Incorporating tone when modeling wordlikeness judgments
Youngah Do, Ryan Ka Yau Lai
November 2020
 

Various phonotactic models have been proposed to predict speakers’ wordlikeness judgments but most have focused primarily on segments. This article aims to model speakers’ wordlikeness judgments incorporating tone. We first show how the two major determinants of wordlikeness judgments, namely phonotactic probability and neighborhood density, can be applied to tone languages. To test the role of the two determinants to wordlikeness judgments in a tone language, judgment data are obtained from speakers of Cantonese. The results are then used to model speakers’ judgments, showing that phonotactic probability, but not neighborhood density, modulates wordlikeness judgments and that the phonotactic probabilities involving nucleus and coda are most relevant to wordlikeness judgments. We also show that phonotactic probability affects the tendency to judge items being absolutely perfect or more or less wordlike, while it does not affect the judgments that an item is absolutely not-wordlike. Implications of these results for phonotactic modeling and processes involved in wordlikeness judgments are discussed.
Format: [ pdf ]
Reference: lingbuzz/005566
(please use that when you cite this article)
Published in: Phonology (to appear in December 2020)
keywords: wordlikeness judgments, phonotactic probability, neighborhood density, tone, phonology
Downloaded:424 times

 

[ edit this article | back to article list ]