Development of a corpus-based online pronunciation learning system for Chinese learners of English

In this talk, we will introduce the major products of our TDG project and discuss the possibilities of engaging students in corpus-based ‘data-driven learning’. A learner speech corpus was established in order to help learners, teachers and researchers inside and outside Hong Kong to (a) identify Hong Kong and Mainland learners’ recurrent segmental and suprasegmental difficulties in learning English pronunciation and (b) suggest possible remedies that may reduce or eliminate such difficulties and enhance intra- and inter-cultural communication. Apart from a learner speech corpus, an online pronunciation learning platform was developed for learners to (a) discover the possible causes of common pronunciation problems spotted in the corpus, (b) practice segmental and suprasegmental aspects of pronunciation through a variety of perception and production exercises that were designed based on the pronunciation problems collected from the corpus, and (c) learn some basic sound analysis techniques via a Praat beginners' manual compiled by the research team.