Sensory acuity and cross-language phonetic similarity jointly predict second language vowel production accuracy

Research on second language (L2) speech perception and production has yielded mixed findings, suggesting that factors beyond perceptual ability influence L2 production outcomes. This study investigated predictors of inter-speaker variability in L2 vowel production, focusing on individual differences in auditory and somatosensory acuity. The roles of phonological awareness and trial-to-trial production variability were also examined.

Forty English-speaking adult late learners of Mandarin produced two Mandarin vowels: /u/ (phonetically similar to English /u/) and /y/ (a novel vowel for English speakers). Production accuracy and trial-to-trial variability were measured acoustically. Auditory acuity was assessed using speech identification and discrimination tasks. Somatosensory acuity was measured through a Phonetic Awareness Task (PAT) and a novel Tongue Placement Task (TPT). Linear mixed-effects models were used to identify predictors of production accuracy.

Predictors of production accuracy differed by vowel. For the perceptually similar vowel /u/, lower production variability was the only significant predictor of higher accuracy. For the novel vowel /y/, higher somatosensory acuity (PAT) was the only significant predictor of accuracy. No predictors significantly accounted for production variability in either vowel.

These findings suggest that sensory acuity and L1-L2 phonetic similarity jointly constrain L2 speech learning. Specifically, somatosensory acuity supports the establishment of accurate articulatory targets for novel L2 sounds, while production variability reflects the stability of learners' phonetic category formation for perceptually similar L2 sounds.