Date of Award
Bachelor of Science
speech perception, false hearing, psychology, listening, rhythm, semantic context, speech in noise
English and other languages such as German are stress-timed languages: the timing of the speech is determined by stressed and unstressed syllables, providing structure for sentences. While natural speech is not generally metrically regular, like in Shakespearean poetry, it still conveys timing cues through stress. Prior research has found that metric regularity enhances the processing of words (Rothermich et al, 2012), potentially because it attunes listeners’ attention to the predictability of stressed, and therefore important, syllables. Other work (e.g., Rogers, 2017) has suggested that predictability in the form of semantic associations (e.g., hearing “barn” facilitates understanding of “hay”) is a driving force for speech understanding, so much so that people falsely “hear” words predicted by semantic context (e.g., hearing “barn” leads to hearing “hay”, even if “pay” was presented).
In the current study, we aimed to examine how stress patterns and semantic associations may interact in listeners’ understanding of speech, as they both provide bases for predictions on the part of the listener. We measured speech understanding by masking the final word of a sentence in noise, then asking participants to identify what that word was (e.g., Jake visits the park to walk his DOG). We manipulated each sentence’s rhythmic predictability (whether the sentence was in natural speech, with a rhythm emphasized, or with a drum beat matching rhythm preceding the rhythmic speech) and semantic predictability (whether the last word made sense with the sentence, e.g. Jake visits the park to walk his dog/log). There was also a baseline condition for each of the rhythmic conditions wherein the sentence predictability was low. The results indicated that the beat prime improved processing of the rhythmic speech in conditions where expectancy effects played a role (semantically congruent and incongruent) but had a negligible impact in the baseline condition.
McLeod, Molly, "Timing is Stressful: Do Listeners Combine Meaning and Rhythm to Predict Speech?" (2022). Honors Theses. 2628.