Processing number agreement in Turkish

This study examines the role of the [animacy] feature in 3rd person plural agreement in L2 Turkish within the framework of the Feature Reassembly Hypothesis (Lardiere, 2009), focusing on detectability in contexts where the target feature is optionally realized and interacts with another feature. We test two groups of L2 learners of Turkish (L1-Persian and L1-English speakers) using a speeded acceptability judgment task to evaluate (i) whether the detectability of the target feature is impacted by learners’ L1 backgrounds and (ii) whether higher proficiency facilitates successful feature reassembly in the detectability of the [animacy] feature. We considered two possible outcomes: learners whose L1 grammar has the [animacy] feature in the same domain may show facilitative effect (for L1-Persian speakers) or shifting from a categorical agreement system to a more flexible one may provide clearer cues that facilitate the detectability of the [animacy] feature in the L2 (for L1-English speakers). Results show that L1-Persian speakers showed persistent transfer effects, even at higher proficiency levels. L1-English speakers, in contrast, showed an emerging Turkish-like pattern as their proficiency increased, suggesting that proficiency can overcome the challenges posed by optionality and feature dependency. These results provide new evidence that the detectability of a feature, particularly when it is optionally realized and its realization depends on another feature, poses challenges for L2 learners.