Investigating vocabulary in academic spoken English of Business lectures in China

Abstract

In recent years, English-medium-instruction (EMI) has experienced substantial growth in educational institutions worldwide, notably in higher education (Macaro et al., 2018). With this fast-growing phenomenon, particular concerns have been raised regarding students' comprehension of EMI lectures, and vocabulary knowledge has been listed as one of the key impediments to comprehension (Ellili-Cherif & Alkhateeb, 2015; Goh, 2013; Y. Wang & Treffers-Daller, 2017; Q. Xie, 2020). This thesis, therefore, aims to investigate vocabulary in EMI lectures in the discipline of Business at an EMI university in China.

To fulfil this aim, the project developed an EMI spoken academic corpus in Business (EMIB) with 120 lectures collected from 54 lecturers with nine different first languages (L1), reaching 1.12 million tokens in total. Based on the corpus, three linked studies have been conducted. Study 1 investigated the lexical complexity of academic lectures. It compared the lexical complexity of EMI Business lectures in China with that in other academic spoken English in Anglophone and non-Anglophone settings, represented by the British Academic Spoken English Corpus (BASE) and the Corpus of English as a Lingua Franca in Academic Settings (ELFA), respectively. Lexical complexity was conceptualised by lexical sophistication (operationalised by vocabulary frequency profile and average frequency score) and lexical diversity (operationalised by the VOCD-D). Results show that there were no significant differences in lexical sophistication between the three corpora, but the ELFA had significantly lower lexical diversity than the other two corpora. Additionally, Study 1 also compared the lexical complexity of speech produced by teachers in the three corpora and explored whether speaker L1, speaker gender, and discipline contributed to the lexical complexity of lectures. The same pattern was found regarding the differences in lexical complexity of teachers' speech. Multiple regression results show that speaker L1 and discipline significantly impacted the lexical complexity of lectures. Study 2 drew on usage-based approaches (Ellis & Wulff, 2020) to investigate the relationship between word difficulty and various word properties, including word prevalence, salience, and contingency. Correlation analysis results show that significant correlations were found between word difficulty and word prevalence/contingency, while word difficulty does not correlate significantly with salience. Similarly, it also examined the relationship between word usefulness and word prevalence, contextual distinctiveness, and keyness. Word usefulness was found to have significant correlations with all the other three variables. To further investigate whether word difficulty and word usefulness can be predicted by the word properties, multiple regression models were built. Results show that word prevalence significantly predicted word difficulty, and prevalence and semantic diversity significantly predicted word usefulness. Using the regression results from Study 2, Study 3 developed an Academic Spoken Word List in Business based on word learning priority, encompassing word prevalence, difficulty, and usefulness. Theoretical, pedagogical, and methodological implications of these findings are discussed.