LINGUIST List 18.1392

Tue May 08 2007

Diss: Phonetics/Phonology/Psycholing: Ogasawara: 'Processing of Spe...'

Editor for this issue: Hunter Lockwood <hunterlinguistlist.org>


Directory         1.    Naomi Ogasawara, Processing of Speech Variability: Vowel reduction in Japanese


Message 1: Processing of Speech Variability: Vowel reduction in Japanese
Date: 08-May-2007
From: Naomi Ogasawara <naomi703email.arizona.edu>
Subject: Processing of Speech Variability: Vowel reduction in Japanese


Institution: University of Arizona Program: Department of Linguistics Dissertation Status: Completed Degree Date: 2007

Author: Naomi Ogasawara

Dissertation Title: Processing of Speech Variability: Vowel reduction in Japanese

Linguistic Field(s): Phonetics                             Phonology                             Psycholinguistics
Dissertation Director:
Merrill F. Garrett Adam P. Ussishkin Timothy Vance Natasha L. Warner Andrew Wedel
Dissertation Abstract:

This dissertation investigates the processing of speech variability,allophonic and indexical variation in Japanese. A series of speechperception experiments were conducted with reduced and fully voiced vowelsin Japanese as a test case. Reduced vowels should be difficult forlisteners to hear because they are acoustically less salient than fullyvoiced vowels, due to the lack of relevant physiological properties. On theother hand, reduced vowels between voiceless consonants represent morecommon phonological patterns than fully voiced vowels. Furthermore,previous studies found that Japanese listeners were capable of hearingcompletely deleted vowels. Listeners intuitively maintain CV syllables inperception, hearing a vowel after each consonant in order to avoidobstruent clusters (a violation of Japanese phonotactics).

It was found that listeners made good use of acoustic, phonological, andphonotactic knowledge of their native language for processing allophonicvariants. In word recognition, listeners performed better when reducedvowels were in the environment where vowel reduction was expected. Thephonological appropriateness of an allophone was judged in relation toadjacent consonants on both sides, and the facilitatory effect ofappropriateness of reduced vowels surpassed the inhibitory effect of theiracoustic weakness. However, in terms of sound detection, listeners foundreduced and fully voiced vowels equally easy to hear in an environmentwhere vowel reduction was expected. Although reduced vowels werephonologically appropriate between voiceless consonants, the phonologicalappropriateness merely balanced out acoustic weakness; it was not strongenough to surpass it. In addition, the phonological appropriateness of anallophone was judged based only on the preceding consonant, which suggeststhat listeners processed sounds linearly. Furthermore, the study found thatphonological appropriateness of the allophone was affected by dialectaldifferences and speech rates. Listeners' preference for a certain allophonewas influenced by the phonology of a listeners' native dialect andexpectation was skewed by fast speech rates.

This study suggests that current speech perception models need modificationto account for the processing of speech variability takinglanguage-specific phonological knowledge into consideration. The studydemonstrated that it is important to investigate at which stagephonological inference takes place during processing.