Contextually dependent cue weighting for a laryngeal contrast in Shanghai Wu

Phonological categories are often differentiated by multiple phonetic cues. This paper reports a production and perception study of a laryngeal contrast in Shanghai Wu that is not only cued in multiple dimensions, but also cued differently on different manners (stops, fricatives, sonorants) and in different positions (non-sandhi, sandhi). Acoustic results showed that although this contrast has been described as phonatory in earlier literature, it is primarily a tone contrast. Phonation correlates only appear in fricatives, and tone sandhi neutralizes the f0 difference. Our perception results were largely consistent with the aggregate acoustic results, indicating that speakers adjust the perceptual weights of individual cues for a contrast according to contexts. These findings support the position that phonological contrasts are formed by the integration of multiple cues in a language-specific, contextspecific fashion and should be represented as such.