Case and Phi-agreement in Laki: Parametrizing split-ergativity in Kurdish

This thesis examines split ergative alignments in five dialects of Kurdish (Northwestern Iranian), with the primary focus on the argument structure and case and agreement patterns in Laki (Southern Kurdish). I show that these Kurdish dialects uniformly follow an aspect-based split where the ergative pattern is obtained only in the anterior aspect structurally encoded on a low Asp head. Along the lines of some previous literature (Bobaljik 1992, 1993; Otsuka 2000; Bhatt 2005; Preminger 2012; Rezac et al. 2014), I take ergative case in Kurdish as a structural case (as opposed to an inherent case), disassociated from theta-role assignment. Extending Preminger’s (2012) proposal, I take the locus of ergative to be the head of a non-thematic ErgP. The main evidence for this proposal is based on the agreement pattern with complex predicates formed by a transitive light verb (a light verb whose heavy verb counterpart is transitive), where we see the ergative pattern with transitives, unaccusatives and unergatives. I further propose that the Erg head selects for a phrase bearing [+ANTERIOR] and [+TRANSITIVE] features. Despite cross-dialectal similarities, however, Kurdish dialects show a high degree of variation arising from various dialect-specific parameters. One source of variation is the behavior of absolutive objects. In dialects with a phasal ErgP, the absolutive agreement with the object always fails, unless, the object undergoes raising to the phase edge. In Laki (Southern Kurdish), a weak/deficient object pronoun raises to the phase edge due to a Clitic Cluster Restriction. In Mukriyani, (Central Kurdish), weak/deficient objects consistently raise to the phase edge for licensing. In Ardalani (Central Kurdish), objects consistently remain in-situ, and the phasehood of the ErgP renders them inaccessible for absolutive agreement. In Northern Kurdish dialects, by contrast, where ErgP is φ-defective and not phasal, supported by the unavailability of a φ-probe on the Erg head, the object is predicted to be accessible for agreement with T. In the Standard dialect, the agreement between T and the object is consistently obtained. In Muş dialect, the object receives accusative case rendering it inactive for agreement with T, leading to a double oblique pattern. These parameters and their interactions with each other result in diverse patterns within a single dialect and across dialects which are otherwise similar. Building upon the insight of previous research (Johns 2000, Deal 2016a, Haig 2017, Carrier 2020), I conclude that ergativity is not a single phenomenon. Rather, it is a constellation of a range of linguistic properties.

Department dc:contributor.department

Linguistics

Year dc:date.issued

2024

Author dc:creator

• Taghipour, Sahar

Advisor dc:contributor.advisor

• Kahnemuyipour, Arsalan

• Agreement
• Case
• Ergativity
• Iranian languages

Handle dc:identifier.uri

http://hdl.handle.net/1807/140070

OAI identifier oai:identifier

oai:utoronto.scholaris.ca:1807/140070