ACOUSTIC CORRELATES OF LEXICAL STRESS IN NATIVE SPEAKERS OF UYGHUR AND L2 LEARNERS

Abstract

Some syllables are louder, longer and stronger than other syllables at the lexical level. These prominent prosodic characteristics of certain syllables are captured by suprasegmental features including fundamental frequency, duration and intensity. A language like English uses fundamental frequency, duration and intensity to distinguish stressed syllables from unstressed syllables; however, a language like Japanese only uses fundamental frequency to distinguish the stressed syllables from unstressed syllables. This study investigates the stress pattern of Uyghur, a Turkic language, as produced by native and non-native speakers. The first three experiments provide a detailed phonetic analysis in order to determine the acoustic cues to stress in Uyghur. In Experiment 1, six disyllabic minimal pairs (e.g., A-cha, a-CHA), contrasting in location of stress, were produced by five native Uyghur speakers with three repetitions in a fixed sentence context. In order to generalize the results from the small set of minimal pairs in the first experiment, Experiment 2 examined the initial syllable of disyllabic nouns that contrasted in first-syllable stress (e.g., DA-ka, da-LA) while syllabic structure (CV versus CVC) was also manipulated. In both experiments, average fundamental frequency, syllable duration, and average intensity were collected in accented and unaccented syllables. The results from both experiments showed that there were significant differences in duration and intensity between stressed and unstressed syllables, with the intensity differences moderated by syllable structure. No difference was found in fundamental frequency. Experiment 3 investigated the role of F0 in lexical stress. Experiment 3 focused on the interaction between sentential intonation and lexical stress in which the declarative assertion sentence (falling F0) and the declarative question sentence (rising F0) were used. The results confirmed the previous experiments. No interaction between sentential intonation and lexical stress indicated that the obtained duration effect was due to lexical stress. There were no effects of fundamental frequency or intensity in terms of stress. While previous studies have classified Uyghur as a pitch-accent and a stress-accent language, the present acoustic data suggest that native speakers make no use of pitch cues to signal stress in Uyghur. Previous research has focused on the acquisition of lexical stress by non-native speakers of English. This study also examined the acquisition of lexical stress by English learners of Uyghur. Five highly advanced English learners of Uyghur produced the six minimal pairs and disyllabic nouns contrasting in the first syllables. The stimuli that were produced by L2 learners were the same as in Experiment 1 and Experiment 2. Highly advanced Uyghur learners used duration as a cue and did not use fundamental frequency and intensity as stress cues. The results indicated that native-like lexical stress can be acquired at the high advanced level.