Speech transformed into song heightens temporal sensitivity but weakens absolute pitch sensitivity

Abstract

The speech-to-song illusion (STS) is a phenomenon in which some spoken utterances perceptually transform to song via repetition. The existence of utterances that do and do not transform makes it possible to investigate similarities and differences between the musical and linguistic modes of listening, with the former elicited by transforming utterances and the latter by non-transforming utterances. In Experiment 1, inter-stimulus interval durations within STS trials were either steady, slightly variable, or highly variable. Participants reported how temporally regular the utterance entrances were. In Experiment 2, participants were first exposed to STS trials and then asked to choose the transposition of the utterances they heard during the exposure phase. Results indicate that listeners exhibit heightened awareness of temporal manipulations to transforming utterances compared to non-transforming utterances, but not to absolute pitch manipulations. This suggests that compared to the linguistic mode of perception, the musical mode entails an increased sensitivity to temporal regularity, but not to absolute pitch. The methodology used here establishes a framework for implicitly differentiating musical from linguistic perception, as well as for behaviorally investigating the different cognitive apparatus that people use when activating the musical or linguistic mode of perception.