Acoustic (and non-acoustic) enhancements of naturally produced clear speech

Talkers enhance their speech according to listeners’ needs, often involving multiple simultaneous articulatory-acoustic enhancements. Previous findings are mixed on whether clear speech elicited through explicit instruction or more natural interactions yields greater enhancements. In the current study, we investigated how talkers modify their speech when introduced to real listeners in a relatively controlled task. Participants were introduced to native and non-native listeners via video and asked to help each intended listener find objects in various places. A carrier phrase was used to ensure that target words appeared in phrase-final position. Analyses show that participants often used multiple words instead of one word to describe target items to both listeners. Comparison of listener conditions across word counts (single-word versus multiple-word) revealed main effects of both factors on overall utterance duration, as well as an interaction between them. Specifically, overall utterances were longer in multi-word phrases and in the non-native listener condition, with a greater difference between the word counts observed in the non-native listener condition. These findings suggest that talkers enhance their speech according to listeners’ needs, when introduced to them via video, by adapting at the global level, particularly in more complex utterances.