Nari Rhee is happy to announce her dissertation proposal defense, which will be held on Monday, May 25th at 1:00pm EST. It will take place virtually.


The proposal document can be found here, and the abstract is included below. 



Title: Enhancement cue integration in many-to-many speech

Advisor: Jianjing Kuang

Committee: Gareth Roberts (chair), Mark Liberman, Gene Buckley



Any phonological contrast is distinguished in the acoustic space by weighted combinations of multiple covarying acoustic cues. A theoretically important question is: where do cue covariations arise in languages?  While some scholars proposed that the covariation of acoustic cues is largely driven by physiological factors in the production of the primary cues, other studies have claimed that covarying cues are language-specifically controlled to enhance phonological contrasts of the language. In this dissertation, I investigate the role and nature of the cue covariations, to answer the question of how much of the covariation is inherently determined, or integrated in and enhanced by the language-specific phonological structures. I propose to approach this question through three studies examining i) cross-linguistic patterns in the correlation between cues, ii) the language-, context-, and feature- dependent role of the covariations, and iii) whether enhancement cue covariations have to be learned. If cue covariation is inherent to the anatomy of speech production (and perception), it should be similar across different languages, and should not have to be learned. Preliminary results based on four languages (Mandarin, Thai, Korean, and English) suggest that while languages share some pattern in the physiologically-triggered covariation of cues, the covariation relationship is fine-tuned by the language's phonological structures, and has to be learned during the process of language development.