Abstract
Introduction Japanese is often characterized as a mora-timed language as opposed to syllable-timed and stress-timed languages (Han, 1962; Port et al., 1987). Mora-timing refers to the kind of temporal organization whereby each mora tends to recur at a regular interval. Many phonetic studies of Japanese, however, fail to prove this isochronous nature of the mora (see Warner & Arai, 2001, for a review of the literature). They found out, instead, that morae in Japanese are not isochronous in a strict physical sense. Moreover, some studies provide evidence that vowels in Japanese exhibit an effect of temporal compensation both within and across morae. In /CV. CV/ sequences, for example, temporal compensation occurs not only between C and V within the same syllable/mora but also between V and C across a syllable/mora boundary (Higuchi & Fujisaki 1981). This suggests that control of segment sequences may be independent of mora boundaries. While the phonetic studies of speech timing in Japanese provide more or less negative evidence for the traditional conception of mora-timing in Japanese, analyses of phonological and phonetic patterns do hint that mora-based control of speech is administered in Japanese. We will consider the following three types of phenomena in the following sections: (a) phonological phenomena such as phonotactics and vowel epenthesis (b) phonetic phenomena including assimilation, vowel devoicing and temporal compensation at the syllable level (c) segmentation phenomena such as those observed in word blending, stuttering and speech errors. All these phenomena show a strong link between segments within a mora but not across two morae.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have