New strategies for relative clauses in Azerbaijani and Apsheron Tati1
|
|  Yüklə 0.74 Mb.
|
.4ConclusionIt is well known that relativizing strategies can be copied, from one language to another, even across linguistic families. For instance, all Kashkay dialects of Southern Iran, though genetically South-West Turkic like Azerbaijani,15 have lost most of the preposed participial strategies typical of the Turkic family and have replaced it with the strategy found in surrounding – and genetically unrelated – Persian, which involves a general subordinating morpheme ki and a pronominal anaphora for all relativized positions other than subject and object. This seems not to have happened in Northern Azerbaijani dialects. Indeed Persian and standard Azerbaijani use two very different strategies to form relative clauses. In Northern Azerbaijan, it is rather the Iranian language Tati that has extended the use of participles under the influence of the dominant Azerbaijani language. The greater capacity of Turkic participles to clarify syntactic roles within the RC certainly accounts for the fact that, unlike Indo-European languages, most Turkic languages have not developed finite relative clause strategies, and when such finite RC occur, they are clearly a result of language contact. Apart from the case of Kashkay mentioned in the introduction, Gagauz and Balkan varieties of Turkish have also copied RC from surrounding Slavic languages.16 Azerbaijani has a finite RC-forming strategy, in which the relative clause contains the head noun, is headed by a finite verb-form, and precedes the matrix clause. While not very widely used, it can be found in all sorts of Azerbaijani texts, as the examples given above have shown. In both the Azerbaijani and Tati head-internal relative clause constructions, the relative clause 1) is preposed to the matrix clause, 2) has its domain noun embedded and marked for its case / function in the relative clause; and 3) the latter is preceded by an inflected interrogative pronoun (‘which’). It is the matrix clause which has either zero anaphora (if the head-noun inside it is subject, transitive or not, or object) or a resumptive pronoun marked for case (for functions other than subject or object). I believed at first that this new head-internal strategy had been copied by Azerbaijani from Tati, because in the past, in all likelihood, Tati, not Azerbaijani, was the dominant language, and the emergence of a new strategy may have been motivated, in Tati only, by the disappearance of the active participle and the need for a new strategy to relativize transitive subjects. But Tati participles have lost their original (Indo-European) S/P or “passive” orientation, and they are now used to relativise transitive subjects as well as other functions. This may have resulted from the influence of East-Caucasian ergative dependent-marking languages with non-oriented participles, but such a development would be far-fetched for the Apsheron dialect we have dealt with, which has not been in contact with any language other than Azerbaijani. Azerbaijani does have, along with the two complementarily oriented participles in –(y)An (S/A or ‘agentive’) and -DIQ- (non agentive), a perfect participle in -mIş (S/P or ‘patientive’): all these options to relativise syntactic roles are conflated in the syntactic behaviour of the only participle of Tati. Only the lowest syntactic function on the accessibility scale prompted the other strategy as the only grammatically correct choice. However, the head-internal relative clauses have extended their domain: since they preserve the same structure as independant clauses, they can be used for any non-factual restrictive relative clause modifying a topicalized argument. While most relative clauses belong to the presupposed part of a sentence , Azerbaijani and Tati head-internal RCs entail outright topicalization. They signal high topicality of the head noun and the subsequent matrix clause is in focus, while the preposed RC is usually presupposed (most often in the Azerbaijani conditional or Tati eventual form). And contrary to most previously described head-internal relative clauses, 1) a case marker signals the relativized function, 2) they are used to relativize any syntactic functions of the well-know accessibility scale, but probably they are most widely used for Agents (transitive subjects), dative and possessor functions, which are naturally most topical. Abbreviations
ReferencesIzgijaeva, E. 2005 Tatsko-Russkij slovar’, Maxachkala Authier, G. 2009 Grammaire kryz, Paris in preparation Le judéo-tat Keenan, E. & Comrie, B. 1977 Noun phrase accessibility and universal grammar. Linguistic Inquiry 8.63-99. Haiman, J. 1978 Conditionals are topics In Language, 54, 3, 564-89 Gorçu T. 2008 Səhrli nağıllar, Baku Grjunberg A. L. 1963 Grünberg, A. 1963. Jazyk severoazerbajdzhanskix tatov. Moskva: Nauka Hale, K. 1976 The adjoined relative clause in Australia, In Grammatical categories in Australian Languages, ed. by Dixon R., Camberra Lazard, G. 1975 La catégorie de l'éventuel In Mélanges linguistiques offerts à E. Benveniste, Paris, Peeters. p. 347-358. Matras Y. & Tufan, S. 2008 Contact-induced change in Macedonian Turkish. In: Grammatical Borrowing in Cross-Linguistic Perspective. Edited by Yaron Matras and Jeanette Sakel. Mouton, Berlin Samvelian, P. 2006 L’enclitique –i introducteur de relative en persan : déterminant, allomorphe de l’ezâfe ou autre chose encore ? , Studia Iranica 34/2, 7-34.
1973 Focus and relativization In Language 49, 19-46. 1 The data for this article are either drawn from Azerbaijani published texts, most of them available on the internet (reference after each example) or from data collected by myself (field notes). Field trips were funded by the Centre National de la Recherche Scientifique and the Institut National des Langues et Civilisations Orientales. I wish to thank my main informant Mrs Naiba Mammadova for sharing with me her knowledge of such a rare language as Tati, and Mr Elvin Abbabeyli for the Azerbaijani elicited examples. I am also indebted to V. Gast for improving my English and to him and G. Lazard for valuable remarks. 2 cf. Izgijaeva 2003. 3 Original literary works, translations, primers, and a newspapers have been published in Derbent from the early 20’s. 4 cf. Authier, in preparation. 5 cf. Comrie & Keenan 1977. 6 Hereafter SN. 7 This would be totally incorrect in standard Turkish. 9 cf. Hale 10 The name Tati is sometimes applied (especially by Yarshater and Stilo) to residual North-West Iranian languages of Southern (Iranian) Azerbaijan; we choose to stick to the European tradition and will call Tati only the language spoken in the North-Caucasus and the North-Eastern Part of the Republic of Azerbaijan. 11 Note that the ezâfe is found in Tati only on words ending with a vowel. 12 Recall that there is no ezafe on words ending with a consonant ! 13 See of course Haiman 1978. 14 cf. Lazard 1975 15 There is an almost perfect mutual intelligibility between Azerbaijani and Kashkai speakers. I tested this personally by submitting recordings to both audiences. Northern Azerbaijani people understand in fact better Kashkai than ‘Stambuli’ Turkish. 16 cf. Matras & Tufan 2007 |