Contrast preservation and constraints on individual phonetic variation

Ferdinand de Saussure, one of the founders of modern Linguistics, described language as a system where everything holds together. Regarding the sounds of language, this has led to the current view that the phonology of a language consists of a complex system of relations between contrastive phonemes. In this dissertation, I test whether there are constraints on individual phonetic variation from a multivariate perspective due to this system of relations, and how these constraints interact with contrast preservation. Two main views of contrast preservation are considered. The first view is that contrast preservation is merely an outcome of other regular phonetic processes that affect multiple consonants simultaneously. The second view is that contrast preservation acts as a constraint on the phonetic realization of phonemes. To this end, two phonetic experiments are performed. In both experiments, multiple acoustic measures of intervocalic consonant strength are taken, and PCA is used for dimensionality reduction, resulting in measures of overall consonant strength. These measures are then analyzed with Bayesian linear mixed effects regression (using weakly informative priors and maximal random effects structures) in order to obtain distributional information about both populations and individual speakers. In the first experiment, word-medial intervocalic /s/ and /f/ are compared for Valladolid Spanish and Barcelona Catalan. Both Catalan and Spanish have the fricatives /s/ and /f/, neither has /v/ contrasting with /f/, and only Catalan has /z/ contrasting with /s/. The results show that Catalan /s/ is stronger than Spanish /s/, but there is no evidence for a difference between the two language’s /f/ strengths, with strong evidence that the magnitude of the difference between Catalan and Spanish /s/ is larger than the magnitude of the difference between Catalan and Spanish /f/. I argue that these results are consistent with a role for contrast preservation as a constraint, with Catalan having stronger /s/ than Spanish because lenition of Catalan /s/ causes phonetic overlap with a contrasting phoneme, while lenition of Spanish /s/ does not. In the second experiment, the simultaneous lenition of Spanish intervocalic /ptk/ and /bdg/ in three dialects (Cuzco, Peru; Lima, Peru; and Valladolid, Spain) is examined. Cuzco is found to have the strongest productions for both /ptk/ and /bdg/, Lima the weakest for both, and Valladolid in between for both. That is, the same hierarchy of strength applies in both cases, though the evidence for the difference between Valladolid and Lima /ptk/ is considerably weaker than the evidence for the other differences. I argue that the results are consistent with constraints on multivariate variation at the dialectal level, but that further research is required to see how constraints at the individual level relate to population differences. Examining individual variation in both experiments, I find that the degree to which an individual speaker lenites /f/ is correlated with the degree to which they lenite /s/, and that the degree to which they lenite /ptk/ is correlated with both the degree to which they lenite /bdg/ and the degree to which they lenite /sf/. These correlations represent a significant constraint on individual phonetic variation from a multivariate perspective. While a connection between individuals’ /ptk/ and /bdg/ lenitions can be explained by both the constraint and outcome views of contrast preservation, the correlation between /sf/ and /ptk/ and the correlation between /s/ and /f/ lend support to the outcome view, and Catalan having stronger /s/ than Spanish but not stronger /f/ lends support to the constraint view. I argue for a framework in which acoustic lenition in a variety of intervocalic consonants may share a common articulatory source of lenition, giving rise to constraints on individual phonetic variation that may lead to contrast preservation as an outcome, but where there may additionally be a role for contrast preservation as a constraint. I conclude by discussing the importance of further acoustic studies that use the methodologies employed here, and studies that explore the articulatory and perceptual implications of the results. In the first experiment, word-medial intervocalic /s/ and /f/ are compared for Valladolid Spanish and Barcelona Catalan. Both Catalan and Spanish have the fricatives /s/ and /f/, neither has /v/ contrasting with /f/, and only Catalan has /z/ contrasting with /s/. The results show that Catalan /s/ is stronger than Spanish /s/, but there is no evidence for a difference between the two language’s /f/ strengths, with strong evidence that the magnitude of the difference between Catalan and Spanish /s/ is larger than the magnitude of the difference between Catalan and Spanish /f/. I argue that these results are consistent with a role for contrast preservation as a constraint, with Catalan having stronger /s/ than Spanish because lenition of Catalan /s/ causes phonetic overlap with a contrasting phoneme, while lenition of Spanish /s/ does not. In the second experiment, the simultaneous lenition of Spanish intervocalic /ptk/ and /bdg/ in three dialects (Cuzco, Peru; Lima, Peru; and Valladolid, Spain) is examined. Cuzco is found to have the strongest productions for both /ptk/ and /bdg/, Lima the weakest for both, and Valladolid in between for both. That is, the same hierarchy of strength applies in both cases, though the evidence for the difference between Valladolid and Lima /ptk/ is considerably weaker than the evidence for the other differences. I argue that the results are consistent with constraints on multivariate variation at the dialectal level, but that further research is required to see how constraints at the individual level relate to population differences. Examining individual variation in both experiments, I find that the degree to which an individual speaker lenites /f/ is correlated with the degree to which they lenite /s/, and that the degree to which they lenite /ptk/ is correlated with both the degree to which they lenite /bdg/ and the degree to which they lenite /sf/. These correlations represent a significant constraint on individual phonetic variation from a multivariate perspective. While a connection between individuals’ /ptk/ and /bdg/ lenitions can be explained by both the constraint and outcome views of contrast preservation, the correlation between /sf/ and /ptk/ and the correlation between /s/ and /f/ lend support to the outcome view, and Catalan having stronger /s/ than Spanish but not stronger /f/ lends support to the constraint view. I argue for a framework in which acoustic lenition in a variety of intervocalic consonants may share a common articulatory source of lenition, giving rise to constraints on individual phonetic variation which may lead to contrast preservation as an outcome, but where there may additionally be a role for contrast preservation as a constraint. I conclude by discussing the importance of further acoustic studies which use the methodologies employed here, and studies which explore the articulatory and perceptual implications of the results.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Spanish

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2017

Author dc:creator

• Eager, Christopher David

Contributors dc:contributor

• Hualde, José Ignacio
• Cole, Jennifer S.
• Escobar, Anna Maria
• Shosted, Ryan K.
• Roy, Joseph

• Spanish
• Phonetics
• Phonology
• Contrast preservation
• Chain shift
• Individual variation
• Dialect differences

Statement dc:rights

• Copyright 2017 Christopher David Eager

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/97345

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/97345