Mismatches in subject verb agreement: The processing perspective

Thesis Type: Postgraduate

Institution Of The Thesis: Middle East Technical University, Graduate School of Social Sciences, Turkey

Approval Date: 2020

Thesis Language: English


Supervisor: Martina Gracanin Yüksek


This study investigates the processing of mismatches in subject-verb agreement in Turkish by L1 speakers. Two online experiments were conducted in which agreement mismatches in Turkish were tested with two kinds of subjects: numeral quantifiers and comitative constructions, which did not cause ungrammaticality in the sentence. Both experiments were self-paced reading tasks on the Ibex Farm platform. Experiment I tested which subject-verb agreement pattern was easier to process in items with numeral quantifier subjects, e.g., üç kişi ‘three persons’; the matching pattern was 3rd person singular (or plural) morphology on the verb, but two other patterns, 1st and 2nd person plural, were possible as a mismatch condition. It indicated that 1st person plural verbal agreement was easiest to process with numerally quantified subjects, and this goes counter to the common assumption that numeral quantifiers agree with 3rd person singular (or plural). Experiment II compared the processing of two interpretations of comitative construction, e.g., Ali’yle ben ‘I and/with Ali’: symmetrical and asymmetrical. The symmetrical interpretation, in which the comitative phrase is interpreted as a coordinated phrase, was deemed a mismatch condition, while the asymmetrical interpretation, in which the prepositional phrase Ali’yle ‘with Ali’ is analyzed as an adjunct on the subject ben ‘I’ was a match condition. It demonstrated that comitative construction subjects were mostly interpreted as a coordinate structure, which is surprising because Turkish has bona fide coordination headed by the conjunction ve ‘and’. Both experiments showed that mismatch conditions were favored by the participants. To discover whether this preference for agreement mismatches is specific to these two phenomena, other agreement mismatches should be studied.