An acoustic analysis of lexical stress in non-words in Uyghur

Abstract

Previous research examined the stress pattern in Uyghur, a Turkic language, using real words and found that Uyghur used the duration as a stress cue but not F0, and the intensity was moderated by syllable types. In order to avoid vowel lengthening in the real words, the non-words (DOSdos vs. dosDOS or DOdo vs. doDO, capitalized as stressed) were used, in which the syllable types (CVC syllable and CV syllable) and vowels (a, u, i, o, ø, y) were controlled. In the production experiments, speakers were clearly instructed that the capitalized ones were stressed and lower-cased as unstressed syllables. All target words were embedded in the same carrier sentence (e.g. ‘I will say X now.'). In the production of ten female native Uyghur speakers, average fundamental frequency, duration, average intensity, and first and second formant frequencies for vowels were collected in the accented and unaccented syllables. The results showed that there were significant differences in duration and intensity between stressed and unstressed syllables. The fundamental frequency differences were associated with word final positions rather than stress positions. Vowels are centralized in the unstressed position. The present acoustic data suggest that native Uyghur speakers use of duration and intensity rather than F0.