Related Experiment Videos
Suprasegmental and segmental timing models in Mandarin Chinese and American English
1Lucent Technologies Bell Laboratories, Murray Hill, New Jersey 07974-0636, USA. vansanten@ece.ogi.edu
The Journal of the Acoustical Society of America
|February 25, 2000
Summary
This study challenges suprasegmental timing assumptions in speech. Syllable timing is not independent of its segments, and segment durations are influenced by prosodic context, not just the syllable unit.
Area of Science:
- Phonetics
- Linguistics
- Speech Science
Background:
- The concept of suprasegmental timing relies on segmental independence and suprasegmental mediation.
- These assumptions are central to isochrony and syllable timing hypotheses in phonology.
- Previous models often assume segments adjust to fit suprasegmental units.
Purpose of the Study:
- To formalize and empirically test the assumptions of segmental independence and suprasegmental mediation.
- To investigate the validity of these assumptions in American English and Mandarin Chinese.
- To determine the role of suprasegmental units in speech timing.
Main Methods:
- Used meaningful, nonrepetitive sentences in American English and Mandarin Chinese.
- Measured syllable duration in relation to segmental composition to test segmental independence.
- Analyzed the influence of prosodic factors (phrasal position, stress, tone) on segment durations within syllables to test suprasegmental mediation.
Main Results:
- Segmental composition significantly affects syllable duration, contradicting segmental independence.
- Prosodic factors exert systematic, differing influences on specific locations within syllables, challenging suprasegmental mediation.
- Found interactions between consonantal identity, syllable position, and stress affecting segment duration.
Conclusions:
- The tested assumptions of segmental independence and suprasegmental mediation are not supported by the data.
- Suprasegmental units play a role in phonology but the concept of suprasegmental timing as a predictive model is questionable.
- Segmental durations are not simply compressed or elongated to fit a predetermined syllabic time.