Visual cues in sound change: A cross-modal perceptual account for the typological rarity of labial palatalization

Abstract

Auditory perceptibility is widely recognized as a major contributing factor to listener-based sound change. Although speech perception is also known to be influenced by vision, the role of visual cues in facilitating or inhibiting sound change is not well understood. Palatalization of /tʲ/ and /kʲ/ to /tʃ/ is a typologically common change that is frequently attributed to misperception, while full palatalization of labial consonants is exceedingly rare. We test whether visual cues contribute to this asymmetry by way of the visual distinctiveness of labial closure. An audiovisual perception experiment was conducted in which 40 native Korean speakers were presented with synthetically altered tokens along 5-step continua from /pʰ/, /tʰ/, or /kʰ/ to /t?ɕ/ in audio-only, audiovisual-congruent, and audiovisual-incongruent conditions. Acoustic cues to place of articulation, F2 onset, and spectral frequency during aspiration, were manipulated. Results show that presence of the visual lip closure cue significantly inhibited perception of labials as palatals: the /pʰ/-/tɕʰ/ boundary was shifted toward /tɕʰ/ in the AV-congruent group, and labials were more likely to be perceived as palatals in the absence of lip closure. This finding suggests that visual speech cues may contribute to the typological asymmetry and that sound change can be modulated by visual cues.