Extracting multiword expressions from texts with the aid of online resources

Abstract

Research in recent years has emphasized the ubiquity and pedagogical significance of multiword expressions (MWEs) (e.g. Siyanova-Chanturia & Pellicer-Sánchez, 2018). However, accumulating a sizeable repertoire of MWEs remains a challenge to language learners. Aiming to shed light on one potential pedagogical path, this classroom-based research project investigates the effectiveness of ‘text-chunking’, an activity where learners are asked to extract MWEs from authentic texts. The participants were two intact classes of Vietnamese EFL learners in their second tertiary year. Over ten weeks, the experimental group (n=26) engaged in weekly text-chunking practice, while a comparison group (n=30) used the same reading materials for content-related activities. The experimental group was taught to consult online dictionaries and an online corpus to verify their identification of MWEs. Their discussions were audio-recorded every week. A pre-test and two post-tests (immediate and delayed) on the form recall of 52 target items revealed significantly greater learning gains in the experimental group, with a medium effect size. Post-treatment questionnaires also suggested these students’ heightened awareness of MWEs and autonomous learning strategies. This presentation will highlight these main findings and, more importantly, discuss the implications for language teachers on how to better scaffold learners in their MWE acquisition. Pedagogical suggestions on how extracting MWEs from texts can be done using online resources to adapt the changes caused by Covid-19 conclude the talk.