Conference Paper Published

Abstract

Lexical tone normalization enables Mandarin speakers to maintain stable tone categories despite substantial pitch variability across speakers and communicative contexts. While previous research has established that speech contexts reliably facilitate tone normalization, non-speech and purely musical contexts do not, supporting the view that this mechanism is speech-specific. However, genres such as rap, which blend speech and musical elements, challenge this dichotomy. This study systematically examined whether rap music and related vocal contexts can induce lexical tone normalization in Mandarin listeners. Native Mandarin speakers categorized target syllables following six types of auditory contexts: natural speech, clear speech (elocution), typical rap, melodic rap, song, and cello. All vocal materials were produced by professional Mandarin-speaking rappers and pitch-matched to the targets. Results revealed that natural speech, elocution, and typical rap contexts robustly elicited tone normalization, as indicated by significant shifts in categorical boundaries and improved identification accuracy. In contrast, melodic rap and song produced only marginal effects, while the cello context had minimal impact. These findings indicate that speech-like cues—particularly prosody and articulation—are critical for tone normalization, whereas increasing musicality, especially melodic structure, can inhibit this process even when some speech-like features remain. Our results refine current models of speech perception by demonstrating that lexical tone normalization depends on the presence and prominence of linguistic cues, and that melody can impose clear boundary conditions on this perceptual adjustment.