2

Many languages and dialects still spoken in Iran have no certain status and could become extinct. These languages must be preserved before it is too late. The main problem is that these languages and their endangerment status have not been sufficiently studied. There is little or no information available about which languages are actually endangered and to what extent. The first important step is to identify and locate these languages. The purpose of International Symposium on Endangered Iranian Languages (ISEIL) is to provide an opportunity for scholars coming from different disciplinary backgrounds to discuss the challenges, problems and new developments regarding documenting and encourage fieldwork on these languages and dialects. ISEIL aims also to develop an Atlas of endangered Iranian languages. Scientific committee SaloumehGHOLAMIGoetheUniversityPolletSAMVELIANUniversitéSorbonneNouvelleCNRS-UMR“Mondesiranienetindien”GeoffreyHAIGBambergUniversityAgnesKORNCNRS-UMR“Mondesiranienetindien”JohnnyCHEUNGINALCOCNRS - UMR“Mondesiranienetindien”MatteoDICHIARAINALCOCNRS-UMR“Mondesiranienetindien”

Organizing team SaloumehGholamiGoetheUniversityPolletSamvelianUniversitéSorbonneNouvelleCNRS-UMR“Mondesiranienetindien”OphélieGANDONUniversitéSorbonneNouvelleCNRS-UMR“Mondesiranienetindien”EmmanuelGIRAUDETCNRS-UMR“Mondesiranienetindien”FarzanehZAREIECNRS-UMR“Mondesiranienetindien”

3

PROGRAM

July 8

9:00 Welcome

9:30 - 10:00 Head-marking and dependent-marking patterns in Western Iranian languages

Masoud Mohammadirad & Mohammad Rasekhmahand (Bu-ali Sina University, Hamedan)

10:00 - 10:30 Inflectional Categories of Verb in Sorkhi Dialect

Zahra Karimi Bavarani, Negar Davari Ardakani & Parsa Bamshadi (Shahid Beheshi University,Tehran)

10:30 - 11:00 Pronominal clitics in Zoroastrian Dari (Behdīnī) of Kerman

Saloumeh Gholami (University of Frankfurt)

11:00 - 11:30 Coffee Break

11:30 - 12:00 Steps being taken to reverse language shift in the Wakhi language in Tajikistan

Jaroslava Obrtelova & Raihon Sohibnazarbekova (Uppsala University & Russian-Tajik Slavonic University)

12:00 - 12:30 Argument indexing in Yaghnobi from an Iranian perspective

Adam Benkato & Ergin Öpengin (Berlin-Brandenburg Academy of Sciences and Humanities & University of Bamberg)

12:30 - 13:00 Bezeynî – An Endangered Kurdish Language

Cemile Celebi (University of Frankfurt)

13:00 - 14:30 Lunch

14:30 - 15:00 Forms and Meanings of the Ezafe in Zazaki

Brigitte Werner (SIL International)

4

15:30 - 16:00 Non-Canonical Subject Construction as a Relict in Iranian Languages: Typological and Diachronic Implications

Mohammad Dabir-Moghaddam (Allame Tabataba’i University)

16:00 - 16:30 Coffee Break

16:30 - 17:00 Anbarāni and Lankarani dialects of Talyshi: a comparative analysis of verbal systems

Hakob Avchyan (Yerevan State University)

17:00 - 17:30 The Sociolinguistic perspectives of Endangered Minority (non)Iranian Languages in Northwest Iran

Hiwa Asadpour (University of Frankfurt)

17:30 - 18:00 Dikin Marāqei of Alamut: an undocumented conservative Tati language

Donald Stilo (Department of Linguistics, Max Planck Institute for evolutionary Anthropology Leipzig)

20:00 Dinner

5

July 9

9:30 - 10:00 Participial Relative Clauses in Şirvan Tat

Murad Suleymanov (École pratique des hautes études)

10:00 - 10:30 Considerations in filming the languages: a means for language preservation and maintenance

Barakatullo Ashurov & Matt Merritt (Rudaki Institute of Language, Literature, Oriental Studies and Written Heritage of the Tajikistan Academy of Sciences)

10:30 - 11:00 Mapping Cross-reference systems

Thomas Jügel (Labex EFL Paris, CNRS - UMR “Mondes iranien et indien”)

11:00 - 11:30 Coffee Break

11:30 - 12:00 Kanduleh Hawrami – report about an ongoing documentation project

Zaniar Naghshbandi & Ludwig Paul (University of Hamburg)

12:00 - 12:30 ELDP and endangered Iranian languages

Mandana Seyfeddinipour (SOAS, University of London)

12:30 - 14:00 Lunch

14:00 - 16:00 Workshop “Special Session on Language Mapping”: Mapping endangered languages of Iran

Erik Anonby & Amos Hayes (Linguistics, Carleton University & Geomatics and Cartographic Research Centre, Carleton University)

16:00 - 17:00 Closing Discussion

6

ABSTRACTS

Hiwa ASADPOUR (University of Frankfurt)

The Sociolinguistic perspectives of Endangered Minority (non)Iranian Languages in Northwest Iran

The current study is devoted to the problems of endangered languages, in particular those spoken by minorities in northwest Iran. By northwest, I refer to Azerbaijan Qarbi which is the most conflict-ridden region of Iran and one of Iran's most multicultural, multi-religious and multi-ethnic areas. It achieved a strategic key role due to its borders with several countries. The involvement of varied territorial, ethnic, religious and political conflict formations with the internal and external economic and geo-strategic situation of the region had a great influence on languages and cultures in the this region.

Left map: Azerbaijane Qarbi (central part of province is Urmia region) Right map: position of Azerbaijane Qarbi in Iran © Hiwa Asadpour

The region has been overlooked in linguistic studies and there is little documentation of the languages in this region for example Öpengin (2013), Asadpour (2011), Khan (2008). There are also several studies which describe linguistic features of languages in the area like for Kurdish Mann (1906), MacKenzie (1961), Minorsky (1957) to name a few. So far, with the existing literature, there is no comprehensive language documentation and language mapping of the region. This paper will focus on these objectives (i) to write a community profile of the Azerbaijan Qarbi languages, where Kurdish Sorani, Kurdish Kurmanji, Azeri, Neo-Aramaic Jewish, Neo-Aramaic Christian and Armenian are spoken in past and present, (ii) to report on the diminishing vitality of endangered languages, the sociolinguistic aspects of the causes, circumstances and results which led to endangerment and attrition in the languages of this area. In this objective, I will point to the previous and current positive or negative impact of policies

7

and institutional intervention in the loss or survival of endangered minority languages (EMLs) in Azerbaijan Qarbi and (iii) to point out possible ways to revitalize these languages. In multilingual situations the change in language functional roles are very important to be studied. These functional roles can help the revitalisation of an endangered language. In addition, the linguistic attitudes of speech communities are also a significant parameter. These attitudes encompassed historical and cultural constructions which directly relate to the EMLs prestige. The sociocultural and ethnic contexts, as well as sociolinguistic indicators of language use and attitudes of the speakers, have been at the center of this research. This study is based primarily on interviews I conducted with native speakers of the languages during my fieldwork in several years. The main aim of field work was documenting and description of languages of Azerbaijan Qarbi and providing the linguistic Atlas of this region.

Key Words: Endangered Languages, Azerbaijan Qarbi, Revitalisation

References Asadpour, Hiwa. 2011. The Computer Developed Linguistic Atlas of Azerbaijan-e Qarbi: Notes on

Typological-Perceptual approaches in Geolinguistics. MA thesis, Science and Research branch, Islamic Azad University, Tehran, Iran.

Khan, Geoffrey. 2008. The Jewish Neo-Aramaic dialect of Urmi. Piscataway: Gorgias Press. (Gorgias Neo-Aramaic Studies 2.) Mann, Oskar. 1906. Die Mundart der Mukri-Kurden. Berlin: Georg Reimer.

Minorsky, V. 1957. Mongol place-names in Mukri Kurdistan. Bulletin of the School of Oriental and African Studies 19(1). 58-81.

MacKenzie, David N. 1961. Kurdish dialect studies I. Oxford: Oxford University Press. Öpengin, Ergin. 2013. Clitic/affix interactions: A corpus-based study of person marking in the Mukri

variety of Central Kurdish. Doctoral thesis, Université de Paris III Sorbonne Nouvelle. Barakatullo ASHUROV & Matt MERRITT

(Rudaki Institute of Language, Literature, Oriental Studies and Written Heritag of the Tajikistan Academy of Sciences)

Considerations in filming the languages: a means for language preservation and maintenance

Tajikistan is home to several ‘remnant’ Iranian languages, which are designated after the geographical zones where the speakers of these languages live. One of the major groups is that Badakhshani, or commonly known as Pamiri languages and are divided into the Shughni-Rushni language group consisting of closely related languages: Shughnani, Rushani, Bartangi, Roshorvi, Khufi, and Sarikoli. This language group is also related to Yazghulami, Sarikoli, Wakhi, Ishkashimi, Sanglechi, and Zebaki. Pamiri languages have very wide geographical zone encompassing Tajikistan, Afghanistan, Pakistan, and China. Another important eastern Iranian language is Yaghnobi, whose native speakers originate from the Yaghnob Valley in the Northern Tajikistan, but today Yaghnobi speakers are dispersed in many different regions of north, central Tajikistan, including the capital city.

History of the study of languages of Tajikistan commenced in the late 18th and in 19th centuries. However, the systematic research on these languages was carried out after the annexation of the Central Asia to the Russian Empire and during the Soviet era, resulting in multiple academic publications. However, considering current socio-cultural and economic changes, all the existing material needs to be reviewed and reexamined against current situations. As of today each of these languages has been designed varied degrees of endangeredness and has been included into the UNESCO Atlas of the World's Languages in

8

Danger. Considering this new status and categorization the academic community together with the members of these language communities ought to apply new methods and approaches in language documentation, maintenance and creating accessible data sources, which in turn would help current and future generation of speakers of these endangered eastern Iranian languages. It can be said, according to contemporary methodologies and funding paradigms that historically language documentation has been primarily carried out for the benefit of the linguistic community. However, there is a shift towards language communities interacting with that same language data, even though it often was not intended for them nor is available in a format that they can access. It is important to correct this approach to language documentation and to understand the cognitive and psychological effects of language communities seeing themselves through the perspective of outsiders.

This paper seeks to provide evidence and suggest within a research paradigm that the narrative film is the highest form of language and cultural documentation and provides the greatest return to the language community. The most prevalent of this type of interaction could be found in the narrative form that is told within the sphere of the community, the actors of which are identifiable to the language community. The authors have conducted a study of a narrative film told within the sphere of the community, by observing the effects and influence of that level of documentation.

Especially, attention was given to evaluate the effect of the documentary in an endangered minority language in the following levels:

1) Social linguistics – How do people interpret recorded speech, in this case within the dialogue of the narrative, since it is no longer oral, but is now archived. How do people perceive intonation, grammar and discourse structures in speech acts contained in narrative films. This study would be done through a focus group, comprised of two sub-groups watching a short narrative film and have them interact with a facilitator. The subgroups would be 1) a displaced group, use of mother tongue is sporadic and heavily influenced by regional language 2) a group that uses the mother tongue language almost exclusively. 2) Psychological linguistics – how do people perceive themselves, when seeing themselves acting in a role that is different than real life. How does this apply to kinship relationships, to dialectical boundaries and selfperception. This study is done through the actors watching the short narrative film and then interacting with a facilitator about their thoughts and feelings. How do the actors feel that the film, and their part in it is showcased, nationally and possibly globally. Since this community is a minority language community, how do they feel that this is being shown in a context that might be antagonistic towards them? Does this cause any consideration for the process of informed consent…. 3) Anthropological – How does the filming of the narrative cause the community that was filmed to see themselves, within the larger language community as well as nationally. Does the film cause a rise or fall of identity? Does the lens of the camera increase, decrease the significance of a community’s worth? How does the village community feel after knowing that the film is being shown in other villages within the same language community? This study is accomplished through considering the tangible ways in which a community is changed, once the camera has captured a point in time and archived it for the community to watch, discuss and embrace or cast aside. How does the community refer to the literary heritage, as contained in the narrative film as opposed to the oral heritage? Do community members begin to refer to the archived literature, as a form of influence on ongoing generations? These are questions that cannot immediately but should be considered in the scope of a narrative film strategy for language documentation.

9

Hakob AVCHYAN (Yerevan State University)

Anbarāni and Lankarani dialects of Talyshi: a comparative analysis of verbal systems

Talyshi, a Northwestern Iranian language, is traditionally divided into three dialectal clusters (Northern Talyshi, Central Talyshi, Southern Talyshi) based on phonological, grammatical and lexical phenomena [Bazin 1980; Stilo 1981; Yarshater 1996]. Although Northern Talyshi is mainly spoken in the Republic of Azerbaijan (in Lankaran, Lerik, Masally, Yardymly regions [Clifton, Tiessen, Deckinga, Lucht 2005: 3]), some its dialects can be found in the neighboring regions of Iran, in the provinces of Gilan and Ardabil. The Anbarāni dialect of Talyshi is spoken in Anbarān district of Namin county, Ardabil province. Though it belongs to the Northern group of Talyshi dialects, Anbarāni demonstrates some particularly local characteristics distinguishing itself from the Lankarani dialect which is also a part of the same dialectical group.

This paper, which is based on the field-work carried out in Minābād and Aminjān (now a part of Anbarān city [Bazin 2012]) villages of Anbarān districts, aims to analyze the differences between the verbal systems of Anbarāni and Lankarani1.

While Anbarani and Lankarani dialects can be distinguished from Central and Southern Talyshi dialects with their common features in the verbal system (for example, the absence of progressive forms which are found in Central and Southern dialects, or the collapse of the two-stem based verbal system in Anbarāni and Lankarani, while Central and Southern dialects preserve the two-stem based pattern), they also show some strict differences within the same verbal structure. For example, though both dialects tend to lose Iranian two-stem verbal paradigm, in two dialects the same verb can show different development (doe – “to give” have two stems in Lankarani` da- (present stem) and do- (past stem) [Pireyko 1976: 86], while it has preserved only the past stem (do-) in Anbarāni, etc.).

The next difference in the verbal system of two dialects is found in the formation of the present tense of the indicative mood: in Lankarani it is formed from the infinitive, by adding a postposition and person suffixes [Miller 1953: 144], while in Anbarāni the present stem of the verb (in case of two-stem verbs, or the universal stem for one-stem verbs) is used in the formation of the present tense. Besides that, in this tense two dialects use variations of the same postposition.

Lankarani and Anbarāni dialects also strictly vary during the examination of the past imperfect of the indicative mood. In Lankarani there are two ways for the formation of this tense which are equally used. We find these formations also in Anbarāni, but at the present stage of the language development the Anbarāni dialect has splitted these forms in two moods: one is used in the indicative; the other is used in the past unreal of the conditional.

Another difference between the verbal systems of two dialects is the sole pattern for the negative form of the future tense of indicative in Anbarani while both Miller and Pireyko describe two means of its formation in Lankarani [Miller 1953: 157; Pireyko 1991: 156].

References 1. Bazin, Marcel (1980). Le Tâlech. Une region ethnique au nord de l‟Iran. Tome II, Paris, 314pp. 2. Bazin, Marcel (2012). ‘Anbarān // Encyclopædia Iranica, http://www.iranicaonline.org/articles/anbaran 3. Clifton, John, Tiessen, Calvin, Deckinga, Gabriela, Lucht, Laura (2005). Sociolinguistic Situation of the Talysh in Azerbaijan, SIL international, 33pp. 4. Miller, Boris (1953). Talyšskij jazyk (The Talyshi language), Moscow, 267pp. 5. Pireyko, Lia (1976). Talysško-Russkij slovar‟ (Talyshi-Russian dictionary), Moscow, 352pp.

1 The data concerning the verbal system of Lankarani is borrowed from the works of B. Miller [Miller 1953] and L. Pireyko [Pireyko 1976; Pireyko 1991].

10

6. Pireyko, Lia (1991). Talyšskij jazyk (The Talysh language) // Osnovy Iranskogo Jazykoznanija. Novoiranskie Jazyki: Severo-Zapadnaya Gruppa, I, Moscow, pp. 91-175 7. Stilo, Donald (1981). The Tati language group in the sociolinguistic context of Northwestern Iran and Transcaucasia // Iranian Studies, vol. XIV, nos. 3-4, Summer-Autumn, 137-187pp. 8. Yarsheter, Ehsan (1996). The Taleshi of Asālem // Studia Iranica, vol. 25.1, pp. 83-113 Zahra Karimi Bavarani, Negar Davari Ardakani & Parsa Bamshadi

(Shahid Beheshi University,Tehran)

Inflectional Categories of Verb in Sorkhi Dialect

The region of Kuhmara Sorkhi is one of the most ancient residencies of the province of Fars. According to administrative subdivisions of the county, this region is a part of Shiraz county. Five dialects are spoken in Kuhmara Sorkhi: Sorkhi, Soqolmači, Bakaki, Jaruqi and Karači. According to Hesampour and Jabbareh (2012a) Sorkhi is considered a west-southern Iranian dialect which is spoken in several villages such as Bagdane, Shourab, Romqan and Siyakh and among the mentioned dialects, Sorkhi is the most endangered one. Due to the influences from Shirazi accent and Standard Persian, the ongoing process of urbanization and the decline in the number of elderly speakers of the dialect, many Persian words have entered into Sorkhi and have replaced the native ones. There is no exact information about the number of Sorkhi speakers. However, it is estimated that the number of people who speak Sorkhi is less than 1500. Hesampour and Jabbareh (2012a) have documented some parts of Sorkhi grammar and its lexicon. However, they have not thoroughly investigated the details of its inflectional morphology. In another book, they have collected legends and tales of the region (2012b). Accordingly, the present research is aimed at describing, documenting and analyzing Sorkhi’s inflectional morphology of verbs. Inflection is defined as being concerned with sets of word forms which have the same lexical meaning but different grammatical (or morphosyntactic) properties (Carstairs-McCarthy, 2000: 595) and Inflectional categories are the categories of morphosyntactic properties which are expressed in the inflectional system (Stump, 2001). Both field work and descriptive method are used in this research. The data were gathered using local interview with elder speakers and available corpora and books. The theoretical basis for this research -as a part of a larger project on western and southern dialects of Iran- is Stump’s classification according to which morphosyntactic properties are divided into three main classes: 1) inherent properties, 2) agreement properties and 3) governed properties which are also called relational properties (Stump, 2001). The inflectional categories and the way of displaying them on verb and possible interactions among these categories in Sorkhi will be surveyed in this article. An interesting point about the dialect is that there are suffixes and prefixes used to represent morphosyntactic category of person/number. A thorough description of Sorkhi inflectional categories for verbs is pictured in the following table: inflectional properties

morphosyntactic categories

morphosyntactic features

exponent example

Inherent properties

Tense past -d, -t, -i, -ø xwar-d (he ate), gere-t (he took), pors-i (he asked), dā (he gave)

non-past mi- present: mi-zey (you beat) future: om-mi-jɑ be-zom (i will beat)

infinite -a daned-a (to know)

Aspect perfective - -

11

progressive mi- ɑsej mi-šom (I am going), ɑsej tu-mi-saxt (you were making)

habitual mi- mi-xor-en (they eat [everyday])

Mood indicative mi- mi-ku-ej (you pound) subjunctive be-, bo-,

hā-, hu be-rez-om (I pour), bo-xor-en (they eat), hā-d-om (I give), ), hu-šom (I go)

imperative bo-, be-, hā-, hu-, vā-

bo-xo (eat), be-rez (pour), hā-da (give), hu-nes (put), vā-dez (sew)

Polarity positive - - negative na-, -ne dede vɑ-na-b-om (I am not

seen), ne-mi-z-i (he does not beat)

Agreement Properties

Person/Number

1sg -om, -am, om-

miz-om (I beat), ešt-am ( I went), om-rexte (I poured)

2sg et-, ed-, -ey

et-dede (you have seen), ed-di (you saw), miz-ey (you beat)

3sg -i, eš- miz-i (he beats), eš-dede (he has seen)

1pl -um, -mu miz-um (we beat), mu-dede (we have seen)

2pl -i, tu- miz-i (you beat), tu-dede (you have seen)

3pl -en, šu- mi-z-en (they beat), šu-dede (they have seen)

Relational properties

Voice active - passive vā- deda vā-m-b-om (I am seen),

šošte vɑ-m-bu (it has been washed)

causative -on ček-on-d (she caused to drop) References Carstairs-McCarthy, Mathew (2000). “Inflection”. In Geert Booij, Christian Lehmann & Joachim

Mugdan (eds.), An International Handbook on Inflection and Word-Formation: Berlin/New York: Walter de Gruyter.

Hesampour, Saeed and Jabbare, Azeem (2012a). Grammar and Lexicon of Kuhmara Sorkhi. Shiraz: Avand-e-Andishe Press.

Hesampour, Saeed and Jabbare, Azeem (2012b). Legends and Tales Belonging to kuhmare Sorkhi. Shiraz: Sivand Press.

Lieber, Rochelle (2009). Introducing Morphology. Cambridge: Cambridge University Press. Shahbazi, Abdollah (1988). The Forgotten Tribe. Tehran: Ney Press. Stump, Gregory T. (2001). “Inflection”. In Andrew Spencer & Arnold M. Zwicky (eds.), The

Handbook of Morphology (pp. 13-43). Oxford: Blackwell Publishing.

12

Adam BENKATO & Ergin ÖPENGIN

(Berlin-Brandenburg Academy of Sciences and Humanities & University of Bamberg)

Argument indexing in Yaghnobi from an Iranian perspective

Yaghnobi is an eastern Iranian language spoken in Tajikistan by communities totaling a few thousand people. Though threatened, it has now been given legal status as the Tajik government seeks to connect itself with its Sogdian past, with which Yaghnobi is considered to be the link. Despite its great interest for (eastern) Iranian linguistics, it is actually underdescribed and understudied.

Yaghnobi morphosyntax is shaped by the verbal stem being distinguished into a present and a past stem. Crucially, this is a purely morphological distinction and does not correspond to a strict division of tense and aspect. Furthermore, Yaghnobi has two cases and a complex array of person marking sets, including pronominal clitics, two sets of verbal person suffixes, and copular endings. The use of case and person markers in indexing clausal arguments is reminiscent of the well-known tense-sensitive ergative alignment, however, with some peculiarities that set it apart from other Iranian languages. The present stem is used both in present and simple past tenses, where the alignment is accusative, as in (1) and (2).

(1) na=m-vof-či NEG=1SG-speak.PRS-3SG.DUR ‘He does not speak to me’

(2) man divar a-pen-im 1SG door IPF-open.PRS-1SG ‘I opened the door’

The past stem, going back to the old past participle in -ta, is used for a range of perfect tenses. In constructions with the past stem, the alignment is mostly along ergative lines, but the specific form of encoding the clausal arguments is also sensitive to the discourse. Thus, in (3) a pronominal clitic indexes an A-past, while the verbal personal ending marks the O.

(3) divar na=š-peta-x door NEG=3SG-open.PST-COP.3SG ‘He did not (yet) open the door’

(4) mox=šint a-wēn-im 1PL=2PL IPF-see.PRS-1SG ‘We saw you.’

Despite these known basic facts of Yaghnobi grammar, the existing sketches of Yaghnobi do not provide information on the remaining aspects of argument indexing. For instance, the grammatical status of pronominal clitics and their placement/sequencing properties are yet to be investigated. For instance, although the pronominal clitic is placed within the verb complex in (3), skipping the “second” position, in (4) it occurs on the clausal subject, thus in the “second” position. The factors behind this and similar variation will be examined in this paper. Furthermore, it is unclear how an overt A-past argument is indexed, via pronominal clitics, oblique case, or both.

In order to account for this and a number of other features of argument indexing in Yaghnobi, in this paper we will (i) provide an overview of the Yaghnobi verbal system, (ii) describe the use of case marking and different person marking sets, (iii) and examine their implications for the Iranian languages. Our research is based on previously published material as well as our analysis of new unpublished material.

13

Cemile CELEBI (University of Frankfurt)

Bezeynî – An Endangered Kurdish Language

The Language Bezeynî (aka. Šêxbizinî, Šawzenî, Šahuzendî, etc.) is spoken by small Kurdish communities, who are spread over present-day Turkey with few language islands in Iraq. Since they consider themselves Kurds and call their language ‘Kurmancî’, they were usually overlooked in linguistic descriptions. These designations also make it difficult to acquire funding for language documentation, because ‘Kurmancî’ is not considered a highly endangered language.

However, Bezeynî represents an independent language that shows close genetic relations with Laki and so-called Southern Kurdish (Fattah 2000). The linguistic situation (small language communities among a Turkish, Zazakî, or Kurmancî speaking majority) and the lack of awareness of the speakers that their language is different from Kurmancî minimises the chances of survival of Bezeynî.

Bezeynî constitutes an important example for the relative chronology of Kurdish languages. The speech community probably migrated from Iraq-Kurdistan, which seems likely (Sykes 1908: 481). Time and exact location of their settlement are unclear, though, their migration route is up to speculative hypotheses about similarity of names (e.g. Izady 1992: 42, 46, 82, 83, 85). Sykes (1908: 455, 477, 481-82) claims that, all parts of Bezeynî tribes in Central and Eastern Anatolia were driven from northeast of Kirkuk (Iraq) by Sultan Selim. Edmonds (1957: 38, 322) reports that, according to local sources, a part of Bezeynî speaking tribes migrated from Mosul (Iraq) or further north to northeast of Kirkuk (Iraq) at the beginning of the 18th century. Traces of language contact in this variety during the migration of communities can be identified in the lexicon and probably in the grammar as well. A systematisation of these linguistic features may allow to stipulate a relative chronology and thereby a possibly migration route.

14

The presentation of Bezeynî will contain an overview of its linguistic situation and illustrate its genetic relation with so-called Southern Kurdish as well as language contact phenomena with Kurmancî.

1. Phonology: B[ezeynî], K[urmancî], S[oranî] ■ Preservation of *#w: “memory” B. wīr < *wīra- vs. K. bīr, cf. Goranî. wīr. ■ Middle Iranian b > w in all positions: “bound” B. wıstā < *bastag vs. K. bastī; “water” B. āw

< *āb < *āp- as with K. āv, cf. Southern Kurdish āw. ■ Preservation of Middle Iranian m: “midday” B. nīmarū < nēmag-rōǰ, K. nīvrō; S. nīwařō;

“earth” B. zamī < zamīg, K. zavī, S. zawī; “step” B. gām < gām, K. gāv, Zazakî gām. ■ Alternation of d~r~y: “I do” B. dı-/rı-/yı-karım vs. K. dı-kım. ■ Palatalisation: “far” B. dür vs. K. dūr, cf. Southern Kurdish dür. “old” B. köne vs. K. kone.

2. Morphology: In contrast to Kurmancî and Zazakî, Bezeynî exhibits enclitic pronouns, which are identical with the ones in Sorani (-m, -t, -y, -mān, -tān, -yān). A peculiar exception appears when enclitic pronouns are attached to a verb in 3sg.: awī īma dıwīnī or dıwīnī-ke-mān “He sees us.”. The orthotone personal pronouns match the Southern Kurdish forms (e.g., 1pl. īma). Noteworthy is the 2sg. variant tun, which is also attested for Southern Kurdish (ton, Fattah 2000: 275-276). The demonstrative pronouns of near deixis in the singular ay (+ variants) link Bezeynî to Southern Kurdish as well (Fattah 2000: 314-319).

A feature that Bezeynî shares with Kurmancî and Zazakî is a special future formation with a particle and the verb in subjunctive mood, e.g. “I will come” mın ho w-ā-m (Sakarya/Adapazari), mın dā w-ā-m (Ankara/Haymana), vs. Kurmancî az dē b-ē-m, Zazaki az do b-ēr-ī. Bezeynî represents an interesting example of a migrated language for studies on genetic relations and areal linguistic effects.

Bibliography: Bedir Khan, E.C. / Lescot, R.: Kurdische Grammatik, Verlag für Kultur und Wissenschaft, Bonn, 1996. Edmonds, C. J.: Kurds, Turks, and Arabs. Politics, Travel, and Research in North-Eastern Iraq, 1919–

1925, London Oxford University Press, New York, Toronto, 1957. Fattah, Ismaïl Kamandâr: Les Dialectes Kurdes Méridionaux: Etude Linguistique Et Dialectologique,

Peeters, Leuven, 2000. Izady, Mehrdad: The Kurds, a concise handbook, Taylor & Francis, USA, UK, 1992. MacKenzie, D. N.: A Concise Pahlavi Dictionary, London Oxford University Press, New York,

Toronto, 1971. Soane, E. B.: To Mesopotamia and Kurdistan in Disguise, Small, Maynard and Company, Boston, 1912. Sykes, Mark: The Kurdish Tribes oft he Ottoman Empire, in: The Journal oft he Royal Anthropological

Institute of Great Britain and Ireland, Vol. 38 (Jul. – Dec., 1908), 451- 486. Thackston, W. M.: Sorani Kurdish, A Reference Grammar with Selected Readings, in:

http://www.fas.harvard.edu/~iranian/Sorani/sorani_1_grammar.pdf, accessed 19th Jan. 2016. Todd, Terry Lynn: A Grammar of Dimili, also known of Zaza, Univertsity of Machigan, 1985. Mohammad DABIR-MOGHADDAM (Allame Tabataba’i University)

Non-Canonical Subject Construction as a Relict in Iranian Languages: Typological and Diachronic Implications

A number of Iranian languages still contain a relict construction whose predicator requires two core arguments, an Oblique logical subject and a Direct morphological subject. The predicators which appear in this construction convey the core meaning of ‘to exist; to have; to want; to be

15

able; to be obliged; to (dis)like’. It may be argued that these predicators share an experiencer Oblique logical subject. What makes the study of the Iranian Non-Canonical Subject Construction more interesting is the fact that this construction shares the peculiarities of the indexation of its arguments with the indexation of the arguments of the transitive past tense predicators in the so-called split ergative Iranian languages. The examples which I will discuss in the paper are mostly from my own corpus of the Western (both the Southwestern and Northwestern) Iranian languages of Iran though my first representative example of the Non-Canonical Subject Construction is cited from an Eastern Iranian language. Example (1) is from Ormuri, a South-East Iranian language spoken in Afghanistan and North-West Pakistan. I quote this example from Stassen (2009). The example was originally reported in Grierson (1921: 229).

(1) ta-sa sarai dyo kullān bukin GEN-one man two son be.3PL.PST ‘A man had two sons.’ Stassen (2009: 108, ex. (1)a)

It should be noted that the second argument in example (1) which is plural, i.e. dyo kullān, is encoded in the copula. This argument is the NP-Direct. The first argument in example (1), which shows the Genitive case, is the NP-Oblique. As further illustration, below I have provided and analyzed examples from a 260 year old letter in the Lari dialect of Larestani language. The letter was first published in Eghtedari (1951: 281-282).

(2) xub m=æ-vayade ke boa=m š=æ-got well 1SG.OBL=INCOMPL-remember that father=1SG.POSS 3SG.OBL=INCOMPL-said ‘I remember well that my father used to say…’ (Lit. I have in mind well that my

father used to say…).

(3) … mo=æ-vi æ eškal o-č-æm 1PL.OBL=INCOMPL-want to hunting SBJV-go-1PL ‘…We want to go hunting.’

(4) æma xæbær mo=ni we news 1PL.OBL=not. exists/is ‘We do not know.’ (Lit. to us news not exists/is).

(5) … æz mækim ændišæ=š e [<ændišæ oš=e] from Makim worry=3SG.OBL have/is worry 3SG.OBL=exists/is ‘He is afraid of the king’. (Lit. to him there is worry from Makim [a secret name

to refer to the king]).

(6) æifoš ke om=na-ša alas that 1SG.OBL=not-can ‘Alas that I can not.’ (Lit. to me it is not possible).

(7) … omid om=e dobaræ ot=vabin-em hope 1SG.OBL=exists/is again 2SG.OBL=see-1SG ‘I hope to see you again.’ (Lit. to me there is hope I see you again).

In example (2) the NP-Oblique is the first person pronoun which is dropped but encoded by the proclitic m= and the NP-Direct is the subordinate clause which begins with the complementizer ke ‘that’. It may be noted that in this example the past tense transitive predicator ‘said’ hosts the A-clitic. However, in example (7) the present tense transitive predicator ‘see’ hosts a proclitic which indexes the O and it also indexes the A by a verbal agreement suffix. On the other hand, in example (8) below from the same text we witness an instance of the ergative-absolutive construction based on agreement. In this example, the light verb of the compound transitive past tense predicator hosts the A-clitic as well as the O-suffix.

16

(8) gül šo=zæt-æm cheating 3PL.OBL=hit-1PL ‘They cheated us.’

Similar examples will be presented from my corpus of Davani, Vafsi, Naini, Kurdish (of Baneh), Laki, and Tati. Finally, I will mention examples of the occurrence of the Non-Canonical Subject construction in Old Persian, Middle Persian, Early New Persian, and to a lesser extent Modern Persian. The range of options allowed by these languages in the indexation of the two core arguments of the Non-Canonical Subject construction as well as the range of predicators found in this construction will be relied on to discuss the typological and Diachronic implications of this study. From a diachronic perspective, I will also assess the relevance of the findings of this paper for the arguments and debates on the genesis of ergativity in the Iranian languages. It is needless to point out that the presence of the Non-Canonical Subject construction as a relict in a number of Modern Iranian languages before it is reanalyzed as a canonical subject construction (as it has mainly happened in Modern Persian) not only necessitates research on endangered Iranian languages but it also strengthens the conjecture that endangerment should not merely be interpreted as the loss of a language. It might simply be the loss of a relict. Selected bibliography Alikhenvald, Alexandra, Robert M. W. Dixon, and Masayuki Onishi (eds.). 2001. Non-Canonical

Marking of Subjects and Objects, Amsterdam: John Benjamins Publishing Company. Barðdal, Jóhanna and Smitherman, Thomas. 2013. The Quest for Cognates: A Reconstruction of

Oblique Subject Constructions in Proto-Indo-European. Language Dynamics and Change 3: 28-67. Bhaskararao, Peri and Karumuri Venkata Subbarao (eds.). 2004. Non-nominative Subjects, Volumes 1

and 2, Amsterdam: John Benjamins Publishing Company. Cole, Peter, Wayne Herbert, Gabriella Hermon, and S.N. Sridhar. 1980. The acquisition of subjecthood.

Language 56:719-743. Eghtedari, Ahmad. 1951. færhæng-e larestani (Larestani Dictionary). Tehran: Farhang-e Iranzamin

Publisher. Grierson, George. A. (ed.) 1921. Linguistic Survey of India, Vol. X. Specimens of Languages of the

Eranian Family. Calcutta: Superintendent Government Printing Office. Haspelmath, Martin. 1999. External possession in a European areal perspective. In Doris Payne and

Immanuel Barshi (eds.) External possession, 109-135. Amsterdam: Benjamins. Seržant, Ilja and Leonid Kulikov. 2013. The Diachronic Typology of Non-Canonical Subjects,

Amsterdam: John Benjamins Publishing Company. Stassen, Leon. 2009. Predicative Possession. Oxford: Oxford University Press. Saloumeh GHOLAMI (University of Frankfurt)

Pronominal clitics in Zoroastrian Dari (Behdīnī) of Kerman

Zoroastrian Dari, a critically endangered language, is one of the unique Iranian languages on account of historical background and its large number of sub-dialects. There are two main dialects of Dari: Kermani and Yazdi. The dialect of Yazdi has about 30 sub-dialects, while there appears to be only one dialect of Kermani. The situation of the Kermani dialect is especially grave. The language has fallen completely out of use and no one retains a sense of ethnic identity associated with the language. The situation of the Yazdi dialect of Dari is comparatively better. The only fluent users are older than child-bearing age.

Dari has generally been considered to be closely related to Central Dialects (Windfuhr 1989). Dari shares features with languages of the NW group, especially with Zazaki and Sorani and SW for example with Persian (Gholami 2016). Sharing of both NW and SW features draws our

17

attention to the fact that the Zoroastrian immigrants to Yazd and Kerman came from different provinces of Iran.

Most contemporary Western Iranian languages make use of enclitic pronouns. They are used in all oblique functions including the marking of objects, the possessor, and, in those Ir. varieties that show ergative patterns , the agent of ergative constructions (Korn 2009: 159). In some Iranian languages and dialects, for example, in Fars dialects, the pronominal enclitics become proclitics. Proclitic pronominals are used in post-ergative patterns in Zoroastrian Dari of Kerman. Their existence can imply potential contact of Kermani Dari with Fars dialects. The pronominal enclitics are used in other oblique functions including the possessor. Examples (1), (2) and (3) illustrate the use of a proclitic in Kermani Dari:

(1) dūnā ra o savzī šūn=gerafta šūn=ārta āteškeda Beans Post.Pos and herbs PC.3P.OBL =bought PC.3P.OBL=brought Fire Temple šūn=šošta. PC.3P.OBL =washed “They bought beans and herbs and brought them to the Fire Temple, and they washed them.”

(2) dendūn-šūr ro koyā dī=yāšt. tooth-brush Post.Pos where PC.2S.OBL=put “Where have you (sg.) put the tooth-brush?”

(3) me to-ro om=dī. 1S.DIR 2S-Post.Pos PC.1S.OBL=saw “I saw you.”

Example (4) and (5) illustrate the use of an enclitic as possessor:

(4) ked=om penğ genza dāra. house=EC.1S.OBL five room has “My house has five rooms.”

(5) del=om dard ekera. stomache=EC.1S.OBL ache does, makes “I have a stomachache.”

ABBREVIATIONS: DIR=Direct case, EC=Enclitic, OBL=oblique case, P=Plural, PC=Proclitic, Post.Pos=Postposition, S=Singular

Keywords: Zoroastrian Dari, pronominal clitics, proclitics, enclitics

Selected literature: Gholami, S. 2016. Zoroastrians of Iran vi. Linguistic Documentation. Encyclopædia Iranica, online

edition, 2016, available at http://www.iranicaonline.org/articles/zoroastrians-in-iran-06 (accessed on 27 January 2016).

Haig, G. 2013. Alignment change in Iranian: What happened to person agreement? Paper held at the Workshop on Patterns of alignment in the Indo-Iranian languages: Towards a typology. LSI, Ann Arbor, July 13-14 2013

Korn, A. 2009. Western Iranian Pronominal Clitics. Orientalia Suecana, LVIII, 159-171. Windfuhr, G. 1989. Behdinān Dialect. Encyclopædia Iranica, IV/1, 1989, pp. 105-8; online edition,

1989, available at http://www.iranicaonline.org/articles/behdinan-dialect.

Informants for Kermani Dari: Farangis K̠adivi, 83 years old, born in Kerman, lives in Kerman, was a teacher Zemorrod Sohrābiān, 80 years old, born in Kerman, lives in Kerman, was a teacher

18

Thomas JÜGEL (Labex EFL Paris, CNRS - UMR Mondes iranien et indien, Paris)

Mapping Cross-reference systems

Cross-reference shows peculiar developments in the Zagros region across linguistic boarders. In some languages pronominal and verbal cross-reference (i.e. pronominal reference and verbal agreement) seem to switch their formal representations: Pronouns are grammaticalised as agreement markers and verbal endings adopt pronominal reference. Besides regular subject-verb agreement (SV), object-verb agreement (OV) appears and even double agreement (SV and OV). In contrast to lexical items, which can easily be borrowed, cross-reference is deeply embedded in the syntactic structure so that similar patterns are probably due to close language contact. In the case of Northeastern Neo-Aramaic (NENA), the Iranian influence is obvious; all the more since peculiar patterns only appear in Neo-Aramaic dialects in Kurdish speaking areas.

In this talk, I will discuss the various cross-reference patterns and parallel development of genetically unrelated languages, several of them minority languages. Moreover, I intend to present the linguistic atlas project that is currently developed in Paris and I will show how such patterns may be mapped (work in progress in collaboration with Emmanuel Giraudet).2

In Sorani-Kurdish as well as in Neo-Aramaic, special pronominal forms were grammaticalised as agreement suffixes, i.e., agreement of verb and logical subject, in the past tense domain. In Sorani (as with Gorani and Hawrami), enclitic pronouns indicate the logical subject, in Neo-Aramaic enclitic pronouns amalgamated with the preposition l-, which then again is cliticised (cf. present páṯx-ǝn “I open” vs. past pṯǝ́x-li “I opened”, Qaraqoš, Khan 2002: 88, 92).

In addition, the enclitic pronouns in Sorani display all kind of oblique functions. They can express a possessor (possessive pronoun), a complement of an adposition, and a direct object (the latter only in the present tense domain). The l-pronouns of Neo-Aramaic can also be used for objects in present tense (Khan 2002: 137).

The inherited verbal endings, which derive from the enclitic copula for the past stem, originally agreed with the logical object. Since the copula of the 3SG was usually omitted, gender distinction of the participle is, if at all, preserved for this form in several Iranian languages (e.g. Hawrami). Some Neo-Aramaic dialects show OV agreement by means of gender/number marking as well, even if the object is oblique (here marked by ʾil-). The enclitic pronouns (EP) function as agreement markers.

gora qṭil- -∅ -le naše qiṭl- -i -le man killed 3SG.M 3SG.EP people killed 3PL 3SG.EP ‘(he) killed the man’ ‘(he) killed the people’ gora (ʾil-) baxta qiṭl- -a -le man OBJ woman killed 3SG.F 3SG.EP ‘the man killed the woman’ KHAN (2004: 88) Jewish Dialect of Suleimaniya and Ḥalabdja

The only Iranian language in the region that seems to display regular OV agreement is Hawrami. However, MACKENZIE’s data only contains clear examples for the 3rd person:

ā kınāče =m -à di -éna čewàł DEM girl =1SG:EP DEF see:PST -3SG.F. before ‘I have seen that girl before this’ (MACKENZIE, 1966: 61)

2 http://axe3.labex-efl.org/?q=fr/LC3f

19

In other languages, the verbal endings encode the object. Thus either the object is expressed by nouns or pronouns (if not omitted), or it is expressed by the verbal ending. Cf. the following example from Sorani (the second phrase dīmīt is attested, e.g., in OMAR, 1993: 98).

mın tō =m dī or mın dī =m =īt 2SG. I you =1SG:EP see:PST I see:PST =1SG:EP

‘I saw you.’

Note that enclitic pronouns, if they attach to the verb, may be inserted inbetween verbal stem and verbal ending (somehow turning the latter into an enclitic, cf. Haig 2008: 292 for a detailed description). In addition to object encoding, verbal endings can also encode other oblique arguments of the verb like the indirect object, cf. the following example from Hawrami (for Sorani see Jügel 2009: 153f. exx. 22-31).

hič =šā na- d̯ā -ne nothing =3PL:EP not give:PST -1SG ‘They gave me nothing.” (MACKENZIE, 1966: 78)

In our project at the Labex EFL, we explore the possibilities of mapping linguistic features in a linguistic atlas, cf. the example of the functional variety of agreement markers below. One set of agreement markers (e.g. verbal endings) may refer to various arguments depending on the domain (by means of agreement or pronominal reference).

Feature Example “Functional Variety” 3 ● Kurmanji VE→A+S ║ VE→S+O ○ Persian VE→A+S, EP→O+OBL ○ Turkish VE→A+S, poss. VE→S, EP→S+POSS ♦ NENA VE→A+S, EP→O+OBL║ VE→S+O, EP→A+OBL ♦ Hawrami VE→A+S, EP→O+OBL║ VE→S+O, EP→A+OBL ◊ Gorani VE→A+S, EP→O+OBL║ VE→S+O, EP→A+O+OBL ◊ Sorani VE→A+S, EP→O+OBL║ VE→S+O+OBL, EP→A+OBL

The increase of variation is represented by an increase of colour intensity. The symbols visual-ise an alignment type (● ergative alignment, ○ accusative alignment, ◊ accusative alignment with reflexes of ergativity, ♦ mixed alignment). Kurmanji lost enclitic pronouns. The verbal endings agree with the subject in the present stem domain (VE→A+S), and with subjects (intransitive verbs) and objects (transitive verbs) in the past stem domain (VE→S+O), e.g. ez ket-ım “I fell”, te ez dīt-ım “you saw me” (Bedir Khan/Lescot 1970: 136, 174). In Persian verbal endings agree with the subject only. Enclitic pronouns can encode objects and obliques. In Turkish both sets of verbal endings agree with the subject. Possessive pronouns encode posses-sors and function as agreement markers in participle clauses, e.g. senin geleceğ-in-i duydum “I heard that you will come”. In all other languages listed above, verbal endings agree with the subject in the present stem domain and enclitic pronouns encode objects and obliques. In the past stem domain the verbal endings agree with the subject (intransitive verbs). Enclitic pro-nouns encode obliques and they function as agreement markers with transitive verbs. In NENA and Hawrami the verbal endings agree with the object (transitive verbs), while in Gorani and Sorani they can encode the object. Transitive agreement markers are also attested for intran-sitive verbs in NENA, e.g. tpíl-le “he sneezed” (Khan 2004: 303). See Hopkins (1989: 432) for an overview. Similar developments are attested for Iranian languages, but not in this region.

3 Abbreviations: VE: verbal ending, A: transitive subject, S: intransitive subject, O: object, EP: enclitic pronoun, poss. VE: possessive verbal ending, POSS: genitive participant, OBL: oblique participant.

20

Bibliography Bedir Khan, E. Dj., and Lescot, R. (1970): Grammaire kurde (dialecte kurmandji), Paris. Haig, G. (2008): Alignment Change in Iranian Languages – A Construction Grammar Approach, Berlin. Hopkins, S. (1989): “Neo-Aramaic Dialects and the Formation of the Preterite”, pp. 413-432 in: Journal of

Semitic Studies 34/2. Jügel, T. (2009): “Ergative Remnants in Sorani Kurdish?”, pp. 142-158 in: Orientalia Suecana 58. Khan, G. (2002): The Neo-Aramaic Dialect of Qaraqosh, Studies in Semitic Languages and Linguistics 36,

Leiden. ——(2004): The Jewish Neo-Aramaic Dialect of Sulemaniyya and Ḥalabja, Leiden. MacKenzie, D. N. (1966): The Dialect of Awroman (Hawrāmān-ī luhōn), København. OMAR, Feryad Fazil (1993): Roşnayî le dengewe – Leuchten aus der Stimme, Berlin. Mesut KESKIN (University of Frankfurt)

Notes on endangered Iranian Language Zazaki with a particular Focus on its northern Dialect

Zazaki is a relative recently attested and written northwest Iranian language. For centuries oral transmission was the only medium for the survival of the Zaza language hence it has a rich corpus on oral poetry and in musical form as well as also in fairy tales, proverbs and short stories. The oral Zazaki literature to this day preserves regional phonetic, morphological, lexical and semantic sections specific to different regions where Zazaki is traditionally spoken. In the past 3 decades the literature in Zazaki has finally begun to be recorded and produced in written format though this is not happening in the place of origin where this language is spoken but mostly in Diasporas (Germany, Sweden etc.) where the speakers of the language have migrated in post 1950s.

The declining of language is the result of Turkey’s repressive linguistic policies; the Republic of Turkey was established as a nation-state with no due consideration given to the multilingual diversity of the population of Anatolia. Furthermore with “tevhid-i tedrisat” (The Law for the unification of Education), Turkish was established as a formal and the only language for teaching in schools. Speakers of all languages were forced onto education programmes that taught (through the medium of Turkish). This has resulted on Zazaki among other minority languages such as Kurdish, Laz, Syriac etc. to become victim of Turkey’s repressive language politics and to face extinction. All the languages that were in use at the time of the national linguistic change in Turkey have now under the threat of becoming extinct and Zazaki among them is recognised as an endangered language (“vulnerable”) by UNESCO.

This paper aims to demonstrate declining use of Zaza language since the repressive and monistic based linguistic policy of Turkey has come to an effect through looking at the use of Zazaki language for members of several generations in the author’s family. The author will attempt to follow declination in the use of language as well as knowledge of the Northern Zazaki through how it is experienced by seven uncles and aunts and seventy cousins on his maternal and paternal sides of the family. By doing this the author will focus on the northern dialect of Zaza Language as a variety spoken by Alevi Zazas. The position of this dialect compared to the central or southern dialects is more endangered.

The findings of the author’s paper revealed that Northern Zazaki is increasingly losing its status as a communication language and also losing of some grammar and lexical components of the language as a result of this. In detailing the journey of this particular dialect the author will also detail linguistic aspects of the language such as the phoneme inventory, as observed in the presence of alveolar affricates [tsh], [ts] and [dz] which led to renewal in phonology in the occurrence of the post-alveolar sibilants [ʃ], [ʒ], [tʃh], [tʃ] and [dʒ] as complementary distributed allophones, as well as existence of aspirated and unaspirated pairs of the voiceless plosives /č,

21

k, p, t/ like in Armenian and Georgian or some morphologic or lexical peculiarities of the Northern Zazaki can be mentioned.

Keywords: Iranian Languages, endangered languages, assimilation politics, Zazaki, Turkey, Northern-Zazaki, Alevi References: Asatrian, Garnik 1995: „Dimlī“. In: Encyclopedia Iranica. Online:

http://www.iranicaonline.org/articles/dimli (September 2011) Azbak, Dilek 2013: Eine vom Aussterben bedrohte iranische Sprache: Kırmancki (Zazaki).

Açıklanmamış bitirme tezi. Freie Universität, Institut für Iranistik, Berlin. Çağlayan, Hüseyin 1995: Die Schwäche der türkischen Arbeiterewegung im Kontext der nationalen

Bewgung (1905-1945). Frankfurt. Çelik, Hıdır Eren: Almanya’da br göçmen toplum: Dersimliler: Herkesin bildiği sır: Dersim (Derleyen: Şükrü Aslan). İletişim Yayınları. İstanbul.

Ceyhan, Pınar 2010: Die Rolle der Erziehung und Bildung in der Türkischen Republik am Beispiel der Dorfinstitute zum Mittel der Zwangstürkisierung der ethnischen Minderheit Zaza. (“Türkiye Cumhuriyeti’nde Köy Enstitüleri örneğinde Zaza etnik azınlığı türkleştirme aracılığında eğitimin rolü”). Açıklanmamış bitirme tezi. Münster Westfalya Wilhelm Üniversitesi.

Demir, Gülistan 2015: Muttersprache Zaza - Warum ist Zaza als Sprache bedroht? Bachelorarbeit. Fulda. Online: http://www.zazaki.de/deutsch/aufsaezte/BAguelistanDemir.pdf

Fırat, Gülsün 2010: Dersim’de etnik kimlik: Herkesin bildiği sır: Dersim. İstanbul. Gippert, Jost 2007/2008: „Zur dialektalen Stellung des Zazaki“ In: Die Sprache. Zeitschrift für

Sprachwissenschaft. Goldstein-Schulz, Esther 2013: Die Sonne blienb stehen. Band I. Neckenmarkt (novum publishing

gmbh), Avusturya. Gülensoy, Tuncer 1983: Kurmanci ve Zaza Türkçeleri Üzerine Bir Araştırma. Ankara. Gündüzkanat, Kahraman 1997: Die Rolle des Bildungswesens beim Demokratisierungsprozeß in der

Türkei unter besonderer Berücksichtigung der Dimili (Kirmanc-, Zaza-) Ethnizität. Münster. Jacobson, C.M. / M. Sandonato 1997: Zazaki se beno? Dergi: Tija Sodıri – Perloda Zon u Zagonê

Kırmanc-Zazay, 4. Sayı, s. 9-12. Frankfurt. Keskin, Mesut 2008: Zur dialektalen Gliederung des Zazaki.

http://publikationen.ub.unifrankfurt.de/volltexte/2009/6284/ (Januar 2010). Keskin, Mesut 2010: Zazaca üzerine notlar: Herkesin bildiği sır: Dersim (Derleyen: Şükrü Aslan). İletişim Yayınları. İstanbul. Online: http://zazaki.de/turkce/makaleler/zazacauzerinenotlar.pdf

Keskin, Mesut 2012: Zazacanın İrani diller arasındaki yeri ve özellikle Dersim’de olan durumu ve yarını. 1.Uluslararası Tunceli (Dersim) Sempozyumu Bildiriler Kitabı, Tunceli.

Keskin, Mesut 2013: Asimilasyon Politikasının Nesiller Üzerindeki Etkisine Bir Örnek. Tunceli Üniversitesi, II. Uluslararası Tunceli (Dersim) Sempozyumu bildiriler kitabı, Tunceli. Online: http://www.tunceli.edu.tr/dersimsempozyumu/Dersim%20Sempozyum%20Kitab%C4%B1%20Bitmi%C5%9F%2018.09.2014%20(1).pdf

Paul, Ludwig 1998a: Zazaki. Grammatik und Versuch einer Dialektologie. Wiesbaden. Selcan, Zülfü 1998a: Grammatik der Zaza-Sprache. Nord-Dialekt (Dersim-Dialekt). Berlin. Masoud MOHAMMADIRAD & Mohammad RASEKHMAHAND

(Bu-ali Sina University, Hamedan)

Head-marking and dependent-marking patterns in Western Iranian languages

This paper investigates head-marking and dependent-marking patterns of fourteen western Iranian languages as stated in their descriptive grammars. For some languages, ex. Central Kurdish, Taleshi, and Tati, dialectal differentiations are also surveyed. Following Nichol’s (1986) seminal work, morphological markings were investigated both at phrase and clause

22

level; At phrase level, morphological markings were surveyed for these constituents: 1) adpositional phrase (PP) with pronoun object, 2) PP with noun object, 3) NP with dependent possessive pronoun, 4) NP with dependent possessive noun, and 5) NP with attributive adjective, while at clause level, marking of three core actants: subjects, direct objects, and indirect objects were investigated when they are present as nominals in the clause. Following examples show morphological marking of constituents at phrase level in New Persian:

1.a) bā Man 2.a) az Madreseh 3) barādar-e u 5. havā-ye xub ‘with me’ ‘from school’ brother-EZ 3SG weather-EZ nice ‘his brother’ ‘nice weather’

1.b) barâ-ye man 2.b) alayh-e Maryam 4) dar-e xāne for-EZ man against-IZ PN door-EZ house ‘for me’ ‘against Maryam’ ‘the door of the house’

As it can be seen New Persian constituents are mostly head-marked at phrase level. At clause level, however, direct objects (6) and indirect objects (7) are dependent-marked, while subjects (8) are head-marked:

(6) barādar-e xod=aš-rā košt brother-EZ self=3SG-OBJ kill.PST.3SG 'He killed his brother' (Windfuhr & Perry 2009:437)

(7) Ahmad ketāb-rā be man dād PN book-OBJ to 1SG give.PST.3SG 'Ahmad gave me the book' (Windfuhr & Perry 2009:479)

(8) man (be) sinemā mi-rav-am 1SG (to) Cinema IPFV-go.PRS-1SG 'I'm going to the movies'. (Windfuhr & Perry 2009:441)

Morphological marking of constituents in other western Iranian languages, however, show remarkable differences with that of New Persian: CONSTITUENT PP PP NP NP NP SUBJ OBJ I.OBJ TOTALS TYPE DEPENDENT PRONOUN NOUN PRONOUN NOUN ADJECTIVE Persian H//0 H//0 H H H H *D/h D 2/6 -4 Muk. Kurdish 0//H D//0 H//H/d H/D H H//N D//0//h D 4/5 -1 Kurmanji D D H/D H/D H H//D D//h D 7/4 3 Gilaki D D//h//0 D D//h D//(H) H/(D) *D D 8/2 6 Table 1. marking patterns in Persian, Mukri Kurdish, Kurmanji, and Gilaki. H: Head-marking pattern; D: Dependent-marking pattern; H/D: Double marking pattern; H//D: two patterns: H or D; 0: No marking; *: Accusative case counted, although used only for definite direct objects; ( ): salient partial pattern; not known whether this pattern is primry or secondary, marked or unmarked, etc.; h, d: Minor (marked) patterns; N: Neutral marking; Total: Sum of D (D), H (H), enteries plus figures from Clause column

In totals column, the first figure shows the number of dependent-marked constituents, while the second figure refers to the number of head-marked constituents. In the type column, the status of languages as head-marked (-) or dependendent-marked (+) is given. It is clear from the table that Iranian languages form a continuum with respect to the marking patterns with Persian being mostly head-marked and Gilaki mostly dependent-marked while Mukri Kurdish and Kurmanji fall between these two ends. The data for other languages also confirms the continuum of morphological marking. Figure 1 plots the D values against the H values for each language:

23

It can be seen that Western Iranian languages cluster especially in upper left corner (8/2 for Gilaki, 7/1 for Maandarani, Tati (Eshtehardi, Chali), Northern Taleshi, and Turkemen Balochi) and lower right corner (New Persian (2/6), Central Kurdish (1/6), Bakhtiari (2/5), Gorani (1/4)), which correspond to dependent-marking and head-marking languages repectively. Hawrami (7/6) and Zazaki (6/8) which cluster on the upper right corner can be best considered as double marking languages, while also Kurmanji (7/4), and Mukri Kurdish (4/5) show a tendency for double marking patterns at phrase level. Southern Taleshi (7/3) and Koroshi (6/4) are other languages which prefer dependent-marking patterns. The sole similarity between languages in terms of marking can be loosly stated as ‘to mark nominal indirect objects as dependents through case markers or adpositions’ while for other constituents languages show variations in morhological marking.

The figures we observe for head-marking and dependent-marking patterns in under-investigated languages correlates with their word order correlations stated in Dabir-Moghaddam (2006) in that Gilaki, Mazandarani, and Vafsi (a variety of Tati) exhibit more correlations with OV languages and in terms of marking they are dependent-marked, while New Persian and Central Kurdish show more correlations with VO languages and in terms of marking patterns they are head-marked. Thus, an implication can be loosly stated for Western Iranian languages according to which Languages which exhibit more OV correlations show a tendency to be dependent-marked, while those which exhibit more VO correaltions tend to be head-marked. There are also areal correlations for marking patterns with languages spoken around caspean sea and Tati are mostly of dependent-marking type, while as one moves towards west and southwest of Iran, head-marking patterns are abound.

Figure 1 is based on major marking patterns in these languages. Languages, however, show minor marking patterns. The study of minor marking patterns reveals that languages spoken in Iran are developing head-marking patterns, as a minor pattern, at different degrees due to the status of New Persian as a superstratum. Consequently, one can predict that, at least for those languages spoken in Iran, the direction of change is toward head-marking patterns, which are cross-linguistically preferred (Nichols 1986).

24

References Nichols, J. 1986. Head-Marking and Dependent-Marking Grammar. Lnaguage, Vol.62, No1. pp.56-

119. Dabir-Moghaddam, M. (2006) Internal and external forces in typology: evidence from Iranian

languages. Journal of Universal Language 7: 29-47. Zaniar NAGHSHBANDI & Ludwig PAUL (University of Hamburg)

Kanduleh Hawrami – report about an ongoing documentation project

The village of Kanduleh (the middle /d/ is not generally pronounced) is located in the eastern part of the Kermanshah province. Kanula is situated about 52 kilometers west of Sahneh and about 75 kilometers to the north-east of Kermanshah (the capital of the Kermanshah province). According to the last national census in 2006, Kanula has the approximate population of 1500. All inhabitants of the village belong to the Shia faith. Traditionally, the populace lives from agriculture, gardening and animal husbandry. The performance of Ta'zieh is the most important cultural and religious event in Kanduleh which takes place once a year for glorifying the martyrdom of the third Imam is Shia faith.

It is generally assumed that the name of the village, Kanduleh, is originally taken from the name of its ruler, Roknolldole Deilami in the Al-e-Booyeh era. [Roknodowle > Kanula]. However, another scenario concerning the etymology of the name of the villages posits that Kandouleh is essentially composed of two parts: "Kand" meaning "village" in Turkish and "Doula" meaning government or governmental in Persian.

Kanduleh with the small and almost deserted neighboring villages of Parian and Sharifabad are the only Gurani-Speaking spots in the eastern part of Kermanshah province. The sociolinguistic situation in Kanula area involves three language groups: the primary language is a variety of Gurani which is somehow different from the other varieties of Gurani in some respects); standard Persian, as the official languages of Islamic Republic of Iran, is the language of education and governmental offices; a variety of Southern Kurdish which is called Kolyaii. It is worth mentioning that Kolyaii is mainly spoken in the county of Songhor and is different from Kalhori which is spoken in most parts of Kermanshah as the standard variety of Southern Kurdish. Persian and Kolyaii are getting more popularity among young generations since it enables them to find more life chances in Kermanshah and other Kalhori- speaking areas. Moreover, the marital relations between the inhabitants of Kanduleh and the neighboring Southern-Kurdish-speaking villages has resulted in generations of multilingual speakers. Based on these observations it can be claimed that the total number of the native speakers of Kandulei is actually less than the number of the actual inhabitants of Kanduleh.

Speakers of "Kandulei" have very interesting myths concerning the origin of their language. Some of them believe that they had been in Iraqi Kurdistan before their migration to this area. Based on this view, Kandulei is proudly regarded as the purest variety of Gurani and Kurdish and is also different from the so-called "corrupted variety of Hawraman". Still others posit that the Gurani variety of Kanduleh is originally taken from the variety of Paveh.

Before the official start of the project of the documentation of the Gurani dialect of Kanduleh in 2015, the fieldworker of the project, Zaniar Naghshbandi, took some training courses in

25

ElAN and ARBIL and other archiving tools in Hamburg University, Max Planck Institute for Psycholinguistics in Nijmegen, and Bamberg University. The whole project is sponsored by Volkswagon Stiftung and is supposed to be achieved in three phases till July 2017. The first phase, recently ended in June 2016, consists of the annotation of 30-minute recording in the ELAN software. The annotation process is basically achieved in three tiers in ELAN: one tier for transcription of the audio recordings, another one for free translation, and a third one for adding comments. However, some of the annotated files include additional tiers for more intensive morphological and syntactic analyses.

Although we haven't yet reached firm results concerning the overall structure of Kandulei, we try to present our findings in the first phase of the project in the form of a structural comparison between some linguistic aspects of the Gurani of Kanduleh and that of the more-studied variety of Paveh.

1. Gender: Unlike the variety of Paveh which still manifests the masculine-feminine distinction in the structure of verb and noun phrases, the variety of Kanduleh lacks gender distinction whatsoever.

2. Encoding A-PAST function with Oblique Marker: The variety of Paveh still uses the oblique marker for encoding the A-past function of noun phrases in some special situations. This function of the oblique marker is totally lost in the variety of Kanduleh and it's only used for signaling definite direct objects and also complements of prepositions. Non-accusative alignment in the past-tense constructions in the variety of Kanduleh is represented just by means of subject-referring clitics.

3. Using Cognate Constructions for Encoding Progressive Aspect: The variety of Paveh makes use of a special cognate construction (infinitive + conjugated present and past verbal forms) for showing the progressive aspect. This type of construction is not attested. Jaroslava OBRTELOVA & Raihon SOHIBNAZARBEKOVA

(Uppsala University & Russian-Tajik Slavonic University)

Steps being taken to reverse language shift in the Wakhi language in Tajikistan

Wakhi, one of the Eastern Iranian minority languages in the Pamir area, spread across Tajikistan, Afghanistan, Pakistan and China, has been classified as an unwritten endangered language. The degree of endangerment varies depending on the country and on the evaluating method. Thus, the Unesco scale of endangerment (UNESCO 2010, 2015) classifies Wakhi as a whole as ‘definitely endangered’ (grade 4 on the 6 grades scale) and the Ethnologue’s (2015) classification using the 10 grades Expanded Graded Intergenerational Disruption Scale (EGIDS) ranks Wakhi as ‘threatened’ (6b) in Tajikistan, ‘vigorous’ (6a) in Afghanistan and Pakistan and ‘shifting’ (7) in China. This paper will focus on the recent language development activities in the Wakhi language in Tajikistan such as introduction of new alphabet and orthography (2011) as well as first publications using the new alphabet (2012 and further).

Although the language itself shows strong vitality (Muller 2008) and is used on a daily basis by the members of the community, several field trips to the Wakhan valley with the aim of collecting traditional oral literature indicated that many of these genres are dying out fast

26

with mostly the older generation still remembering them but not passing them on to younger generations. There is a strongly felt need by the Wakhi speakers themselves to document their traditional oral literature and preserve it for future generations in a written form.

This paper will briefly outline the history of Wakhi language documentation and the attempts to set up a functional writing system that preceded the creation of the Wakhi alphabet in Tajikistan in 2011. It will further discuss key questions related to the creation of the alphabet such as:

-Arguments for choosing the Cyrillic script based on the current Tajik alphabet -Phonetic/Phonological complexity of the Wakhi language and the necessity to introduce 15 new letters that don’t exist in the Tajik alphabet, graphic form of the new letters -Definition of basic orthography rules, such as omitting or inserting the semi-vowels, orthography of borrowed words, considering the dialectal pronunciation variations, orthography of grammatical forms, etc.

All these steps were discussed at length with the major Wakhi linguists (native Wakhi speakers) in Tajikistan and the resulting alphabet proposal was tested during a field trip through the Wakhan valley in summer 2011. The final form of alphabet and basic orthography rules is therefore the result of detailed phonological analysis, grammar study, sociolinguistic factors, linguistic observations and advice by Wakhi scholars and intuition of non-linguist Wakhi native speakers.

With these activities and the first published books for the Wakhi speakers-readers in the Wakhi language in Tajikistan (Short stories for children, 2012; Folklore literature, 2016; Fairy-tales, 2016) the Wakhi language expands into a new functional category – the written form.

References: Edelman, J. I.; Dodykhudoyeva, L. R. The Pamir Languages. (2009) In: Windfuhr, Gernot, ed., The

Iranian Languages, London & New York: Routledge, 773-786 Gryunberg, Alexandr Leonovich & Stebline-Kamenski, Ivan Mikhailovich (1976) Ваханский язык

(Wakhi language), Moscow: Nauka Karamshoev, D. (1992) Алифбои фаврии забонҳои помирӣ бо намунаҳои адабӣ (A prompt alphabet

of Pamiri languages with the examples of literature), Khorog. Lashkarbekov, B.(2006) ‘К этнолингвистической истории ираноязычных народов Памира и Восточного Гиндукуша' (About the ethnolinguistic history of the peoples speaking Iranian languages in Pamir and Eastern Hindukush), in Памирская экспедиция, Moscow, 111-131

Lashkarbekov, B. (2013) ‘Изучение ваханских общин в условиях языкового сдвига’ (Study of Wakhi communities under the conditions of language shift), Journal of Russian humanitarian academic fund 1 (70),

Lewis, M. P. (2010) ‘The Sustainable Use Model and Expanded Graded Intergenerational Disruption Scale (EGIDS)’, ILAC VI, October 2010, Penang, Malaysia

Lewis, M.Paul, Gary F. Simons, and Charles D. Fenning (eds.). 2015. Ethnologue: Languages of the World, Eighteenth edition. Dallas, Texas: SIL International, http://ethnologue.com

Moseley, Christopher (ed.) 2010. Atlas of the World’s Languages in Danger, 3rd edn. Paris, UNESCO Publishing, http://www.unesco.org/languages-atlas/ (information about Wakhi language updated in September 2015)

Muller, K., Abbess, E., Tiessen, C., Tiessen, G. (2008) ‘Language Vitality and Development among the Wakhi People of Tajikistan’, SIL International, www.sil.org/silesr/2008/silesr2008-011.pdf

Obrtelova, J.; Sohibnazarbekova, R. (2012) ‘Проблемы редуцированных гласных и сочетания согласных в Ваханском языке’ (Issues of reduced vowels and consonant clusters in the Wakhi language in Tajikistan). In Journal of Tajik National university, 4/4 (91), Dushanbe: Sino, 149-154.

Pakhalina, Tatiana Nikolaevna (1975) Ваханский язык (Wakhi language), Moscow: Akademia Nauk SSSR

27

Donald STILO

(Department of Linguistics, Max Planck Institute for evolutionary Anthropology Leipzig)

Dikin Marāqei of Alamut: an undocumented conservative Tati language

Dikin Marāqei (henceforth DM), spoken in the village of Dikin in Alamut-e Sharqi District (formerly Rudbar-e Alamut District), is one member of the so-called Marāqei group of Tati. Marāqei dialects represent an enclave of the Central Tati type dispersed through some 16 villages in a wide-spread, mountainous area of Alamut and as such are completely unexpected for the area. This factor is clearly a result of a population movement from a much more westerly region in Iran, as ‘Marāqei’ implies. The speakers of these varieties only coexist with non-Marāqei Tatoid speakers and have no villages exclusively of their own. These dialects are generally of a highly conservative Tati type and are unintelligible to the neighboring Tatoid varieties even within their own villages.

Dikin had some 850 inhabitants in 170 families, at the time of my visit in 1976. This figure was given to me by the educated elite of the village themselves but the 2006 census (https://en.wikipedia.org/wiki/Dikin) now lists the population at 334 in 111 families. In 1976 (and surely now still) Dikin village had three different linguistic groups reflecting a divide among: A) adherents of a secret sect in the Alamut area who call themselves ‘Marāqei’ or ‘Kalle-bozi (Goathead)’ (ca. 500 people in 1976), and the regular Shiite villagers (called ‘Pashei (Mosquito)’ by the Marāqeis) who speak either B) ‘Dikin-Pashei’, a typical local Alamuti (Tatoid) dialect, or C) standard Tehrani Persian as a first language. The latter groups do not speak the Marāqei dialect, but the Marāqeis are all trilingual.

The Tatic family (Tati and Talyshi) as a whole is a highly conservative group among Northwest Iranian languages. My evaluation of this conservativeness is based on the following 12 factors (among others) that are also found in other NWI languages outside Tatic. Most varieties of Tatic still have at least a majority of these features:

CONSERVATIVE FEATURES OF THE TATIC FAMILY, ALSO FOUND IN MARĀQEI 1. Retention of grammatical gender, which is still robust in about half the Tati family, but

absent from Talyshi; 2. Retention of Direct vs. Oblique cases for nouns and pronouns, reflecting both gender and

number:

Dikin Marāqei:

Masc. (‘boy’) Fem. (‘girl’) ‘I’ ‘he’ ‘she’ ‘they’

Direct Sg. zollIg-Ø kellIg(-an) æz æv ǽv-an Oblique Sg. zollIg-i kellIg-ǽ man æv-í æv-ǽ Direct Pl. zollIg-ín kellIg-ín æv-ín Oblique Pl. zollIg-ón kellIg-ón æv-ón

3. The retention of distinct Direct and Oblique Person Agreement Markers (PAMs); the

Direct and Oblique PAMs are both used in the verbal system while the Oblique PAMs are also used as possessives in the NPs;

4. Retention of the 2nd singular Direct PAM -iš in most Tati and Talyshi varieties:

Dikin Marāqei: 1st sg. 2nd sg. 3rd sg.M. 3rd sg.F. 1st pl. 2nd pl. 3rd pl. Direct PAMs -en [-In] -iš -æ ~ -(æ)g -(ég)an -om -o -æn Oblique

PAMs =em =et =eš =emon =eton =ešon

28

5. A distinction of Intransitive and Transitive conjugations in the tenses built on the past stem of the verb:

Intrans. (‘I etc. fell’) genéstin genéstiš genéstæ (M) genéstian (F) genéstiom genéstio genéstiæn Trans. (‘I, etc. saw’) víndemæ víndetæ víndešæ víndemonæ víndetonæ víndešonæ

6. Ergative alignment of overt arguments in past tenses and Nominative-Accusative alignment in present tenses; The Agent argument is generally co-indexed by the Oblique PAMs in the past and the Patient is co-indexed by the Direct PAMs. In the Nominative-Accusative alignment of the present tenses the opposite is true:

æz æ jænig-ǽ m-in-In man æ jænek víndi-an man tö=m vínd-iš I.DIR that woman-O.F DUR-see-1S.DIR I.O that woman see.PAST-3S.F.DIR I.O you=1S.OBL see.PAST-2S.DIR ‘I see that woman’ (Nom.-Acc.) ‘I saw that woman’ (Erg.) ‘I saw you’ (Erg.)

Both sets of PAMs may occur simultaneously on the verb: Present (Nom-Acc.) Past (Ergative) m-in-en=eš vindi=m-ian DUR-see-1S.DIR=3S.OBL see.PAST=1S.OBL-3S.F.DIR ‘I see him/her’ ‘I saw her’

7. 3rd sg. forms in the Present and Subjunctive but only Intransitives in the Past agree in gender with subjects:

æ jænég-an či m-ej-égan? kinæ=tan æzínnæ kejow šiá bi-àn? that woman-DIR.F what DUR-say-3S.F.DIR daughter=2P.OBL yesterday where go.PAST.PPL AUX.PAST-3S.F.DIR ‘What does that woman say?’ ‘Where had your daughter gone yesterday?’

8. Most Tati varieties retain a synthetic passive morpheme. In DM this marker is -i- and seems to occur only in the Present while in the Past tenses a dedicated intransitive past formant -est is added to the active present stem forming a past intransitive root:

Present Active Past Active Present Passive Past Passive a-m-vaj-æ a-vat=eš-æ a-m-vaj -i-æg a-vaj-ést-æ PVB-DUR-pull.out-3SM.DIR PVB-pull.out.PAST=3S.OBL-AUX PVB-DUR-pull.out-PASS-3SM.DIR PVB-pull.out=PAST.INTRANS-3S.DIR ‘he uproots’ ‘he uprooted’ ‘it becomes uprooted’ ‘it became uprooted’

9. =(r)a only has a lexical sense, mostly benefactive: DM: öštö=ra ‘for you’.

In addition to the above nine features, there are three more features that generally typify the conservative nature of Tatic languages but these features are not present in DM (as far as is known): 10. A special Oblique formant -ær is used for terms of kinship, e.g. Vafsi xák-æ/xak-ǽr ‘sister (Dir./Obl.)’; 11. A dedicated negative Imperative (Prohibitive) in mæ- (Note DM: a-né-næn PVB-NEG-put ‘dont put!’) 12. On a more restricted basis, certain Tatic languages have retained an independent Optative paradigm. MAJOR INNOVATIONS FOUND THROUGHOUT TATIC (INCLUDING DM):

1. Dedicated genitival forms for personal pronouns and the interrogative pronoun ki ‘who’ via a čV- ~ Vš- prefix from Old Iranian hača ‘from’. These forms in fact represent the most salient diagnostic of the Tatic family. This prefix is added to the Oblique pronoun form but the DM 3rd person forms show some reductions/elisions:

29

DM: 1st sg. 2nd sg. 3rd sg.M. 3rd sg.F. 1st pl. 2nd pl. 3rd pl. ‘who’ Direct æz tö æv ǽv-an æmæ šömæ æv-ín ki Oblique man tö æv-í æv-ǽ æmæ šömæ æv-ón ki Possessive če-man öš-tö č-e-y č-ǽ č-æmæ šömæ č-ón iš-ki

2. Various languages have undergone alterations to their ergative alignment. Some only lose agreement with P, others have innovated a Double Oblique alignment and some have no Patient marking whatsoever (among other possibilities). While DM generally has canonical ergative alignment, there are some regular deviations:

a. As we see in the above table, 2nd sg/pl, 1st pl pronouns and ki lack the Direct-Oblique distinction and thus have neutral alignment (but retain the ergative alignment in coindexing the Patient in the verb):

æmæ tö vind-iš tö æmæ vindi-on we you.S see.PAST-2S.DIR you.S we see.PAST-1P.DIR ‘we saw you (sg)’ ‘you (pl) saw us’

šömæ æmæ vindi-on ki tö vind-iš you.P we see.PAST-2P.DIR who you.S see.PAST-2S.DIR ‘you (sg) saw us’ ‘Who saw you?’

b. When æz/man ‘I’ functions as a past Patient, an Oblique is used, yielding Double Oblique alignment: æ killig-ǽ man=eš vindi-n that girl-O.F I.OBL=3s.OBL see.PAST-1S.DIR ‘That girl saw me.’ (see also Point 6 above for examples of man as past Agent)

INNOVATIONS SPECIFIC TO DM (and some other Tati varieties but not universally) Phonology:

The most distinctive phonological development of DM is a push chain of vowel shifts: *a > ā, * ā > o, *o > ö: *a > ā: man ‘I’, čašm ‘eye’, haft ‘7’, da ‘10’, dard ‘ache’, varf ‘snow’, bast- ‘close, tie (past stem)’ *ā > o: bero ‘brother’, po ‘foot’, nom ‘name’, yonzæ ‘11’, pošna ‘heel’, šona ‘shoulder’, mion ‘middle’ *o > ö: dö ‘2’, nö ‘9’, tö ‘you (sg)’, šömæ ‘you (pl)’

Morphology and Syntax 1. Some singular case endings (masculine Oblique, feminine Direct) – but not feminine

Oblique or any plural forms – become separated from the root by Oblique PAMs (as fronted Agents in the past and as possessives):

æsíf=em-an a-gáti-an zemin=de. beró=t-i dǽss=em-an dæss-ǽr=et čašm-ón=et apple=1S.OBL-D.F PVB-took-3SF.DIR land=from brother=2S.OBL-O.M hand=1S.OBL-D.F hand-O.F=2S.OBL eye-O.PL=2S.OBL ‘I got the apple (f.) off the ground.’ ‘your brother

(Obl.)’ ‘my hand (Dir.)’

‘your hand’ (Obl.)

‘your eyes’(Obl.)’

2. The Fem. Direct case marker -àn only occurs when the noun is definite. This marking of definiteness does not occur in masculine or any plural nouns and does not cooccur with possessive PAMs; this area of DM grammar remains enigmatic. More field would be necessary to resolve these issues.

3. 3rd sg. masc./fem. Direct PAM in -k/g-: mi-b-ik ‘it becomes’; bé-š-ik/bé-š-Ígan ‘that he/she goes (Subjunct.) and in the negative of ‘to be’ nÍggion/nÍgæ/nÍggian/nÍggio ‘I am not/he isn’t/she isn’t/you (pl) aren’t’, etc.

4. -k/g also occurs in Perfects before PAMs: bag-in/bag-iš ‘I/you have become’ votag=iš biæ ‘s/he had said’.

5. Postpositions are most common but DM has a few prepositions: Postpositions: =de ‘from’, =ro ‘for’, hæmro ‘with’, sa ‘on’; mion ‘in, into’, pey ‘after’; Prepositions: to ‘until’, bo ‘with (alternate lexeme)’, bi ‘without’;

30

6. Present and imperfect are formed with the unstressed prefix me-: me-zon-ǽ ‘he knows’/me-zonést-æ ‘X knew’

In addition to a general sketch of the morphology and syntax of DM (for which the handout will be essential as a reference), I will devote a segment of the talk to discussing my view of the history of the language and its relationship to other forms of Tati. This segment will also require an examination of the phonological innovations and especially lexical relationships to other Tati varieties to determine both the provenance of DM and its genealogically closest relatives. In addition, the effect of language contact and areal phenomena (especially in the lexical domain) typical of the Taleqan and south Caspian area that have acted upon DM will also be demonstrated.

ABBREVIATIONS: AUX = copula as auxiliary, D = Direct case, DIR = Direct PAM, DUR = Durative, F = Feminine, M = Masculine, NEG = Negative, O = Oblique case, OBL = Oblique PAM, P = Plural, PASS = Passive, PPL = Past Participle, PVB = Preverb, S = Singular Murad SULEYMANOV (École pratique des hautes études)

Participial Relative Clauses in Şirvan Tat

Tat spoken in the Caucasus is a series of related dialects or even languages spoken mainly in the Republic of Azerbaijan which belong to the southwestern branch of Iranian languages. They are not to be confused with Tati, a cluster of Northwest Iranian languages spoken in Iranian Azerbaijan. Tat derives from an early form of New Persian but has developed independently for at least the past millennium. It is divided into two main dialect groups with little to no mutual intelligibility: Judaeo-Tat (JTat), historically smaller in the number of speakers, but relatively well studied and with quite an abundant secular literary tradition; and the non-written Muslim Tat (MTat), much less studied, with no credible data on the current number of speakers, and classified by the UNESCO Red Book of Endangered Languages as ‘severely endangered’.

For centuries, Tat has been in contact with Azeri, an Oghuz Turkic language, as well as with East Caucasian languages. The dominant role of Azeri has resulted in profound Turkic influence on Tat, including its phonology, common vocabulary, derivational morphology and subordinate sentence structures.

This presentation, which is based on data from my fieldwork, will focus on relative clauses in the Muslim Tat dialect of the region of Şirvan (central Azerbaijan).

Along with the Persian-like relative strategy using the clitic ki (not covered in this presentation), all Tat varieties possess a set of participial relative constructions which could be due to Turkic and partly also to East Caucasian influence. MTat does not have an equivalent of the Persian present participle in -ande, leaving the past participle in -t/d/rA to cover the domain of both past and present (its range is thus parallel to the Azeri participles in An). This is also the participle used to relativise subjects (note that in Tat noun phrases, the attribute precedes the head):

(1) bä häzrät yaraš nä=do-rä či=yi LOC saint allure NEG=give-PTCP thing=3 “This is something that does not suit a prophet.”

In the northern sub-dialects of Şirvan Tat, which are less exposed to Azeri influence, relatives with passive semantics may or may not contain passive marking (usually expressed by participle + birän ‘to be’).

31

(2a) di-rä xuna älmun=i (2b) di-rä bi-rä xuna älmun=i see-PTCP house ours=2/3 see-PTCP be-PTCP house ours=2/3 “The house that is seen (i.e. visible from here) is ours.”

Unlike Azeri, which makes use of participles in -dXQ, MTat does not have a separate participle form for relativising elements other than the subject. Instead, MTat has expanded the use of the participles in -t/d/rA to express non-subject relative clauses. Where the domain nominal (DN), i.e. relativised term, is not indexed, the relative participle takes a possessive marker which cross-references, as in Azeri, either the implicit (ex. 3) or explicit (ex. 4) possessor. The examples below confirm that in MTat, the equivalent of the Persian past participle has lost its passive orientation (Authier 2012a: 248):

(3) zä-re=šun hä tojir=ä xištän=i=yi hit-PTCP=POSS3PL FOC merchant=OBL self=POSS3=2/3 “The one they were beating was the merchant himself.”

(4) Čingiz Abdullayev=ä bä nȫš-te=yi oxšar=i PN=OBL LOC write-PTCP=POSS3 alike=2/3 “This is similar to what Çingiz Abdullayev writes.”

At first glance, MTat appears to have copied the Turkic strategy. However, in the cases where the DN is indexed, the relative clause takes the form of a possessive construction, which in MTat is double-marked, with the adjunct in the oblique case and the head marked by a possessive suffix, leaving the participle unmarked:

(5) Cečäl=ä vo do-ra pül-ho=yi näne=yi ye Baldman=OBL with give-PTCP money-PL=POSS3 grandmother=POSS3 one gözäl-ä imorät-i dux-tabɨ beautiful-ATTR mansion-INDF sew-PQP3 “With the money given by Baldman, his grandmother had built a beautiful mansion.”

This head-marked relative clause in MTat differs significantly on the one hand from Azeri (where the possessive marking remains on the participle regardless of whether the DN is indexed or not) and on the other hand from the Iranian languages spoken in the region, notably from JTat and Talyshi (where possessive marking in this case is altogether absent; see Authier 2012b: 246–250, Kaye 2015).

Finally, Tat exhibits unmarked non-subject relative constructions where the participle is morphologically identical to the one found in the subject relative clause structure in ex. 1. This particular structure, which is not systematic in MTat (nor in Azeri or JTat where it also occurs), seems to be triggered by specific semantic factors which remain to be studied.

(6) män, tɨ nä=di-rä či-ho=rä ü di-re I you NEG=see-PTCP thing-PL=OBL s/he see-PRF2/3 “He saw things that you and I would not have seen.”

Furthermore, despite the absence of a cross-reference, the subject of the relative clause can even be dropped if it is deemed unimportant or obvious from the context.

(7) män tä bä di-rä jö umo-rom I till LOC see-PTCP place come-PRF1 “I have come to the place that (you) see.”

(8) hazor manot do-ra märd=ä soğ-i=yi=rä mu-xoh-um thousand manat give-PTCP man-OBL alive-NMLZ=POSS3=OBL EVT-want-1 “I will wish for the well-being of the person whom (I have) given one thousand manats.”

32

The presentation will aim at explaining constraints on possessive marking of participles in clauses where the DN is not indexed, as well as offer a hypothesis regarding factors triggering a subject relative construction for non-subject relativisation.

References: AUTHIER, Gilles (2012a). “New strategies for relative clauses in Azeri and Apsheron Tat.” In Gast, Volker and

Holger Diessel (eds.). Clause Linkage in Cross-Linguistic Perspective. Data-Driven Approaches to Cross-Clausal Syntax, 225-252. Berlin, Boston: De Gruyter

AUTHIER, Gilles (2012b). Grammaire juhuri, ou judéo-tat, langue iranienne des Juifs du Caucase de l'est. Wiesbaden: Reichert

KAYE, Steven (2015). Relative constructions in Northern Talyshi. Presentation at the Sixth International Conference on Iranian Linguistics. Tsereteli Institute of Oriental Studies/Ilia State University. Tbilisi, June 23–26

Brigitte WERNER (SIL International)

Forms and Meanings of the Ezafe in Zazaki

Zazaki is a Northwestern Iranian language spoken by 2-3 million people; the homeland of the Zaza people is in eastern Turkey. The language is regarded as marginalized because a slow language shift towards the prominent national language, Turkish, can be observed, and Northern Kurdish (Kurmanji) has a growing influence in these regions. This research focuses on the variation of the ezafe form, especially as it is found in noun phrases that consist of a masculine head noun plus an adjective. The variation is examined in data from three different locations: the Cermik/Siverek region, the Dersim/Varto region, and the Palu/Bingöl region. The results of this research show that in the data from Cermik/Siverek region, there is more diversity of forms of the ezafe than in data from the other regions. In Persian grammar, the ezafe is often described as a linking morpheme. In contrast, in Zazaki, the ezafe has many forms and a range of syntactic functions. In the corpus of around sixty or more noun phrases for this research, the ezafe forms are examined according to the syntactic relationship in which the noun phrase stands: nominative, accusative, dative, genitive, or allative. The marked ezafe forms (oblique) are used to mark a noun phrase as definite accusative, dative, or genitive. It is noteworthy that in the Cermik/Siverek variety, the “non-nominative-non-accusative” relation is marked by a form that has a prefixed /d/ on the ezafe. This can occur in noun phrases in which the head noun is masculine or feminine, and also in plural noun phrases. That feature resembles the form of oblique pronouns used for recipient or dative in the Dersim/Varto variety. There, the prefix /d/ is attached to the oblique pronouns: /d-ey/ ‘3.masc.’, /d-ae/ ‘3.fem.’, /d-inu/ ‘3. pl.’. These forms (“d-pronouns”) do not exist in the other varieties. Below, two example sentences are presented with prefixed /d/ on the pronoun (1), and prefixed /d/ on the ezafe (2): (1) Dersim/Varto Variety:

ez kitab don dae 1S book give.1S 3SF.OBL (RECIPIENT)

‘I give her the book.’

(2) Cermik/Siverek Variety: ez-o kitab dan-a mar-da mı 1S- COP.1S book give-1S mother-EZ.F 1S.OBL (RECIPIENT)

‘I give my mother the book.’

Based on these results, a generalization can be stated that prefixed /d/ refers to the recipient/goal relation. This supports the hypothesis that the ezafe has evolved from pronouns. Another feature

33

that also supports this hypothesis is the existence of the “absolute” ezafe, an independent form of the ezafe, which expresses an anaphoric or demonstrative meaning. An example of the absolute ezafe is given in sentence (3).

(3) Cermik/Siverek Variety: a-ya kı en xasek-a ê mın-a 3SF-EZ(F) COMPL most pretty-COP.3SF EZ 1S.OBL-COP.3SF ‘She, the (one) that is the prettiest, is the (one of) me.’ ‘The one who is prettiest, she is mine.’

This study is based on my own fieldwork and also on material in published grammar descriptions by several authors, including many who are speakers of Zaza. The grammatical studies by Terry Todd (1985), MacKenzie (1993), Ludwig Paul (1998), and C. M. Jacobson (2001) are the main academic foundations for this research.

WORKSHOP “SPECIAL SESSION ON LANGUAGE MAPPING”: MAPPING ENDANGERED LANGUAGES OF IRAN

Erik ANONBY & Amos HAYES

(Linguistics, Carleton University & Geomatics and Cartographic Research Centre, Carleton University)

Language mapping is an important tool for language documentation, providing insight into the geographic and social distribution of languages and linguistic forms. Its application to endangered languages brings with it a specific set of methods and outcomes. This portion of the ISEIL 2016 Special Session on Language Mapping, which focuses on endangered languages of Iran, will address the following topics:

Mapping endangered languages. What role does language mapping play in establishing priorities for further documentation? How does the mapping of endangered languages differ from mapping of other languages? How might the mapping of a small language community be beneficial or harmful to its speakers?

The Iranian context. What work has already been done on mapping endangered languages in Iran? How have institutional responses to language endangerment shaped this work? Is it important or even feasible to map the languages of the Iranian diaspora?

The Nunaliit Atlas Framework. What is Nunaliit, and how is used for language mapping? What work is being done on mapping endangered languages within the Nunaliit-based Atlas of the Languages of Iran? How are new atlases set up with Nunaliit?

Participants will explore these questions through collaborative activities including the development of a questionnaire for mapping endangered languages of Iran, and through practice in linking linguistic data to maps in the context of a live online language atlas.

34

Université Sorbonne Nouvelle – Paris 3 Maison de la Recherche, salle Prestige

4, rue des Irlandais 75005 Paris

Métro(subway):ligne7,CardinalLemoine

RER:ligneB,LuxembourgBus:89,84,47,27