The role of discourse context in Chinese L2 learners' processing of English relative clauses

Abstract

 This study investigated whether and to what extent Chinese L2 learners were able to exploit discourse-level information in processing English relative clauses. Twenty-eight intermediate-to-advanced Mandarin-speaking L2 learners of English as well as twenty-eight native English speakers were administered a self-paced reading experiment, in which they were required to read sentences whose subjects were modified by two types of relative clauses with different extraction sites, while the information status of the subject or object NPs within the relative clauses was independently manipulated by virtue of the preceding context sentences. Online reading times and offline comprehension question response accuracy show that in line with the native speakers, L2 learners experienced significantly more processing difficulties with object-extracted relative clauses than their subject-extracted counterparts. The reading time penalty of object-extracted relative clauses with topical embedded subjects was significantly reduced in native speakers compared to those with non-topical embedded subjects, but the information status of the embedded subjects did not significantly affect the processing complexity of object-extracted relative clauses in L2 learners. The contrasting patterns of results observed in native vs. non-native speakers are consistent with the prediction of Sorace’s Interface Hypothesis and Cunnings’ retrieval interference hypothesis, suggesting that as compared to the native speakers, the L2 learners might be less efficient in integrating information sources from across the syntax and discourse domains or more prone to memory retrieval interference from discourse prominent items during online sentence processing, but contradict the prediction of Clahsen & Felser’s SSH, whereby discourse contextual factors would exert a considerable influence on L2 sentence comprehension.