Production and Perception of Mandarin Laryngeal Contrast: The Role of Post-plosive F0

Abstract

This study examines the relation between plosive aspiration and post-plosive f0 (fundamental frequency) in the production and perception of the laryngeal contrast in Mandarin. Production data from 25 Mandarin speakers showed that, in word onsets, VOTs (voice onset time) of aspirated and unaspirated plosives were different, as expected. At the same time, the speakers produced different post-plosive f0 between aspirated and unaspirated plosives, but the difference varied according to the lexical tones – post-aspirated f0 was higher than post-unaspirated f0 in high-initial tones (i.e., lexical tones with high onset f0), but the pattern was the opposite and less robust in low-initial tones. In the perception of the same participants, VOT was the primary cue to aspiration but, when VOT was ambiguous, high post-plosive f0 yielded more aspirated responses in general. We claim that the asymmetry in f0 perturbation between high-initial and low-initial tones in production arises from different laryngeal maneuvers for different tonal targets. In low-initial tones, in which the vocal folds are slack and the glottal opening is wider, aspirated plosives have a lower subglottal air pressure than unaspirated plosives at the voicing onset, resulting in lower post-aspirated f0 than post-unaspirated f0. But in high-initial tones, the vocal folds are tense, which requires a higher trans-glottal pressure threshold to initiate phonation at the onset of voicing. As a result, the subglottal pressure does not decrease as much. Instead, the faster airflow in aspirated than unaspirated plosives gives rise to the pattern that post-aspirated f0 is higher than post-unaspirated f0. Regardless of this variation in production, our perception data suggest that Mandarin listeners generalize the f0 perturbation patterns from high-initial tones and associate high post-plosive f0 with aspirated plosives even in low-initial tone contexts. We cautiously claim that the observed perceptual pattern is consistent with the robustly represented production pattern, as high-initial tones are more prevalent and salient in the language and exhibit stronger f0 perturbation in the speakers' productions.