摘要
This study investigates the roles of Mandarin subsyllabic units in spoken word recognition by examining the neural processing of two phonotactic anomalies: (1) segmental gaps, which contain a non-existing combination of segments (e.g., *[ki1]); and (2) tonal gaps, which refer to a nonword comprised of possible segment combinations with an incongruous tone (e.g., *[tau2]; cf. [tau1] “knife”). Event-related potentials were recorded while participants performed an auditory lexical decision task. The response to segmental gaps differed from the other stimuli types in the amplitudes and scalp distributions of several components, including the P350, the N400, and the late positive complex. The P350 effect occurred around 370 ms before the entire syllable was revealed, indicating that lexical processing is not based solely on syllable representations. Furthermore, the overall differences between segmental and tonal gaps suggest that tones and vowels are dissociable. These results thus provide converging evidence for the view that Mandarin syllables are processed incrementally through phonemes.