Innovative use of teachable machine infused with corpus data to enhance the acquisition of mandarin tone

With the burgeoning artificial intelligence (AI) technology in recent years, computer-assisted language learning (CALL) is gradually evolved into Intelligent CALL (ICALL), and such a shift effects a change to the quality of the interaction between learners and technologies (Kannan & Munday, 2018). However, it poses an enormous challenge to teachers to adopt ICALL approach effectively to Mandarin teaching. Therefore, powerful tools and best-practice guidance on how to take pedagogical advantage of AI are urgently needed (Lindner et al., 2019).

The purpose of this proposed study is threefold: to create a Mandarin tone assessment tool using the data from a learner corpus developed by Chen (2020) via a user-friendly web-based platform – Teachable Machine (Carney et al, 2020); to introduce an innovative approach which first provides learners with explicit instruction on the pronunciation features, particularly the pitch contour of each tone with the aid of a website constructed by Chen (2020) and then instructs learners to use the AI-powered tool and an online real-time pitch analyzer – AmPitch (Huckvale, 2016); and to collect views on the teaching approach from learners. It is worth mentioning that to create the customized assessment tool, the correct pronunciation of reading of monosyllabic words by 40 Hong Kong learners was extracted and uploaded to corresponding classifications (i.e., 4 tones separately). In other words, the norms of the tone assessment are based on learners whose pronunciation is acceptably correct. Two implications can arise from this study. First, teachers are encouraged to select data themselves and adapt the tools for their students. Second, different technologies which are compatible could be applied together. Copyright © 2022 The British Chinese Language Teaching Society.