A corpus-based study of discourse markers in secondary student talk

This study examines and compares the usage of discourse markers (DMs) in student talk in three secondary schools in Hong Kong, representing different English proficiency levels. The data were collected from a self-made learner corpus that comprised around 70,000 tokens from 44 academic discussions by 172 adolescent English learners. Drawing on the corpus data, this study investigates 35 groups of interpersonal, referential, textual/structural and cognitive DMs based on two widely adopted conceptual frameworks: Castro (2009) and Fung and Carter (2007). The results suggest that DMs have limited diversity when used by second language (L2) speakers in Hong Kong. However, close data analysis indicates that the students with higher English proficiency used more DMs than those with lower ability. Furthermore, the more proficient L2 learners were able to use more sophisticated DMs at the interpersonal and cognitive levels, whereas those with lower English proficiency were often restricted to using DMs at the referential and textual/structural levels. These observations can be explained in relation to speakers’ assimilation into their English-speaking environment. Finally, this study highlights the necessity of increasing L2 speakers’ awareness of using DMs in conversations to prepare them to become more proficient speakers.