To investigate the perception-production link, researchers often train second language (L2) learners’ perception and track gains in production, rather than the reverse. In this study, we examined whether production-based training yield perception gains and whether implicit learning occurs without targeted instruction. Adult L2 learners participated in three sessions of 20-min, production-focused pronunciation training. One group received explicit instruction on prosodic cues (intonation and pausing for high versus low attachment in English relative clauses); the other group focused on segmental cues (vowel duration for English coda voicing contrasts). We assessed participants’ performance in both explicitly instructed and uninstructed domains using a pre-/post-test design involving perception (two-alternative forced-choice identification) and production (reading) tasks. Results revealed production accuracy improved in the explicitly trained domain for each group. Furthermore, we observed implicit learning effects in the production of segmental contrasts. Surprisingly, the strongest perceptual gains for segmental contrasts appeared in the prosody-instructed group, suggesting that segmental perception can improve “for free” when production training targets prosody. In contrast, prosodic perception showed only modest improvement across groups. Our findings suggests that production and perception are linked, but not symmetrically; perceptual benefits of production-based training depend on both learning mode and contrast type.
Kim et al. (Wed,) studied this question.