The role of visual speech cues in sound change: A study of the cot-caught contrast among Michigan speakers

Abstract

Interspeaker articulatory variation has been argued to be a driving force in sound change (Baker et al. 2011), yet the factors governing such variation are not entirely understood. Fronted /O/ in the Northern Cities Shift is characterized by an increase in F2, and can be produced through either tonguefronting or lip-unrounding (Havenhill 2015), while fronted /u/ in British English is achieved through tongue-fronting alone (Harrington et al. 2011). We investigate the hypothesis that visual speech cues restrict the ways in which articulatory patterns may vary between speakers. Participants were exposed to congruous and incongruous audiovisual nonce word stimuli containing /A/ and /O/, produced by a Michigan speaker. The perceived vowel was identified by selecting a rhyming English word (e.g., the choices for [zOt] were “cot” and “caught'). When paired with visual lip-rounding cues, participants perceived auditory /A/ as /O/. Yet the same auditory stimulus presented with unround lips was perceived as /A/, suggesting that the visual cue may be sufficient to maintain a contrast, despite acoustic similarity. However, conditions are favorable for merger when speakers produce both / A/ and /O/ with unround lips. As a result, visibly distinct articulatory variants may be preferred when pressure to maintain a contrast is high.