The Acquisition of Null and Overt Subjects in L2 Turkish at the Syntax Discourse Interface Across Different Languages

Abstract

This study investigates the acquisition of null and overt subjects at the syntax-discourse interface across discourse null subject and non-null subject languages in L2 Turkish and focuses on L1 transfer effects. Considering this, the study aims to understand whether discursive constraints residing at this interface is problematic to acquire as predicted by the Interface Hypothesis in contrast to Full Transfer / Full Access Hypothesis. Framed within the Syntax before Discourse Hypothesis considerations, it also inquires whether syntactic features are acquiered earlier than discursive principles regarding the use of null and overt subject distribution. The participants of the study consist of two learner groups and a control group. Intermediate and advanced L2 Turkish speakers of Korean and Japanese (n=42, intermediate=15, advanced=27) constitute the discourse null subject group. English and German L2 Turkish speakers at the intermediate and advanced level (n=26, intermediate=11, advanced=15) form the non-null subject group. The control group of the study is composed of native Turkish speakers (n=26). Three tasks were employed to fetch data. The Overt Pronoun Constraint Task was designed to assess the syntactic knowledge of the L2 Turkish speakers on the formal distribution of null and overt subjects in complex sentences with quantified/wh-word or referential DP antecedent contexts. As for the latter two tasks, the Contextualized Grammaticality Judgement Task and Question-Answer Task were framed to evaluate the acquisition of discursive constraints regulating the null and overt subject distribution at the syntax-discourse interface. For each task in the study, descriptive analysis and three statistical analyses were carried out. With respect to the findings of the study, the data taken from the the Contextualized Grammaticality Judgement Task and Question-Answer Task found that the L2 speakers of both learner groups and proficiency levels had problems in acquiring discursive constraints (topic continuity and topic shift) which govern the use of null and overt subjects. This finding is in line with the assertions of the Interface Hypothesis. Since the interface between syntax and discourse presents learnability problems, features residing at this interface cannot be fully acquired. Drawing on the languages with different typologies, the study also suggested that L1 transfer does not hold at the syntax-discourse interface since no difference was observed between two learner groups even though the speakers of discourse null subject group have the same discursive rules in their mother tongues as in Turkish, which necessarily refutes the claims of the Full Transfer models. This study also found that topic continuity and topic shift constructions were not violated equally by the L2 Turkish speakers. Since overt subjects are more salient to perceive and carry more information than null subjects in discourse, violating overt subject constructions in topic-shift contexts is less tolerable than using redundant (overt) subjects in topic-continuity contexts. Finally, the results of the Overt Pronoun Constraint Task found that the formal constraints that govern the co-indexation between subjects of the embedded and matrix clause were acquired by the advanced speakers of both groups despite some target deviant performances in the context of overt subjects. Since the L2 participants were more successful in this task than the discourse-bound tasks, this finding is in line with the Syntax before Discourse Hypothesis.