Listeners’ adaptation to unreliable intonation is speaker-sensitive

Abstract

Variable linguistic environments require the ability to quickly update expectations and behavior including speech comprehension. This adaptive capacity is key to understanding how listeners successfully recognize speaker intentions in light of the ubiquitous variability in speech. The present study investigates how listeners' real-time sentence comprehension adapts to speaker-specific prosodic variability. In two forced choice mouse tracking experiments, listeners had to identify a visual referent guided by pre-recorded instructions. When exposed to a speaker that uses unconventional pitch accent placement, listeners discard intonational information for that speaker, but keep using intonation to resolve the referential ambiguity for another speaker that places pitch accents conventionally. These results show for the first time that intonationally guided sentence comprehension adapts in a speaker-sensitive way. The study further provides valuable first insights into the temporal unfolding of this adaptation process. Listeners first attribute unconventional patterns to the context, thus discarding the informational value of intonation for both speakers. After sufficient evidence, however, listeners start attributing unexpected patterns to only the unconventional speaker. Materials, data, and scripts can be retrieved here: http://osf.io/fdpg4

How prior experience with pitch accents shapes the perception of word and sentence stress

“…Importantly, we were able to show that the effect of exposure (high vs. low-pitched stressed syllables) on perceived word and sentence stress generalises to different speakers. Hence, differently from what has recently been shown on the adaptive processing of intonational variation for sentence comprehension (Roettger & Rimland, 2020), the adjustment of cue weights for the processing of word and sentence stress seems to be speaker-independent. Based on what is known on the adaptation to foreign accents (Bradlow & Bent, 2008), the usage of multiple different speakers for our different speaker exposure condition, instead of one single different speaker, might have been particularly beneficial for listeners.…”

“…Bradlow & Bent, 2008;Xie et al, 2018a;Xie et al, 2018b). While the influence of exposure to different pitch accent types on sentence comprehension is speaker-sensitive and hence restricted to the speaker a listener is currently exposed to (Roettger & Rimland, 2020), the exposure effect we find for correctness rates in stress judgments generalises. It has thus more far-reaching implications, not only for cross-linguistic speech processing, but also for language change and developmental aspects.…”

“…We predict that listeners are generally better at indicating the stressed syllable when it is high-pitched (medial-peak contour) than when it is low-pitched (early-and late-peak contours) and that this effect of intonation condition is stronger after immediate exposure to high-pitched than to lowpitched stressed syllables. A speaker-specific adaptation account would further predict that the effect of experience is limited to (or stronger for) the same-speaker compared to the exposure to various speakers, similar to exposure effects for segmental perceptual learning (Eisner & McQueen, 2006;McQueen et al, 2006) and the intonational marking of information structure (Roettger & Rimland, 2020).…”

Short-term perceptual re-weighting in suprasegmental categorization

“…In the case of word emphasis in English, for which F0 is the primary dimension, we expect listeners to down-weight duration (a secondary dimension; Jasmin et al, 2020) upon introduction of an artificial accent that reverses the canonical F0xDuration covariation in English. Prior research indicates that overall weighting of prosodic information as a cue to speaker intent can change in the short term due to context; for example, exposure to a speaker with idiosyncratic/unreliable use of pitch accents leads listeners to down-weight intonation as a cue to speaker intent (Roettger and Franke 2019, Roettger and Rimland 2020). Moreover, when ambiguous pitch contours are consistently paired with a particular interpretation, categorization functions shift according to whether a statement had a positive or a negative interpretation (Kurumada et al 2018).…”