Fostering learner autonomy in English pronunciation: A convergent, interdisciplinary phonetics approach*

This study adopts a technology-supported, phonetics-based instructional approach to enhance learners’ English pronunciation. The course was designed as an integrated curriculum that combines phonetics, phonology, intonation (ToBI), pronunciation training, and pronunciation assessment within a unified framework. Learners engaged in systematic pronunciation practice, conducted acoustic analyses using Praat, recorded speech samples, and participated in two rounds of multiple-forced-choice pronunciation assessment experiments. In addition, they implemented AI-based automatic speech recognition (ASR) using Python code and evaluated their pronunciation through the Word Error Rate (WER). This course was developed from a convergent, integrative, and experimental perspective and intended to promote learner autonomy. Survey responses indicated that approximately 70% of learners perceived overall improvement in their pronunciation skills. A mixed-effects model analysis revealed that short-term exposure to automated ASR-WER feedback was insufficient to systematically alter participants’ rating behaviors and had only limited influence on their sensitivity to specific evaluation dimensions.