A Study Of The Effects Of Working Memory And Situation-based Context On On-line Predictive Inference Generation During L2 Narrative Text Comprehension

This thesis is designed to explore the effects of the working memory and situation-based context on predictive inference generation during L2 narrative text comprehension. One on-line experiment is carried out by using DMDX 3.0 to measure the reading time on the target sentence, with Chinese EFL learners as participants. The independent design is a 2 (working memory: low and high)×2 (situation-based context: low and high) factorial.The results show that (1) working memory capacity contributes considerably to Chinese EFL learners'predictive inferential processing during L2 narrative text comprehension. Learners with the high working memory span respond significantly faster to target sentences in both high- and low-context conditions, whereas their low-span counterparts do not. (2) situation-based context influences Chinese EFL learners'on-line predictive inference generation. Under the high-context condition, both high- and low-span working memory learners can make predictive inferences faster and more accurately than under the low-context condition. (3) there is a significant two-way interaction between working memory and situation-based context on the generation of predictive inferences. Situation-based context shows greater facilitation for low-span learners in generating predictive inferences than for high-span counterparts.