Observations on consonant types and “tone” in Tibetan

Abstract

This paper is the first to report on some acoustic and physiological characteristics of spoken Tibetan. Phonetic experiments were performed in an attempt to find any association between consonant types and the so-called tones (as has been found to exist on the phonological level, at least [Kjellin, 1975]). Digital pitch extraction revealed that F 0 is maximally separated into two groups at the release of initial consonants—lower for lexically voiced and higher for lexically voiceless obstruents—but that the initially-lower F 0 rises during the course of the syllable to reach the same level as that of the other group, as, for instance, it does in English also. The range between the highest and lowest F 0 was relatively small. A fiberscopic study of glottal gestures during speech showed that the time course of the vocal-fold abduction/adduction for voicelessness was essentially the same for all voiceless consonants including those which are lexically voiced. Electromyograms (EMG) of laryngeal muscles displayed typical patterns associated with F 0 changes in the vowels and/or with the actual state of voicing in the consonants, but there was nothing which could be unequivocally interpreted as representing the lexical voicing of devoiced obstruents. Only the vocalis muscle had an EMG pattern for the closure period of the lexically voiced but actually devoiced and aspirated stop [pH] (followed by lowered F 0 ) which resembled the patterns observed for the true voiced consonants.