UCLA Working Papers in Linguistics, no.2
Papers in Phonology 3
edited by Matthew K. Gordon
| Chai-Shune Hsu | Voicing underspecification in Taiwanese word-final consonants |
1–24 |
| Donca Steriade | Phonetics in phonology: The case of laryngeal neutralization |
25–146 |
| Jie Zhang | Duration in the tonal phonology of Pingyao Chinese |
147–206 |
| Taehong Cho | Intra-dialectal variation in Korean consonant cluster simplification: A stochastic approach |
207–226 |
Back to UCLA WPL publication list
Abstracts
Chai-Shune Hsu – Voicing underspecification in Taiwanese word-final consonants
This paper presents evidence for voicing underspecification from Taiwanese. It is shown that laryngeally underspecified obstruents can be realized in the least-effortful voicing state, and that the acoustic outcome will be contextually variable. Contrary to Lombardi (1995), I argue that VOICE is a binary feature, and that [0 voice] is a phonological category distinct from either [+voice] or [-voice]. In Taiwanese, word-initial voiced stops are [+voice], while word-initial voiceless stops are [-voice]; word-final stops are unspecified for the feature VOICE (i.e., [0 voice]). Results from experiments show that these assumptions correctly predict the observed range of voicing variability of Taiwanese word-final stops.
Up to table of contents
Donca Steriade – Phonetics in phonology: The case of laryngeal neutralization
The study identifies the factors responsible for the loss of laryngeal contrasts and the reflexes of these factors in individual grammars. The main result reported is that the site of laryngeal neutralization can be uniformly identified by reference to phonetic implementation factors. Many of these factors are perceptual: laryngeal categories are neutralized in positions where the cues to the relevant contrast would be diminished or obtainable only at the cost of additional articulatory maneuvers. Conversely, laryngeal contrasts are permitted (or licensed) in positions that are high on a scale of perceptibility. It is argued here that the main factor involved in neutralization and licensing is the distribution of cues to the relevant contrasts. This hypothesis, referred to as Licensing by Cue, is compared here to the idea of Licensing by Prosody (Ito 1986, 1989, Goldsmith 1990, Rubach 1990, Lombardi 1991, 1995) according to which the distribution of features in general - and of laryngeal features in particular - is controlled by their prosodic position. The general idea pursued here is that phonological grammars incorporate knowledge of the conditions under which feature contrasts are physically implemented. The focus in this study is on the empirical evidence supporting such a view.
Up to table of contents
Jie Zhang – Duration in the tonal phonology of Pingyao Chinese
This thesis investigates the effects of phonetic duration on tonal behavior, especially tone sandhi, in Pingyao Chinese. Two types of phonetic duration are under consideration: duration of the rime in a syllable, and the duration between the end of a rime and the beginning of the following rime across a syllable boundary. Rime duration conditions the pitch distance between the tonal targets of a contour tone-the longer the rime, the more pronounced the tonal transition. The overall duration of a disyllabic or multisyllabic word constrains the pitch movements of the word-the pitch can only be allowed to inflect up to a certain complexity in a word. Moreover, given the limited duration between the end pitch of a rime and the starting pitch of the following rime, the tonal grammar prefers a less drastic pitch change across the syllable boundary. A general OT grammar is first constructed based on these duration factors and other traditional phonological considerations such as faithfulness and the OCP. A language-specific OT grammar for Pingyao is then presented. The grammar accounts for the sandhi pattern as well as lexical realization of Pingyao tones. Its advantages over previously-documented rule-based analyses are also discussed.
Up to table of contents
In Korean, a tri-consonantal sequence in VC1C2-C3V is reduced toVC1-C3V orVC2-C3V depending on the type of consonant sequence or dialect. While previous OT work accounted for variations between dialects by reranking contraints, possible intra-dialectal variations have rarely been discussed. This paper provides an Optimality Theoretic account of free variation in tri-consonant cluster reduction. The first goal is to show, based on data collected from 8 Seoul speakers, that there is substantial intra-dialectal variation in consonant cluster reduction, which is far more complicated than has been previously discussed. Second, building on Jun's phonetically-driven OT account, this paper provides further evidence that phonetic factors plays a pivotal role in determining complex patterns of the consonant reduction. In doing so, relationship between tri-consonant reduction and place assimilation will be also discussed. Finally, this paper shows that the relative frequency of the variants can be correctly captured by the Gradual Learning Algorithm (Boersma & Hayes 1999) in which constraints are ranked stochastically on a continuous scale with possibly overlapping values.
Up to table of contents