Why Aren't All Cantonese Tones Equally Confusing to English Listeners?

English listeners often struggle to perceive tones, but some are easier than others. This study examined these phenomena grounded in the feature weighing perspective (FWP) and the Perceptual Assimilation Model for Suprasegmentals (PAM-S). Forty-seven English and Cantonese listeners completed 4,212 trials of Cantonese tone discrimination and sequence recall tasks. The English listeners showed asymmetrical perceptual patterns of discrimination but not sequence recall. Specifically, these English listeners discriminated T1-T5, T3-T5, and T4-T5 more accurately than T1-T4, T3-T4, and T1-T3. However, they recalled the contour tone and level tone sequences with similar accuracies. Results of the discrimination task aligned with the predictions of PAM-S but not FWP. However, results of the sequence recall task did not support PAM-S. Together, these results suggest that PAM-S only applies to simple discrimination, not abstract phonological processing with a high memory load.

Automated summary

This study examined the perception of Cantonese tones by English listeners using tone discrimination and sequence recall tasks. The results showed that English listeners had asymmetrical patterns of discrimination but consistent performance in sequence recall. The findings supported the predictions of the Perceptual Assimilation Model for Suprasegmentals (PAM-S) in tone discrimination, but not in sequence recall. This suggests that PAM-S is applicable to simple discrimination tasks but not to abstract phonological processing with high memory load.