Authored by: Toshiki Osada

The Munda Languages

Print publication date:  March  2008
Online publication date:  April  2015

Print ISBN: 9780415328906
eBook ISBN: 9781315822433
Adobe ISBN: 9781317828860




Mundari is mainly spoken in the state of Jharkhand, which was recently set up by the Government of India on 15 November 2000, and in the adjoining states of Orissa and West Bengal in India. Muɳɖa means ‘village-headman’ in Mundari. But the language name muɳɖārī is given by the neighbouring peoples, the indigenous name is hoɽo jagar ‘human language’ or muɳɖa jagar ‘Munda language’.

 Add to shortlist  Cite


1  Introduction

Mundari is mainly spoken in the state of Jharkhand, which was recently set up by the Government of India on 15 November 2000, and in the adjoining states of Orissa and West Bengal in India. Muɳɖa means ‘village-headman’ in Mundari. But the language name muɳɖārī is given by the neighbouring peoples, the indigenous name is hoɽo jagar ‘human language’ or muɳɖa jagar ‘Munda language’.

As Gregory Anderson shows us in the Introduction of this book, Mundari belongs to the Kherwarian group of the North Munda branch. According to the Census of India 1991, the number of speakers of Mundari is 861,378. The same Census reports the number of speakers of Munda as 413,894. The names Munda and Mundari seem to confuse. It is likely that the census officer did not have perfect criteria for naming the languages in India. There is actually no difference between the Munda language and the Mundari language linguistically. Thus, the total number of the speakers of Mundari is likely to be more than one million. From a linguistic point of view, the designation Munda is used for the language family. Mundari, on the other hand, refers to an individual language, namely the language of Munda people.

As Hoffmann reported in the Encyclopaedia Mundarica, Vol.1, page (6), Mundari has four dialects; that is, Hasada from hasa-daʔ ‘(literally) land water (place name)’ in Mundari, Naguri from naguri (place name), Tamaria from tamaɽ-ia ‘language of Tamar (place name)’, and Kera from keraʔ (perfect ending, instead of keda in another dialect). Munda (1980:kha) has proposed the name Latar dialect (latar means ‘low’) instead of Tamaria. I do not adopt this term here because I have never heard latar jagar in Mundari.

The Hasada dialect is considered as the standard variety among Munda peoples. Hasada speakers are located on the eastern side of Ranchi–Chaibasa Road while Naguri speakers are situated on the western side. The Tamaria dialect is distributed in the Panchpargana area (Bundu, Tamar, Silli, Baranda and Rahe). Further, Kera is mainly spoken by the inhabitants of Ranchi city and the adjacent area, who ethnically belong to the Oraon tribe. According to Pinnow (1959:2), Ho should be considered as a dialect of Mundari from a linguistic point of view 1 We, however, regard the Ho language as a separate language on the basis of the ethnic identity of its speakers (see the chapter on Ho and the other Kherwarian languages in this book).

The study of Mundari started in the nineteenth century; for example, Haldar (1871), Whitley (1873), Nottrott (1882). These works are neither comprehensive, nor reliable from a linguistic point of view. For example, the glottal stops were not described in these works. Linguistically oriented grammars have been written by Hoffmann (1903), Cook (1965) (his data are collected not by him but by Hoffmann), Sinha (1975) (his descriptions contain a lot of self-contradiction and some data are not reliable), Munda (1980) (this is written in Hindi and contains reliable data but is not comprehensive), and Osada (1992) (the section on syntax is very poor); phonology by Gumperz with Biligiri (1957) and Sinha (1974) (the data are not reliable; the same as Sinha 1975); verbal morphology by Langendoen (1966, 1967) (his data are based on the Naguri dialect; he applied Mundari data to the standard theory by Chomsky but unsuccessfully as I show in section 3.2), Munda (1971) (this paper is focussed on aspect but incomplete); morpho-syntax by Osada (1999, 2007). A dictionary of Mundari has been compiled by Hoffmann (19301978), Bhaduri (1931), Prasad (1973, 1976) (in these dictionaries she missed a description of the glottal stops), and Mundu (1995). The most influential work is Hoffmann’s Mundari Grammar (=MG) and Encyclopaedia Mundarica (=EM). The descriptions in MG and EM differ in dialect. MG is mainly based on Naguri while EM mainly on Hasada. The data in EM are more comprehensive and reliable than those in MG. In addition to EM, Munda as a native speaker has given us reliable data. Thus I utilize the data from EM and Munda (1971, 1980).

2  Phonology

2.1  Phonemic inventory

Mundari has a five-vowel system as shown in Table 3.1. Vowel length and nasalization are not phonemic. It is, however, very important to make the distinction phonetically. As regards vowel length, an open and monosyllabic /CV/ is realized as two morae; for example, /ru/ ‘to beat a drum’ [ruː]. 2

Vowel nasalizations are found in the following circumstances:

  • /(C)VɳV(C)/ for example, /ceɳe/ ‘bird’ [ tʃɛ0303;ɳɛ0303;], /aɳeʔ/ ‘to pour out a liquid’ [ãɳẽʔ], etc.
  • /CNV/ (CN means a nasal consonant) for example, /mu/ ‘nose’ [mũ:], /nu/ ‘to drink’ [ nũ:]
  • /jV/ (optionally) for example, /ji/ ‘smell’ [dʒĩː], /ja/ ‘any’ [dʒãː] or [dʒaː], but /jo/ ‘fruit’ [dʒɔː].
  • /oe/, /oa/, /ua/ (optionally) for example, /koe/ ‘beggar’ [kɔ0303;ɛ0303;] /koasi/ ‘fog’ [kõãsi], / cua/ ‘to extract a liquid by fire’ [tʃũã] or [ tʃua ].

If expressives are considered, nasalization becomes (very marginally) contrastive. We note the following minimal pair in the expressives soe soe ‘sound of boiling water’ and sõẽ sõẽ ‘to sit in a slovenly fashion’.

Table 3.1:   Mundari Vowels

Vowel inventory












Mundari has 23 consonants (Table 3.2) including 10 stops: p, b, t, d, ʈ, ɖ, c, j, k, g; one sibilant: s; three liquids: r, ɽ, l; five nasals: m, n, ɳ, ñ, ŋ; two glides: w, y; and two glottals: h, ʔ given in the consonant inventory chart. All stop consonants except the two glottals h, ʔ appear in word-initial and word-medial position. In word-final position, the distinction between voiced stops p, t, k and voiceless stops b, d, g is neutralized and realized as checked consonants b, d and a glottal stop ʔ except in recent loanwords. The retroflex stops ʈ, ɖ in word-final position only occur in loan words from adjoining Indo-Aryan languages; for example, haaʈ ‘market’ from Hindi hāʈ The two stops c, j are phonetically realized as affricates [tʃ, dʒ] and occur in word-final position in recent loanwords; for example, ãc ‘flame’ from Hindi ā̃c, kagoj ‘paper’ from kāgaz/kāgoz in Persian through adjoining Indo-Aryan. The sibilant s appears in all positions but occurs in word-final position only for loanwords; for example, bes ‘good’ from bez in Persian (through adjacent Indo-Aryan). The two liquids r and l can occur in all positions while another liquid ɽ can occur only in word-medial position. Two nasals m and n can appear in all positions. But among nasal consonants, ɳ occurs only in inter-vocalic position and ŋ occurs only in word-final position. The palatal nasal ñ appears only in one word; that is, ‘I (1st person singlar)’. is realized as [aiɲ] or [aiŋ] when used independently but as [aɲ] before genitive suffix -aʔ, añaʔ ‘my’. The frequency of is very high, so I recognize /ñ/ as a distinct phoneme. The two glides w, y never occur in initial position.

As far as the differences among dialects are concerned, the Hasada, on which my description is based, and Tamaria dialects have no aspirated stops while the Naguri and Kera dialects have them. The same goes for intervocalic h. Another difference among the dialects is that ɽ in Hasada corresponds (Table 3.3) to ɖ in others.

Table 3.2:   Mundari Consonants

Consonant inventory







stop voiceless







stop voiced























Table 3.3:   Mundari Dialect Comparison

Differences among dialects





















2.2  Checked consonants

The most peculiar feature of consonants is the so-called checked consonant series. The stop phonemes /b/ and /d/ are realized as checked consonants in morpheme-final position.

The phonetic description of these checked consonants is as follows: first, the glottis is closed and the tongue or the lips simultaneously form an oral closure. The tongue or lip position is the same as that of the corresponding normal stops. Then the glottal closure is released, which is optionally followed by nasal release and voicing. Thus, [ʔm], [ʔn].

It is mentioned that nasal release after the glottal release is optional. In my observation, whether nasal release occurs or not is determined by the syllable structure of the morpheme. There is no nasal release in polysyllabic words but only in monosyllabic ones. For example,

  • /ub/ ‘hair’ [uʔm] but /udub/ ‘to tell’ [uduʔb̥];
  • /rid/ ‘to grind’ [riʔn] but /birid/ ‘to stand up’ [biriʔd̥].

Gumperz (1957) considers checked stops in word-final position as allophones of the voiceless stops /p/ and /t/. But I treat these checked consonants as allophones of voiced stops /b/ and /d/ as Hoffmann did. The following morphophonological change is very clear: /dub-a/ (/a/: IND) ‘will sit’ [duba] not [dupa] and /birid-a/ ‘will stand up’ [birida] not [birita].

Glottal stops are also regarded as checked consonants because these are allophones of /g/ and /y/. Glottal stops are followed by echo-vowel release in monosyllabic morpheme but never in polysyllabic morphemes. For instance, /rag/ ‘to call’ [raʔa], but /racag/ ‘to pull’ [racaʔ]; /poy/ ‘to rinse’ [pɔʔε], but /tukuy/ ‘to saw’ [tukuiʔ].

Recent loanwords in Mundari allow morpheme-final g. Thus the following minimal pair can be found:

  • [nεg] ‘religious feast’
  • [nεʔε] ‘here take it’ (interjection)

I phonemicize the first word as neg and the second one as neʔ. In addition to this final /g/, I describe final /y/ as or .

2.3  Syllable structure and phonotactics

A phonological word in Mundari can be syllabified by a simple rule due to the simplicity of consonant clusters. There are only three types of syllable boundary, that is, (a) between two successive vowels, (b) between a vowel and a following consonant, and (c) between two consonants which form a consonant cluster.

The middle vowels in trisyllabic words can optionally be deleted. It seems to me that a phonological word in Mundari has a tendency to keep two morae. We have already seen the examples of monosyllabic words in the form of CV, CVb, CVd, CVʔ, and CVy in sections 2.1. and 2.2. Hence we discuss here only monosyllabic words of the remaining forms which may be counted as having two morae. In most C 1 VC 2 words, C 2 is either a liquid or a nasal if it is not /b, d/ or /ʔ/. Denoting a mora boundary by %, we have sim [si%m] ‘chicken’, dul [du%l] ‘to pour’. We have exceptions in the following loanwords, for example, bes [bε%s] ‘good’, soj [sɔ%dʒ] ‘straight’. We consider that the consonants which occur as C 2 in all these cases constitute one mora.

We now list all the possible combination of C and V in the syllable structure of a phonological word, as seen in Table 3.4.

Table 3.4:   Mundari Word CV Structure



ub ‘hair’


jo ‘fruit’


joʔ ‘to sweep’



au ‘bring’


uku ‘hide’


udub ‘tell’


bai ‘make’


bulu ‘thigh’


tain ‘live’


buluŋ ‘salt’


eŋga ‘mother’


umbul ‘shade’


ɖoɳɖo ‘fool’


seŋgel ‘fire’



auri ‘not yet’


aosan ‘bring about an improvement’ (EM)


apia ‘three’


asadi ‘feel ennui’


ale-aʔ ‘our’ (plural and exclusive)


eperaŋ ‘quarrel’


aɽandi ‘marriage’


okoɳɖo ‘lift the head whilst lying down’ (EM)


baila ‘deaf’


saitan ‘evil’


balae ‘difficulty’


rasika ‘rejoice’


balae-n ‘worry’


tutukun ‘cold’


salaŋgi ‘tall’


hasaŋgar ‘live coal’


eŋga-o ‘mother also’


oɳɖoka ‘human sacrifice’


inku-aʔ ‘their’


eŋga-teʔ ‘his/her mother’


enbaɳʈa ‘rather’ (EM)


kumbuɽu ‘thief’


nimnaŋge ‘enough (for food)’


pampalad ‘butterfly’

2.4  Intonation/stress

Mundari is not a tone language, unlike some Mon-Khmer languages in the other branch of Austroasiatic. Among Munda languages, Korku, which forms the North Munda language group along with Kherwarian languages (including Mundari, Santali, and Ho), has a tonal contrast (Zide 1960, 1966). 3 Mundari does not have stress but pitch accent. Previous studies have mentioned only stress (Cook 1965: 100, Langendoen 1963: 14–15, N.K. Sinha 1975: 39).

Word accent in Mundari can be described in the following manner:

  • Word accent is not phonemic because it is predictable. In each phonological word an accent is assigned to only one syllable which is marked by a high pitch.
  • The accent patterns are as follows:
    • A monosyllabic phonological word is always accentuated; 4 for example, /ba/ ‘flower’ [baā] or [bāa], daʔ ‘water’ [ʔa] or [daʔā].
    • Accent is normally assigned to the second syllable in disyllabic words; for example, bulū ‘thigh’, bulūŋ salt’, seŋgēl ‘fire’.
    • Exceptions to this rule are the result of syllable weight. When the first syllable in a disyllabic word is heavier than the second syllable, the accent normally falls on the first syllable. When a syllable boundary is located between the nasal and homorganic stop sequences, the accent is assigned not to the first syllable but to the second syllable; for example, sīrma ‘sky, year’, gōmke ‘lord’ but dondō ‘to lift’, campā ‘a kind of flower’.
    • Further, in a trisyllabic word, accent is never assigned to the first syllable even if the first syllable is the heaviest. The second syllable in a trisyllabic word cannot be accentuated unless the element in the last syllable is a suffix. An unaccented vowel in the second syllable can optionally be deleted; for example, pampalād ‘butterfly’, aṛandī ‘marriage’, apī-a ‘three’, kumb(u)ɽū ‘thief’.
    • A quadrisyllable word is divided into two bisyllabic phonological words. Accent is allocated to each phonological word; for example, akādandā ‘to feel astonished’.

As far as sentence intonation is concerned, the major role of intonation is to provide contrast between several sentence types which may be marked by the distinctive use of patterns of pitch. Furthermore, intonation functions as a signal of grammatical structure such as the marking of sentence boundaries. Moreover, intonation conveys paralinguistic features, that is, information about the speaker‘s emotion, attitude, social background, etc. However we do not discuss these features in detail here. We mainly discuss the terminal contour.

Word accent also keeps its high level pitch at the sentence level. Besides high-level pitches, falling /↘/, rising /↗/ and falling-rising /↘ ↗/ contours play a major role in sentence intonation. As for level-pitches there are three; high1, middle2 and low3.

  • In an affirmative (declarative) sentence a falling intonation is allocated to the final syllable.
    • jom1-ke3-d-a2-ko
    • eat-COMPL-TR-IND-3PL
    • ‘They ate something.’
  • In an interrogative sentence a falling–rising intonation is assigned to the final syllable of the sentence.
    • jom1-ke3-d-a2-ko ↘ ↗
    • eat-COMPL-TR-IND-3PL
    • ‘Did they eat something?’
  • When sentence particles occur in sentence-final position, the sentence intonations are different from the pattern of (ii).
    • In sentences with the question marker ci, ci is always high level pitch.
      • jom1-ke3-d-a2-ko3 ci1
      • COMPL-TR-IND-3PL Q
      • ‘Did they eat something?’
    • In sentences with question marker ci + negation marker ka, ci is not marked, but ka has a marked rising intonation.
      • jom1-ke3-d-a2-ko3 ci2 ka
      • ‘I wonder whether they ate something or not.’
    • A sentence with the negation marker ka is characterized by a falling contour.
      • jom1-ke3-d-a2-ko2 ka
      • ‘Did they eat something?’
  • Negative and declarative sentences have the same pitch patterns as (i). Further, negative and interrogative sentences have the same pitch patterns as (ii).
    • ka1 =ko1 jom1-ke3-d-a2
    • ‘They didn’t eat something.’
    • ka1=ko1 jom1-ke3-d-a2 ↘ ↗
    • ‘Didn’t they eat something?’

2.5  Morphophonology

Mundari has a kind of phonological restriction, which divides vowels into two distinct subsets, that is, high vowels and mid-vowels which do not co-occur within a morpheme. This phenomenon is well-known as vowel harmony.

These distinct subsets can be described by their distinctive features (Table 3.5). The subsets 1 and 2 cannot co-occur within a morpheme while 3 can co-occur with 1 and 2. This rule can be extended beyond a morpheme boundary within a phonological word. For instance, in ‘that’ + ko (plural marker) = inku ‘those ones (animate)’; ni ‘this’ + ko (plural marker) = niku ‘these ones (animate)’.

Table 3.5:   Mundari Harmony Feature Sets

  1. [+high]
  2. [-high, -low]
  3. [+low]

Interestingly, not all phonological words can be generated by the rule. Hence it seems that only personal pronominal suffixes, including dual and plural suffixes, undergo the vowel harmony rule beyond a morpheme. Moreover, this rule can be adopted in a newly borrowed word. Thus suri < English sorry. In this case regressive assimilation has occurred, whereas progressive assimilation has occurred in the case of personal pronominal suffixes.

In the verbal morphology morphophonological changes frequently occur. The transitive marker d becomes ʔ with the first and third person singular object. At the same time the completion aspect marker ke becomes ki. Thus,

  1. biŋ coke=ʔ jom-ja-ʔ-i-a. snake frog=3SG:SUBJ eat-INGR-TR-3SG:OBJ-IND

    ‘The snake is eating the frog.’

  2. biŋ coke=ʔ jom-ki-ʔ-i-a. snake frog=3SG:SUBJ eat-COMPL-TR-3SG:OBJ-IND

    ‘The snake ate the frog.’

3  Morphology

3.1  Word class

For Mundari there has been a lengthy discussion of the difficulties in categorizing words into classes in terms of the traditional definitions of the parts of speech since Hoffmann (1903:xxi) declared the following:

Thus the same unchanged form is at the same time a Conjunction, an Adjective, a Pronoun, an Adverb, a Verb, and a Noun, or, to speak more precisely, it may become a Conjunction, an Adjective, and so on, but by itself alone it is none of them. It is simply a vague elastic word, capable of signifying, in a vague manner, several distinct concepts, that is of assuming a variety of functions.

This means that a prototypical lexical verb like jom ‘eat’ can be used as a noun without any morphological change, while a prototypical noun like buru ‘mountain’ can only be verbalized by attaching verbal endings. For example,

  1. buru=ko bai-ke-d-a. mountain=3PL:SUBJ make-COMPL-TR-IND

    ‘They made the mountain.’

  2. saan=ko buru-ke-d-a. firewood=3PL:SUBJ mountain-COMPL-TR-IND

    ‘They heaped up the firewood.’

  3. maɳɖi=ko jom-ke-d-a. food=3PL:SUBJ eat-COMPL-TR-IND

    ‘They ate the food.’

  4. jom=ko nam-ke-d-a. food=3PL:SUBJ get-COMPL-TR-IND

    ‘They got the food.’

In (3) buru is used as an argument, with the meaning ‘mountain’, while in (4) it is used as a two-place predicate with the meaning ‘heap up’. To illustrate the other direction of deployment, in (5) the word jom is used as a two-place predicate with the meaning ‘eat’, while in (6) it is used as an argument with the meaning ‘food’.

Nicholas Evans of Melbourne University and I published a paper titled ‘Mundari: the myth of language without word classes’ in Linguistic Typology in 2005. We introduced three criteria for establishing lack of word class distinctions, that is, equivalent combinatorics (members of both classes should have equivalent combinatorics), compositionality (the semantic results of using a member of one putative class in a constructional slot prototypically associated with the other putative class should be derivable through strict compositional principles) and bidirectionality (members of X should be deployable in the environments associated with Y, and members of Y should be deployable in the environments associated with X). Further, these three criteria should be exhaustive across the lexicon, that is, the same test should yield the same results for all lexemes in the putative class, not just for a few well-chosen ones. In our paper, we have seen that applying these three criteria decisively demonstrates that Mundari is not a monocategorial language. 5

Thus I describe nouns and verbs as follows: nouns can be morphologically marked for certain grammatical categories such as noun class (animate/inanimate) and number (singular/dual/plural). Verbs can be marked for grammatical features such as aspect and mood. Second, they can take affixes for voice and transitivity which are related to grammatical functions such as subject and object. The verb agrees with subject and object in person and number which are marked by a personal suffix.

In addition to noun and verb we set up the following word classes:

Pronoun, adjective, postposition, adverb, numeral, conjunction, particle, interjection, and expressive.

3.2  Nominal morphology

3.2.1  Noun class and number

Nouns are divided into animate and inanimate in terms of a system of concord between subject, object, and verb. Animate nouns refer to human beings and animals. In fact most grammatically animate nouns denote human beings and animals. Besides them the following nouns are considered animate:

  • Heavenly bodies: caɳɖuʔ ‘moon’, siŋgi ‘sun’, ipil ‘star’.

In relation to heavenly bodies the following verbs can be coded by animate marking: gama ‘to rain’, hoyo ‘to blow (the wind)’.

  1. gama-ja-d-a-eʔ


    ‘It is raining.’

  2. hoyo-le-d-a-eʔ


    ‘It had blown.’

  • Supernatural beings: boŋga ‘spirit’, siŋ boŋga ‘supreme God’.

As for gender distinction, some animate nouns can be divided into female and male nouns, marked morphologically by the endings i and a, respectively under the influence of adjoining Indo-Aryan varieties. For instance,









In order to express a distinction of sex in Mundari, a following modifier is preposed to the noun: eŋga (originally means ‘mother’) is used for female while saɳɖi (originally means ‘cock’) is used for male. Thus,













As far as kinship terminology is concerned, koɽa may be used for male and kuɽi for female. For instance,



‘his/her son’









boko-ñ koɽa


‘my younger brother’

younger sister/brother-my




‘my younger sister’

younger sister/brother-my


The number marking system for nouns in Mundari has three tiers, that is, singular-dual- plural. Singular is unmarked, and the dual and plural markers are kin and ko, respectively. Count nouns are marked for number irrespective of their animacy.

hon ‘a child’

hon-kin ‘two children’

hon-ko ‘children’

ipil ‘a star’

ipil-kin ‘two stars’

ipil-ko ‘stars’

kitab ‘a book’

kitab-kin ‘two books’

kitab-ko ‘books’

lijaʔ ‘a piece of cloth’

lijaʔ-kin ‘two pieces of cloth’

lijaʔ-ko ‘pieces of cloth’

3.2.2  Case

Mundari NPs do not inflect for case: both the subject and object of a sentence are morphologically unmarked. The subject and object of a sentence are determined by word order. The unmarked word order is as follows: S + O + Verb.

Examples are given as in (9) and (10).

  1. pusi-kin seta-ko=kin hua-ke-d-ko-a. cat-DL dog-PL=3DL:SUBJ bite-COMPL-TR-3PL:OBJ-IND

    ‘The two cats bit the dogs.’

  2. seta-ko pusi-kin=ko hua-ke-d-kin-a. dog-PL cat-DL=3PL:SUBJ bite- COMPL-TR-3DL:OBJ-IND

    ‘The dogs bit the two cats.’

In addition to two arguments a postpositional phrase or adverb denoting location or time can be inserted in any position before verb. We can illustrate this in (11).

  1. setaʔ-re seta-ko maɳɖi=ko jom-ke-d-a. morning-LOC dog-PL food=3PL:SUBJ eat-COMPL-TR-IND

    ‘In the morning the dogs ate the food.’

Case relations in Mundari are mainly marked by postpositions. Thus, instrumental is expressed by the postposition te following a noun or pronoun. Comitative is expressed by postposing loʔ after a noun or pronoun. Benefactive is expressed by the postposition naŋgen following a noun or pronoun. There are several dialectal variants; nagen/natin/naten. Source is expressed by postposing ate in Hasada and Tamaria dialects or ete in Naguri and Kera dialects after a noun or pronoun.

The possessive is expressed by the suffixes -aʔ, -reaʔ, -raʔ, and -ren. The possessive suffix - denotes alienable possession by an animate noun, while -reaʔ/-raʔ, and -ren indicate alienable possession by an inanimate noun. The distinction between -reaʔ/ -raʔ, and -ren is made on account of the animacy of the head noun. We demonstrate it in Table 3.6 as follows:

Table 3.6:   Mundari Genitive











3.2.3  Pronouns

Personal pronouns exhibit a 3 (First, Second, and Third) × 3 (Singular, Dual, and Plural) system (see Table 3.7).

We have found abin as a variant of second person dual and akiŋ as a variant of third person dual. I used this variant akiŋ in my previous works as Munda (1971) did. The form akin, however, is more common. Thus I, henceforth, use akin for the third person dual.

The possessive pronoun is formed by adding the genitive suffix - to a pronoun. Further, the Mundari equivalent of the independent possessive in English such as ‘mine, yours’, etc. is expressed by postposing the genitive suffix - to the possessive pronoun. The independent possessive is found only in the singular system.

Table 3.7:   Mundari Pronouns

Full form

Short form







1 (inclusive)






1 (exclusive)



















Thus, the possessive pronoun and independent pronoun can be described as follows (Table 3.8):

Table 3.8:   Mundari Possessive Pronouns – I

Possessive pronoun

Independent pronoun





1 (inclusive)





1 (exclusive)













Furthermore, we have an old system of possessive pronouns (Table 3.9). This is performed by the reduced pronominal suffixes which follow the genitive marker -ta.

In colloquial Mundari, this system has been almost completely replaced by the construction, pronoun + genitive suffix -, though the language of poetry in Mundari still retains this system. For instance, we have disum-tabu (disum ‘country’) ‘our country’ in poetry, but abu-aʔ disum ‘our country’ in colloquial speech.

Table 3.9:   Mundari Possessive Pronouns – II

Possessive pronouns




1 (inclusive)




1 (exclusive)











3.2.4  Demonstratives

Demonstratives in Mundari make a 3 × 2 contrast set (Proximate: Intermediate: Remote x marked: unmarked) as follows (Table 3.10):

The variants ne-/ni-, e-/i-, and he-/hi- are defined by the vowel harmony rule. The demonstratives have a rich derivational system. I illustrate them in Table 3.11.

Apart from the demonstratives, we have interrogative and indefinite pronouns as word classes. These three have a similar word process.

Table 3.10:   Mundari Demonstratives – I




Demonstrative bases









Table 3.11:   Mundari Demonstratives – II




Demonstrative adjectives









Demonstrative pronouns (animate)




























Demonstrative pronouns (inanimate)









Definite demonstratives





‘this much’

‘that much’




‘this much’

‘that much’

(more than one’s expectation)







‘this much’

‘that much’






‘this much’

‘that much’

(more than one’s expectation)







‘this so much’

‘that so much’






‘this so much’

‘that so much’

(more than one’s expectation)

Demonstrative adverbials

Adverbs of place


nere, nete, neate

enre, ente, enate

henre, hente, henate

‘here, there, from here’

‘there, thither, from there’

‘yonder, to yonder, from yonder’


nare, nate, naate

enre, ente, enate

‘here, hither, from here’

‘there, thither, from there’

hanre, hante, hanate ‘yonder, to yonder, from yonder’

Adverbs of time




‘this time’

‘that time’




‘this time’

‘this time’

(against one’s expectation)

(against one’s expectation)

Adverbs of manner





‘like this’

‘like that’

‘like yonder’





‘like this’

‘like that’

‘like yonder’

(against one’s expectation)

There are four interrogative bases (Table 3.12):

Table 3.12:   Mundari Interrogatives



ca and ci




The interrogative ci-lika can be derived from the other interrogative base ci plus leka ‘like’. The interrogatives ca and ci may be related to the demonstrative bases a and i. These four interrogatives are derivational bases. These derivational formations are identical to the demonstrative ones. The interrogative bases oko and cilika can be used independently while ca and ci can act only as a derivational base. The interrogative oko functions as a modifier whereas cilika as an adverb of manner.

Mundari has three indefinite bases (Table 3.13):

Table 3.13:   Mundari Indefinites







The distinction between interrogative oko and indefinite oko is somewhat dependent on the context. They, however, have at least one distinguishing syntactic criterion. The indefinite oko can be followed by the topic marker do but the interrogative oko cannot. For instance,

  1. oko-e hijuʔ-aka-n-a. INTER-3SG come-CONT-ITR-IND

    ‘Who has come?’

  2. oko-e do hijuʔ-aka-n-a. INDEF-3SG TOP come-CONT-ITR-IND

    ‘Someone has come (but not all).’

The indefinites ja and jeta, which are nearly synonyms, have the same syntactic function but the indefinite jete implying emphatic is more frequently used with the negative. We should pay attention to the semantic distinction between a (more than one’s expectation) and i/e (less than one’s expectation) here. I think that the semantic feature ‘negative’ may be related to ‘less than one’s expectation’.

Demonstrative, Interrogative and Indefinite pronouns have a similar word formation process. We summarize it here in the following way:

  • Adjectivals DB(=Demonstrative bases), INTB(=Interrogative bases), INDB(=Indefinite bases) + -n. For example, ca-n ‘what a kind of’, ja-n ‘any kind of’.
  • Pronoun (Animate) DB, INTB, INDB + (-n-) + -e/- (for singular), -kin (for dual), -ko/-ku (for plural). For example, oko-kin ‘who (dual)’, ja-n-ku ‘any persons’.
  • Inanimate DB, INTB, INDB + -(n)- + -a. For example, ca-n-a ‘which things’, ja-n-a ‘any things’.
  • Possessive pronoun (Animate) DB, INTB, INDB + (-n-) + -e/- (for singular), -kin (for dual), -ko/-ku (for plural) + (Genitive). For example, ca-n-kin-aʔ ‘whose (dual)’, jeta-n-ku-aʔ ‘of any persons’.
  • Possessive pronoun (Inanimate) DB, INTB, INDB + (-n-) + -a- + -reaʔ/-raʔ (Genitive). For example, oko-a-reaʔ ‘of which thing’, ja-n-a-raʔ ‘of any things.’
  • Definites DB, INTB, INDB + -(i)- + -min/mun- + -(aŋ/uŋ). For example, ci-min-aŋ ‘how much’.
  • Emphatic definites DB, INTB, INDB + -(i)- + mpuruŋ/mpiruŋ/mpiraŋ/mpinaŋ/mpinuŋ/mpunuŋ). For example, ci-mpuruŋ ‘how much exactly’.
  • Adverbs of time DB, INTB, INDB + -(i)- + -mtaŋ/mtuŋ. For example, ci-mtaŋ ‘when.’
  • Adverbs of place DB, INTB, INDB + -(n)- + (saʔ, taʔ) + re, te, ate. For example, oko-saʔ-te ‘to which side’.
  • Adverbs of manner DB, INTB, INDB + -(i)- + -leka/ka. For example, ja-leka ‘any ways’.
  • DB = Demonstrative bases Proximate Intermediate Remote unmarked ne/i e/i he/i marked na a ha
  • INTB = Interrogative bases ca-/ci- ‘what’ oko- ‘which’
  • INDB = Indefinite bases oko- ‘some’ ja- ‘any’ jeta- ‘any’
  • Semantic features i/e unmarked or less than expected a more than expected

3.2.5  Numerals

Table 3.14 presents Cardinal numerals.

As we have seen below, Mundari has a vigesimal counting system. According to Norman Zide (1978:1), ‘presumably Proto-Austroasiatic as well as old Indo-Aryan and Dravidian (old and modern) lacked vigesimal counting systems, but both Munda and modern Indo-Aryan use them. Whether the Indo-Aryan vigesimal systems “come from Munda” – as has been claimed – is questionable’.

The following short forms are used for the modifier of a head noun:

mid/mod ‘one’, bar ‘two’, api ‘three’, upun ‘four’, moɳe ‘five’, turui ‘six’, ee ‘seven’, iral ‘eight’, are ‘nine’, gel ‘ten’.

The counting forms consist of the addition of -ia/ea in postconsonantal position or a in post-vocalic position to the short forms, as is shown below.

Table 3.14:   Mundari Numerals





















gel miad/moyod


mod/mid hisi


mod/mid hisi miad/moyod


bar hisi


api hisi


moɳe hisi or mod/mid sau

‘5×20=100 or 1×100=100’

Distributive numerals are expressed by reduplication of the cardinal numerals. Distributive forms are a partial reduplication of cardinal forms for the numeral forms for ‘one’ to ‘six’ and ‘ten’, while complete reduplication is required for the numeral forms ‘seven’, ‘eight’, and ‘nine’. These coincide with the distributive form of Santali numerals.

mi-miyad ‘one each’

tu-turi-a ‘six each’

ba-bar-ia ‘two each’

ee-a ee-a ‘seven each’

ap-api-a ‘three each’

iral-ia iral-ia ‘eight each’

up-upun-ia ‘four each’

are-a are-a ‘nine each’

mo-moɳe-a ‘five each’

ge-gel-ea ‘ten each’

Ordinal numeral forms are as follows:

  • sida ‘first’
  • eʈaʔ ‘second’

The following variant forms are notable:

  • mi/mo ‘one’ in mi-sa/mo-sa ‘once’ (cf. bar-sa ‘twice’, api-sa ‘three times’, etc.)
  • mu ‘one’ in mu-siŋ ‘one day’ (cf. bar-siŋ; ‘two days’, api-ma ‘three days’, upun-ma ‘four days’, etc.)

As Emeneau (1956/1980: 115) has pointed out, numeral classifiers are an Indian areal feature. Mundari uses hoɽo ‘person’, oɽaʔ ‘house’, booʔ ‘head’ as classifiers. Thus,




‘three children’




The word jan/jon (from Indo-Aryan) is also currently used in Mundari. However, jan/jon always co-occurs with Indo-Aryan numerals. For example,





Numeral Classifier


‘three children’

3.2.6  Postpositions

Postpositions can be placed in a postnominal position and can form a postpositional phrase which may be used as a complement standing in a functional relationship with the verb.

The main postpositions can be divided in the following way:

  • re ‘in’ te ‘to, by’ ate/ete ‘from’
  • saʔ ‘on the side’ taʔ ‘vicinity’ loʔ ‘with’
  • ko ‘approximate’

The postposition ko does not appear independently but with (i) or (ii) and following (i) and (ii). Compound postpositions can be formed in the following order:

(ii) + (iii) + (i)

The process is detailed as follows:

saʔ ‘on the side’

taʔ ‘vicinity’

loʔ ‘with’

re ‘in’

saʔ-re ‘in the side’

taʔ-re ‘in the place’

loʔ-re ‘with’

te ‘to’

saʔ-te ‘to the side’

taʔ-te ‘to the place’

te ‘by’

saʔ-te ‘by the side’

taʔ-te ‘by the place’

loʔ-te ‘along’

ate ‘from’

sag-ate ‘from the side’

tag-ate ‘from the place’

ko-re ‘near in’



loʔ-ko-re ‘with’

‘near the side’

‘near the place’

‘shortly before’

ko-te ‘near to’



loʔ-ko-te ‘along’

‘near to the side’

‘near to the side’

‘shortly after’

Other postpositions will be illustrated below.

  • naŋgen ‘for’.

The semantic functions of this postposition are benefactive and purpose.

  • jaked, habiʔ/hamiʔ ‘until, up to’.

This may refer to location as in the following example:

  1. Ranci-jaked (habiʔ/hamiʔ)=ko sen-ke-n-a. Ranchi-up to=3SG:SUBJ go-COMPL-ITR-IND

    ‘They went up to Ranchi.’

It may also refer to time as in the following instance:

  1. sombar-jaked (habiʔ/hamiʔ) Ranci-re-ko tai-n-a. Monday-until Ranchi-LOC=3PL:SUBJ stay-ITR-IND

    ‘They will stay at Ranchi until Monday’

Moreover, it may refer to quantity as follows:

  1. ne keɽa api-hisi-jaked(habiʔ/hamiʔ)=e gonoŋ-oʔ-a. this buffalo three-twenty-up to=3SG:SUBJ cost-PASS-IND

    ‘This buffalo will cost up to 60 rupees.’

3.2.7  Derivation

We have already discussed the lexical semantic ambiguity involving the distinction between nouns and verbs in section 3.0. This distinction should be maintained because of the fact that a noun can be derived from a verb by a morphological process, namely, infixation:



dub ‘to sit’

du<nu>b ‘a meeting’

rakab ‘to rise’

ra<na>kab ‘a slope’

ol ‘to write’

o<no>l ‘the writing’

eʈeʔ ‘to begin’

e<ne>ʈeʔ ‘an origin’

tebaʔ ‘to arrive’

te<ne>baʔ ‘arrival’

tukuiʔ ‘to sew’

tu<nu>kuiʔ ‘the sewing’

tagoeʔ ‘to chew’

ta<na>goeʔ ‘the molar teeth’

The other nominalizing affixation is the possessive suffix - following a postpositional phrase which consists of a verb and the instrumental postposition -te. Thus,

ol ‘to write’

ol-te-aʔ ‘an instrument for writing: pen, pencil, etc.’

jom ‘to eat’

jom-te-aʔ ‘an instrument for eating: spoon, chopstick, etc.’

dub ‘to sit’

dub-te-aʔ ‘an instrument for sitting: chair, stool, etc.’

3.2.8  Adjective

The distinction between verbs and adjectives is problematic. It seems to me that Mundari is a typical adjectival-verb language; that is, ‘the usual verbal equivalent of a predicate adjective is a predicate verb in a non-relative construction while the usual verbal equivalent of a modifying adjective is a verb in a relative construction’ (Schachter 1985: 18–19). Thus maraŋ ‘big, great’ when used predicatively can be marked for aspect, mood, voice and (in)transitivity like a predicate verb. It might be said that one word class covers two semantically different classes, that is, adjectives and verbs. For example,

  1. en maraŋ hoɽo that great person

    that great person’ or

    ‘that person who will be great’

    en jom hoɽo that eat person

    ‘that person who will eat’

  2. en hoɽo maraŋ-a. that person great-IND

    ‘That person will be great.’

    en hoɽo jom-a. that person eat-IND

    ‘That person will eat.’

In my book I wrote ‘the definition of the adjective is rather notional’ (Osada 1992: 123). I have, however, introduced one morphological criterion: infixation possibilities distinguish verbs and adjectives. There is an infix <pV> as a reciprocal marker which is illustrated in section 3.2.7. A number of adjectives take a formally identical marker but they do not acquire the reciprocal meaning ‘each other’, which is but natural as these adjectives are one-place words. Instead, they acquire the intensive meaning ‘very’. This is the sole test justifying the setting up of adjectives as a distinct word class. It is interesting to note that a head noun modified by an adjective with the infix <pV> takes the plural marker though it may be either singular or plural in meaning:

  1. en maraŋ hoɽo that great person ‘that great person’

  2. en ma<pa>raŋ hoɽo-ko that great<INTENS> person-PL

    • ‘that very great person’
    • ‘those very great persons’

At least seven adjectives denoting size, shape, and the like take the intensifying infix, because of semantic limitations on intensification. We contain seven adjectives in Mundari in terms of this derivation:

  1. maraŋ ‘big, great’ → ma<pa>raŋ ‘very big, great’ huɽiŋ ‘small’ → hu<pu>ɽrŋ ‘very small’ jiliŋ ‘long’ → ji<pi>liŋ ‘very long’ salaŋgi ‘tall’ → sa<pa>laŋgi ‘very tall’ ɖiŋgae ‘short’ → ɖi<pi>ŋgae ‘very short’ cakar ‘wide’ → ca<pa>kar ‘very wide’ moʈo ‘fat’ → mo<po>ʈo ‘very fat’

It is noteworthy that some pronouns: demonstrative (e.g. naminuŋ, etc. ‘this much more than one expects’), interrogative (cimunuŋ, etc. ‘how much’), and indefinite (jaimunuŋ, etc. ‘to any extent, whatever be’), when modifying an intensive adjective also acquire the infix (by way of a kind of ‘pleonastic agreement’):

  1. naminuŋ maraŋ hoɽo-ko ka=ñ lel-aka-d-ko-a. this much big person-PL NEG=1SG:SUBJ see-CONT-TR-3PL:OBJ-IND

    ‘I have never seen such big person(s).’

  2. nam<p>inuŋ ma<pa>raŋ hoɽo-ko ka=ñ this much<INTENS> big<INTENS> person-PL NEG=1SG:SUBJ



    ‘I have never seen all so big (but more than one’s expectation) person(s).’

  3. cim<p>unuŋ hu<pu>ɽiŋ tai-ke-n-a. how much<INTENS> small<INTENS> remain-COMPL-ITR-IND

    ‘How small was it?’

  4. jaim<p>unuŋ ji<pi>liŋ-re-o ka=ñ suku-a. whatever<INTENS> long<INTENS>LOC-also NEG=1SG:SUBJ like-IND

    ‘Anything that is too long I don’t like.’

3.2.9  Adverbials

Adverbs can function independently as verbal complements. Expressions for adverbs of location can be made by constructing postpositional phrases. The number of adverbs is rather small.

Now I illustrate the following adverbs of time:






‘the day after tomorrow’


‘some days ago’






‘next year’


‘two years later’

In addition to the adverbs given above, adverbs of time can be expressed by a postpositional phrase. For instance, setaʔ-re ‘in the morning’, sombar-ate ‘from Monday’, etwar-jaked ‘until Sunday’.

Adverbs of location must always be expressed by a postpositional phrase. For example, Ranci-re ‘in Ranchi’, oɽaʔ-te ‘to the house’, hatu-ate ‘from the village’. Several local semantic functions are expressed mainly by postpositional phrases.

The postposition te (instrumental) can be used for adverbs of manner; for example, rasika-te ‘joyfully’, mani-te ‘slowly’, eskar-te ‘alone’.

3.3  Verbal morphology

Langendoen (1967) tried to describe Mundari verb conjugation based on Chomsky’s standard theory. He confessed the following in a straightforward manner.

The reader who is convinced of the efficacy of morpheme order charts for displaying the facts of a complex morphological system are advised to attempt to formulate such a chart for the Mundari data presented in this paper. I am reasonably convinced that no such formulation will be a match for the generative-transformational statement given here (inadequate as it is at various points) for displaying the intricate interconnections among the various patterns found in the Mundari verb conjugation. And really it must be admitted that the morphology of the Mundari verbal form is not nearly as complex as that of many languages. (Langendoen 1967:57)

I aim to present in this chapter not a theory-oriented but a data-oriented description. The basic verbal structure in Mundari may be described in terms of an order element formula as given in Table 3.15.

A verbal base is formed by affixing to a verbal stem. Verbal bases can be simple or complex; complex bases are formed by reduplication or serializing of the verbal stem. Verbal stems may be either transitive, or intransitive, or labile (i.e. transitive-intransitive, like the English ‘break’). Intransitive verbs are few in number (here belong, inuŋ ‘to play’, ‘to dawn’, ɖoɳɖo ‘to be foolish’, and the like). The intransitive or transitive use of labile verbs is distinguished by means of intransitive and transitive suffixes, -n and -d, respectively.

Table 3.15:   Mundari Verb Template

Verb base + (Aspect marker)

+ (n)

+ a (=Subj)

+ (d/ʔ) (+ Obj)



  a n: intransitive marker, d: transitive marker, with variant ʔ.

  b the transitive marker and the intransitive marker only appear when an aspect marker is present.

  c the suffix -a (indicative marker) is used to indicate the main verb of the clause excepting certain imperative forms.

3.3.1  Subject

The subject and object agreement element can be marked only when the subject NP and object NP are classified as animate nouns. Table 3.16 lists personal pronominal suffixes which are used for subject–object agreement:

Table 3.16:   Mundari Agreement Markers




















The same forms are used for subject, object and indirect object, but occupy different slots. The subject agreement element is attached either to the end of the verb or as a clitic to the preverbal NP, which may be not only the subject but also a non-subject. For example,

  1. hon-ko=ko dub-aka-n-a. child-PL=3PL:SUBJ sit-CONT-ITR-IND

    ‘The children have sat.’

  2. maɳɖi=ñ jom-ta-n-a. food=-1SG:SUBJ eat-PROG-TR-IND

    ‘I am eating the food.’

  3. gapa=ko senog-a. tomorrow=3PL:SUBJ go-IND

    ‘They will go tomorrow.’

Munda people belonging to the younger generation tend to place the subject agreement element at the end of the verb.

As I have pointed out earlier, the subject agreement element can be marked only when the subject NPs are classified as animate nouns. In addition to this principle, transitive subject NPs are basically animate nouns. Thus, the following sentence is not grammatical because the transitive subject is inanimate.

  1. *maɳɖi hon-ko bisi-ja-d-ko-a. food child-PL poison-INGR-TR-3PL:OBJ-IND

    ‘The food has poisoned the children.’

Instead, it is placed in instrumental function by adding the instrumental postposition -te, in a type of passive construction:

  1. maɳɖi-te hon-ko=ko bisi-ja-n-a. food-by child-PL =3PL:SUBJ poison-INGR-ITR-IND

    ‘The children have been poisoned by the food.’

Some inanimate nouns denoting natural objects, which are capable of automatic locomotion, can occupy the transitive subject slot. In that case, the subject agreement element gets marked on the verb like an animate transitive subject NP. This is called animatization. For example,

  1. gaɽa buru=iʔ bai-ke-d-a. river mountain=3SG:SUBJ make-COMPL-TR-IND

    ‘The river made the mountain (by carrying the sands).’

  2. gaɽa hon-ko=eʔ idi-ke-d-ko-a. river child-PL=3SG:SUBJ take-COMPL-TR-3PL:OBJ-IND

    ‘The river took away the children.’

Among inanimate nouns, only natural forces, for example, hoyo ‘wind’, gama ‘rain’, etc. can be animatized.

It is very easy to identify animate transitive subject NPs subjects because an animate subject is always marked on the verb. Apart from this principle, subjecthood and objecthood can be defined by a syntactic test, which I discuss in section 3.2.2.

3.3.2  Object types

As shown earlier, the same pronominal suffix is used for subject and object agreement. But the object agreement element occupies the slot just before the indicative marker -a in the indicative sentence or the slot just before the second person pronominal suffix in the imperative sentence or the slot just before the optative marker -ka-. For instance,

  1. Soma hon-ko=eʔ lel-ko-a. Soma child-PL=3SG:SUBJ see-3PL:OBJ-IND

    ‘Soma will see (take care) the children.’

  2. Soma hon-ko lel-ko-me. Soma child-PL see-3PL:OBJ-2SG

    ‘Soma, please see the children.’

  3. Soma hon-ko lel-ko-ka-eʔ. Soma child-PL see-3PL:OBJ-OPT-3SG:SUBJ

    ‘May Soma see the children.’

The above sentences are unmarked for aspect marker and transitive/intransitive marker. We can differentiate transitive sentences from intransitive sentence by the transitive/intransitive marker, as in the following:

  1. Soma hon-ko=eʔ dub-ke-d-ko-a. Soma child-PL=3SG:SUBJ sit-COMPL-TR-3PL:OBJ-IND

    ‘Soma made the children sit.’

  2. hon-ko ote-re=ko dub-ke-n-a. child-PL ground-LOC=3PL:SUBJ sit-COMPL-ITR-IND

    ‘The children sat on the ground.’

  3. pulis-ko kumbuɽu-kin=ko sab-ja-d-kin-a. police-PL thief-DL =3PL: SUBJ catch-INGR-TR-3DL:OBJ-IND

    ‘The policemen have caught the two thieves.’

  4. kumbuɽu-kin hola=kin sab-ja-n-a. thief-DL yesterday=3DL:SUBJ catch-INGR-ITR-IND

    ‘Two thieves have been caught yesterday.’

Note that the verb dub ‘to sit’ acquires causativity in the transitive sentence, whereas the verb sab ‘to catch’ acquires a passive sense in the intransitive sentence.

In ditransitive sentences, the benefactive marker -a is used in the following:

  1. am seta-ko=ñ om-a-m-ta-n-a. 2SG dog-PL =1SG:SUBJ give-BEN-2SG-PROG-ITR-IND

    ‘I am giving the dogs to you.’

  2. am seta-ko=ñ om-ke-d-ko-a. 2SG dog-PL =1SG:SUBJ give-COMPL-TR-3PL:OBJ-IND

    ‘I gave the dogs to you.’

It is very interesting that only one object, for example, am ‘you’ or seta-ko ‘dogs’ can be cross-referenced. Thus the sentence (42a) is ungrammatical. Further, if you want to encode the beneficiary on the verb in the completive sentence the completion aspect marker ke should change to the ‘cislocative’ or suspended aspect marker a as in (42b). Thus,

    • *am seta-ko=ñ om-a-m-ke-d-ko-a. 2SG dog-PL=1SG:SUBJ give-BEN-2SG-COMPL-TR-3PL:OBJ-IND ‘I gave the dogs to you.’
    • am seta-ko=ñ om-a-d-me-a. 2SG dog-P=1SG:SUBJ give-SUS-TR-2SG:OBJ-IND ‘I gave the dogs to you.’

As I mentioned in the previous section, inanimate NPs cannot normally occupy the transitive subject slot. But with some verbs, inanimate NPs can do this. The semantic range of these verbs is restricted to the following: 6

  • Sensory and mental experiences
  • Emotional experiences.
  • Physical and biological experiences.

I call these experiential verbs. This semantic range almost corresponds interestingly with dative subject predicates in Indo-Aryan (Klaiman 1986). There are two types of sentence in experiential verbal constructions in Mundari, as follows, in one, the experiencer is the subject (43, 45) while in the others it is the object (44, 46):

  1. balbal-te=ñ sowan-ta-n-a. sweat-by=1SG:SUBJ smell-PROG-ITR-IND

    ‘I am experiencing a smell of sweat.’

  2. ne baa maja sowan-ja-ʔ-ñ-a. this flower good smell-INGR-TR-1SG:OBJ-IND

    ‘This flower has made me experience a good smell.’

  3. sida samae susun-te=ko rasika-le-n-a. previous time dance-by=3PL:SUBJ be.joyful-ANT-ITR-IND

    ‘In the previous time they had experienced joy through dancing.’

  4. susun bese rasika-ke-d-ko-a. dance very be.joyful-COMPL-TR-3PL:OBJ-IND

    ‘The dance made them experience joy’

Recall that transitive subject NPs are basically animate nouns. If we assign the NP to grammatical relations by the morphological marking system, as for (44) and (46), the subjects are either nothing, or inanimate nouns and the objects are ñ ‘me’ in (44) and ko ‘them’ in (46). As I have discussed the grammatical relations above, these criteria will be kept throughout this chapter. I give here a new analysis of experiential verbal constructions.

The new analysis adopts the term ‘experiencer’ and ‘stimulus’. The experiencer denotes the human experiencer of sensory, mental, emotional, physical, and biological states expressed by the experiential verbs, whereas the stimulus is the source or cause of experience.

Further, I adopt the notion of ‘experiencer-subject’ from Croft (1991, 1993) for a typological analysis of mental verbs, where experiential verbs assign the experiencer to the subject position in (43) and (45). In (44) and (46), on the other hand, the experiencer is assigned to the object position. I, therefore, consider it as the experience–object construction. This analysis is useful, because it enables us to keep the criteria for subject–object assignment.

Now I give a new analysis of the experiential construction. In general I regard an experiential verb as an intransitive verb (see (43) and (45)). Then I consider the same experiential verb as a causativized form of an intransitive verb in the object-experiencer construction (see (44) and (46)). This analysis fits Croft (1991:215)’s cross-linguistic findings that ‘experiencer-object verbs are causative’.

The stimulus occurs with the instrumental postposition -te in the experience–subject construction shown in (43) and (45), and as a subject in the experience–object construction shown in (44) and (46). Are the NPs ne baa ‘this flower’ in (44) and susun ‘dance’ in (46) really subjects? And are the NPs ñ in (44) and ko in (46) really objects? We can make a syntactic test for subjecthood and objecthood by using relativization, as in the following:

  1. maja sowan-le-n-baa gosoʔ-ja-n-a. good smell-ANT-ITR-flower wither-INGR-ITR-IND

    ‘The good-smelling flower has withered.’

  2. ne baa maja sowan-le-d-(hoɽo)=ko senoʔ-ja-n-a. this flower good smell-ANT-TR-(people)=3PL:SUBJ go-INGR-ITR-IND

    ‘The people whom this flower made experience a good smell have gone.’

  3. bese rasika-ke-n-susun nimir caba-ja-n-a. very be joyful-COMPL-ITR-dance recently finish-INGR-ITR-IND

    ‘The very joyful dance has finished recently’

  4. susun bese rasika-ke-d-(hoɽo)=ko maraŋ-ja-n-a. dance very be joyful-COMPL-TR-(people)=3PL:SUBJ grow-INGR-ITR-IND

    ‘The people whom the dance made joyful have grown up.’

As the intransitive marker -n- appears in (47) and (49), the NPs baa and susun are subjects. On the other hand, the transitive marker -d- occurs in (48) and (50) the NPs (hoɽo)-ko are objects. My new analysis can be supported by this syntactic test of relativization.

I add an important note here. The stimulus is always an inanimate noun. In other words, transitive subject NPs in the object–experiencer construction are inanimate nouns. For instance, the following sentences are not acceptable:

  1. *ne kuɽ maja= sowan-ja-ʔ-ñ-a. this girl good=3SG:SUBJ smell-INGR-TR-1SG:OBJ-IND

    ‘This girl made me experience a good smell.’

  2. *susun kuɽi bese=ʔ rasika-ke-d-ko-a. dance girl very=3SG:SUBJ be.joyful-COMPL-TR-3PL:OBJ-IND

    ‘The dancing girl made them experience joy’

In order to say the equivalent sentences of (51) and (52) in Mundari, these should be replaced by (53) and (54), respectively.

  1. ne kuɽi-aʔ sowan maja sowan-ja-ʔ-ñ-a. this girl-GEN fragrance good smell-INGR-TR-1SG:OBJ-IND

    ‘This girl’s fragrance made me experience a good smell.’

  2. susun kuɽi=ko lel-ki-ʔ-i-ci bese dance girl=3SG:SUBJ see-COMPL-TR-3SG:OBJ-CONJ very



    ‘As they saw the dancing girl she made them experience a joy.’

I, therefore, rewrite the constraint for subject–object assignment here.

  1. Transitive subject NPs are animate nouns except for the object–experience construction.

As illustrated above, an experiential verb in Mundari is considered as an intransitive. That is to say, in the experience–subject constructions, only the experiencer assigned to intransitive subject occurs, and the stimulus occurs with the instrumental postposition -te. However only a few experiential verbs, for example, suku ‘to feel happy, to like’, kairao ‘to feel angry, to get an angry’, giuʔ ‘to feel ashamed, to shame’, can act as transitive verbs. In that case, the animate NPs can occupy the transitive object slot not as stimuli, but as beneficiaries. I show the general benefactive construction in (56) and the experience–subject and beneficiary–object construction in (57) and (58).

  1. maɳɖi am=iñ om-a-m-ta-n-a. food you=1SG:SUBJ give-BEN-2SG:OBJ-PROG-ITR-IND

    ‘I am giving you the food.’

  2. am=iñ giuʔ-a-m-ta-n-a. you=1SG:SUBJ shame-BEN-2SG:OBJ-PROG-ITR-IND

    ‘I am feeling shame at you.’

  3. ne kuɽi-kin=iñ suku-aka-n-a. this girl-DL=1SG:SUBJ like-CONT-ITR-IND

    ‘I like these girls.’

Examples (57) and (58) have two-place constructions, but include the intransitive marker -n. We, therefore, consider them semi-transitive; that is, they are two-place, but intransitive, they have a subject and an indirect object. Further, there are no experience–object constructions paired with (57) and (58) in Mundari. This is a great difference between Indo-Aryan and Mundari with respect to experiential constructions.

In sum, object types in Mundari are three; patient〓object in the unmarked transitive construction, experience–object in the experiential construction, and experience–indirect object in the benefactive construction.

3.3.3–3.3.4  Tense and aspect

The tense system is divided into future (unmarked) and non-future (marked).

The future tense implies the habitual aspect like used to in English. In this case the iterative verbal base is usually used. Further, the future tense also indicates universal truth. For example,

  1. uriʔ-jilu ka=le jo-jom-a. cattle-meat NEG=1PL.EX:SUBJ eat-ITER-IND

    ‘We (excl.) never eat beef.’

  2. setaʔ siɳgi=ʔ rakab-a. morning the Sun=3SG:SUBJ rise-IND

    ‘The sun rises in the morning.’

Present and past tenses are expressed by the aspect marker following the transitive or intransitive marker. Thus,

  • Present
    1. maɳɖi=ñ jom-ta-n-a. food=1SG:SUBJ eat-PROG-ITR-IND

      ‘I am eating the food.’

  • Past
    1. maɳɖi=ñ jom-ke-d-a. food=1SG:SUBJ eat-COMPL-TR-IND

      ‘I ate the food.’

The aspect markers are classified into perfective and imperfective sets. We will look at the perfective first. The following perfective aspect markers, for example, a, ke, le, ja are involved in the verb morphology of Mundari.

  • a.

This is not frequently used. Nobody except Munda takes this aspect marker into consideration. Munda regarded it as a ‘cislocative’ aspect which implies ‘an action which is completed and suspended for an indefinite period of time’ (Munda 1971:29). The ‘cislocative’ a can be followed by both the intransitive marker n and the transitive marker d. The following examples indicate the relationship between a-n and a-d:

  1. diku=ñ itu-a-d-ko-a. Hindi=1 SG:SUBJ teach-SUS-TR-3PL:OBJ-IND

    ‘I have taught Hindi to them.’

  2. diku=ñ itu-a-n-a. Hindi=1SG:SUBJ teach-SUS-ITR-IND

    ‘I have been taught Hindi; I have known Hindi.’

We have used the term suspended for the aspect marker a.

  • ke.

This aspect marker indicates the completion of an action without reference to any other action.

  1. maɳɖi=ñ jom-ke-a. food=1SG:SUBJ eat-COMPL-IND

    ‘I will finish eating the food (without waiting for anything).’

  2. Ranci-te=ñ sen-ke-n-a. Ranchi-to=1SG:SUBJ go-COMPL-ITR-IND

    ‘I went to Ranchi.’

  • le.

This aspect marker signifies the completion of an action in relation to some other action. We call it current relevance of anterior, or simply an anterior. We will compare le with ke as given below.

  1. duɽum-le-n-a-eʔ.


    ‘He/she had slept first (then has already got up).’

  2. duɽum-ke-n-a-eʔ.


    ‘S/he slept.’

  • (iv) ja.

This aspect marker indicates the completion of an action which is relevant to a current situation; in fact it refers to its inception. According to Comrie (1976:19), ‘the other perfect forms of the same verbs can in fact be used to indicate the beginning of a situation (ingressive meaning)’. In other words, this aspect marker is used for ingressive aspect.

  1. senoΔ-ja-n-a-ko.


    ‘They have started going; they have just gone.’

  2. maɽɖi=ko jom-ja-d-a. food=3PL:SUBJ eat-INGR-TR-IND

    ‘They have started eating the food; they are eating the food.’

Within the broad domain of imperfectivity ‘a distinction is made between the terms “progressive” and “continuous,” the former being a situation of the latter (progressiveness is the combination of continuousness with non-stativity)’ (Comrie 1976:12). The following imperfective aspect markers have been treated here:

  • ta.

This aspect marker is labelled ‘progressive’ based on the Comrie’s definition; that is, the combination of continuousness with non-stativity.

  • aka.

This aspect marker is labelled ‘continuous’, on the other hand, slightly different from the progressive aspect according to the definition as follows: ‘durative without the habitual’ (Comrie 1976:26).

The distinction between the aspect markers ta and aka is based on the contrast between a telic and an atelic situation. Telic refers to an event where the activity has a clear terminal point while atelic, where the event has no such natural end-point. For instance,

  1. dub-ta-n-a-ko.


    ‘They are in the process of sitting.’

(This action will have a terminal point when they sit down.)

  1. dub-aka-n-a-ko.


    ‘They are sitting.’

(They have already sat down. They can stand up or continue sitting. It does not matter.)

Hence we will consider the aspect marker aka as continuous in an atelic situation. Unlike English, the stative verb in Mundari has both progressive and continuous forms. Furthermore, transitive verbs with a continuous aspect correspond to the experiential perfect in English.

  1. Ranc do=ñ lel-aka-d-a. Ranchi top=1SG:SUBJ see- CONT-TR-IND

    ‘I have seen Ranchi; I have been to Ranchi.’

  2. Ranci-te=ko sen-aka-n-a. Ranchi-to-3PL:SUBJ go-CONT-ITR-IND

    ‘They have gone to Ranchi (and have not yet come back).’

3.3.5  Mood

There are three moods in Mundari: indicative (unmarked), imperative, and optative. Imperative mood is marked by deleting the indicative marker a in an indicative sentence. Negation of imperative, that is, prohibitive, is marked by preposing alo to a verbal base. Optative mood indicates the attitude of the speaker: it expresses wishes and is marked by ka which precedes the pronominal suffix. Although Munda (1971) called it subjunctive, we adopt the term optative as Hoffmann (l903), Cook (l965), and N.K. Sinha (1975) did. Negation of optative is marked by alo + ka, which are preposed to a verbal base. The modal intensifier koʔ can be added to imperative and optative sentences. It implies politeness.

The basic structures are as follows:

  • Imperative second -m/me singular VB (+ AM) (+ OBJ) + personal -ben dual suffix -pe plural
  • Optative VB (+ AM) (+ OBJ) + NEG ka personal suffix
  • Prohibitive for second person second -m alo person VB (+ AM) (+ OBJ) + -a suffix -pe
  • Prohibitive for first and second persons alo ka personal suffix VB (+ AM) (+ OBJ) + -a

The sentence examples are given below:

  1. maɳɖi jom-e-me. food eat-EPEN-2SG

    ‘Eat the food.’

  2. maɳɖi jom-e-ka-ko. food eat-EPEN-OPT-3PL

    ‘May they eat the food.’

  3. alo=pe jom-a. NEG=2PL eat-IND

    ‘Don’t eat.’

  4. alo-ka=ko jom-a. NEG-OPT=3PL eat-IND

    ‘They must not eat.’

We note that all aspect markers except the suspended aspect marker a can follow a verbal base in the imperative sentence. The semantic difference between aspect markers can be shown below.

  • Completion
    1. jom-ke-m.

      ‘Eat up; if you don’t eat it you won’t get a chance to eat later.’

  • Progressive
    1. jom-ta-m.

      ‘(You should) Eat (it); our life is not immortal. So you should eat it’.

  • Anterior
    1. jom-le-m.

      ‘Eat (it) first (and then).’

  • Ingressive
    1. jom-ja-m.

      ‘Eat (it) along (while on your way to doing something else).’

  • Continuous
    1. jom-aka-m.

      ‘Eat (it) (continuously).’

3.3.6  Orientation/directionality

Not investigated in this study.

3.3.7  Valence/voice

There are three means of decreasing valency; that is, reflexive, reciprocal, and passive.

Reflexive is expressed by the suffix -en after consonants or -n after vowels. For example,

  1. Soma=eₔ lel-en-ta-n-a. Soma=3SG:SUBJ see-RFLXV-PROG-ITR-IND

    ‘Soma is looking at himself’

The reflexive -en/-n is highly productive but the following verbs do not take the reflexive suffix: dub ‘to sit’, kami ‘to work’, gitiʔ ‘to lie down’, ajom ‘to feed’. The reflexive refers to ‘a verb where the subject and the object relate to the same entity’ (Crystal 2003). Causative verbs cannot be reflexivized; for example, ajom ‘to feed’. Interestingly, the class of intransitive verbs which can be causativized is also not reflexivized. Thus, dub ‘to sit’ is intransitive but can take the transitive marker to add causativity, as in the following:

  1. dub-aka-n-a-eʔ.


    ‘He has sat, that is, he is still sitting.’

  2. hon-ko=eʔ dub-aka-d-ko-a. child-PL=3SG:SUBJ sit-CONT-TR-3PL:OBJ-IND

    ‘S/he has caused the children to sit down.’

The reciprocal marker <pV> is monosemous and never attached to non-verbal stems. The reciprocal decreases verb valency. Thus the reciprocal verbal base takes only the intransitive marker -n even with ditransitive verbs. For instance,

  1. Soma seta hon-ko=eʔ om-ki-ʔ-i-a. Soma dog child-PL=3SG:SUBJ give-COMPL-TR-3SG:OBJ-IND

    ‘Soma gave the dog to the children.’

  2. seta-ko=le o<po>m-ta-n-a. dog-PL=1PL.EX:SUBJ give<RECIP>give-PROG-ITR-IND

    ‘We are giving the dogs to each other.’

Passive verbal bases can be formed by suffixing -oʔ to a verbal stem. The passive suffix can be attached to either transitive or intransitive verbs. The passive may imply the sense of possibility, that is, a ‘passive potential’, as is common in other Munda languages.

  1. (a) ayum ‘to hear’ ayum-oʔ ‘to be audible’ (b) lel ‘to see’ lel-oʔ ‘to be visible’ (c) duɽum ‘to sleep’ duɽum-oʔ ‘to feel sleepy’

Further, passivity in Mundari implies non-volitionality Unlike Mundari, passive in Hindi (Pandharipande 1978) and Bengali (Klainman 1986) makes crucial reference to the semantic notion of volitionality For instance, we may look at the following Mundari sentences:

  1. duɽum-oʔ-ta-n-a-eʔ.


    ‘S/he is feeling sleepy (by a non-volitional cause).’

  2. kug-oʔ-ta-n-a-eʔ.


    ‘He is beginning to cough.’

Thus, the following verbs cannot be passivized due to their implication of volitionality. For example, coʔ ‘to kiss’ *cog-oʔ, dulaɽa ‘to love’ *dulaɽa-oʔ.

The means of increasing valency are causative, conjugation change, and benefactive.

Causative is expressed by the unproductive prefix a-:

  1. (a) jom ‘to eat’ a-jom ‘to feed’ (b) nu ‘to drink’ -→ a-nu ‘to give to drink’

Conjugation change affects labile verbs. For convenience we shall consider the transitive use of labile stems as causativization as I have shown above in (85) and (86). Thus,

  1. (a) dub i. ‘to sit’ (with the intransitive marker -n) (b) ii. ‘to cause to sit’ (with the transitive marker -d)

  2. dub-aka-n-a-eʔ.


    ‘He has sat, that is, he is still sitting.’

  3. hon-ko=eʔ dub-aka-d-ko-a. child-PL=3SG:SUBJ sit-CONT-TR-3PL:OBJ-IND

    ‘S/he has caused the children to sit down.’

The benefactive suffix -a (always followed by the beneficiary agreement marker) indicates not only a beneficiary argument added to two-place transitive but also (optionally) the indirect object of ditransitives, and in this case both forms may coincide. Although the benefactive suffix increases valency when added to a two-place transitive, the benefactive verbal base takes the intransitive marker -n only. This may be the reason why reciprocals cannot be derived from the benefactive. Compare:

    • daru=m maʔ-ke-d-a. tree=2SG:SUBJ cut-COMPL-TR-IND ‘You cut the tree.’
    • daru=m mag-a-ñ-ke-n-a. tree=2SG:SUBJ cut-BEN-1SG-COMPL-ITR-IND ‘You cut the tree for me.’

3.3.8  (Non-)finiteness

Non-finite verb forms can be made by the deletion of the indicative marker a, and are used in the formation of relative clauses. They are of the following type:

Verbal Base (+ Aspect Marker + Transitive/Intransitive Marker).

The aspect markers and transitive/intransitive markers can be deleted in the future tense. Aspect markers are obligatorily followed by the transitive/intransitive marker -d/-n in relative clauses, although aspect markers without transitive/intransitive markers can be used in the finite verbal system. Thus,

  1. jom-hoɽo


    ‘the man who will eat’

  2. jom-ke-d-hoɽo


    ‘the man who ate it’

*jom-ke-hoɽo is not grammatical, although jom-ke-a=eʔ (eat-COMPL-IND=3SG:SUBJ) ‘He/she will have eaten the food’ is completely acceptable.

The non-finite form can be followed by a noun or pronoun in a relative clause and by a postposition in a subordinate clause. I discuss these clauses in sections 4.4 and 4.5.

3.3.9  Negation

The negative markers in Mundari are ka and alo.

ka is highly productive for lexical and sentence negation in indicative sentences. It is a morphologically bound form. alo is used for the negation of imperative or optative sentences.

Basically ka is ‘No’ in polar (yes–no) questions. It also functions as a sentence negation marker, and is then fixed in preverbal position, followed by the subject agreement element. If the subject is inanimate the negation marker ka is put in preverbal position as a free form. For instance,

  1. Ranci-te=m senoʔ-ta-n-a ci. Ranchi-to=2SG:SUBJ go-PROG-ITR-IND Q

    Are you going to Ranchi?’

  2. ka, ka=ñ senog-a. NEG NEG=1SG:SUBJ go-IND

    ‘No, I don’t go.’

  3. ne gaɽi Ranci-te ka senog-a. this car Ranchi-to NEG go-IND

    ‘This car will not go to Ranchi.’

As is shown in the section on Mood (section 3.2.5), alo functions as a prohibitive marker in imperative sentences, and indicates the negation of hope and desire in optative sentences.

In sentences with an auxiliary verb we use special forms for negation. These are banoʔ for inanimates, baŋgaiʔ for first person singular and third person singular and baŋ for animate other than first person singular and third person singular. I discuss these in detail in section 3.2.11.

3.3.10  Derivation

Verbal derivation in Mundari is a type of partial reduplication. It is not productive, and is only applied to closed monosyllabic words. Beside partial reduplication, we have full reduplication. This formation is productive. For example,


Partial reduplication

Full reduplication

sab ‘to catch’



jom ‘to eat’



tud ‘to pick’



goeʔ ‘to kill’



taŋgi ‘to wait’


rakab ‘to rise’


The partial reduplication may indicate either:

  • Repetition, implying an emphatic action.
  • Customary act, implying a universal fact.

For instance,

  1. alo=m kumbuɽu-a da-dal-a-ko. NEG=2SG thief-IND strike-ITER-IND-3PL:SUBJ

    ‘Don’t steal it. They strike you repeatedly’

  2. uriʔ-jilu ka=le jo-jom-a. cattle-meat NEG=1PL.EX:SUBJ eat-ITER-IND

    ‘We (EX.) never eat beef’

The following particial reduplication is lexicalized:

  1. tasad tu-tud-aka-n-a. grass pick up-ITER-CONT-ITR-IND

    ‘The grass is ready to pick up.’

  2. uli-joo jo-jom-aka-n-a. mango-fruit eat-ITER-CONT-ITR-IND

    ‘The mango fruit is ready to eat up.’

In addition to the above meaning, we have the meaning of action of limited duration, just in full reduplication.

  1. taŋgi-taŋgi-ke-n-a-le.


    ‘We (excl) waited for a little while.’

I describe full reduplication formation later, in the section on serial verb constructions (section 3.2.12).

3.3.11  Copula

The equivalent of the verb ‘to be’ has two forms in Mundari:

  • The existential copula menaʔ. This refers to the subject’s location in space.
  • The identity copula tan. This refers to the subject’s identity.

Both are irregular verbs. The animate subject of the existential copula menaʔ is encoded in the verb morphology while the inanimate subject is not encoded. Thus,

  1. Soma oɽcaʔ-re menaʔ-i-a. Soma house-LOC COP-3SG-IND

    ‘Soma is in the house.’

  2. Soma tan-iʔ. Soma COP-3SG:SUBJ

    ‘It is Soma.’

Both the copula menaʔ and tan are defective verbs, occurring only in the present. These forms merge into tai ‘to stay’ in the past and future (the past form tai-ke-n, the future form tai-n). For instance,

  1. Soma oɽaʔ-re=ʔ tai-ke-n-a. Soma house-LOC=3SG COP-COMPL-ITR-IND

    ‘Soma was in the house.’

  2. Soma=eʔ tai-ke-n-a. Soma=3SG:SUBJ COP-COMPL-ITR-IND

    ‘It was Soma’ or ‘Soma was there.’

  3. Soma oɽaʔ-re=ʔ tai-n-a. Soma house-LOC=3SG:SUBJ COP-ITR-IND

    ‘Soma will be in the house.’

  4. Soma=eʔ tai-n-a. Soma COP-ITR-IND

    ‘It is going to be Soma who will be there’ or

    ‘Soma will be there.’

Negation in copula sentences is slightly more complicated. The negative of menaʔ has three variants. For example,

  1. Soma oɽaʔ-re baŋgaiʔ-i-a Soma house-in COP NEG-3SG-IND

    ‘Soma is not in the house.’

  2. parkom oɽaʔ-re banoʔ-a. bed house-LOC COP NEG-IND

    ‘A bedstead is not in the house.’

  3. hon-ko oɽaʔ-re baŋ-ko-a. child-PL house-LOC COP NEG-3PL-IND

    ‘Children are not in the house.’

The negative of tan is formed by just adding to the negative marker ka before tan as in a regular verb.

  1. Soma ka tan-iʔ. Soma NEG COP-3SG

    ‘It is not Soma.’

The identity copula tan may be related to the progressive aspect marker ta. The distinction between the existential copula and the identity copula may be considered to be an Indian areal feature. 7

3.3.12  Auxiliary or serial verb constructions

Serial verb constructions have attracted the attention of linguists who are concerned with South Asian languages. Ever since Masica (1976) considered compound verbs with the ‘explicator verb’ as a typological areal feature, studies on compound verbs in South Asian context have been seriously attempted.

In Indo-Aryan languages, there are two types of verb–verb constructions; one is a verb with a conjunctive participle and another is a compound verb with an ‘explicator’ verb. For example, in Hindi a conjunctive participle is the so-called -kar construction whereas a compound verb is the combination of a main (polar) verb and an explicator (vector, operator, or intensifier) which indicates completion, passivity, permission, etc. of the action or process expressed by the main verb. Thus, we consider this type as an auxiliary verb construction.

In Mundari there are two types of the serial verb construction:

  • Main verb + Main verb. Its meanings are similar to the conjunctive participle construction in Hindi, that is, they are perfective and simultaneous.
  • Main verb + Explicator. Unlike other South Asian languages, most of the operators retain their original meaning.

I illustrate the following examples in (i) here:

  • Pair action

    1. hatu-re=liŋ jom-nu-ke-d-a. village-in=1DL.EX eat-drink-COMPL-TR-IND

      ‘We two (excl) ate and drank in the village; we two took dinner in the village.’

These are pair actions like senoʔ-hijuʔ ‘to keep company with’, isin-basaŋ ‘to cook and boil, that is, to make food’, etc.

  • Sequential action

    1. ne saan saɽima-cetaŋ-te=bu dondo-rakab-e-a. this firewood roof-over-to-1PL.INC lift-go up-it-IND

      ‘We will lift this firewood and take it up to the roof’

This class of serial verbs are goʔ-ader ‘to carry (something) on the shoulder and take it into the house’, dul-pereʔ ‘to pour and fill up’, etc.

  • Result

    1. ne hon=le asul-maraŋ-ki-ʔ-i-a. this child=1PL.EX:SUBJ feed-grow-COMPL-TR-3SG:OBJ-IND

      ‘We (excl) fed and raised this child up.’

Examples of this type of serial verb are: sen-nam ‘to go and meet (somebody)’, nir-laga ‘to run and get tired’, etc.

  • Cause

    1. añ-aʔ hon-kin banda-re=kin ɖumbuiʔ-goeʔ-ja-n-a. my child-DL pond-LOC=3DL:SUBJ. be drown-die-INGR-ITR-IND

      ‘My two children have died by drowning in a pond.’

Other examples are: maʔ-goeʔ ‘to kill with an axe’, haka-goeʔ ‘die by hanging’, etc.

  • Simultaneous action or events

    1. ne gaɽa poʈopoʈia-te=ko har-parom-ke-d-a. this river motorbike-by=3PL:SUBJ drive-cross-COMPL-TR-IND

      ‘They drove the motorbike and crossed the river.’

This class of serial verbs are: dub-hape ‘while sitting to keep a silence’, duraŋ-au ‘to come along singing’, etc.

The second type (ii) Main verb + Explicator is a common feature in South Asian languages. The verbs meaning ‘come’ and ‘go’ are common explicators in Indo-Aryan and Dravidian languages (Kachru and Pandharipande 1980:115). In Mundari the equivalents of these verbs are never used as explicators, and never appear as the second member of serial verbs either. U.N. Singh et al. (1986) listed the equivalents of the following eighteen vectors (=explicators) for their analysis for classifying polar verbs in selected South Asian languages.


Among them the following verbs can be considered the second member of serial verb constructions in Mundari.

English gloss

Mundari verbs

Meaning as a second verb of the series

Serial verbs



motion onward or away continuation

nir-idi ‘to run away’

kuli-idi ‘to go on asking a question’



to the last degree

rasika-goeʔ ‘to rejoice excessively’

landa-goeʔ ‘to be convulsed with laughter’



motion up

racaʔ-rakab ‘to pull up’

nir-rakab ‘to run up’



to exceed

laga-giɽi ‘to get tired excessively’

pereʔ-giɽi ‘to full to excess’



to send

raʔ-kul ‘to send to call’

kiriŋ-kul ‘to send to buy’



motion out of the place

apir-uɽuŋ ‘to fly out’

dul-uɽuŋ ‘to pour out’



motion from a given point towards the speaker (to and fro)

nir-au ‘to come running this way’

jom-au ‘to go for taking food, and then come back’

We will illustrate below the second verbs of the series whose meaning is slightly different from their basic meaning as main verbs, that is, the ones that have been (partially) grammaticalized.

Basic meaning

Mundari verbs

Semantic function

Serial/Auxiliary verbs




gitiʔ-jom ‘to lie down for one’s benefit’

nam-jom ‘to get for oneself’




om-pereʔ ‘to give completely’

leka-pereʔ ‘to count completely’




lel-bai ‘to look carefully/attentively’

ayum-bai ‘to listen carefully’

PLUCK (a fruit)


for a moment

ayum-god ‘to hear for a moment’

lel-god ‘to see for a moment’

PLUCK (a potherb)


to stop

jagar-sid ‘to stop talking’

sayad-sid ‘to stop breathing’

The following verbs, when used as second verbs of the series, can act not only as modal and aspectual auxiliaries but also as adverbials:

Basic meaning

Mundari verbs

Semantic function

Serial verbs




om-caba ‘to finish giving’

lel-caba ‘to finish seeing’




jom-eʈeʔ ‘to start eating’

ol-eʈeʔ ‘to start writing’




ol-daɽi ‘to be able to write’

lel-daɽi ‘to be able to see’




ol-rika ‘to cause to write’

bai-rika ‘to cause to make’



again, back

lel-ruɽa ‘to see again’

jom-ruɽia ‘to eat again’




duṛum-laaʔ ‘to sleep excessively’

jom-laaʔ ‘to eat excessively’




senoʔ-ayar ‘to go ahead’

nir-ayar ‘to run ahead’




jom-tayom ‘to eat later’

sen-tayom ‘to go later’

Every serial verb can be marked by affixation for mood on the basis of semantic and grammatical constraints operable on the second verb. In the Main verb + Main verb construction, reciprocal infixation of <pV> is applied to each of the main verbs; for example, dondo-rakab ‘to lift and go up’, do<po>ndo-ra<pa>kab ‘to lift and go up each other’.

Verbal intensifiers follow a verbal base. This construction is similar to a serial verb construction, but the second element is not a free form but rather a bound form. Unlike Indo-Aryan, the system of verbal intensifiers is very rich in Mundari. For instance,

  • V-baʔ/tab ‘V quickly’
    1. maɳɖi jom-baʔ-e-me. food eat-quickly-it-2SG

      ‘Eat the food quickly’

  • V-bapad/goroeʔ ‘V by all means’
    1. en kuɽi au-bapad-i-me. (EM) that girl bring-by all means-3SG-2SG

      ‘Marry the woman by all means.’

  • V-baɽa/balaŋ ‘V here and there’
    1. aɽandi-naŋgen=ko sen-baɽa-ta-n-a. marriage-for=3PL:SUBJ go-here and there-PROG-ITR-IND

      ‘They are going here and there for a marriage.’

  • V-coʈeʔ ‘almost V’
    1. kaji-coʈeʔ-ke-d-ci=ko senoʔ-ja-n-a say-almost-COMPL-TR-CONJ=3PL:SUBJ go-INGR-ITR-IND

      ‘They began saying and stopped in the middle, then have gone.’

  • V-garaŋ ‘intensified V’
    1. kakala-garaŋiʔ-me.


      ‘Shout at him/her loudly’

  • V-hantaɽa ‘engage in V’
    1. jom-hantaɽa-e-me.

      eat-engage in-it-2SG

      ‘Be engaged in eating it.’

  • V-kate/kuca ‘V repeatedly’
    1. maɳdi=ko jom-kuca-ke-d-a. food=3SG:SUBJ eat-repeatedly-COMPL-TR-IND-3PL

      ‘They ate it repeatedly’

  • V-noʔ ‘V a little (while)’
    1. saman goʔ-noʔ-le-m. luggage carry on the shoulder-a little-ANT-2SG

      ‘Carry it on the shoulder for a while.’

  • V-torsa ‘V along’
    1. hijuʔ-torsa-me.


      ‘Come back at once.’

  • V-tuka ‘V and return 8
    1. idi-tuka-ñ-ka-eʔ.


      ‘He/she may take me and return.’

  • V-utar ‘V entirely’
    1. Ranci-te=ko senoʔ-utar-ja-n-a. Ranchi-to=3PL:SUBJ go-entirely-INGR-ITR-IND-3SG

      ‘They went away to Ranchi for ever.’

3.4  Expressives

Mundari has a rich system of expressives. The term ‘expressive’ was suggested by Diffloth (1976:263–264) and adopted by Emeneau (1980:7) in the South Asian context in the following:

‘(E)xpressive’is the most inclusive term for a form class with semantic symbolism and distinct morphosyntactic properties; ‘ideophones’ are a subclass in which the symbolism is phonological; ‘onomaptoetics’ are ideophones in which the reference of the symbolism is acoustic (i.e. imitative of sounds). Since the ideophones may have reference not only to sounds, but to any other objects of sense, including internal feelings as well as external perceptions (sight, taste, smell, etc.), and since the Indo-Aryan/Dravidian items already examined have this very wide type of reference, the broadest term ‘expressive’ seems appropriate.

I have already written about Mundari expressives in my grammar (Osada 1992:140–144). However, I could not touch the syntactic and semantic properties of expressives. Thus I will discuss here (1) morphology (2) syntax, and (3) semantics of expressives.

3.4.1  Morphology of expressive

Expressives can be divided into the following types on the basis of their word formation pattern:

  • Full reduplication
  • Partial reduplication
  • Vowel mutation  Full reduplication

This type of expressive should be distinguished from verbal reduplication, which is clearly derived from the verbal base. It is a salient feature that a basic unit of the reduplicational element has no meaning. Thus,

Expressive form


cakob cakob

‘to eat noisily’

lugum lugum

‘to mumble (something)’

hayam hayam

‘to talk in whispers’

gusu gusu

‘an inactive character’

suyuŋ suyuŋ

‘lean and small (person)’

kase kase

‘to look askance at (a person)’

mondor mondor

‘a smell of rice beer’

mogo mogo

‘a smell of flowers’

kata kata

‘to roar with laugh by many people’

muguiʔ muguiʔ

‘smiling cheerful’  Partial reduplication

Partial reduplication can be formed by two elements. The second element is a partial reduplication of the first element. We can subcategorize this type according to the exact formal pattern. So far it has been impossible to find common meanings associated with each partial formal type:

  • CVX pVX Expressive form Meaning riti piti ‘very small leaves as those of tamarind’ risuɽi pisuɽi ‘the act of showing the teeth again and again’ rasa pasa ‘a continuous rustle of dry leaves, paper, or straw as produced by the gliding of a snake or the passage of a rat or other small animals’ laʈa paʈa ‘to make a stew thick, pasty’ laʈar paʈar ‘a mixture of truth and lies wherein one does not know what to believe’ ledeŋ pedeŋ ‘so fat that in walking he has difficulty’ leco peco ‘connoting several recurvations, or twisting of the mouth to the right and left’ loso poso ‘a loosely limbed body with soft or flabby muscles’
  • CVX bVX Expressive form Meaning kau bau ‘to do uncomfortably or uneasily’ kered bered ‘a quarrelling and fighting disposition’ cere bere ‘chattering and twittering of numerous birds’ cali bali ‘trickiness’ lada bada ‘the thuds of things soft, as mud, falling in succession’ ladi badi ‘to put things in a disorderly manner, more or less one over another’ sador bador ‘the act of letting bits fall while eating of strewing bits all around by pecking’ rada bada ‘onomatopoeia of hail, dry fruit or other hard and dry things falling all about in rapid succession, also of rather numerous drops of water falling all about’
  • CVX mVX Expressive form Meaning celloŋ meloŋ ‘naughty boy’ ceŋgol meŋgol ‘shamelessness’ jaka maka ‘shining with a flashy dress (sari with gold)’ jiki miki ‘shining with leather’ rigi migi ‘a cloth variegated with parallel lines or squashes of various colour’ keoŋ meoŋ ‘a feeling of loneliness and fear in the middle of the forest’ kaɳɖu maɳɖu ‘indigestion and pain in the pit of the stomach after eating or drinking something acid or sour or unripe’ seled meled ‘mixture of different kinds of grain, etc.’ gero mero ‘a shamed face or a crying face’
  • CVX kVX Expressive form Meaning ɖale kale ‘negligent (of taking care)’ haʈi kuʈi ‘to do the opposite action’
  • CVX gVX Expressive form Meaning rain gain ‘good or bad principles of conduct’ (EM) mane gane ‘want of punctuality in starting dilatoriness’ (EM)
  • CVX cVX Expressive form Meaning repo cepo ‘shrivelled’ dukur cukur ‘uneasiness of mind’
  • CVX jVX Expressive form Meaning reŋge jeŋge ‘the condition of getting bothered or being subjected to trouble or annoyance’ hauɽu jauɽu ‘desultory talk or conversation, passing from one subject to another without order or natural connection’ runu junu ‘to go or walk with difficulty due to a handicap’
  • CVX dVX Expressive form Meaning rawa dawa ‘opportunity to do someting reprehensible, because there is nobody to interfere’
  • CVX tVX Expressive form Meaning ribuiʔ tibuiʔ ‘the act of fat people, walking with the buttocks rubbing against each other’ roka toka ‘quickly’
  • CVX sVX Expressive form Meaning rahan sahan ‘the use of dress, furniture, plate and utensils by more civilized people’ boro soro ‘cowardice’
  • CVX rVX Expressive form Meaning tiri riri ‘the sound of a flute’

(i), (ii), and (iii) are very common.  Vowel mutation

This type of expressives are divided into six on the basis of vowel mutational patterns.

  • (C)aC[(C)a(C)] (C)uC[(C)u(C)] Expressive form Meaning ɖala ɖulu ‘a fat and short person’ lada ludu ‘a fat child’ ladar ludur ‘a wrinkled old person’ aŋgar uŋgur ‘to look around restlessly’ baʔ buʔ ‘holes here and there’ cabaʔ cubuʔ ‘the splashing sound made by repeated poking with a stick in water or mud’ ʈapaʔ ʈupuʔ ‘baby tries to walk’ tagam tugum ‘a fat person who cannot walk swiftly’ lada ludu ‘a fat baby’
  • CaC[a(C)(a)] CoC[o(C)(o)] Expressive form Meaning sar sor ‘to eat away with a savage appetite’ karaeʔ koroeʔ ‘a gurgling breathing of one being strangled’ kal kol ‘a buzzing of the ears’ ragaɽa rogoɽo ‘soil mixed with stones so large that it cannot be ploughed’ rakaɽa rokoɽo ‘the rattling of something in a box or in a bottle or the like’ ɖdaŋ ɖoŋ ‘a deep and big hole’ pagad pogod ‘a swollen state of the whole body’ ʈarad ʈorod ‘a sound of frog’
  • CaC[aC] CiC[iC] Expressive form Meaning palad pilid ‘the act of shining in various places’ par pir ‘the act of dispersing’
  • CaC[(C)aC] CeC[(C)eC] Expressive form Meaning paŋgad peŋged ‘a glitter of light appearing and disappearing now here, then there’ caʔ ceʔ ‘used for the cry of babies’ (EM)
  • CiCa(C) CoCo(C) Expressive form Meaning kidar kodor ‘a rooster with a long upright comb and long wavy feathers on the neck and tail’ (EM) kiɽaŋ koɽoŋ ‘a tall and lean person’ gida godo ‘semi-liquid things’ pica poco ‘to empty a soft or pasty substance by compression’
  • CiC CoC Expressive form Meaning bir bor ‘tall and straight’ lir lor ‘a long and weak sapling’

The formal analysis of expressives has been done.

3.4.2  Syntax of expressives

The syntax of expressives has never been described. Expressives can occupy any place, that is, in a predicate, complement, or argument slot. As the head of a predicate, expresives can take derivational suffixes, for example, passive, reflexive, benefactive, and aspect markers. Expressives can also form serial verb constructions. Thus,

  1. busuʔ-re seta-hon=e utul-putul-ta-n-a. straw-LOC dog-child=3SG:SUBJ EXPR-PROG-ITR-IND

    ‘The puppy is playing in the straw then the straw is shaking.’

  2. nir-nir-te=ʔ aŋgor-saŋgor-giɽi-aka-n-a. run-run-to=3SG:SUBJ EXPR-throw away-CONT-ITR-IND

    ‘S/he is running and running then s/he is totally getting out of breathe.’

Some expressives require an experiencer object like in the experiential constructions. For instance,

  1. rua-te alae-balae-ki-ʔ-ñ-a. fever-to EXPR-COMPL-TR-1SG:OBJ-IND

    ‘I got a trouble by a fever.’

An expressive alone or an expressive with the progressive aspect marker ta and the intransitive marker -n can occupy in the complement slot as an adverbial phrase in the following:

  1. kata-kata=e landa-ta-n-a. EXPR=3SG:SUBJ smile-PROG-ITR-IND

    ‘S/he is laughing uproariously’

  2. iriʔ-iriʔ-ta-n=(e)-m landa-ta-n-a. EXPR-PROG-ITR=EPEN-2SG:SUBJ smile-PROG-ITR-IND

    ‘You are smiling like you are mocking somebody’

An expressive can occupy in the argument slot to modify a noun or noun phrase. For example,

  1. iniʔ do janao akoʔ-bakoʔ hoɽo-ge. that person TOP always EXPR person-EMPH

    ‘S/he is always a stupid person.’

An expressive can occupy in the head of noun phrase in the following instance:

  1. iniʔ-aʔ isiɽi-sikiɽi ka=ñ suku-a. that person-GEN EXPR NEG=1SG like-IND

    ‘I don’t like her coquettish laughing.’

As is seen above, expressives have a reduplicated form. Although the single form has usually no meaning, some single forms which are followed by the completive aspect marker ke and intransitive marker n occupy the complement slot as an adverbial phrase:

  1. tii=ʔ cadʈa-cadʈa-ke-d-a. hand=3SG:SUBJ clap:EXPR-COMPL-TR-IND

    ‘S/he clapped her/his hand.’

  2. cadʈa-ke-n=eʔ tabɽi-li-ʔ-i-a. clap-COMPL-ITR=3SG:SUBJ slap-ANT-TR-3SG:OBJ-IND

    ‘S/he slapped him/her like clapping.’

3.4.3  Semantics of expressives

Nobody has ever described the semantics of expressives in Mundari. Hoffmann has just described the several expressive forms as variants in EM. For example, the following thirteen forms are the sole entry for ‘a smile to smile etc.’:

mogoeʔ, mogoeʔ-mogoeʔ, mergoeʔ, mergoeʔ-mergoeʔ, merloŋ, merloŋ-merloŋ, mirluŋ, mirluŋ-mirluŋ, moeʔ-moeʔ, muguiʔ, muguiʔ-muguiʔ, musuiʔ, musuiʔ-musuiʔ.

According to my informants some forms such as mogoeʔ-mogoeʔ, mirluŋ-mirluŋ, moeʔ-moeʔ are not known by them because of dialectal differences. They, however, can differentiate meanings in the following:

mergoeʔ mergoeʔ

‘smiling in mouth’

merloŋ merloŋ

‘smiling by children or aged-persons who have no teeth’

muguiʔ muguiʔ

‘smiling cheerful’

musuiʔ musuiʔ

‘smiling in eyes shyly’

Apart from these, there are a lot of expressives to express the action of laughing, etc. I demonstrate the semantic field of laughing, smiling and chuckling below. 9

hada hada

‘to roar with laughter successively’

kata kata

‘to roar with laughter (less than hada-hada) by many people’

kaʔ kaʔ

‘to laugh like a hen’s clucking’

keʔ keʔ

‘to laugh like a jackal’s howling’

keteʔ keteʔ

‘to laugh innocently (by children)’

kõẽ kõẽ

‘to laugh without sound’

kere kete

‘to laugh while talking’

isiɽi isiɽi

‘to ridicule one’s action or talk’

isiɽi sikiɽi

‘to laugh coquettishly’

iriʔ iriʔ

‘to laugh like a mock at’

I give another example of expressive for light reflection in the following:

jaka jaka

‘shining with gold’

jaka maka

‘shining with a flashy dress (sari with gold)’

jiki miki

‘shining with leather’

caka maka

‘shining with steel or silver’

jili mili

‘shining with building’

jilib jilib

‘dazzle with electric light’

bijir bijir


jilab jolob

‘glimmering with a firefly’

jolob jolob

‘glimmering with many fireflies’

jaraŋ jaraŋ

‘glittering in the sun’

paŋgad peŋged

‘a glitter of light appearing and disappearing now here, then there’

piɽid piɽid

‘glimmering on the sand’

palad pilid

‘the act of shining in various places’

pilid pilid

‘twinkling with stars’  Sound symbolism

As far as sound symbolism is concerned, ‘it is often said that if vowel quality is used for size symbolism, [i] will symbolize smallness, and the lower vowels, especially [a], will symbolize largeness, with degrees in between’ (Diffloth 1994:107). Diffloth, however, has suggested a counter-example (i: big, a: small) from Bahnar, which also belongs to the Austroasiatic language family.

In Mundari, it seems to me that i symbolize smallness while a symbolize largeness in the following:

saʈa saʈa

‘a passing rain for a long time’

siʈi siʈi

‘a passing rain’

jaɽam jaɽam

‘a heavy rain (the water in the river is full)’

jiɽim jiɽim

‘a heavy rain (the water in the rice-field is full)’

kaca kaca

‘to scold somebody with action’

kici kici

‘to scold somebody only by mouth’

The following cases should be taken into consideration in our future study:

baya baya

‘to act lazily’

buyu buyu

‘to act, especially walk lazily (more lazy than baya-baya)’

pisir pisir

‘to drizzle (not enough to get wet, even without an umbrella)’

pusur pusur

‘to drizzle (but to get wet)’

4  Syntax

4.1  Syntax of the simple sentence

As we have seen in section 3.2.2, the subject and object of a sentence are determined by word order. The unmarked word order is as follows: S + O + Verb.

The word order is not fixed for subject NP and object N P. Subject and object agreement, therefore, is very important for the signalling of grammatical relations. But in some cases ambiguity cannot be excluded. When the subject NP and object NP have the same person and number, the sentence is ambiguous. For instance,

  1. Soma seta=eʔ hua-ki-ʔ-i-a. Soma dog=3SG:SUBJ bite-COMPL-TR-3SG:OBJ-IND

    • ‘Soma bit the dog.’
    • ‘The dog bit Soma.’

On pragmatic grounds, meaning (a) may be less likely. But if pusi ‘cat’ is placed in the first position instead of Soma, the sentence is totally ambiguous.

  1. pusi seta=eʔ hua-ki-ʔ-i-a. cat dog=3SG:SUBJ bite-COMPL-TR-3SG:OBJ-IND

    • ‘The cat bit the dog.’
    • ‘The dog bit the cat.’

One construction in which subjects and objects are distinguishable is in relative clause constructions. The head NP in relative clauses can be a subject or an object in Mundari. But the transitive marker and intransitive marker can be used to distinguish the subject head NP from the object head NP: the intransitive marker is used when the pivot is the subject. Thus,

    • pusi hua-ke-n-seta=eʔ goeʔ-ja-n-a. cat bite-COMPL-ITR-dog=3SG:SUB die-INGR-ITR-IND ‘The dog who bit the cat has died.’
    • pusi hua-ke-d-seta=eʔ goeʔ-ja-n-a. cat bite-COMPL-TR-dog=3SG:SUB die-INGR-ITR-IND ‘The dog whom the cat bit has died.’

The subject and object agreement elements can be marked only when the subject NP and object NP are classified as animate nouns.

In addition to two arguments, a postpositional phrase or adverb denoting location or time can be inserted into any position before the verb. In this case the orders of NPs and PP are rather free except the last position, which is reserved for the verb. We illustrate this in (143).

    • setaʔ-re seta-ko maɳɖi=ko jom-ke-d-a. morning-LOC dog-PL food=3PL:SUBJ eat-COMPL-TR-IND
    • maɳɖi seta-ko setaʔ-re=ko jom-ke-d-a. food dog-PL morning-LOC=3PL:SUBJ eat-COMPL-TR-IND
    • maɳɖi setaʔ-re seta-ko=ko jom-ke-d-a. food morning-LOC dog-PL=3PL:SUBJ eat-COMPL-TR-IND ‘In the morning the dogs ate the food.’

4.2  Typological features

I list the constituent order below: 10

  • S + O + V
    1. Soma maɳɖi=ʔ jom-ke-d-a. Soma food=3SG:SUBJ eat-COMPL-TR-IND

      ‘Soma ate the food.’

  • S + O + V or O + S + V
    1. pusi seta=eʔ hua-ki-ʔ-i-a. cat dog=3SG:SUBJ bite-COMPL-TR-3SG:OBJ-IND

      • ‘The cat bit the dog.’
      • ‘The dog bit the cat.’

  • Oblique + S + O + V or S + Oblique + O + V or S + O + Oblique + V
      • setaʔ-re seta-ko maɳɖi=ko jom-ke-d-a. morning-LOC dog-PL food=3PL:SUBJ eat-COMPL-TR-IND
      • maɳɖi seta-ko setaʔ-re=ko jom-ke-d-a. food dog-PL morning-LOC=3PL:SUBJ eat-COMPL-TR-IND
      • maɳɖi setaʔ-re seta-ko=ko jom-ke-d-a. food morning-LOC dog-PL=3PL:SUBJ eat-COMPL-TR-IND ‘In the morning the dogs ate the food.’

  • Noun Phrase + Postposition For example, oɽaʔ-re ‘at home’
  • Genitive + Noun Phrase For example diri-reaʔ oɽaʔ ‘a stone house’
  • Adjective + Noun For example maraŋ oɽaʔ ‘a big house’
  • Demonstrative + Noun For example, ne oɽaʔ ‘this house’
  • Numeral + Noun For example, baria oɽaʔ ‘two houses’
  • Relational clause + Head Noun
      • pusi hua-ke-n-seta=eʔ goeʔ-ja-n-a. cat bite-COMPL-ITR-dog=3SG:SUBJ die-INGR-ITR-IND ‘The dog who bit the cat has died.’
      • pusi hua-ke-d-seta=eʔ goeʔ-ja-n-a. cat bite-COMPL-TR-dog=3SG:SUBJ die-INGR-ITR-IND ‘The dog whom the cat bit has died.’

  • Degree word + Adjective For example, bese maraɳ ‘very big’
  • Final position of polar question particle
    1. Soma maɳɖi=ʔ jom-ke-d-a ci. Soma food=3SG:SUBJ eat-COMPL-TR-IND Q

      ‘Did Soma eat the food?’

  • First position of interrogative in content questions

    1. oko-e hijuʔ-aka-n-a who come-CONT-ITR-IND

      ‘Who has come?’

  • Clause + Adverbial subordinator
    1. maɳɖi jom-tayom-te=ko senog-a. food eat-after-LOC=3PL:SUBJ go-IND

      After eating the food they will go.’

4.3  Complex sentence structure

4.3.1  Coordination

Coordination is expressed by the following particles:

  • oɽoʔ/ad ‘and’

This coordinating conjunction can conjoin not only noun phrases but also clauses.

  • Noun Phrases
    1. araʔ sim oɽoʔ hende merom red fowl and black goat

      ‘red fowl and black goat’

  • Clauses
    1. jom-ke-d-a-eʔ ad=eʔ senoʔ-ja-n-a. eat-COMPL-TR-IND-3SG and=3SG:SUBJ go-INGR-ITR-IND

      ‘He/She ate and went away.’

  • ci ‘or’

This coordinating conjunction can connect not only noun phrases but also clauses.

  • Noun Phrases
    1. araʔ sim ci hende merom red fowl or black goat

      ‘red fowl or black goat’.

  • Clauses
    1. haga-m Ranci-te-ʔ senoʔ-ja-n-a ci hatu-re menaʔ-i-a. brother-your Ranchi-to-3SG go-INGR-ITR-IND or village-LOC COP-3SG-IND

      ‘Your brother has gone to Ranchi or he is at home.’

  • ca ‘or’

This coordinating conjunction is not used to connect two noun phrases but to connect two clauses.

  1. Soma hijuʔ ca=eʔ senoʔ ka-ñ itu-a-n-a. Soma come or=3SG:SUBJ go NEG-LSG:SUBJ know-SUS-ITR-IND

    ‘I don’t know whether Soma comes or goes.’

  • ci-aʔci ‘because’

According to Hoffmann in EM, ‘this conjunction was introduced into the translation of the Bible made by the first Lutheran Missionaries’ (p. 843). It has been made by the calque of Hindi kyõki or cũki.

  1. Ranci-te nida=le tebaʔ-ke-d-a. ci-aʔci bas Ranchi-to night=1PL.EX:SUBJ reach-COMPL-TR-IND because bus



    ‘We (excl.) arrived at Ranchi at night, because the bus was out of order.’

  • mendo ‘but’

This coordinating disjunction mendo can be analyzed into men ‘to say’ and the particle do. This can connect two sentences.

  1. maɳɖi jom-mone-ja-ʔ-ñ-tai-ke-n-a. mendo ka-ñ food eat-want-INGR-TR-1SG-COP-COMPL-ITR-IND but NEG-1SG



    ‘I wanted to eat the food, but I have not got it.’

  • karedo ‘otherwise’

This consists of the negator ka the postposition -re and the particle do. This connects two sentences, especially an indicative sentence.

  1. maɳɖi jom-le-m. karedo loyoŋ-te=bu senog-a. food eat-first-2SG otherwise rice field-to=1PL.INC go-IND

    ‘Eat the food first, or we will go to the rice-field.’

4.3.2  Relative-type clauses

I have already illustrated the non-finite form in Mundari. The non-finite form can be followed by a noun or pronoun in a relative clause. For instance,

  1. Ranci-te sen-ke-n-hoɽo=eʔ hijuʔ-ruɽa-ja-n-a. Ranchi-to go-COMPL-ITR-person=3SG:SUBJ come-return-INGR-ITR-IND

    ‘The person who went to Ranchi has just returned.’

  2. Ranci-te sen-ke-n=iʔ hijuʔ-ruɳa-ja-n-a. Ranchi-to go-COMPL-ITR=3SG:SUBJ come-return-INGR-ITR-IND

    ‘The one who went to Ranchi has just returned.’

  3. abu jom-ke-d-sim-do=eʔ sibil-ge-tai-ke-n-a. 1PL.INC eat-COMPL-TR-chicken-TOP=3SG:SUBJ tasty-EMPH-COP-COMPL-ITR-IND

    ‘The chicken that we (INC) ate was tasty’.

In a relative clause the following points can be noted:

  • The head noun can be the object (161) or the subject (159) of the non-finite verb. It is very clear that the object follows the transitive marker d and the subject follows the intransitive marker n. As was shown above, this is a syntactic test to distinguish objecthood from subjecthood.
  • The head noun can be omitted; in this case the third person singular form which follows is = rather than =.
  • The personal pronoun other than the third person cannot be allocated in the postverbal position as a head noun in a relative clause. Thus, *jom-ke-d-pe ‘you who ate it’ is ungrammatical (-pe is second personal plural suffix). If we change ‘one’ to ‘you’ in the sentence (160), we should paraphrase it with two sentences.
    1. Ranci-te=m sen-ke-n-a. ena-te=m Ranchi-to=2SG:SUBJ go-COMPL-ITR-IND that-by=2SG:SUBJ



      ‘You went to Ranchi and (then) you have just returned.’

4.3.3  Subordinate clauses

The non-finite form can be followed by a postposition or a question marker in a subordinate clause. The postposition re and the question marker ci are most frequently used. The difference between re and ci is complicated. The basic meaning re is ‘in that time’ while ci is ‘while’. The usage of the aspect marker is also complicated. I give an example of each below:

  1. maɳɖi jom-ta-n-ci=ko jagar-ta-n-a. food eat-PROG-ITR-CONJ=3PL:SUBJ talk-PROG-ITR-IND

    ‘While they are eating the food, they are talking simultaneously’

  2. maɳɖi jom-ta-n-re=ko jagar-ta-n-a. food eat-PROG-ITR-LOC=3PL:SUBJ talk-PROG-ITR-IND

    ‘When they are eating the food they are talking at that time.’

  3. hoɽo-ko ukuʈa jagar-ta-n-ci=ko people-PL backbiting talk-PROG-ITR-CONJ=3PL:SUBJ



    ‘While we two were backbiting, people heard us.’

  4. maɳɖi jom-ta-n-re hoʈoʔ-re=ñ ɖiʔ-ja-n-a. food eat-PROG-ITR-LOC throat-LOC=1SG:SUBJ stick-PROG-ITR-IND

    ‘When I was eating the food it stuck in my throat.’

  5. paisa nam-aka-d-ci rasika-ja-ʔ-ñ-a. money get-CONT-TR-CONJ joy-INGR-TR-1SG:OBJ-IND

    ‘I am joyful because I have got money’

  6. paisa nam-aka-d-re uriʔ=le kiriŋ-li-ʔ-i-a. money get-CONT-TR-LOC cattle=1PL.EXC:SUBJ buy-ANT-TR-3SG:OBJ-IND

    ‘By the time we got money, we had already bought the cattle.’

As for the personal marking, the subject marking in the subordinate clause can be omitted when the same subject occurs in the main clause in the following:

  1. hon-ko masɖi jom-ke-d-ci=ko senoʔ-ja-n-a. child-PL food eat-COMPL-TR-CONJ=3PL:SUBJ go-INGR-ITR-IND

    ‘When the children had eaten the food they went.’

  2. hon-ko maɳɖi=ko jom-ke-d-ci child-PL food=3PL:SUBJ eat-COMPL-TR-CONJ=3PL:SUBJ kami-te=ñ senoʔ-ja-n-a. work-to=1SG:SUBJ go-INGR-ITR-IND

    ‘As the children ate the food I have gone to work.’

The -re clause also means conditional ‘if’. The combination -re-do (locative and topic marker) is more common for the conditonal clause:

  1. bisi jom-ja-n-re go-goeʔ-a. poison eat-INGR-ITR-LOC die-ITER-IND

    ‘If the poison will be taken everybody shall die.’

  2. maɳɖi=m jom-ke-d-re-do piiʈi-te=laŋ senog-a. food=2SG:SUBJ eat-COMPL-TR-LOC-TOP market-to=1DL.INC:SUBJ go-IND

    ‘If you finish eating the food let us go to the market.’

Instead of re, sida-re ‘before’ and tayom-te can be used in the subordinate clauses:

  1. maɳɖi jom-sida-re=ko senoʔ -ja-n-a. food eat-before-LOC=3PL:SUBJ go-PROG-ITR-IND

    ‘Before eating the food they have gone.’

  2. maɳɖi jom-tayom-te=ko senog-a. food eat-after-LOC=3PL:SUBJ go-IND

    ‘After eating the food they will go.’

The following examples denote supposed reality. In this sentence te-raʔ or honaŋ should be added:

  1. iniʔ-loʔ aɽandi-ja-n-re-do naaʔ-loʔ-do=ñ that person-with marry-INGR-ITR-LOC-top now-with-top=1SG:SUBJ



    ‘If I would marry him I should get old (I am still young because I didn’t marry him).’

  2. añ-honaŋ seta bai-ja-n-re-do cadlom 1SG-if dog become-INGR-ITR-LOC-TOP tail



    ‘If I were a dog I may sway my tail repeatedly (I flatter my master).’

Another postposition loʔ can follow the non-finite form in a subordinate clause:

  1. maɳɖi jom-ta-n-loʔ=ko duraŋ-ta-n-a. food eat-PROG-ITR-LOC=3PL:SUBJ sing-PROG-ITR-IND

    ‘While they are eating the food they are singing.’

It is very similar to (163). But jom-ta-n-loʔ is used in a wider range of contexts than jom-ta-n-ci.

The postposition ate ‘from’ cannot follow the non-finite form directly. But ke-ate in the indicative sentence and le-ate in the imperative sentence are very common when denoting successive action:

  1. maɳɖi jom-ke-ate=ko senoʔ-ja-n-a. food eat-COMPL-from=3PL:SUBJ go-INGR-ITR-IND

    After taking the food they have gone.’

  2. maɳɖi jom-le-ate senoʔ-me. food eat-ANT-from go-IND

    ‘When they are eating the food they are talking in that time.’

In addition to ke-ate and le-ate there are ta-n-ate, ke-n-ate, 11 ja-n-ate, le-n-ate, and aka-n-ate, as in the following:

  1. maɳɖi jom-ta-n-ate=m paɽao-ta-n-a. food eat-PROG-ITR-from=2SG:SUBJ study-PROG-ITR-IND

    ‘You are still eating but starting to study.’

  2. maɳɖi jom-ke-n-ate=ko aboŋ-en-ja-n-a. food eat-COMPL-ITR-from=3PL:SUBJ wash-RFLXV-INGR-ITR-IND

    ‘After taking the food they have washed their hands’.

  3. eŋga-m goeʔ-ja-n-ate bar-sirma hoba-ja-n-a. mother-your die-INGR-ITR-from two-year happen-INGR-ITR-IND

    ‘It has taken two years after your mother’s death.’

  4. dub-aka-n-ate=ñ jom-ta-n-a. sit-CONT-ITR-from=1SG:SUBJ eat-PROG-ITR-IND

    ‘I am eating while sitting on the ground.’

The concessive clauses are as follows:

  1. jom-e-ca ka=eʔ jom-e maɳɖi bai-ta-m. eat-it-or NEG=3SG:SUBJ eat-it food make-PROG-2SG

    ‘Whether s/he eats or not, you must be ready for the food.’

  2. senoʔ-re-o ka=m senoʔ-re-o kami-do calao-a. go-LOC-also NEG=2SG:SUBJ go-LOC-also work-TOP go well-IND

    ‘Whether you go or not, the work goes well’

A negative clause is illustrated in the following:

  1. ne-hoɽo auri nu-bairi-do ka=eʔ jagar-a. this-person yet drink-only-TOP NEG=3SG:SUBJ talk-IND

    ‘He doesn’t talk until he drinks liquor.’

4.4  Switch reference

According to Anderson and Boyle (2002:48), based solely on some of the very limited data in (Osada 1992), it was speculated that ci in Mundari might be a same subject marker. Thus,

  1. jom-ke-d-ci=ko senoʔ-ja-n-a. eat-COMPL-TR-=3PL:SUBJ go-INGR-ITR-IND

    ‘They went away as soon as they had eaten.’

In our observation it is not the same subject marker as is shown below.

  1. jom-ke-d-ci=ko senoʔ-ja-n-a. 1SG eat-COMPL-TR-CONJ=3PL:SUBJ go-INGR-ITR-IND

    ‘They went away as soon as I had eaten.’

5  Semantics/Discourse

5.1  Semantics

There are many verbs in Mundari which are equivalent to the meaning of ‘cut’ in English.

  • had ‘to cut with sawing motion by a knife, a saw, an axe, etc’
  • ged ‘to cut meat by bãiʈi (a large meat cutter)’
  • laʈab ‘to cut a paper, hair, etc. by scissors’
  • maʔ ‘to cut a tree with a striking motion by an axe’
  • od ‘to cut through and through a tree and so fell it’
  • samaʔ ‘to cut something (meat bone, jackfruit, etc.) in small pieces with a koɳɖe (axe)’
  • ir ‘to cut the stalk of grain by ɖatrom (sickle) for reaping’
  • gaɳduiʔ ‘to cut the long things (rope, etc.) in pieces of a given length’
  • ona ‘to cut timber into log’
  • dula ‘to cut the top portion of trees with duɳu (long axe)’
  • paɽaʔ ‘to cut the long things (firewood, etc.) into two portions’

According to Suwilai (2002), there are numerous verbs in the same semantic field in Kham, which belongs to the Austroasiatic language family spoken in Cambodia.

5.2  Discourse

As far as discourse is concerned, there are the topic marker and emphatic marker in Mundari. I describe these here.

The particle do follows the noun phrase or postpositional phrase that it marks as the topic in discourse. It singles out the element about which the comment is made in the sentence. The following elements can be marked by the topic marker do:

  • Subject
    1. do=ñ senog-a. 1SG TOP=1SG:SUBJ go-IND

      ‘I will go.’

  • Direct object
    1. maɳɖi do=ñ jom-ke-d-a. food TOP=1SG eat-COMPL-TR-IND

      ‘I ate the food.’

  • Location
    1. Ranci-re do=ñ tai-ke-n-a. Ranchi-LOC TOP=1SG:SUBJ live-COMPL-ITR-IND

      ‘I lived in Ranchi.’

  • Source
    1. Ranci-ate do=m hijuʔ-aka-n-a. Ranchi-from TOP=2SG:SUBJ come-CONT-ITR-IND

      ‘You have come from Ranchi.’

  • Instrumental
    1. ne daru hake-te do=pe maʔ-ke-d-a. this tree axe-by TOP=2PL:SUBJ cut-COMPL-TR-IND

      ‘You cut this wood with an axe.’

  • Benefactive
    1. Soma-naŋgen do nakiʔ=ñ kiriŋ-a-iʔ-a. Soma-for TOP comb=1SG:SUBJ buy-BEN-3SG-IND

      ‘I will buy the comb for Soma.’

  • Comitative
    1. añ-loʔ do han-te=laŋ sen-ke-n-a. 1SG-with TOP yonder-to=1PL.INC:SUBJ go-COMPL-ITR-IND

      ‘We two (inc) went there together.’

The possessive is not marked for the topic by the particle do. For example, *diri-reaʔ /raʔ do oɽaʔ (diri ‘stone’, -reaʔ/-raʔ GEN oɽaʔ ‘house’) is ungrammatical. While the independent possessive is marked for the topic. Thus, añ-ag-aʔ do menaʔ. (añ-ag-aʔ ‘mine’, menaʔ COP) ‘There is mine.’

In addition to the topic of a sentence, do in Mundari marks the contrast as well. For instance,

  1. am do Ranci-te, do Kuŋʈi-te=ñ sen-ke-n-a. 2SG TOP Ranchi-to 1SG TOP Khunti-to=1SG:SUBJ go-COMPL-ITR-IND

    ‘You (went) to Ranchi, but I went to Khunti.’

Further, as we have mentioned in the indefinite oko can be followed by the topic mark do as in (198) but the interrogative oko cannot.

  1. oko-e hijuʔ-aka-n-a Who come-CONT-ITR-IND

    ‘Who has come?’

  2. oko-e do hijuʔ-aka-n-a. someone TOP come-CONT-ITR-IND

    ‘Someone has come, (but not all).’

The particle ge may function as an emphatic marker in discourse. The following elements can be marked by the emphatic marker ge:

  • Subject
    1. ge=ñ senog-a. 1SG EMPH=1SG:SUBJ go-IND

      ‘It is I who will go.’

  • Direct Object
    1. maɳɖi ge=ñ jom-ke-n-a. Food EMPH=1SG:SUBJ eat-COMPL-ITR-IND

      ‘It is the food that I ate.’

  • Benefective
    1. Soma-naŋgen ge nakiʔ=ñ kiriŋ-a-iʔ-a. Soma-for EMPH comb=1SG:SUBJ buy-BEN-3SG-IND

      ‘For Soma, I will buy the comb.’

  • Source
    1. Ranci-ate ge=m hijuʔ-aka-n-a. Ranchi-from EMPH=2SG:SUBJ come-CONT-ITR-IND

      ‘From Ranchi, you have come.’

  • Instrumental
    1. ne daru hake-te ge=pe maʔ-ke-n-a. this wood axe-with EMPH=2PL:SUBJ cut-COMPL-ITR-IND

      ‘With axe, you cut this wood.’

  • Independent possessive
    1. nea do añ-ag-aʔ ge menaʔ. this TOP 1SG-GEN-GEN EMPH COP

      ‘This is mine (not any other persons’).’

While the topic marker do is never allocated in the postverbal position, the emphatic marker ge can be used for the verbal phrase in postverbal position. Thus,

  1. maɳɖi=ko jom-ta-n-ge-a. food=3SG:SUBJ eat-PROG-ITR-EMPH-IND

    ‘They are taking food indeed.’

As seen above, the interrogative oko cannot be followed by the topic marker do as in (198) but it can be followed by the emphatic marker ge. The indefinite oko, on the other hand, cannot be followed by the emphatic marker ge. For instance,

  1. okoe ge her-le-d-a mani do who EMPH sow-ANT-TR-IND mastard TOP

    ‘Who has sown the mustard seed indeed?’

6  Lexicon

6.1  Austroasiatic/Munda components

It is easy to list the cognated words as Schmidt (1906) and Pinnow (1959) have proposed. I list here some words with Pinnow’s index number.

Mundari meaning

Pinnow’s index number

baba ‘paddy’


bir ‘forest’


buru ‘mountain, hill’


daʔ ‘water’


hon ‘child’


jaŋ ‘seed, bone’


japid ‘to shut eyes’


kula ‘tiger’


lutur ‘ear’


luʔ ‘to ladle out’


muiʔ ‘ant’


mid/miyad, mod/moyod ‘one’


nu ‘to drink’


raʔ ‘to call, cry’


sim ‘fowl’


ti ‘hand’


6.2  Loan strata

There are many loanwords in Mundari mainly from the adjoining Indo-Aryan languages, notably Sadani. I note the following points:

  • As for nouns, Mundari-speaking area is in a boundary of vowel variant a/ɔ in the Indo-Aryan. So some word forms coexist in the following: man/mon ‘mind’ < Indo-Aryan man/mɔn, badnam/bodnam ‘infamous’ < Indo-Aryan badnām/bɔdnam, etc.
  • As for verbs, those with the endings -ao in Mundari are borrowed from the Indo-Aryan languages; For example, bujao ‘to understand’, dekao ‘to show’, hatao ‘to remove’, etc.
  • In some dialects, the dative marker ke has been introduced from Sadani. Thus, ape-ke joar (2PL-Dative greeting) ‘Joar to you’.
  • From a phonetic point of view, the final b in some loanwords is realized as a so-called checked consonant; For example, /kitab/ ‘book’ [kitaʔb̥].

7  Brief Analysed Texts

7.1  Riddles

  • Q. aɽakaʈa palad-pilid. rafter flickering:EXPR ‘Over the rafters of the roof flickering here and there.’ A. kaʈea ‘mouse’
  • Q. aʈa-maʈa bir-ko tala-re boŋga hon-ko dense jungle-PL middle-LOC spirit son-PL caʈu=ko harub-ta-d-a. the earthen pot=3PL:SUBJ cover-PROG-TR-IND ‘The sons of spirits are covering the earthen pot in the middle of jungle.’ A. ud ‘mushroom’
  • Q. aʈa-maʈa bir-ko tala-re boŋga hon-ko dense jungle-PL middle-LOC spirit child-PL ub=ko raɽa-ta-d-a. hair=3PL:SUBJ untie-PROG-TR-IND ‘The sons of spirit are untieing their hair in the middle of jungle.’ A. badcom. ‘a kind of grass which looks like a child who untied his hair’
  • Q. catoma-re oɽaʔ lo-ta-n-a. Catoma (Village name and umbrella)-LOC House fire-PROG-ITR-IND gagara-re=ko raʔ-ta-n-a. Gagara (Village name and brasspot)-LOC=3PL:SUBJ cry-PROG-ITR-IND ‘The house in the umbrella is on fire while they are crying in the brasspot.’ A. huka ‘hookkah (a kind of pipe for smoking)’
  • Q. da ta-ñ kokoʔ ɖaɳɖaʔ. give me GEN-1SG hockey stick ‘Give me my hockey stick.’ A. seta cadlom. ‘dog’s tail’
  • Q. dub-me daru ɖemka daru deʔ-a. sit-2SG tree dwarf tree climb-IND ‘Sit down please tree, dwarf tree will climb.’ A. haɽad booʔ ‘a yam of the jungles’
  • Q. edel daru rau-ta-n-a cotton tree felt down-PROG-ITR-IND bajuɳia-ko hijuʔ-ta-n-a. musician-PL come-PROG-ITR-IND ‘The cotton tree is falling down then the musicians are coming.’ A. ‘excrement’ (the raw cotton tree smells. the musician means a fly.)
  • Q. gaɽa-gaɽa-te ledera atu-ta-n-a. river-river-to cloth float-PROG-ITR-IND ‘The clothes are floating towards rivers.’ A. geded ‘algae’
  • Q. gaɽa-gaɽa-te laka koɽa kapi goɽ-aka-d-a. river-river-to thickset boy axe carry-CONT-TR-IND ‘The thickset fellow has carried the axe on shoulder towards rivers.’ A. kaɽakom ‘crab’
  • Q. gaɽa-gaɽa-te laka koɽa paciri tapaɽ-ta-n-a. river-river-to thickset boy wall raise-PROG-ITR-IND ‘The thickset fellow is raising the walls towards rivers.’ A. kaɽakom ‘crab’

7.2  Kande’s story

This is a part of seɳa jumbaɽa ‘Tree of intelligence in the Munda way’ written by the late Kande Munda who was a good writer and musician.

  • kalam ku<nu>li utrao-re=ñ met-a-i-ta-n-a. pen ask-Nom-ask rise-LOC=1SG:SUBJ say-BEN-3SG-PROG-ITR-IND ‘When the question rises why you are writing this I am saying to him/her.’
  • jumbaɽa mane ɖoɽoʔ daru-ko-raɽ a bunch of branch and leaf mean sponge gourd tree-PL-GEN koto-ko-re sakam ar sakam-koto-ge branch-PL-LOC leaf and leaf-branch-EMPH jumbaɽa-giɽi-aka-n-jumbaɽa. dense-totally-CONT-ITR-a bunch of branch ‘Jumbaɽa, a dense bunch of branches and leaves means jumbaɽa which is a dense bunch of flowers in the branch of the tree.’
  • jaan jetan bir-ko-re anaʔ-manaʔ naɳi-ko, any any forest-PL-LOC various:EXPR climber-PL daru-cupad-ko-te juuɳu-giɽi-aka-n-jumbaɽa tree-bush-PL-to bush-totally-CONT-ITR-a bunch of branch ‘In any forest jumbaɽa in which various creepers overgrow in the trees and bushes.’
  • jumbaɽa ʈuŋ hijuʔ-ta-n-a. a bunch of branch meaning come-PROG-ITR-IND ‘The meaning of jumbaɽa is coming clear.’
  • ne vacak-re mosa-te-ge bariya oroto hambud-aka-n-a. this significance-LOC once-by-EMPH two meaning embrace-CONT-ITR-IND ‘This significance embraces two meanings.’ miyad-do huldub-huldub daru jumbaɽa-ko-re one-TOP dense:EXPR tree a bunch of branch-PL-LOC nana-boron ceɳe-cipurub=ko dub-aka-n-ci varous:EXPR bird-feather=3PL:SUBJ sit-CONT-ITR-that niral cere-bere, kere-ɖore-ta-n=ko jagar-a. clean chatting-EXPR screeching:EXPR-PROG-ITR=3PL:SUBJ talk-IND ‘First, when the various birds are sitting in the dense branch they are talking clearly to each other.’
  • ne jagar-ta-n-ko-re ako-aʔ duku-suku-ko-raɽ this talk-PROG-ITR-PL-LOC 3PL-GEN grief-joy-PL-GEN jagar-ko-ge daŋ jaaʔ=ko ja<pa>gar-ta-n-a. talk-PL-EMPH indeed perhaps=3PL:SUBJ talk<RECIP>PROG-ITR-IND ‘They are perhaps talking each other’s grief and joy in their talk.’
  • en leka-ge ne puti paɽao-ko-yo. that like-EMPH this book read-PL-also ‘To read this book is also in the same way like a bird.’
  • jidan-jumbaɽa-re=ko dub-aka-n-a. life-a bunch of branch-LOC=3PL:SUBJ sit-CONT-ITR-IND ‘They are sitting in the dense bunch of branch.’
  • ad duku-suku-raʔ heɽem-haɽad oncar-re=ko tain-a. and grief-joy-GEN sweet-bitter care-LOC=3PL:SUBJ remain-IND ‘And their desire remains a sweet or bitter memory in their suffering and joyful life.’
  • neya seɳa-jumbaɽa cae this intelligence-jumbaɽa or duku-suku-jumbaɽa men-ke-re-yo bai-oʔ-a. grief-joy-jumbaɽa say-COMPL-LOC-also make-PASS-IND ‘You can call it the tree of intelligence or the tree of grief and joy’
  • paɽao-ta-n-re gaʈa-gaʈa-ko-re med-daʔ study-PROG-ITR-LOC various:EXPR-PL-LOC ear-water(=tear) joro=e leka-n-kaji-ko nam-og-a. shed=3SG:SUBJ like-RFLXV-word-PL meet-PASS-IND ‘S/he is faced with the critical words because of which s/he sheds tears.’
  • ar laiʔ capu-capu-te landa=e leka-n-aʔ-o. and belly touch-touch-INS laugh=3SG:SUBJ like-RFLXV-GEN-also ‘And s/he is also convulsed with laughter.’
  • ne puti seɳa-jumbaɽa duku-suku-jumbaɽa this book intelligence- jumbaɽa grief-joy-jumbaɽa ar prem-prit-jumbaɽa=bu men-ta-ʔ. and love-love-jumbaɽa=1PL.INC say-PROG-TR ‘This book is the tree of intelligence or the tree of grief-joy or the tree of love.’
  • ol-oʔ-do neya haya-te-o mone-ja-n-a. write-PASS-TOP this desire-INS-also desire-INGR-ITR-IND ‘I am writing because of the following intention.’
  • jug badli-idi-badli-idi-ta-n-a. time change-continue-change-continue-PROG-ITR-IND ‘The time is changing and changing continously’
  • muɳɖa daŋgɽa-daŋgɽi-ko isukul-kolej-ko-re Munda boy-girl-PL school-college-PL-LOC isu=ko paɽao-n-ja-n-a. very=3PL:SUBJ study-RFLXV-INGR-ITR-IND ‘The Munda boys and girls have studied at school and college.’
  • isu=ko paɽao-n-ta-n-a ar very=3PL:SUBJ study-RFLXV-PROG-ITR-IND and ayar-te-o=ko paɽao-n-ge-a. future-INS-also=3PL:SUBJ study-RFLXV-EMPH-IND ‘They are studying very much and will study in the future.’
  • paɽao-seɳa-ja-n-isu-hoɽo-ko jagar-ta-n ayum-le-ko-re study-intelligent-INGR-ITR-many-person-PL talk-PROG-ITR hear-ANT-PL-LOC eŋga-aʔ jagar muɳɖari hirum-hirum-te-do=ko mother-GEN talk Mundari co-wife-co-wife-INS-TOP=3PL:SUBJ lusuɳɖu-giɽi-ja-ʔ-i-leka=ko ayum-oʔ-a. despondent-totally-INGR-TR-3SG-like=3PL hear-PASS-IND. ‘When educated persons speak and hear in their mother tongue, Mundari, they hear totally in a despondent way to face their co-wives.’
  • sen-hora seŋaɳɖe, paidal cae moʈor-ko-re, go-road go-about walk or motorcycle-PL-LOC ‘Along the path and travelling way by walk or by motorcycle,’
  • buru piiʈi-piɽi-ko-re, jeta hoʈel-ko-re, festival market-place-PL-LOC any hotel-PL-LOC ‘in the festival and weekly market place, or in any hotel,’
  • ne samae huɽiŋ-huɽiŋ paɽao-aka-n-muɳɖa-daŋɡɽa-daŋɡɽi-ko this time little-little study-CONT-ITR-Munda-boy-girl-PL ‘The Munda boys and girls who educated a little bit recently.’
  • na<pa>m-ja-n-loʔ-ge hindi-te-ge meet<RECIP>INGR-ITR-COMIT-EMPH Hindi-INS-EMPH joar-jagar=ko eʈeʔ-torsa-e-a. greeting-talk=3PL:SUBJ begin-quickly-EPEN-IND ‘(They) are quickly greeting and talking in Hindi.’
  • Buɳɖu-re mid sirma sarna-samaj-raʔ du<nu>b ʈaharao-le-n-a. Bundu-LOC one year Sarna society-GEN sit<NMLZ> open-ANT-ITR-IND ‘One day there was a meeting of the Sarna society in Bundu.’
  • kaji-do muɳɖa-ko-aʔ du<nu>b word-TOP Munda-PL-GEN sit<NMLZ> ‘The name is for the Mundas’ meeting.’
  • en-re seɳa-garaŋ-aka-n-leka-n muɳɖa-daŋgaɽa-ko-ge=ko there intelligent-highly-CONT-ITR-like-ITR Munda-boy-PL-EMPH=3PL:SUBJ jama-le-n-a. gather-ANT-ITR-IND ‘The highly intelligent Munda boys have gathered there.’
  • aliŋ-o ɖakɖar Ram Dayal Muɳɖa-loʔ=liŋ 1DL.EX-too Dr Ram Dayal Munda-with=1DL:SUBJ sen-aka-n-tai-ke-n-a. go-CONT-ITR-COP-COMPL-ITR-IND ‘I had been with Dr Ram Dayal Munda there.’
  • en-re je jagar-ta-n-ko hindi-te-ge there all talk-PROG-ITR-3PL Hindi-INS-EMPH jagar-la<pa>ʔ-ta-n-ko ayum-oʔ-ta-n-a. talk-excess<RECIP>PROG-ITR-3PL hear-PASS-PROG-ITR-IND ‘All speakers speak and compete to talk in Hindi.’
  • ɖakʈar sayob-aʔ jagar paɽi tebaʔ-le-n-a, Doctor Lord-GEN talk turn reach-ANT-ITR-IND ‘It had been a time to speak, Dr Munda’s turn.’
  • do=eʔ met-a-ko-ta-n-a. then=3SG:SUBJ say-BEN-3PL-PROG-ITR-IND ‘Then he is saying something for them.’
  • hela, neya-do muɳɖa-ko-aʔ du<nu>b tanaʔ. alas this-TOP Munda-PL-GEN meeting COP ‘Alas! This is the Mundas’ meeting.’
  • ar diku-te-ge soben=pe jagar-ja-d-a. and Hindi-INS-EMPH all=2PL:SUBJ talk-INGR-TR-IND ‘And everyone speaks in Hindi.’
  • neya-do muɳɖa du<nu>b tanaʔ. this-TOP Munda meeting COP ‘This is the Mundas’ meeting.’
  • are=ñ ayum-ja-d-pe-a, hindi-raʔ ʈekano banoʔ. well=1SG:SUBJ hear-INGR-TR-2PL:OBJ-IND Hindi-GEN certainty NEG.COP ‘I am listening to you but your Hindi is not good.’
  • niʔ neka=e ʈokoeʔ-le-d-ko-a-te this person like this=3SG:SUBJ scold-ANT-TR-3PL:OBJ-IND-to ‘This person had scolded them like this.’
  • muɳɖa-te jagar-ja<pa>gar-se-sen-ja-n-a. Munda-INS talk-talk<RECIP>go-ITER-INGR-ITR-IND ‘Then people start talking in Mundari.’


I would like to express my sincere thanks to the editor, Gregory D.S. Anderson who encouraged me to complete my paper and to Nicholas Evans and Yoshiharu Takahashi who gave me useful comments on the early version of my paper. I am also indebted to my wife, Maki who gave me a lot of data as a native speaker of Mundari. Naturally, the description in this chapter will overlap with my previous work to some extent.

According to Gregory Anderson, his Ho informants in Mayurbhanj claim only about 80–85% intelligibility with Mundari which is only a little higher than that for Santali. In my own experience, however, when I spoke to Ho speakers in Mundari in the Chaibasa area they considered me to be speaking in Ho. Thus, the Ho of Chaibasa is obviously closer in the dialect/language continuum to Mundari than the Ho of Mayurbhanj.

According to Anderson and Zide (2002:55), ‘a minimally bimoraic shape appears to have been obligatory for free-forms of nouns in Proto-Munda and many of its daughter languages’. Our data on Mundari supports this.

According to Nagaraja (1999), he never reported a tonal contrast in Korku.

We use the upper bar V for the pitch accent here.

According to our definition, we classified into four types among languages proposed to lack a noun–verb distinction to distinguish morphological from syntactic evidence and to leave open the possibility that word classes distinguishable by morphological criteria could be indistinguishable by syntactic criteria; that is, (i) Omnipredicative languages (all major word classes are able to function directly as predicates without derivation, and with no change of meaning), (ii) Precategorial languages (we will restrict ‘precategorial’ to the case where – as in omnipredicative languages – open-class lexemes can occur in any syntactic position. However, in precategorial languages, it is not possible to state a predicate-type meaning for the lexeme directly; rather there is an increment that is made, according to the functional position it is plugged into), (iii) Broschartian languages (the semantic result of placing lexemes in referring or predicating environments depends not on a high-level word class category like noun or verb, but rather is sensitive to much more specific semantic categories, each characterized by their own particular pattern of semantic incrementation), and (iv) Rampant zero conversion languages (the vast majority of lexical items of a given form may appear in both predicating and referring syntactic environments with no formal signalling of conversion, but unlike in a Broschartian language, the semantic effects of syntactic environment are far less predictable). It is only monoprecategorial languages, from among the types above, that can truly be claimed to lack a noun–verb distinction. As the space of my chapter is limited in this book I cannot touch on the details of this issue. Please see (Evans and Osada 2005a), three commentaries (Croft 2005, Hengeveld and Rijkhoff 2005, Peterson 2005) and our response (Evans and Osada 2005b).

These categories are due to Klaiman (1986). I give a list of verbs in Osada (1999).

See my paper (Osada 1991) for details.

tuka may be derived from the noun tuka ‘nest’.

I do not repeat the above-mentioned expressives here.

This list owes to Haspelmath, Dryer, Gil and Comrie (eds) (2005) WALS.

ke-ate and ke-n-ate are slightly different. The former is the successive action while the latter is the reverse action. It means they must wash their hands before taking the food.


Anderson, Gregory D.S. and John P. Boyle (2002) ‘Switch reference in South Munda’, in M.A. Macken (ed.) Papers from the Tenth Annual Meeting of the Southeast Asian Linguistics Society 2000, Tempe: Program for Southeast Asian Studies Monograph Series, Arizona State University, 39–54.
Anderson, Gregory D.S. and Norman H. Zide (2002) ‘Issues in Proto-Munda and Proto-Austroasiatic nominal derivation: The bimoraic constraint’, in M.A. Macken (ed.) Papers from the Tenth Annual Meeting of the Southeast Asian Linguisticociety 2000. Tempe: Program for Southeast Asian Studies Monograph Series, Arizona State University, 55–74.
Bhaduri, M.B. (1931) A Mundari–English dictionary, Calcutta: Calcutta University Press.
Census of India 1991 (2004) Language Atlas of India 1991, General direction by Jayant Kumar Banthia . Delhi: Controller of Publications, Civil Lines.
Comrie, Bernard (1976) Aspect: An Introduction to the Study of Verbal Aspect and Related Problems, Cambridge: Cambridge University Press.
Cook, Walter A. (1965) A Descriptive Analysis of Mundari: A Study of the Structure of the Mundari Language According to the Methods of Linguistic Science, PhD Dissertation, Georgetown University, Ann Arbor: Microfilms International.
Croft, William (1991) Syntactic Categories and Grammatical Relations: The Cognition of Information, Chicago: The University of Chicago University.
Croft, William (1993) ‘Case marking and the semantics of mental verbs’, in J. Pustejovsky (ed.) Semantics and the Lexicon, Dordrecht: Kulwer Academic, 55–72.
Croft, William (2005) ‘Word classes, part of speech and syntactic argumentation’, Linguistic Typology, 9: 431–441.
Crystal, David (2003) A Dictionary of Linguistics and Phonetics, London: Blackwell.
Diffloth, Gérard (1976) ‘Expressives in Semai’, in P.N. Jenner , Laurence C. Thompson and Stanley Starosta (eds) Austroasiatic Studies Part 1, Honolulu: The University Press of Hawaii, 249–264.
Diffloth, Gérard (1994) ‘i: big, a: small’, in L. Hinton , J. Nichols and J. Ohala (eds) Sound Symbolism, Cambridge: Cambridge University Press, 107–114.
Emeneau, Murray B. (1980) Language and Linguistic Area, selected and introduced by Anwar S. Dil , Stanford: Stanford University Press.
Evans, Nicholas and Toshiki Osada (2005a) ‘Mundari: The myth of a language without word classes’, Linguistic Typology, 9:351–390.
Evans, Nicholas and Toshiki Osada (2005b) ‘Author’s response: Mundari and argumentation in word-class analysis,’ Linguistic Typology, 9:442–557.
Gumperz, John in collaboration with H.S. Biligiri (1957) ‘Notes on the phonology of Mundari’, Indian Linguistics Taraporewala Memorial Volume: 6–15.
Haldar, Rakhal Das (1871) ‘An introduction to the Mundari language’, Journal of Asiatic Society of Bengal, 40(1): 46–67.
Haspelmath, Martin , Matthew S. Dryer , David Gil and Bernard Comrie (eds) (2005) The World Atlas of Language Structure (=WALS), Oxford: Oxford University Press.
Hengeveld, Kees and Jan Rijkhoff (2005) ‘Mundari as a flexible language’, Linguistic Typology, 9: 406–431.
Hoffmann, John (1903) Mundari Grammar (=MG), Calcutta: Government Press.
Hoffmann, John (1930–1978) Encyclopaedia Mundarica (=EM), Patna: Government Press.
Kachru, Yamuna and Rajeshwari Pandharipande (1980) ‘Toward a typology of compound verbs in South Asian languages’, Studies in the Linguistic Sciences, 10(1): 113–124.
Klaiman, M.H. (1986) ‘Semantic parameters and the South Asian linguistic area’, in Bh. Krishnamurti (ed.) South Asian Languages Structure, Convergence and Diglossia, Delhi: Motilal Banarsidass, 179–194.
Langendoen, D. Terence (1963) Mundari Phonology. Unpublished paper.
Langendoen, D. Terence (1966) ‘The copula in Mundari’, in J.W.M. Verhaar (ed.) The Verb ‘Be’ and Its Synonyms, Dordrecht: Reidel, 75–100.
Langendoen, D. Terence (1967) ‘Mundari verb conjugation’, Linguistics, 32: 39–57.
Masica, Colin (1976) Defining a Linguistic Area: South Asia, Chicago: University of Chicago Press.
Munda, Ram Dayal (1971) ‘Aspects of Mundari verb’, Indian Linguistics, 32: 27–49.
Munda, Ram Dayal (1980) muɳɖārī-vyākaraɳ (Mundari Grammar in Hindi), Ranchi: Muɳɖārī sāhitya pariśad.
Mundu, M.M. (1995) muɳɖārī ʈuŋkoɖhāri (Mundari-Hindi-English dictionary), Ranchi: Catholic Press.
Nagaraja (1999) Korku Language: Grammar, Texts, and Vocabulary, Tokyo: Institute for Study of Languages and Cultures of Asia and Africa.
Nottrott, A. (1882) Grammatik der Kohl-Sprache, Gütersloh.
Osada, Toshiki (1991) ‘Notes on convergence features in the Chotanagpur area’, in S. Basu-Mullick (ed.) Unity in Diversity in Chotanagpur, Delhi: Uppal, 99–119.
Osada, Toshiki (1992) A Reference Grammar of Mundari, Tokyo: Institute for Study of Languages and Cultures of Asia and Africa.
Osada, Toshiki (1999) ‘Experiential construtions in Mundari’, in Gengo Kenkyu: Journal of Japanese Linguistic Society, 115: 51–76.
Osada, Toshiki (2007) ‘Reciprocals in Mundari’, in V. Nedjalkov (ed.) Reciprocal Construction, Philadelphia: John Benjamin, 1575–1590.
Pandharipande, Rajeshwari (1978) ‘Exceptions and rule government: The case of the passive rule in Hindi’, Studies in the Linguistic Sciences, 8(1): 153–173.
Peterson, John (2005) ‘There’s a grain of truth in every “myth,” or, why the discussion of lexical classes in Mundari isn’t quite over yet’, Lingistic Typology, 9: 391–405.
Pinnow, Heinz-Jürgen (1959) Versuch einer historischen Lautlehre der Kharia Sprache. Wiesbaden: Otto Harrassowitz.
Prasad, Swarnalata (1973) muɳɖārī-hindī śabdakoʂa. (Mundari–Hindi Dictionary in Devanagari Script), Ranchi: Bihar Tribal Research Centre.
Prasad, Swarnalata (1976) hindī-muɳɖārī śabdakoʂa. (Hindi–Mundari Dictionary in Devanagari Script), Ranchi: Bihar Tribal Research Center.
Schachter, Paul (1985) ‘Parts of Speech System’, in Timothy Shopen (ed.) Language Typology and Syntactic Description, Cambridge, 1:3–61.
Schmidt, Wilhelm (1906) Die Mon-Khmer Völker, ein Bindeglied Zwischen Völkern Zentralasiens und Austronesian. Archiv für Anthropologie, 5:59–109.
Singh, Udaya Narayan , K.V. Subbarao and S.K. Bandyopadhyay (1986) ‘Classification of polar verbs in selected South Asian languages’, in Bh. Krishnamurti (ed.) South Asian Languages Structure, Convergence and Diglossia, Delhi: Motilal Banarsidass, 244–269.
Sinha, N.K. (1974) Mundari Phonetic Reader, Mysore: Central Institute for Indian Languages.
Sinha, N.K. (1975) Mundari Grammar, Mysore: Central Institute for Indian Languages.
Suwilai Premsrirat (2002) Thesaurus of Kham Dialects in Southeast Asia, Nakhon Pathom: Institute of Language and Culture for Rural Development, Mahidol University at Salay.
Whitley, J.C. (1873) Mundari Primer, Calcutta: Government Printing Office.
Zide, Norman H. (1960) Korku Phonology and Morphophonemics, PhD Dissertation, University of Pennsylvania.
Zide, Norman H. (ed.) (1966) Studies in Comparative Austroasiatic Linguistics, The Hague: Mouton.
Zide, Norman H. (1978) Studies in the Munda Numerals, Mysore: Central Institute of Indian Languages.
Search for more...
Back to top

Use of cookies on this website

We are using cookies to provide statistics that help us give you the best experience of our site. You can find out more in our Privacy Policy. By continuing to use the site you are agreeing to our use of cookies.