LINGUIST List 23.4646
Tue Nov 06 2012
Diss: Phonetics/ Phonology/ Catalan-Valencian-Balear/ Dutch/ Polish/ Spanish; Strycharczuk; 'Phonetics-Phonology Interactions...'
Editor for this issue: Lili Xia
Patrycja Strycharczuk <pstrycharczuk
Phonetics-Phonology Interactions in Pre-sonorant Voicing
E-mail this message to a friend
Institution: University of Manchester
Program: Linguistics and English Language
Dissertation Status: Completed
Degree Date: 2012
Author: Patrycja Strycharczuk
Dissertation Title: Phonetics-Phonology Interactions in Pre-sonorant Voicing
Dissertation URL: http://personalpages.manchester.ac.uk/staff/patrycja.strycharczuk/Research_files/PhD_2_sided.pdf
Subject Language(s): Catalan-Valencian-Balear (cat)
This dissertation is a study of phonetics-phonology interactions in pre-sonorant voicing. I use the term 'pre-sonorant voicing' to refer to a process in which a word-final obstruent becomes voiced when a sonorant sound follows in the next word. An example of this is the realisation of word-final /s/ in the Quito dialect of Spanish, pronounced as [z] when the next word begins in a vowel or a sonorant consonant. For instance, Spanish 'gas acre' ('acrid gas') and 'gas noble' ('noble gas') are pronounced with a [z] in the Quito dialect. However, in the absolute word-final position /s/ surfaces as voiceless, and so 'gas' is pronounced as [gas]. In my study of Quito Spanish, I address the question of whether this /s/-voicing process is most accurately analysed as belonging to the domain of abstract phonological computation, or whether the occurrence of voicing can be explained by the coarticulatory influence of the neighbouring sound, in which case the process should be interpreted as phonetic. I approach this question empirically, based on purpose-collected experimental data from native speakers of the Quito dialect, using speech rate manipulations. The results of the speech rate test indicate that at least some speakers actively target a voiced variant in their realisation of /s/-voicing, which is consistent with a phonological interpretation. Such an interpretation, however, is difficult to reconcile with what is predicted by any contrast-driven test for phonological features (e.g. Minimal Pairs Test, Successive Division Algorithm, see Currie-Hall (2006) for an overview), since [s] and [z] are not lexically contrastive in Quito Spanish, and thus cannot be considered phonemes. Based on these diverging predictions, I consider an alternative bottom-up model where abstract features emerge from the continuous phonetics.
Other languages analysed in this dissertation deliver further examples of phonological pre-sonorant voicing which I identify using a variety of diagnostics. These include statistical distributions of acoustic parameters, with bimodality pointing to phonological voicing in West-Flemish and Poznan Polish. Further, a systematic comparison of voicing in different environments performed by means of mixed-effects regression supports the phonological status of prevocalic sibilant and cluster voicing in Central Catalan. These empirical findings bear on a number of current issues within phonological theory. The data from West-Flemish and Catalan are of relevance to the discussion of the relationship between phonology and phonetics, and the role of synchrony and diachrony in explaining sound patterns. The results from Poznan Polish contribute to the debate on phonological neutralisation, while the findings on Quito Spanish are of consequence to models of phonology-morphosyntax interactions.
Page Updated: 06-Nov-2012