14th International Congress of Phonetic Sciences (ICPhS-14)San Francisco, CA, USA |
This study provides evidence against the hypothesis that Japanese
has word level mora-timing. Unlike previous studies which used
careful speech, this paper evaluates timing in a corpus of
spontaneous Japanese speech from 11 speakers.
Correlations between word duration and number of moras in
the word are shown to be much lower than in careful speech
studies. Furthermore, if there were durational normalization at
the level of the word, then there should be some unit across
which this normalization takes place. However, we show that
there is no consistency across speakers as to whether the lexical
word or the prosodic word serves as such a unit. Analyses of
arbitrarily truncated words further confirm that a linear
accumulative model of variance can explain the data, but a model
with mora compensation cannot.
Bibliographic reference. Arai, Takayuki / Warner, Natasha (1999): "Word level timing in spontaneous Japanese speech", In ICPhS-14, 1055-1058.