False hearing in L2-accented sentences

Unfamiliar accents complicate the task of speech perception, leading to reductions in intelligibility and increases in the effort listeners must put forth. One adaptive mechanism people may use to accommodate this challenge is to rely more on top-down processing during comprehension. To test this hypothesis, this study employed a false hearing paradigm (Sommers et al., 2015). Participants listened to sentences, half produced by an American English L1-accented speaker and half by an L2-accented speaker. Each sentence ended with a target word that was either highly predicted by the semantic context (e.g., “The captain steered the SHIP”) or was a phonological neighbor of the predicted word (e.g., “The captain steered the SHEEP”). We hypothesized that listeners would rely more on semantic context and less on bottom-up acoustic information for L2- versus L1-accented speech and thus show higher false hearing rates. Preliminary data show this to be the case: in incongruent sentences, young adults showed significantly higher false hearing rates in L2-accented sentences than in L1-accented sentences.