Language-specific effects in speaker and rate normalization

Abstract

This study concerns the effects of context on the perception of Mandarin tones. Mandarin tones are primarily distinguished in terms of F0 range and temporal location of inflection point. The most relevant context effects therefore involve changes in speaker and speaking rate. In this study, we compared speaker and rate normalization of Mandarin tones 2 (mid–rising) and 3 (low–falling–rising) by Mandarin and English listeners. Synthetic tone 2–tone 3 continua varying in turning point (TP, the inflection point of the tone), ΔF0 (the difference in F0 between onset and turning point), or both were appended to natural precursor phrases representing high- and low-pitched speakers, as well as fast and slow speaking rates. Results of the rate experiments with Mandarin listeners confirmed that normalization occurs only when precursors and stimuli vary along the same acoustic dimension, a finding previously observed for speaker normalization [C. B. Moore and A. Jongman, J. Acoust. Soc. Am. 102, 1864–1877 (1997)]. English speakers showed speaker and rate normalization effects as well, but only for the continuum in which TP and ΔF0 varied simultaneously. We propose that limitations on perceptual resources allow English listeners to attend to extrinsic information only when intrinsic acoustic differences become more perceptually salient.