Abstract
Japanese EFL learners’ difficulty with accurately decoding connected English speech motivated this mixed methods study. The aural decoding capacities of 63 first-year Japanese university students, with low to intermediate level English proficiency, were first measured with a battery of paused transcription tests (PTT). The transcriptions were clusters of three-words that each possessed attributes typical of co-articulated speech. In addition, after each test, a subgroup of 10 participants individually listened to the same PTT and recounted introspective self-observations of their perceived difficulties with the aural decoding tasks in their L1. These quantitative and qualitative data were used to identify four trends in decoding errors which were categorized as follows: limited collocation familiarity, syntactic knowledge constraints, difficulties utilizing co-text, and L1 phonological influence. This study investigates some of the difficulties associated with aural decoding, highlights the challenges of identifying the origins of decoding errors and suggests that more focus is needed on developing decoding skills as well as knowledge of formulaic language in L2 listening education.
A Dependency Syntactic Knowledge Augmented Interactive Architecture for End-to-End Aspect-based Sentiment Analysis