ABSTRACT
This study investigated how native speakers and second language (L2) learners process a Chinese construction “ba NP1 Verb1 de NP2 Verb2” (e.g., “ba brother bring over de friend praise-PERF”), which is temporarily ambiguous between a subject-extracted relative clause (SRC, “the friend who brought over the brother praised [someone/something]”) and an object-extracted relative clause (ORC, “[someone] praised the friend whom the brother brought over”). In an on-line self-paced reading experiment, 36 Chinese native speakers and 52 advanced L2 learners from both head-initial and head-final first languages (L1s) were presented with four types of sentences – temporarily ambiguous SRC, unambiguous SRC, temporarily ambiguous ORC, and unambiguous ORC. The results showed that both native speakers’ and L2 learners’ default interpretations of the construction was an SRC, but SRCs were read more slowly than ORCs by L2 learners. Additionally, L2 learners speaking head-final L1s processed head-final Chinese RCs faster than those speaking head-initial L1s. These findings are consistent with the predictions of the linear distance theory, the dependency locality theory, and the canonical word order theory. Pedagogical implications of the study are also discussed.
本文采用移动窗口范式考察汉语母语者和汉语二语学习者对歧义短语“把NP1 Verb1的NP2 Verb2”（把哥哥带来的朋友表扬了）的加工。该短语可按主语关系从句（把哥哥带来的朋友表扬了大家）或宾语关系从句（把哥哥带来的朋友表扬了以后）进行续接。实验发现母语者和二语学习者均将其解歧为主语关系从句，但二语学习者加工主语关系从句的速度慢于加工宾语关系从句。实验结果与线性距离理论、附属位置理论及词序理论的预测相符。本文还讨论了研究结果在对外汉语教学中的应用。
