Wang, C;
Xu, Y;
Zhang, J;
(2022)
The invalidity of rhythm class hypothesis.
In: Froto, Sónia and Vigário, Marina, (eds.)
Proceedings of Speech Prosody 2022.
(pp. pp. 347-351).
International Speech Communication Association (ISCA): Lisbon, Portugal.
Preview |
Text
wang22b_speechprosody.pdf - Published Version Download (313kB) | Preview |
Abstract
Languages are said to be stress-timed, syllable-timed or mora-timed. In a stress-timed language, inter-stress intervals are or tend to be constant, hence, isochronous, while in a syllable-timed or mora-timed language, successive syllables or morae are or tent to be equal in duration. Empirical research has failed to find evidence of isochrony in any language, yet the hypothesis is now sustained by perception accounts or phonetic metrics that do not measure isochrony. We have re-examined the rhythm class hypothesis by looking for evidence of at least a tendency toward isochrony, through a comparison of English, an alleged stress-timed language, and Mandarin, an alleged syllable-timed language. The results show that in English, segments are not compressible to allow equal syllable duration, and syllables are incompressible to enable equal inter-stress interval duration and phrase duration. In contrast, Mandarin shows a small tendency toward both equal syllable duration and equal phrase duration. These findings are exactly the opposite of what would be predicted by the rhythm class hypothesis. We therefore argue that the hypothesis is not just flawed, but simply untenable, and the so-called rhythm classes should no longer be held as a basic fact of human language.
Archive Staff Only
 |
View Item |
