Lexical Stress in Mandarin Second-Language Speakers of English: An Electromagnetic Articulography Study

The present study focused on the acoustic and articulatory realization of English lexical stress in Mandarin second-language (L2) speakers of English. We aimed to understand (a) how suprasegmental and segmental features were used in the acoustic domain and (b) how lingual and nonlingual articulators were manipulated in the articulatory domain during the production of lexical stress.

Production of stress minimal pairs (e.g., OBject-obJECT) was analyzed. The data were drawn from a publicly available data set consisting of time-synchronous acoustic-articulatory data from 20 first-language (L1) speakers of English and 20 Mandarin L2 speakers. Acoustic features included duration, intensity, fundamental frequency (F0), and vowel quality. Articulatory properties involved positional information of the tongue, lips, and jaw.

All suprasegmental cues investigated (F0, intensity, duration) were found to be involved in the production of lexical stress by the two speaker groups, although in L1-specific ways in the case of F0. In contrast, the segmental cue (vowel quality) was used to distinguish lexically stressed and unstressed syllables by the L1 speakers only. Both groups demonstrated increased displacements in nonlingual articulators (jaw and lip) in lexically stressed vowels, and a significant positional difference in the lingual articulator (tongue dorsum) was found for some (but not all) of the L1 speakers' productions.

Mandarin L2 speakers were found to use some of the same acoustic and articulatory cues as English L1 speakers to realize lexical stress in English. In the L2 group, however, it was the suprasegmental cues rather than segmental cues that most consistently distinguished lexical stress contrasts, and nonlingual articulators were weighted more heavily than the lingual articulator.