This study investigates the merger of the Mandarin alveolar-retroflex contrast in two vowel contexts (a vs u) among bilingual speakers of Quanzhou Southern Min first language (L1) and Mandarin second language (L2), with a focus on the interaction between linguistic and social factors. Sixty-one bilingual speakers' productions were evaluated by L1 Mandarin listeners in a perceptual identification task, and k-means clustering analysis based on each fricative's perceptual accuracy was used to classify speakers as "distinctive" or "merged." Acoustic analysis of the spectral center of gravity (CoG) confirmed that L1 Standard Mandarin speakers maintained a larger CoG separation between /s/ and /ʂ/ than bilinguals, especially in the u context, reflecting individual variability within the bilingual group. For bilinguals, acoustic analysis further corroborated the perceptual classification: "Distinctive" bilinguals maintained clear CoG separation between /s/ and /ʂ/ categories, whereas merged bilinguals showed little to no difference in either vowel context. Furthermore, gender and age significantly influenced this contrast production. Notably, these effects were largely mediated by L2 exposure levels. Women and younger speakers tended to have greater exposure to Mandarin, which correlated with more distinctive productions. This result highlights the dynamic relationship between linguistic and social factors in shaping phonological contrast variability in bilingual speakers.
Weng et al. (Sun,) studied this question.