Perceptual Style-Shifting Across Singing and Speech: Music Activates Pop Song English for NZ Listeners

American singing accents are prevalent in popular music throughout the English-speaking world. Singing with an American-influenced phonological style is a supralocal norm, referred to here as Pop Song English (PSE). This article presents two perception experiments that explore New Zealand (NZ) listeners' speech processing in musical and non-musical contexts. An analysis of the Phonetics of Popular Song corpus provides the foundation for the first experiment, revealing that sung dress and spoken trap have similar values for F1 in NZ. Experiment 1 then examines the categorization of these phonemes for words that fall on a continuum between <i>bed</i> and <i>bad</i>. In Experiment 2, a lexical decision task, NZ listeners hear words and nonwords produced by a New Zealand and an American speaker. In both experiments, results show that listeners are influenced by the presence of music, undergoing a <i>perceptual style-shift</i>. In Experiment 1, their perceptual phoneme boundary shifts to a more open position in the Music condition, and in Experiment 2, they exhibit a facilitation in reaction time to the US voice in the musical compared with the non-musical conditions. PSE is thus not only the norm for singing in NZ, it is also a norm for listening to song, represented in the minds of the general music-listening public. This finding extends our understanding of how speech perception depends on context. Speech and song are two highly distinct and perceptually contrastive contexts of language use, and listeners employ knowledge of how linguistic variation maps onto these contexts to resolve ambiguities in the speech signal.