Perception, Production, and the Implementation of Phonological Opacity in the Bangla Vowel Chain Shift
General Material Designation
[Thesis]
First Statement of Responsibility
Nagle, Traci Christine
Subsequent Statement of Responsibility
Berkson, Kelly H.
.PUBLICATION, DISTRIBUTION, ETC
Name of Publisher, Distributor, etc.
Indiana University
Date of Publication, Distribution, etc.
2020
PHYSICAL DESCRIPTION
Specific Material Designation and Extent of Item
256
DISSERTATION (THESIS) NOTE
Dissertation or thesis details and type of degree
Ph.D.
Body granting the degree
Indiana University
Text preceding or following the note
2020
SUMMARY OR ABSTRACT
Text of Note
This dissertation explores the productivity of a phonologically opaque synchronic chain shift in Bangla, a language spoken in India and Bangladesh by 265 million people. Bangla exhibits a vowel-harmony chain shift in which certain verb-stem vowels alternate [(i ~ e), (e ~ æ), (u ~ o), and (o~ ɔ)] based on the identity of a vowel in the neighboring suffix. This pattern is called a chain shift because harmonization with affix vowels causes stem vowels to move only one step along an ordered scale of vowel heights. Previous work has suggested that opaque patterns are difficult to learn and may not be extended to nonce words (e.g., Zhang 2016). Using production and perception experiments conducted with native speakers of Bangla, this dissertation investigates the nature of this alternation pattern in the grammar of native Bangla speakers. The central question is whether speakers can extend this chain-shift pattern to nonce words. Though the chain shift pattern is exceptionless in real verbs, the results indicate that Bangla speakers do not freely extend this pattern to nonce words. Instead, they apply the chain shift only about half of the time. Analyses conducted to probe possible acoustic explanations for these results (e.g., incomplete neutralization, vowel-to-vowel coarticulation) suggest that the patterns observed cannot be explained by phonetics alone. Rather, they seem to point to underlearning (Becker, Ketrez, & Nevins 2011) of the chain shift pattern. As Bangla remains under-examined in modern theoretical linguistic and phonological literature, this dissertation contributes to the general literature on Bangla. Its findings also carry implications for formal models of phonology and language learning, and for the conceptualization and modeling of synchronic chain shifts. The results indicate that chain shifts, and possibly opaque patterns more generally, may operate very differently in cognitive systems than some theoretical models have assumed.