Conversation Analysis (CA)-Informed Algorithmic Approach to Augmenting Closed Captions with Stressed Syllables to Help Thai EFL Learners Improve Their English Pronunciation

Closed captions have been proven to be effective resources for English as a Foreign Language (EFL) teaching. Informed by Conversation Analysis (CA) research, this study seeks to enhance closed captioning features to assist EFL learners in recognizing stress and intonation patterns in spoken English. Using an algorithmic approach, CA-informed closed captions (CC) were generated by underlining stressed syllables and capitalizing pitch accents. Our goal is to enable EFL learners, particularly those from non-stress-timed languages such as Thai, to articulate English with greater clarity and speak with higher intelligibility. Evaluation results demonstrate that the proposed approach achieved an accuracy of 63 %, with a precision of 90 % and recall of 61 %. Further work is in progress to evaluate and improve the usability of the application by EFL teachers and learners.