The co-variance of acoustic cues and lip gestures in the production of the Mandarin sibilant contrast

Phonological contrasts in speech production and perception are frequently studied with acoustic and/or auditory cues. Less is known about how visual properties are incorporated in a contrast and how/if they can take part in a merger-in-progress, especially for changes that can be initiated by different articulatory gestures. This paper investigates the lip gestures and acoustic properties associated with the production of Mandarin sibilant contrast, which is undergoing neutralization in some varieties (e.g., Taiwan), with retroflexes merging with alveolars. We analyzed audiovisual production data using Computer Vision articulography and found that speakers range from having an almost complete lip gesture merger to a complete distinction. Roundness scores derived from lip measures were found to strongly predict and correlate with acoustic properties. The additional rounding gesture associated with retroflexes peaked at or before the sibilant onset and continued into the vowel. We extend previous acoustic findings of the sibilant contrast/merger to a shared audiovisual domain in production, providing the necessary background for future work investigating audiovisual cue trading in perception and in the context of an ongoing merger. We further discuss implications for the directionality of sound change.