LINGUIST List 33.2689

Sun Sep 04 2022

Diss: Germanic; Western Nilotic; Computational Linguistics; Historical Linguistics; Phonetics; Phonology: Stefon Michael Flego: ''The Emergence of Vowel Quality Mutation in Germanic and Dinka-Nuer: Modeling the Role of Information-Theoretic Factors Using Agent-Based Simulation''

Editor for this issue: Sarah Goldfinch <sgoldfinchlinguistlist.org>



Date: 24-Aug-2022
From: Stefon Flego <stefonflegogmail.com>
Subject: The Emergence of Vowel Quality Mutation in Germanic and Dinka-Nuer: Modeling the Role of Information-Theoretic Factors Using Agent-Based Simulation
E-mail this message to a friend

Institution: Indiana University Bloomington
Program: Linguistics
Dissertation Status: Completed
Degree Date: 2022

Author: Stefon Michael Flego

Dissertation Title: The Emergence of Vowel Quality Mutation in Germanic and Dinka-Nuer: Modeling the Role of Information-Theoretic Factors Using Agent-Based Simulation

Dissertation URL: https://www.proquest.com/pqdtglobal/docview/2705390335/6718BF4F6AE64B8BPQ/1

Linguistic Field(s): Computational Linguistics
                            Historical Linguistics
                            Phonetics
                            Phonology
Language Family(ies): Germanic
                            Western Nilotic

Dissertation Director:
Kelly Harper Berkson

Dissertation Abstract:

This dissertation uses computational modeling to examine the relationship between information-theoretic factors and the actuation and direction of language change. It models the emergence of vowel quality mutation (VQM), meaning stem vowel alternations that participate directly in morphosyntactic exponence and which have their historic origin in long-distance coarticulation with former suffix vowels (e.g. early Germanic *fōt-s ~ *fōt-iz --> English ‘foot’ ~ ‘feet’; Surkum kùɟ-ɛ̀ ~ kùɟ-ʌ̀, cf. Agar Dinka kṳ̀uc ~ kwò̤oc, ‘not know’.3SG~1SG). Given the cross-linguistic ubiquity of suffixing systems, long-distance coarticulation, and erosion of suffixal material over historical time, it is unclear why VQM has not developed more frequently across the world’s languages. Nor is it well understood why VQM would develop at all, as it acquires several typologically rare properties not shared with its antecedent system. Using agent-based modeling, I show that we can better understand the evolution of VQM by simulating how the contextual predictability of vowel quality contrasts in stems and suffixes affects the magnitude of long-distance coarticulation and erosion of suffixal material over historical time. The results indicate that VQM is only likely to emerge from a stem-suffix system under particular information-theoretic conditions, namely when stem vowel contrasts have relatively high average contextual predictability and suffix vowel contrasts have relatively low average predictability, an asymmetry consistent with data from early Germanic and Dinka-Nuer. The model predicts that these information-theoretic factors also influence the internal structure of the resulting mutation system and the size of its vowel quality inventory. This type of approach enables us to begin tracing the seeds of language change to quantifiable aspects of communication and language use, and gives insight into how linguistic systems evolve holistically— from the physics of speech articulation up through higher-level morphosyntactic expression.




Page Updated: 04-Sep-2022