Noname manuscript No. (will be inserted by the editor) On Iranian Case and Agreement Faruk Akkuş Received: date / Accepted: date Abstract This paper investigates case and agreement patterns in Iranian languages, mainly focusing on Zazaki and Kurdish varieties. Empirically, the paper discusses the typologically rare double-oblique pattern, along with a novel way of splitting the oblique. On the basis of the syntactic behavior of oblique-bearing arguments, the paper argues that the term ‘oblique’ corresponds to distinct cases, ranging from structural accusative case to nonstructural dative case and ergative case. Oblique number agreement is case-sensitive, targeting only ergative-oblique out of the oblique cases. In order to capture the facts, I adopt a Multiple Agree account (Hiraiwa 2005), in which partial number agreement is a process that takes place in the morphology via Impoverishment, and not in the syntax proper. The study proposes to capture the case patterns in Iranian languages along the lines of Svenonius (2006), in which arguments bearing nonstructural case get their licensing from a combination of two heads (cf. Chomsky 1993), one of which is Stem, the locus of split-ergativity in Iranian. A chain is established between Stem and v, which yields the nonstructural case on internal arguments, whereas the ergative case on external arguments is the result of the chain between Stem and Voice. Keywords ergativity · oblique · case · (partial) agreement · Zazaki · Kurdish · Wakhi 1 Introduction Iranian languages are currently spoken across a vast stretch of Asia, ranging from the westernmost provinces of China to Central Anatolia in Turkey. Their speakers inhabit several distinct geographic and cultural areas and have been in long-standing contact with numerous genetically diverse languages. Naturally, the Iranian languages have come to manifest a substantially divergent typological spectrum. In this paper, I investigate alignment patterns, focusing on varieties of Kurdish, Zazaki and to some extent Wakhi. University of Pennsylvania Department of Linguistics, 3401-C Walnut Street, Suite 300, C Wing, Philadelphia, PA 19104 E-mail: [email protected] 2 Faruk Akkuş Iranian languages are of theoretical interest since they can inform us about the phenomenon of split-ergativity, displaying microvariation in case and agreement patterns both intra- and inter-dialectically. Furthermore, in Iranian languages overt case and overt agreement are correlated, thus bearing implications for the relationship between case and agreement. Several accounts have been given for case and agreement alignments as well as alignment changes in Iranian languages (e.g. Haig 2008; Baker and Atlamaz (B&A, 2014) Atlamaz and Baker 2016, 2018; Karimi 2013, 2015; Kareem 2016; Stilo 2009). Haig (2008) provides a Construction Grammar approach, while Baker and Atlamaz (2014) combine the agreement-assigned case system with dependent-case theory in order to derive the cross-dialectal variation. Atlamaz and Baker (2018) adopt a recent interpretation of the dependent case hypothesis, primarily to explain partial number agreement in Kurdish. Karimi (2013), on the other hand, assumes an inherent case approach coupled with an intervention account. I undertake a careful investigation of the split case system and agreement patterns in Kurdish and Zazaki varieties, including the previously unstudied Mutki Zazaki (and Wakhi), which have typologically rare case patterns. This empirical contribution leads to theoretical insights, and calls for a different theoretical picture from what has been argued for Iranian languages. I investigate the nature of the term ‘oblique’ since it is widely used in the context of Iranian languages, and suggest a tripartite split for the term ‘oblique’. I demonstrate that the oblique encompasses cases ranging from structural ACC case to nonstructural DAT case and nonstructural ERG case. This investigation reveals several issues with case assignment analyses that rely on the notion of oblique case. Given the different syntactic behaviors of oblique-marked NPs, I draw a parallel with absolutive marking, which has been argued to be non-uniform (Legate 2006, 2008). I demonstrate that ergative case has a transitivity condition, and that number agreement is case-sensitive, targeting only ergative-oblique out of the oblique cases under well-defined conditions. I account for these facts by employing Multiple Agree, combined with Impoverishment rules, to yield the correct output. I argue for a bipartite case-assignment system, in which nonstructural-case bearing arguments (including ergative case bearing arguments) get their licensing from a combination of two heads in the projection of the verb à la Svenonius (2006). I define the locus of split-ergativity in Iranian languages to be what I call StemP, located lower than TAM projections in the structure. A chain is established between the projection headed by Stem and v to yield nonstructural case on internal arguments, whereas the ergative case on external arguments is the result of a chain between Stem and Voice. The paper is organized as follows: Section 2 introduces the case and agreement patterns in Kurdish and Zazaki. Section 3 investigates the nature of oblique arguments and demonstrates that oblique corresponds to various distinct cases. Moreover, it describes the partial agreement pattern in these languages. This is followed by an account of the partial agreement facts in Section 4. Section 5 proposes a case assignment mechanism for Iranian languages based on Svenonius (2006). Section 6 summarizes and concludes the paper. 2 Alignment in Iranian Languages This section mainly describes the case and agreement patterns across Kurdish and Zazaki languages. As a terminological note, Iranian languages show a basic On Iranian Case and Agreement 3 morphological case distinction between ‘marked’ and ‘unmarked’ cases. Following the tradition in Iranian linguistics, I call the unmarked case direct (DIR) and the marked case oblique (OBL) in this paper. The OBL is commonly used for objects in the present stem (NOM-ACC alignment), and transitive subjects in the past stem (ERG-ABS alignment). In the following discussion, I refer to various dialects of Kurdish and Zazaki. From the Kurdish varieties, I focus on Adıyaman Kurdish (AK) and Muş Kurdish (MK). Zazaki dialects can be categorized into two groups: (i) dialects that exhibit the canonical ergative alignment in past tense, such as Dersim, Siverek and Varto, which are treated as Standard Zazaki (StZ) although I sometimes refer to the names used in the original sources, (ii) nonstandard varieties that diverge from the canonical ergative alignment, such as Mutki Zazaki (MZ). 2.1 The alignment patterns in Kurdish varieties Adıyaman Kurdish (Atlamaz 2012) has a split-ergative system.1 In the past tense, Adıyaman Kurdish’s case system follows an ergative-absolutive pattern. That is, in a transitive structure, the object receives the unmarked (nominative) case, and the verb agrees with the object in person and number. The subject receives ergative case. Subjects of intransitive clauses, on the other hand, receive unmarked case. Examples in (1) illustrate this. (1) a. ez rıvi-m 1sg.dir run.past.1sg ‘I ran.’ (AK, B&A 2014, 3a) b. mı tı di-yi 1sg.obl 2sg.dir see.past-2sg ‘I saw you.’ (AK, B&A 2014, 4a) In the present tense (which can be characterized as non-past; i.e. referring to ongoing as well as future events), AK has a nominative-accusative system. In these environments, it is the subject that displays DIR marking and agrees with the verb, and the object appears with the OBL marking, as in (2). (2) a. ez dı-rv-ım-e 1sg.dir impf-run.pres-1sg-cop ‘I am running.’ (AK, B&A 2014, 1) b. ez te dı-vun-ım-e 1sg.dir 2sg.obl impf-see.pres-1sg-cop ‘I am seeing you.’ (AK, B&A 2014, 2a) Therefore, the alignment pattern can be summarized as in Table 1.2 1It is identical to the standard/literary Kurdish variety described in Thackston (2006) in most respects, including the alignment system. 2Following Dixon (1994), S stands for the subject of an intransitive verb, A for the subject of a transitive verb and O for the object in a transitive verb. 4 Faruk Akkuş Case Marking Agreement DIR OBL Present S, A O S, A Past S, O A S, O Table 1 Alignment in Adıyaman Kurdish Muş Kurdish (MK) shows an important deviation from the canonical pattern: the so-called double-oblique pattern (Gündoğdu 2011). This is the most widespread type of “deviant ergative pattern” in Kurdish languages (Haig 1998, 2004, 2008, 226; Dorleijn 1996; Gündoğdu 2011). This dialect is exactly like AK in present clauses, but it is different in past clauses, where the transitive subject and the direct object are both oblique-marked. The verb then bears invariant third singular agreement by default, and there is no DP in direct case in the clause. Thus, (3a) and (3b) are like AK, but (3c) is notably different.3 (3) a. ez te di-bin-im 1sg.dir 2sg.obl impf-see.pres-1sg ‘I see you’ (MK, Songül Gündoğdu, p.c.) b. ez ket-im 1sg.dir fall.past-1sg ‘I fell down.’ (MK, Gündoğdu 2011, 77) c. min te dît 1sg.obl 2sg.obl see.past.3sg ‘I saw you.’ (MK, Gündoğdu 2011, 81) The alignment pattern in Muş Kurdish is illustrated in Table 2. Case Marking Agreement DIR OBL Present S, A O S, A Past S A, O S Table 2 Alignment in Muş Kurdish Note that although AK and MK differ in their ergative alignment in past transitive clauses, they exhibit the same pattern in the S arguments. The subject of an intransitive clause bears DIR. Unergative and unaccusative sentences behave identically, in the sense that their subjects are unmarked, as shown in (1a) and (3b) respectively. 2.2 The alignment patterns in Zazaki varieties Zazaki (another north-west Iranian language) is often regarded to have most robustly retained the historically canonical ERG-ABS alignment in the past (e.g.
Details
-
File Typepdf
-
Upload Time-
-
Content LanguagesEnglish
-
Upload UserAnonymous/Not logged-in
-
File Pages55 Page
-
File Size-