Production and perception of a merger: The case of [l] and [n] in Eastern Min Chinese

Abstract

We identified six acoustic correlates to the phonemic contrast between word-initial [l] and [n] in both English and Mandarin (Cheng and Jongman, 2019, JASA). All six cues suggested the phonemic contrast between prescriptive L and N (phonemes in the unmerged system) disappeared in Eastern Min (EM) in production. The current study investigated the phonological and social factors that modulate the merger-in-progress. In production, a classifier was trained with acoustic parameters from Mandarin (F3, F1 bandwidth, A1-P0) to categorize EM tokens. Linear discriminant results illustrated acoustically distinct [l] and [n] tokens as the phonetic realizations of the merger. Both prescriptive L and N were variably realized as acoustic [l] or [n] in production. Low vowels correlated with acoustically more nasal-like onsets. Speakers of all ages and genders leaned toward [l]. Perceptually, AX discrimination results (ISI = 750 ms) suggested EM speakers did not distinguish prescriptive L from N (mean A′ = 0.46), providing converging evidence of the merger. EM listeners did not perceive the difference between acoustic [l] and [n] either (mean A′ = 0.47). However, lower vowels with nasal codas illustrated a higher degree of sensitivity than other phonological conditions, which was an effect of the larger acoustic distance between stimuli.