Learning the pronunciation of English words from textual input: Should we listen first?

This study investigated factors influencing incidental English word pronunciation acquisition by upper-intermediate L2 learners through exposure to spoken discourse. Due to inconsistent English spelling-sound correspondences, silent reading is likely to leave learners with inaccurate pronunciations. This study explored whether these inaccuracies could be easily corrected through listening. Two sequences were compared: silent reading followed by listening and listening followed by silent reading.\nIn a counterbalanced within-participant design, 50 upper-intermediate ESL learners at a research-intensive University in Ontario engaged with a text containing 16 target words. The text was divided into to parts. Participants either read a part silently, then aloud, followed by listening, or they listened first, then read silently and aloud. The sequence was reversed for the other part of the text. Post-tests assessed pronunciation improvements and interviews explored individual differences.\nThe results indicated that a single audio exposure was insufficient for accurate pronunciation acquisition. Both the trial-and-error and retrieval approaches yielded comparable final outcomes. However, the Input-Output-Input sequence (listening, reading, and listening again) showed potential as a more effective teaching strategy, combining the benefits of both approaches to enhance learning outcomes.