pf3); the second part as -nEn / -n. (p. 270)

• ma- -nEn: Ø-ma-dE-Ø-nEn 3sS-NEG-go-NPT-NEG 'He does not go.' (p. 271)
• ma- -n (second part after a vowel): ma-si-Ø-ŋa-n NEG-die-NPT-1e-NEG 'I do not die.' (p. 271)
• Portmanteau man- … -ban/-pan = NEG + 1st person singular OR 1pe agent, past: man-lok-pan NEG-run-1eS/PT-NEG 'I did not run.'; man-lɔm-ban 'I/we did not beat him.' (p. 270)
• The negative prefix assimilates: m- before bilabials, n- before dentals, N- before velars (p. 247, p. 273).
• The negative suffix copies (-n reappears) after the non-singular object -si (p. 264, p. 273): ma-l ɔ ps-Ø-u-n-si-n 'they do not beat them.'
• Existential negation: wa 'be' → hop / hopma 'not be' (p. 327); identity copula negated by ekHan (p. 326).

6. Syntax & word order

• Verb-final, head-final, modifier-before-head — "almost a head-right language" (Abstract, p. 8; Ch. 13, p. 319).
• Complement precedes head throughout (p. 319):
  • Subordinate clause before main clause.
  • NP before VP; subject before object before verb (SOV).
  • Noun before postposition (paŋ-No 'in the house').
  • Main verb before vector/auxiliary (pat-u biy-u 'spoke for him'; kʰunE pHEn-ro wa 'he is coming').
  • Root before suffix; indirect object before direct object (p. 320–321).
• Modifier before head (p. 321–322): Adjective–Noun (kemba napmi 'tall man'), Adverb–Adjective (calik nuba 'very good'), Numeral–Noun (sumsi mendak 'three goats'), Genitive–Noun, Adverb–Verb.
• Head before modifier — the exceptions (p. 323): in kinship phrases, the adjective FOLLOWS the kinship head (a-bHu cukpa 'my youngest brother'), and a pronominal head is followed by its modifier. Kinship noun precedes proper noun (annE parbati 'my elder sister Parbati').
• Relatively free constituent order due to case marking, but movement is intra-clausal only — you cannot move an item out of its clause without changing meaning (p. 324). Limbu is a marker-right language: sentence particles, XP-markers and focus particles all come on the right (p. 325).
• Split ergativity (p. 337–338): a noun subject of a transitive verb is ergative-marked (lahaN-ŋa napmi sEr-u 'Lahang killed a man'), but a pronoun subject is NOT (a napmi sEr-u-ŋ 'I kill a man'). Demonstratives, however, DO take ergative (kumba-ŋa napmi sEr-u 'this killed a man'). First & second person intransitive subjects and transitive objects are marked alike on the verb; third person intransitive subject and transitive object are marked differently (so the ergative–absolutive pattern is not maintained throughout).
• Copular system — Limbu has ~10 semantically specialised copulas (p. 326): identity operator (suffixal, e.g. a yakthuŋba-ŋa 'I am a Limbu'), existential wa / negative existential hop, locational yuŋ, adhesive keŋ (hanging/elevated), attributive cuk, inchoative puŋ 'become', descriptive lo, vertical locational yep, horizontal locational nEn.
• Interrogatives: yes/no by suffix -i (allomorphs -mi/-ni/-Ni) without word-order change (kʰEnE tɔk ka-jɔ-i 'do you eat rice?'); doubtful yes/no by particle bi; constituent (wh-) questions by nominaliser -pa/-ba (p. 331, p. 339).
• Coordination: nominal 'and' particle nuŋ (appa nuŋ amma 'father and mother'); clausal 'and' by sequential -aŋ (kʰunE bo tah-a-aŋ teg-a 'he came here and went'). There are no native words for 'but' or 'or' (p. 340).

7. Dialectal differences vs. van Driem's Limbu (Phedappe/Panthare)

This is the load-bearing section. Tumbahang's own comparative chapter (Ch. 2, §11, p. 51–61) and his findings list (§11.4.1, p. 58–59) systematically compare Chhatthare against the three non-Chhatthare varieties that make up the "standard / van Driem" Limbu: Panthare (the recognised standard, Wiedert & Subba 1985), Phedappe (van Driem 1987), and Taplejungge/Tamarkhole/Mewakhole (Mikhailovsky 2003). Do not mix these inventories.

7.1 Phonology (p. 52, §11.1)

So Chhatthare has the richest consonant inventory (adds voiced /b/, /g/, aspirated /cʰ/, and a phonemic trill /r/) but the simplest vowel system (7 vowels, no length). For a translator: a Phedappe/Panthare lexical entry that relies on vowel length will not map to Chhatthare, and Chhatthare words with /g/, /cʰ/ or phonemic /r/ have no direct phonemic counterpart in Panthare/Taplejungge.

7.2 Lexicon (Swadesh sample, Table 5, p. 53–54)

Chhatthare diverges heavily in basic vocabulary. Selected contrasts:

The verb 'to beat' is lɔmma in Chhatthare but hipma in all three other dialects (p. 54) — the root used throughout van Driem-style paradigms is a different lexeme.

7.3 Demonstratives (Table 4, p. 53)

Entirely different set:

7.4 Verb agreement / affixes — the morphological differences

From the findings list (p. 58–59) and the morpheme chapters. These are the items most likely to trip up a translator mixing paradigms:

(Comparative paradigm tables: Table 6 non-past p. 54–55, Table 7 past p. 55–56, Table 8 negative non-past p. 56–58.)

7.5 Things that are SHARED across all dialects (p. 59, §11.4.2)

So you don't over-correct: these features are common to Chhatthare AND van Driem Limbu and need no dialect adjustment:

• Three-number system (singular / dual / plural) marked on the verb.
• Third-person non-singular object marked on the verb.
• Inclusive / exclusive distinction in 1st-person non-singular.
• Both agent and patient marked on the verb in 2→1, 2→3, 1→3.
• Third-person singular subject/agent is unmarked (zero) in all dialects.
• 2nd person form unchanged across 3→2, 2→3 (except 1→2, where the portmanteau differs — -na Chhatthare vs -ne elsewhere).
• -a is the past morpheme in all dialects.
• Negation is a discontinuous morpheme in all dialects.

7.6 Classification & intelligibility caveat

Tumbahang argues on linguistic grounds that Chhatthare is "very different" and "hardly intelligible to speakers of other dialects" (van Driem 1987:xxii calls Chhatthare "virtually wholly unintelligible to the Phedappe speakers"), but is kept as a dialect of Limbu for cultural/political reasons (p. 49–50, p. 60). Watters' proposal: Chhatthare and non-Chhatthare both descend from Proto-Limbu, which first split into Chhatthare vs. non-Chhatthare, the latter then diversifying into Phedappe/Panthare/Tamarkhole (p. 61).

Bottom line for the toolkit: treat Chhatthare and standard (Panthare/Phedappe = "van Driem") Limbu as effectively separate inventories at the level of phonemes, basic lexicon, demonstratives, and several core verb affixes. The broadcast/standard variety the Sighak (2024) article describes is Panthare/Panchthare, i.e. the van Driem-aligned standard — NOT Chhatthare.

Appendix: page-citation index (Tumbahang 2007, printed footer pages)

• Dialect inventory & sociolinguistics: 40–50
• Comparative phonology / words / verb paradigms (Tables 2–8): 52–58
• Findings — differences & similarities: 58–59
• Morphophonology (assimilation, voicing): 111–116
• Nominal morphology: gender 117–119; numerals/classifiers 119–121; number 121–124; pronominalisation 124–125
• Case: absolutive 126–128; ergative 129–130; genitive/collapse 134–135; vocative/locative 134–136; comitative/mediative/ablative 136–138; directive/allative/comparative 139–140; full case table (Table 14) 140–141
• Pronouns: personal 142–145; case of pronouns 145–146; possessive 146–148; demonstrative 149–150; interrogative 150–151
• Verb agreement: covert affixes / animacy 234–239; affix slots 240–241; person markers 242–253; reflexive/reciprocal 264–267; inclusive/exclusive 267–269; negation 269–274; tense/aspect/mood 275–284
• Syntax: constituent order 319–325; copular 326; existential/locational 327–330; interrogative 331; participant coding & case markers 332–337; split ergativity 337–338; nominalisation/coordination/particles 339–343