This study examines how expectations for prosodic prominence, as related to information structure, influence speech perception in line with how segmental cues are modulated by prosodic prominence in so-called “prominence strengthening” in production. Following the general methodology of a previous study, which elicited listeners’ explicit judgements of prominence, listeners heard target sentences that were preceded by contexts that set up the critical utterance-final word to be relatively prominent or not, based on the information-structural relation it held with the preceding context. We first replicated that listeners shift explicit prominence ratings (of identical speech stimuli) on the basis of preceding context. We then show that these effects translate to the perception of vowel duration as a cue to voicing: listeners appear to effectively compensate for expected vowel lengthening under prominence, in line with previous studies that have sought to manipulate prosodic prominence within the speech signal itself. Finally, we find that explicit judgements for length are influenced by actual/veridical vowel duration, but not by preceding context. Results are discussed in terms of how they inform our understanding of prosody, and more specifically prominence, in speech perception.
Steffman et al. (Sun,) studied this question.