As reported in Yip (2004), the Possible-Word Constraint (PWC), a mechanism proposed by Norris, McQueen, Cutler, and Butterfield (1997), could effectively operate in Cantonese speech segmentation. In this study, a word-spotting experiment was conducted to further examine how the mechanism works in segmenting Cantonese speech. In the word-spotting experiment, listeners were asked to spot out the target Cantonese word from a series of nonsense sound strings. In general, results suggested that listeners found it more difficult to spot out the target Cantonese word [p:k³si:⁶] from the nonsense sound strings, which attached or embedded with a single consonant context [p:k³si:⁶], than from the nonsense sound strings, which attached or embedded with either a vowel context [a:¹p:k³si:⁶] or a pseudo-syllable context [khi:m1p:k³si:⁶]. Different locations of the context (initial; intermediate; final) produced different degrees of interference effects. Together with my previous findings, the present results further supported that the PWC appears to be a useful mechanism in segmenting Cantonese speech.

Key words: Cantonese speech, segmentation, Possible-Word Constraint