Abstract/Summary

This thesis reports on a mixed methods study which investigated the ability of native-Chinese speaking readers of academic English to identify incongruent words in coreferential noun phrases, as well as the extent to which such an ability may vary as a function of reading proficiency. Additionally, the study also explored the extent to which the presence or absence of a grammatical determiner (e.g. this/these) may contribute towards such an ability. The study was composed of two major elements: an eye-tracking experiment and follow up think-aloud interviews. Reading proficiency was measured according to the participants’ IELTS scores. In addition, individual differences, in terms of vocabulary knowledge, grammar knowledge and working memory capacity were explored in relation to the participants’ ability to detect the incongruencies. The results of the study suggest that the ability to detect an incongruent word in a text depends primarily on the reader being able to establish a situation model mental representation of that text. Furthermore, the results also suggest that this ability is chiefly, although not entirely, confined to upper-intermediate or advanced-level readers. While the measures of individual differences used in this study did not significantly predict for the ability to identify the incongruent words in the text, analysis of the transcripts from the think-aloud interviews found that the main impediment to the formation of situation model representations of academic text was difficulty with linguistic processing. The implications of these findings are discussed in relation to both the practice and theory of second language reading pedagogy. In terms of practice, the implications include (i) the need to develop metacognitive strategy training designed to help L2 readers establish situation model representations of the texts they read; (ii) the need for specific instruction on the grammatical construction and cohesive function of complex coreferential noun phrases in academic writing; (iii) the potential for using anomaly detection as a tool for measuring L2 reading comprehension; (iv) the need for university managers to provide teaching staff with training on how to help L2 readers establish situation model representations of the texts they read. With regards to theory, the results ix of the study lend weight to Walter’s (2007) suggestion that L2 reading should be understood more in terms of accessing fundamental cognitive processes through the L2, than of a transference of pre-existing reading skills from the L1 to the L2