Does Perceptual Assimilation Model for Suprasegmentals Apply to Tone Sequence Recall? Evidence from Punjabi Learners of Cantonese

Abstract

<p>Building on prior research that indicates PAM-S’ ability in predicting non-native tone<br>discrimination, this study investigates the applicability of the Perceptual Assimilation Model<br>for Suprasegmentals (PAM-S) to sequence recall. Twenty-two Punjabi learners of Cantonese<br>completed a Cantonese tone assimilation task and a Cantonese tone sequence recall task.<br>Concerning assimilation, the results showed that the high level tone and the mid-level tone<br>were assimilated as a single category (SC) pair, and the low falling tone and the low rising tone<br>were assimilated as an uncategorized-categorized pair without overlap categories (UC-no).<br>Regarding sequence recall, the recall accuracy was higher for the UC-no pair compared to the<br>SC pair for varying sequence lengths. The findings are in line with the prediction of PAM-S on<br>the sequence recall task. In particular, PAM-S applies to the sequence recall tasks under<br>different memory loads at the phonological level.</p>