Abstract
<p>Why does a listener's linguistic background shape whether repeated speech comes to sound like music? Previous studies have reported that speakers of tonal languages experience a weaker Speech-to-Song (STS) transformation, in which repeated speech comes to sound increasingly musical. It remains unclear, however, whether this reduction reflects a general property of tonal-language listeners or arises specifically when they hear their native language. We compared native Mandarin Chinese, English, and Dutch listeners' responses to English speech, Mandarin Chinese speech, and non-speech stimuli (data collected 2024– 2026) before and after repeated presentation. All groups reliably experienced the STS transformation. Critically, however, the transformation was markedly reduced only when native Mandarin Chinese listeners heard Mandarin Chinese speech. The same listeners showed a typical transformation for English speech, and English and Dutch listeners showed comparable transformations for both languages. Furthermore, this reduction disappeared when linguistic information was removed while preserving the original pitch contours. These findings indicate that reduced transformation is not a general characteristic of tonal-language listeners but depends on whether pitch remains available for lexical interpretation. The results suggest that the linguistic function of pitch constrains its reinterpretation as melody, a pattern consistent with predictive-coding accounts of speech and music perception.</p>