Technology-assisted consciousness-raising for L2 prosody: Effects on oral performance and the underlying cognitive-metacognitive mechanisms of thought-group chunking

This study, using an embedded mixed-methods design, investigated how technology-assisted consciousness-raising training (CRT) optimizes prosodic processing in L2 speech. Specifically, we examined whether CRT targeting thought-group prosody enhances intelligibility, comprehensibility, fluency, and accentedness in pre-service English teachers' oral discourse. Drawing on Baddeley's (2012) working memory model, Levelt's (1999) perceptual loop theory, and Zimmerman's (2000) self-regulated learning (SRL) framework, the study explored the underlying cognitive and metacognitive mechanisms of prosodic learning. Sixty-four participants (30 experimental, 34 control) completed 11 weeks of training and pre-post tasks. The experimental group received visual feedback via Audacity, which externalized pitch contours and pause patterns after each recording, allowing learners to reflect and adjust their performance before reattempting. This visual scaffolding, while not acoustically analyzed in the study, may have alleviated working memory demands through dual coding, providing a plausible explanation for the observed gains. Pre-post evaluations based on expert rater judgments revealed significant improvements for the experimental group across all four dimensions, with the largest gains in fluency and comprehensibility. Interaction effects showed distinct performance trajectories compared to the control group, which demonstrated smaller, practice-based gains. Semi-structured interviews corroborated these findings, revealing increased confidence, reduced anxiety, and heightened metacognitive awareness. Notably, participants consistently underestimated their accentedness relative to external ratings, reflecting common SRL challenges in self-assessment. These findings suggest that technology-assisted CRT facilitates systematic adjustments in prosodic organization and may reduce-or be perceived by learners as reducing-self-monitoring burdens. Furthermore, CRT bridges top-down and bottom-up approaches by combining visual feedback (bottom-up) with goal-oriented prosodic automatization (top-down). The study demonstrates CRT's potential to enhance both linguistic performance and learner autonomy. Future research could incorporate direct acoustic measures, active placebo controls, and spontaneous speech tasks to further examine the underlying mechanisms and improve ecological validity.