Many phonological processes originated as phonetic reduction phenomena. An informative way to understand these processes and the related issue of the phonology/phonetics interface is to build a phonetic implementation model that simulates these processes. This paper investigates production data of Mandarin tone2 (a rising tone) in an environment that may induce Mandarin tone2 sandhi, and shows that the sandhi rule is a phonetic reduction phenomenon occurring in prosodically weak positions. The data supports a model where speech variation arises as the speaker balances the trade-off relationship between ease of articulatory effort and communication accuracy. A weight representing prosodic strength is introduced in the simulation model to capture this trade-off relationship. The shifting prosodic weight from unit to unit controls surface phonetic variations which match the production data.
Cite as: Shih, C. (2005) Understanding phonology by phonetic implementation. Proc. Interspeech 2005, 2469-2472, doi: 10.21437/Interspeech.2005-757
@inproceedings{shih05_interspeech,
author={Chilin Shih},
title={{Understanding phonology by phonetic implementation}},
year=2005,
booktitle={Proc. Interspeech 2005},
pages={2469--2472},
doi={10.21437/Interspeech.2005-757}
}