Conferences : World Congress of KURDISH STUDIES : Sandrine TRAIDIA


Section PRESSE
World Congress of

Irbil, 6-9 September 2006

Organized by the Kurdish Institute of Paris in partnership with
Salahadin University (Irbil) and with the support of the
Kurdistan Regional Government and of the
French Ministry for Foreign Affairs

The Syntaxic Status of the Noun in (N-V) Kurdish compound verbs

Par Sandrine TRAIDÏA (*)

This work focuses on kurdish (N-V) compound verbs where N is a noun and V a simple transitive verb, which is a very effective and productive way of verb formation in sorani kurdish. The kurdish compound verbs do not show homogen syntactic properties : we can list different syntactic constructions.
For instance, here are four sentences containing a compound verb :

(1) Hawîn, Dara zor areq deka
    (summer / Dara / a lot / sweat / makes)
    'Summer, Dara sweats a lot'
[No N V]
(2) Awaz arezû-y xanû deka
    (Awaz / hope of house / makes)
    'Awaz wants a house'
[No N-Ez-N1 V]
(3) Karwan hez le penîr deka
    (Karwan / liking / of / cheese / makes)
    'Karwan likes cheese'
[No N Prep N1 V]
(4) Nîgar qij-î kiç-eke şane deka
    (Nigar / the girl / comb / makes)
    'Nigar combs the girl'
[No N1 N V]

Each of this sentences illustrates a particular syntactic construction of compound verb. These different syntactic behaviours have been listed in square brackets.
The perspective adopted in this search is a syntactic one, our purpose is to know if, besides the semantic unity which is one of the caracteristics of the compound verbs, there is also a syntactic unity between its N and V elements.
Indeed, compound verbs show semantic unity, they've got a non-compositional meaning, that is to say the compound verb meaning is not the sum of the N meaning plus the V meaning as the following scheme shows :
(N-V) meaning ? N meaning + V meaning

Can we speak of syntactic unity about this verbs ? These (N-V) compound verbs do they act as a single word, as a lexical unity or as a VP verbal phrase (V + N d.o.) ?
Is N a d.o. of the verb or does it make part of a " frozen (N-V) block " in which it is amalgamed to the verb and doesn't have any specific status ?
As we know, prototypic kurdish d.o. allow some syntactic manipulations having respectively its specific interest, between them : compatibility with the indefinite suffix -êk 'a' or with the definite suffix -eke 'the'), relativation, replacement of N by a pronoun and passive manipulation (when an active sentence containing a nominal d.o. is passived, the prototypical nominal d.o. generally becomes the subject of the correspounding passive sentence).

Let's now confront our different types of compound verbs with these tests :

I)Type (1) : [No N V]

(1) Hawîn, Dara zor areq deka
'Summer, Dara sweats a lot'

a) Compatibility with -eke and -êk determiners :
*Hawîn, Dara zor areq-êk deka
*litt. (summer / Dara / a lot / (a /one) sweat / makes)
*Hawîn, Dara zor areq-eke deka
*litt. (summer / Dara / a lot / the sweat / makes)

b) Relativation :
*Hawîn, areq ke Dara deka zor e
*litt. Summer, the sweat that Dara makes is much

c) Replacement of N by a pronoun :
Hawîn, Dara zor areq deka. *Dara zor de-y-ka
litt. 'Summer, Dara sweats a lot. * Dara makes it a lot.'

d) Passivization :
*Hawîn, zor areq dekirêt.
*litt. 'Summer, a lot of sweat is made'.
This passive sentence is not semantically correct. It is not possible to passivate the compound verb areq kirdin : it functions as a syntactic " frozen " block.

Conclusion : The syntactic nominal d.o. manipulations listed before are not possible on N.

II)Type (2) : [No N-Ez-N1 V]

(2) Awaz arezû-y xanû deka

'Awaz wants a house'

a) Compatibility with -eke and -êk(-yek) determiners :

Awaz arezû-y xanû-ke deka

'Awaz wants the house'

Awaz arezû-y xanû-yek deka

'Awaz wants a house'

b) Relativation :

?arezû ke Awaz deka kon e

?'The wish that Awaz makes is old'

c) Replacement of N by a pronoun :
Awaz arezû-y xanû deka. * Awaz de-y-ka.
'Awaz wants a house. * Awaz makes it.'

d) Passivization :
*arezû dekirêt
*litt. 'wish is made.'

Conclusion : N is compatible with determiners (definite and undefinite suffix -eke and -yek) and it seems to be possible to proceed to the relativation of N.

III)Type (3) : [No N Prep N1 V]

(3) Karwan hez le penîr deka
'Karwan likes cheese'

a) Compatibility with -eke and -êk(-yek) determiners :
*Karwan hez-eke le penîr deka
*litt. 'Karwan the liking of cheese makes.'
*Karwan hez-êk le penîr deka
*litt. 'Karwan a liking of cheese makes'

b) Relativation :
* hez ke Karwan le penîr deka taze ye
* litt. 'liking that Karwan makes of cheese is recent'

c) Replacement of N by a pronoun :
Karwan hez le penîr deka. *Karwan de-y-ka.
litt. 'Karwan likes cheese. *Karwan makes it'

d) Passivization :
* hez le penîr dekirêt.
*litt. 'liking of cheese is made'

Conclusion :
N does not allow any of the syntactic nominal d.o. manipulations listed before.

IV)Type (4) : [No N1 N V]

(4) Nîgar kiç-eke şane deka
'Nigar combs the girl'

a) Compatibility with -eke and -êk(-yek) determiners :
* Nîgar kiç-eke şane-ke deka
* litt.'Nigar the girl makes the comb'
* Nîgar kiç-eke şane-yek deka
* litt.'Nigar the girl makes a comb'

b) Relativation :
*şane ke Nîgar deka biçûk e
*litt. 'the comb that Nigar makes is small.'

c) Replacement of N by a pronoun :
Nîgar kiç-eke şane deka. *Nîgar kiç-eke de-y-ka
'Nigar combs the girl. *Nigar the girl makes it.'

Conclusion : It seems that N can't be subjected to this operation.

d) Passivization :
The correspounding passive sentence is :
kiç-eke şane dekirêt.
'The girl is combed.'

Conclusion : The N şane does not become the subject of the correspounding passive phrase, but it remains " amalgamated with the verb ", making a block with it as also remarks Haig (2002 :25) for the Ns of certain kurmancî kurdish compound verbs.

Remark : In addition, this group of verbs illustre another phenomenon : the presence of two " potential " direct objects 'kiçeke' and 'şane' in the same sentence. Or the standard kurdish canonic phrase only contains one direct object ! We will come back later to this interesting phenomenon.

The kurdish compound verbs differently react to this battery of tests, some of them do not satisfy any of them, others can satisfy a few of them : this is a case of arezû kirdin which seems to be compatible with determiners and which can be subjected to relativation.
Some compound verbs seem to be more 'frozen' than others : there is no homogeneous way of syntactic behaviour between the Ns of compound verbs. A possible syntactic analysis for N would be to consider it as a d.o. of the verb, but with a special status, as a particular d.o. which doesn't have all the properties of prototypical direct objects : we could speak as Samsam Bakhtiari (2000) of atypic or non-prototypic d.o.
Let's turn to our data : as we have already seen, the following sentence :

(4) Nîgar kiç-eke şane deka
'Nigar combs the girl'
lightlights a new phenomenom : the " cohabitation " of two d.o. (or two candidates for this status) in the same sentence :

(4) Nîgar kiç-eke şane deka
(No = subject) (N1 = d.o.?) (N = d.o.?) (V)

The question induced by this sentence is : does N still remain d.o. of the verb when it is in competition for this status with another noun N1? And what about the syntactic status of N1 ?
Let's recapitulate this point :
1) We have seen that the N of this sentence şane " comb " do not satisfy any of the prototypical kurdish d.o. properties.
2) We have in the same sentence two potential candidates in competition for the d.o. status : N (şane) and N1 (kiç-eke).
Let's confront N1 to our d.o. tests in order to see its reactions :
(4) Nîgar kiç-eke şane deka

a) Compatibility of N1 with -eke and -êk(-yek) determiners :
Nîgar kiç-eke şane deka
'Nigar combs the girl.'
Nîgar kiç-êk şane deka
'Nigar combs a girl.'

b) Relativation of N1:
ew kiç-e ke Nîgar şane deka cuwan e.
'The girl that Nigar combs is beautiful.'

c) Replacement of N1 by a pronoun :
Nîgar kiç-eke şane deka. Nîgar şane de-y-ka
'Nigar combs the girl. Nigar combs her.'

d) Passivization of N1:
kiç-eke şane dekirêt.
'The girl is combed.'

The NP (kiç-eke) is the subject of the passive corresponding sentence : (kiç-eke) is the d.o. of the active sentence.
Conclusion : N1 satisfies this selection of nominal d.o. tests : N1 is the d.o. of the sentence.
What about the syntactic status of N ?
At this point, 3 ways of analysis are possible :

1) Both N1 and N are d.o. but there is a sensible difference between them : N1 is the " prototypic " d.o. of the sentence thereas N is a non-prototypic one.
However, we can't accept this analyse because it would mean that two d.o. can " cohabite " in the same sentence, so that the d.o. relation can repeat in kurdish sentence or we have seen it is not the case : kurdish canonic sentence only accept one d.o.
2) N1 is the prototypic d.o. of the sentence and N is not a d.o., moreover, N is syntactically incorporated, it hasn't any syntactically status.
However, we can't accept this hypothesis of nominal incorporation to analyse this phenomenon because we can't say that N is totally prived of syntactical status. Indeed, the fact that N can be in some cases, a support for functional affixes has led us to reject the syntactic incorporation hypothesis to analyse this phenomenon :

    Nîgar kiç-eke şane-y kird.
    (Nigar / the girl / comb-Clit.3s / Make-Pret.3s)
    'Nigar combed the girl.'

3) The relation between N and the verb is something different : we have seen that it is neither a d.o. relation, nor a syntactic incorporation.
In fact, it is something between the prototypical d.o. relation and the syntactic incorporation. I will speak to definite this status of « quasi syntactic incorporation ».

References :
Alonso Ramos, Margarita, 1998, Étude sémantico-syntaxique des constructions à verbe support, Thèse de Doctorat, Université de Montréal.

Haig, G., 2002, Complex predicates in Kurdish : Argument sharing, incorporation, or what? Sprachtypologie und Universalien forschung / Language typology and universals 55(1) , 15-48.

Karimi-Doostan, Gh., 1997, Light Verb Constructions in Persian. Ph.D. dissertation, Essex University, England.

Samsam Bakhtiari, Fatemeh, 2000, La composition dans la relation verbe-objet: problématique générale et application au persan, Thèse de doctorat sous la dir. de Denis Creissels, Université Lumière (Lyon).

(*) Linguist and Ph.D. student