Presentations

Abstract

This study aimed to investigate the association between statistical learning (SL) and morphosyntactic processing in Turkish. In an eyetracking-while-listening experiment, 52 adults listened to Turkish reversible canonical, reversible scrambled and non-reversible sentences, while viewing two pictures (target/foil). The participants had to use morphological cues (case markers) on the noun phrases (NP) to identify the target picture. Gaze proportions to target picture were calculated for each interest period: NP1, NP2 and verb phrase (VP). In addition, auditory and visual SL tests were implemented utilizing triplets of auditory (synthetized syllables) and visual (alien pictures) stimuli with higher transitional probability within than between triplets. The participants’ short-term memory capacity and nonverbal intelligence scores were also measured using digit span tests and Raven’s matrices. Linear mixed-effects models showed that canonical sentences were associated with more gaze proportions to target than scrambled sentences on NP1. Also, an interaction was found between auditory SL and sentence type on VP, which stemmed from greater gaze proportions to target in canonical sentences than scrambled sentences for participants with higher ASL scores. This finding suggests that individual differences in SL ability correlate with morphosyntactic processing, possibly due to differences in sensitivity to distributional and transitional statistics of case markers.