An Investigation Into The Lexical Inferences Process Of Chinese EFL Learners

Lexical inference is usually adopted as an efficient way to cope with unknownwords in the reading process by readers. The inference process of L2learners hasalways been a topic in the researches of L2reading and researchers have put forwarddifferent knowledge sources L2readers may employ to infer word meaning. Amongvarious factors related to the inference process, learners’ L2proficiency received greatattention in the previous researches. However, few researches presented acomprehensive measurement of learners’ L2proficiency and the role of word classesand knowledge sources in inferences did not receive enough attention.The present research investigates the relationship between Chinese ESL learners’L2proficiency and inferences process. The theoretical models adopted in this studyinclude Hasstrup’s (1991) Taxonomy of Knowledge Sources for Inferences and theCross-talk Continuum Model, McKoon&Ratcliff’s (1992) Minimalist Hypothesisand Graesser’s (1994) Constructionist Theory of Inference Generation.64universitystudents involved in the present study were divided into high-and low-proficiencygroups according to their L2vocabulary size and reading comprehension ability. Testsof lexical inferences together with questionnaires were employed to investigate L2learners’ choices of knowledge sources and accuracy in the inference process. Datacollected from the tests were analyzed by Statistical Package for Social Science(SPSS)17.0.The present research finds that there exists a positive correlation betweenlearner’s L2proficiency and the accuracy of inferences. L2learners withhigh-proficiency used knowledge sources more effectively than those withlow-proficiency. For the low-proficiency group, there are significant differencesamong the accuracies in their use of different types of knowledge sources, but for thehigh-proficiency group, no significant difference is found. The data also reveals thatthere exists no significant difference among the accuracy of three word classes both in high-and low-proficiency groups.