1. General Overview
Kambera is a Malayo-Polynesian language spoken in the East of Sumba, an island in Indonesia. Kambera has a high dialectal variation among the 150,000 speakers and no standard variant is defined.
Kambera has a simple syllable structure. Syllables consist of at least one vowel, and onset consonant is optional. All syllables are open and roots have to be built up by exactly two of them. If there is an onset then it cannot be complex; the language allows only one initial consonant: C (V).
Kambera is a head-marking language. The subject and object(s) are marked on the verb by clitics that are cumulative exponents expressing number and person of the arguments as well as case. Only definite arguments trigger agreement on the verb and can be optionally realized, as pro-drop is possible in Kambera. All of the definite arguments are marked on the verb as long as no constraints on possible vcombination of two object clitics concerning their person are violated. Clusters of two third people object clitics; two times a first or second person object clitic, as well as a third person followed by a first or second person clitic is disallowed. All other sequences are grammatical.
The language distinguishes the four cases nominative, genitive, accusative, and dative. Exponents expressing nominative are realized in preverbal position as proclitics and exponents expressing one of the other cases are enclitic appearing in post verbal position. In addition, the verbs bear’s markers for aspect and mood that shows up between the verb stem and the arrangement marker for non-nominative exponents. Kambera has two numbers, singular and plural. In addition to first, second, and third person it distinguishes between first person inclusive and exclusive in the plural.
Table 2.1
Overview of All Clitics That Marked Arguments on the Verb
Nominative Genitive Accusative Dative
1st singular ku- -ŋgu -ka -ŋga
2nd singular (m)u- -mu -kau -ŋgau
3rd singular na- -na -ja -nja
1st incl. plural ta- -nda -ta -nda
1st excl. plural ma- -ma -kama -ŋgama
2nd plural (m)i- -mi -ka(m)i -ŋga(m)i
3rd plural da- -da -ha -nd a
By looking at these agreement markers is important to note that the forms get more and more complex from left to right. The dative markers consist of two syllables, the nominative and genitive forms of only one syllable and the accusative markers are mixed, only the first person exclusive and the second person plural are polysyllabic. The markers for nominative and genitive in the 1st person exclusive and 2nd person plural resemble the corresponding markers in the accusative and dative, so the first conclusion is that at least the dative and accusative markers are not only built up of one vocabulary item, but that vocabulary insertion has taken place several times.
Another striking fact is that apart from the third person plural the dative forms seem to be derived from their accusative counterparts by adding a nasal triggering assimilation of voice of the following consonant that itself triggers assimilation of the nasal concerning its place at the same time. All the dative markers begin with a nasal, fused to a prenasalised consonant if the following segment is a plosive or a semivowel. A similar situation occurs in the genitive, but here only first person singular and first person inclusive plural show the nasal and assimilation processes.
2. Clitics in Kambera Language
Kambera is a head-marking language; verbal arguments are commonly marked on the verb by pronominal clitics. The agent argument of a simple declarative sentence and the single argument of an intransitive predicate are commonly marked with a nominative proclitic, a patient object is communicably marked with an accusative enclitic.
1) Na tau wútu na- palu –ka nyungga
ART person be.fat 3sN-hit -1sA I
‘The big man hit me’
2) Na ài na- tambuta dàngu amung
ART wood 3sN- drop.out with root
‘That tree is uprooted’
The core function of a genitive clitic is to mark nominal possesseors, as in
3) Na ama-mu
ART father-2sG
‘Your father’
As a subject marker, the genitive is commonly used in syntactically embedded clauses; the relative clauses in (21) – (22) and the complement clause in the next example have genitive subject;
4) Nda ku- pi –a –nya na ngàndi-mu kuta
NEG 1sN-know-MOD -3sD ART take-2sG pepper.plant
‘I didn’t know that you would bring kuta ’
But also genitive subject in syntactically non-embedded clauses:
5) Ba meu-meu-na, ban na-imbu-nya
While RDP-roar-3Sg as 3sN-search-3sA
‘And it roared while it went after him’
This type of ‘nominal clause’ functions to provide the background information for the clauses that carry the main narrative. They express irrealis mood, which explains why they are often used in questions or in expressions of concession, foir example:
6) Hangu butang –butang –ma-a-na bai manila,
Straight.away RDP –pull out-EMP-just-3sG real peanut
Nda na- hili karai
NEG 3sN- again ask
‘He just began to pull out peanuts straight away, he didn’t even ask’
Though syntactically not embedded, nominal clauses are functionally dependent. This is also reflected in the fact that adverbials expressing tense, aspect, mood and degree have scope over them, as illustrated by hangu ‘straight away’ in (6) and lundu ‘until’ in (7), which are both obligatory followed by nominal clauses.
7) Lundu njili-nggu ba ku-yaulu-ya na wei
until be.tired-1sG as 1sN-chase-3sA ART pig
‘Till I got tired I chased the pig’
Kambera complements (including clausal ones) are cross-referenced on the verb with a pronominal object clitic. The regular form for patients/themes is an accusative clitic, as illustrated in (1) and (8).
8) Da- ngàndi –ya na uhu
3pN- take -3sA ART rice
‘They take the rice’
9) Da- ngàndi –nya na uhu i Ama
3pN- take -3sD ART rice ART father
‘They bring father the rice’
10) Jàka na-paàra-nya-ka nggi –ya-ka i Umbu
if 3sN-ask-3sD –PRF where-3sA-PRF ART Lord
‘If he asks him where the Master is, if he asks that….’
The default marker in these constructions is the quotative verb wà. It has always at least one pronominal clitic attached to it; a genitive enclitic that marks the speaker. The quote itself does not doffer from any other declarative clause in Kambera (neither in the morphological form of the verb, or in its mood, aspect or pronominal marking). Usually, there is no intonational break between the quote and the quotative verb. These following sentences are some illustrations. The quotative wà is glossed as ‘report’ in the examples below.
11) “Ku-ngangu –ma duku” wà –na-ma
1sN-eat –EMP EMP.1s report -3sG-EMP
“I have eaten (it)”, she said
12) “Budi mini-a nú” àmbu wà –nda-i!
new male-MOD DEI NEG.IRR report -1pG-ASP
‘Don’t say it’s just the young guys!’
13) Njadi u-u nda wà-na, ndia nda wà-na,
so yes NEG report-3sG NEG.emp NEG report-3sG
‘So he neither consents nor protests,
hi hi –bia –nanya –ka duna
RED –cry –just -3s –PRF EMP.s3
he just keeps on crying’
The verb wà is an intransitive root verb, and is mostly used in quotative constructions, though it can also function as a main speech verb:
14) Wà-nggu ba wà-na hama tu-na-I nú
Report-1sG while report-3sG be.same put-3sG-ASP DEI
‘I tell (it) as it was told’
Wà is a morphologically regular regular in that it can be derived with an applicative suffix –ng. This suffix licenses an additional addressee agreement: wà ‘report’ > wà-ng ‘report, to say X’. However, the other major Kambera word formation process, causativization, is related by wà, so we may say its morphological properties are reduced.
The applicative suffix –ng is not visible when the verb is inflected, but is part of its citation form. The addressee is commonly expressed by a dative object clitic. Illustrations with the applicative form wà-ng are:
15) E, wà –nggu –nya na ama –mu!
EXC report -1sG -3sD ART father -2sG
‘Hey, I was talking to your father!’
16) “Mài –kai -wa” wà –na-nggai
Come -2sA-HORT report -3sG –2pG
‘He says that you must come’ (lit. “You come, he tells you”)
Unlike ordinary complement clauses as the one in (4), quotes are not cross-referenced with clitics on the quotative verb, but are simply juxtaposed to the quotative clause. This is another indication that a quote is not a syntactic complement of the quotative verb. Sentence (17a) contains two quotes: “Kill the foal” and “Kill the foal, I said”. Neither can be cross-referenced on the quotative verb, as shown in (17b) and (17c).
17) a. Tobu –nya na ana njara wà-nggu
slaughter -3sD ART child horse report-1sG
ba wà –mi nú
as report -2pG DEI
‘”We’ll kill the foal ”, you (pl.) said’
b. [Tobu-nya na ana njara] wà-nggu -*nya …
slaughter -3sD ART child horse report-1sG -3sD
c. [Tobu-nya na ana njara wà-nggu] ba wà-mi -*nya ...
slaughter -3sD ART child horse report-1sG as report-2pG-3sD
Indeed, coordinating conjunctions such as hi ‘and, so’, ka ‘so that’, ba ‘and, as, while, because’ and jàka ‘if, when’ may always appear optionally between the quote and the quotative clause, also suggesting that they are two independent, coordinated clauses. In (18) the conjuction in the quotative clause is hi, in (19) ka and in (20) jàka. Note that the quotes themselves also contain initial conjunctions: ka, hi and hi, represently.
18) ‘Ai Umbu, ka nda u- mila-ngga nú eti’
EXC Lord so.that NEG 2sN-be.poor-1sD DEI liver
hi wà –na –nya
and report -3sG -3sD
‘”Oh sir, if you would take pity on me,” he said to him’
(Lit. ‘”Wouldn’t you have poor liver for me,” he said to him’)
19) ‘…hi na- ana hàmu na wài ngera-mu’
AND 3sN- DIM be.good ART water spirit-2sG
ka wà-da –du-nya-ka nú
so.that report-3pG –MOD-3sD-PRF DEI
‘”…so that you will fare well” (lit. so that your fate will be a little better), they said to him’
20) …hi nda rongu hàmu-bia-da-nya-i-ka uda
and NEG hear be.good-MOD-3pG-3sD-ASP-ASP EMP.3p
‘”…and they won’t hear it clearly once again”,
jàka wà-na-ka i Umbu Mbara, …
when report-3sG-PRF ART Lord Mbara
when Lord Mbara said that…’
Another important difference between Kambera quotes and true verbal complements is the fact that though the unmarked constituent oreder in Kambera is verb-object, without exception quotes precede the quotative verb (i.e. represent OV word order).
In addition, the fact that Kambera quotes may appear without the quote verb wà also suggest that they are not verbal complements but embedded at the discourse level.
Evidence for the non-transitive status of wà is that this verb, unlike normal transitive verbs, cannot appear in object relativizations. Object relativizations in Kambera are marked by the morpheme pa-. paients/themes and beneficiaries/addressee/recipients undergo the same relativization. Below this is illustrated for the theme of ngàndi ‘take something’ and the recipient of ngàndi-ng ‘take something for someone’:
21) Na nggula na pa- ngàndi-nggu
ART sugar ART RMO- take 1sG
‘The sugar that I took (along)’
22) Da makaweda da pa- ngàndi-nggu nggula
ART old.woman ART RMO- take.to-1sG sugar
‘The old ladies whom I brought the sugar’
Relativatizations are standardly used in quotation,. As in (23), and may function like passive, as in (24) and (25).
23) Nggàra pa- ngàndi-mu, Rambu?
What RMO- like-2pG Lady
‘What did you bring, ma’am?’
24) a. [Nggula [pa- ngàndi-naj]] –nya
sugar RMO- take-3sG -3sD
‘Itk (is) sugar that (is) brought by herj’
b. [Pa- ngàndi (-naj)] -nyak
RMO- take (-3sG) -3sA
‘Itk is brought (by herj)’
If we now consider the quotative verb wà, we find that this verb vever appears in object relativization, or in such passive-like structures:
25) *[Pa- wà -mu] –nya
RMO- report -2sG -3sD
‘It is said (by herj)’
Does this menas that a Kambera speaker cannot question what is being said? The answer is negative: in questions about the content of a quote, a bare form of wà is used – the object relative marker pa- is absent:
26) Nggàra wà –mu, Rambu?
what report -2pG Lady
‘What did you say/think, ma’am?’
Interestingly, the same construction is used to question adjuncts of wà, such as pira ‘how much’, and nggiki ‘how, in what way’:
27) “Pira wà-mu-nja mú?” “Ana hau ndui” wà- na
How.much report-2sG-3sG-3pD DEI DIM one money report-3sG
“How much did you ask from them?” “Just one coin”
28) Nggiki wà -nggu ba ku- karai –nya?
how report -1sG- while 1sN- ask 3sD
‘How should I ask him?’
29) Nggiki wà –mu?
how report -2sG
‘How do you think about it?’
The absence of the relative marker in these questions in expained when we assume that the questioned elements are all adjucts: like pira ‘how much’, and nggiki ‘how’, nggàra ‘what’ does not question an argument but an adjunct of wà.
Interestingly, there are indication that wà with a third person singular genitive marker may become reanalyzed as a monomophemic verbal root. The motivation for this reanalyzed comes from the prosodic structure of the verbal root wa. Phonologically, this root does not comply with the minimal word requirements of Kambera, which state that the roots of content words must be bimoraic feet. The sequence wà-na, however, is a bimoraic foot, and is therefore a good candidate to become analyzed as a verbal root. In this reanalyzed the genitive clitic loses its referential function, so the subject must be marked otherwise. The nominative may take over the job:
30) Ba na- wàna –nya i Darius ama -na…
when 3sN- report -3sD ART Darius father -3sG
‘When he told his father Darius…’
Evidence that wà-na is treated as a morphological unit comes from the relative position of emphatic clitics and adverbs. Normally emphatic clitics such as –ma and –du must precede the genitive clitic, see (33) (and (19)-(20). With wà, however, the order is obligatory reserved, as t5he grammatical contrast in (34ab) shows:
31) E, ba namu –ma –na –nya na ana njara
EXC as remember –MOD-3sG-3sD ART child horse
‘Hey, he loves the foal’
32) a. * “Na-palu-ka i Ina” nda wà –na –ma –nya –i
3sN-hit-PTF ART mother NEG report -3sG-3sD-ASP
‘He never tells him that his mum hit him again’
