Phonetic cues influence judgment of syntax

Abstract

This speech perception experiment provides evidence of interaction between phonetic and syntactic cues associated with socio-/ethnolects. These phonetic and syntactic (socio)linguistic cues were crossed in stimuli recorded by speakers of African American English (AAE), Southern Standard British English (BrE), and General American English (GAE). Thus, some of the sentences listeners heard contain 'matching' cues—for example, a have raising question with BrE phonetic features—while the others contain “mismatching” cues much less likely to have been experienced together by listeners. Listeners (n = 30) rated the spoken utterances' acceptability on a scale of 1 to 5; response time was recorded. The mixed-effects linear regression fitted to the Z-transformed data shows BrE and AAE constructions are more acceptable when the phonetic cues match. The effect is most robust for the BrE accent; compared to estimates for the GAE accent ( β = 0:9538, p < 0.001), those for the BrE accent are overall lower (β = −0.106, p = 0.048). The penalty is effectively counteracted, however, when the BrE voice is producing the associated syntactic construction (β = 0:181, p = 0:017). Response time data also suggests stimuli with matching cues are processed faster (t = -2:3586, p = 0:01856).To my knowledge, this is the first experimental evidence of an effect of accent on syntactic acceptability.