| Sentence processing is the basis of language comprehension. It is of great significance to probe into the mechanism of sentence processing for the study of language comprehension. Yet, there is still a controversy on how semantic and syntactic units in one sentence are immediately integrated into a single coherent explanation during online sentence processing. Traditionally, researches on this question are commonly based on behavioral approaches which study the language processing system through observing and analyzing the reaction time and accuracy measure. According to behavioral studies, two basic models of sentence processing system have been set up:Modular model and Interactive model. The former insists that sentence processing system begins with independent syntactic processing and then follows with lexical and semantic processing; the latter believes that syntactic, semantic as well as other information interact with each other at the initial stage of sentence processing.However, neither reaction time nor accuracy measure could accurately describe the real-time and continuous language processing system, because they both reflect the comprehensive results of language processing. Therefore, researchers have switched their focus into ERP (event related potentials) technique which enables them to examine the timing of neural cognitive activities of human brain under a natural condition.With the facilitation of ERP technique, studies on sentence processing have been advanced from behavioral level to neuropsychological level and yielded new theories and models. Two basic views concerning ERP-evidenced sentence processing mechanism have been brought up:Syntactic-centered and Semantic-centered theory. The former argues that syntactic analysis is highly automated and it starts from the very initial stage of online sentence processing, while semantic analysis is overruled by the syntactic analysis. The most representative neurocognitive model supporting Syntactic-centered theory is the three-phase sentence processing model proposed by Friederici; the latter believes that semantic memory-based analysis or possibly a semantically-driven combinatorial thematic analysis temporarily dominate online comprehension. The most representative neurocognitive model supporting Semantic-centered theory is the eADM (extended Argument Dependency Model) proposed by Bornkessel-Schlesewsky and Schelesewsky. So far, debate on whether sentence processing is syntactic-centered or semantic centered has not yet been settled.Currently, most studies concerning the mechanism on sentence processing are conducted in the field of L1 (first language) including English, Dutch, Italian and other Indo-European languages. Explorations on psychological mechanism of L2 (second language) sentence processing, especially at a contextual level which seems to better reflect the natural dynamic language comprehension process, are still very limited.Thus, the present study aims to explore the psychological mechanism of L2 sentence processing at a contextual level relative to sentential level, using ERP components to tentatively describe it Furthermore, the present study devises two groups of L2 learners with either high or low L2 proficiency, namely, Group H and Group L, aiming to find out their similarities or differences in hope of providing evidence to improve L2 teaching or learning. Major findings are summarized as follows:(1) At the sentential level, the sentence processing mechanism of L2 learners with higher language proficiency is different from that of L2 learners with lower language proficiency. L2 learners with higher proficiency develop a combinatory processing stream when processing difficulties in semantic occur, while L2 learners with lower proficiency terminate the processing system when facing the same kind of processing problems.(2) Within the context, the sentence processing mechanisms of L2 learners with lower language proficiency and higher language proficiency have both been changed. Context seems to lighten the processing costs of L2 learners with higher language proficiency while aggravate that of L2 learners with lower language proficiency. It is found that the processing mechanism of L2 learners with lower language proficiency tends to be similar to that of L2 learners with higher language proficiency. Yet, the question of how and in what way the context tweaks the processing system of L2 learners with lower language proficiency needs further exploration.(3) The present study seems to have found out that N400 effect, which is traditionally considered to be related to semantic integration, seems to reflect the processing costs caused by memory retrieval in L2 researching context.1P600 effect, which is traditionally reported to be related to syntactic integration, is possibly in response to the increasing processing costs caused by the integrating combinatory processing streams in L2 researching context2.(4) The sentence processing mechanism of L2 learners with higher language proficiency at both sentential and contextual level tends to be consistent with the explanation of eADM model which is brought up on the basis of L1 studies. While the sentence processing mechanism of L2 learners with lower language proficiency at both sentential and contextual level is only partly consistent with eADM model. The sentence processing mechanism of L2 learners tends to be more inclined to the Semantic-centered theory. |
PDF Full Text Request