ABSTRACT
The majority of words in Cantonese are compounds, which seems likely to burden the process of identifying words in running speech. Cantonese is also a stress-timed language, which reduces the potential for durational contrasts to distinguish embedded constituents from self-standing words. The current study demonstrates the challenge of identifying words in spoken Cantonese. As a compound unfolds, listeners are more likely to consider an onset-embedded constituent as the intended word than the actual word they are hearing – a result that seems poorly adapted to the prevalence of compounds. However, the results also show these challenges are offset by sentence-based cues, such as those provided by noun classifiers. This occurs despite variability in classifier-noun pairings and the fact that adult speakers often show incomplete mastery of these pairings. Together the results demonstrate how even highly biased cases of lexical competition are overcome by sentence-level constraints that may be only moderate in strength.
Acknowledgements
The authors thank Dr. James Myers and the anonymous reviewers for their many helpful comments and suggestions.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. See Myers and Gong (Citation2002) and Packard (Citation2000) for discussion and evidence supporting the idea that spoken word recognition in Chinese languages operates with the end goal of identifying words rather than individual morphemes.
2. Romanized spellings are given in Jyutping. Numbers denote the tone of the word.
3. One of the items in the same classifier condition was excluded from the analyses due to a design error discovered only after testing was complete (the classifier could in fact only occur with the noun for the target and not the depicted competitor). Measures in this condition are therefore based on 11 observations per participant rather than 12.