Syllable as the segmentation unit in perceiving spoken Chinese

Abstract

How spoken Chinese is segmented by native speakers has been addressed by the author in recent years. Evidence has accumulated to support the view that the syllable is the level at which speakers segment their spoken language. Earlier studies found that Mandarin–Chinese phonemes were detected faster in real syllables than in nonexistent syllables, that they were faster for level-tone and falling-tone syllables than for rising-tone or dipping-tone syllables, and that structural complexity had no effect on phoneme detection. These effects have preempted phoneme and word as the segmentation unit in real-time processing. This study was designed to investigate the role of stress foot in segmentation, to discover the locus of the tone effect, and to replicate the tone and complexity effects with Taiwanese speech materials. Stress assignment and phoneme detection tasks were employed as the experimental paradigms. The results indicated that stress assignment in Mandarin is not consistent enough to warrant a stress-foot segmentation procedure, that the tone effect originates from syllable duration variations, and that earlier effects were replicated with Taiwanese.