Stephen Bodnar & Roy Lyster, McGill University, Canada

Title: Choose Your Own Content: a technology-based approach for automatically creating tailored grammar practice exercises from the web

Keywords: computer-assisted language learning; grammar practice; technology; content-based language learning; L2 French

Abstract:

Much is now known about how second and foreign language (L2) instruction can be most effective, but practical constraints often limit the widespread adoption of pedagogical best practices: teachers are busy professionals with limited time for preparation, and learners' time in the classroom is limited. Research at the intersection of Computer-Assisted Language Learning and Artificial Intelligence investigates how human language technologies can make it easier to provide learners with more access to high-quality L2 instruction. This workshop introduces an innovative new Content-based Language Learning Instant Exercise (COLLIE) tool, which automatically creates pedagogically state-of-the-art grammar practice exercises for French L2 learners. Drawing on open source technologies and linguistic resources, and adopting French grammatical gender as a pedagogical target, we show how a scaffolded sequence of practice exercises can be quickly generated automatically, from content that an instructor or learner chooses. This workshop begins by presenting background on Lyster's (2018) notice-awareness-output practice sequence, and patterns in French that predict grammatical gender and which have been shown to be useful to learners (Lyster & Izquierdo, 2009). Next, we provide a demonstration of COLLIE, showing concretely how it can be used to generate reading, sorting, judging, writing, speaking and riddle activities that stimulate learners to acquire knowledge of French gender-predicting patterns, using words from the documents they choose, and to generalize these patterns to new words they have not yet seen. The third part of the workshop will follow a structured focus group format in which we invite the audience to provide feedback on different aspects of the tool. In particular, we would like to hear how educators envision a tool such as COLLIE fitting into their language teaching practices, with the aim of increasing the tool's usefulness outside of the lab to authentic learning settings.