Shaping phonetic performance in second language learners
Leung, Hiu-Nam Jaime
This study aimed to evaluate the efficacy of a software-administered shaping procedure in guiding English monolinguals to acquire accurate Mandarin pronunciation. A single-subject reversal ABAB design was used to evaluate treatment effects. A purposely-developed algorithm generated an accuracy score defined as the similarity between a participant’s utterance and the target pronunciation. The shaping procedure provided performance-dependent reinforcement, while the control condition provided performance-independent reinforcement at a density yoked to the shaping procedure. A no-feedback condition assessed spontaneous language learning ability prior to treatment. Data were evaluated via visual analysis and complemented with effect size analyses and repeated-measures ANOVAs. There were no overall treatment effects. However, three individuals demonstrated a statistically significant difference between treatment and control. A follow-up study compared shaping to no feedback using a simplified procedure and simpler stimuli. A multiple-baseline design was used. The results showed no treatment effects. Possible contributing factors and directions for future research are discussed.
second language learning, computer-assisted language learning, shaping