Lishui Wu tone and tone sandhi : an acoustically-based description

Abstract

This thesis describes tone and tone sandhi in the Lishui {u4E3D}{u6C34} variety of Wu Chinese from the Chuqu {u5904}{u8862} subgroup in southern Zhejiang, using data from three speakers. The description is both auditory and (normalised-) acoustic, and includes also a derivational account of the tonology, and a description of Lishui segmental phonology and common lexicon. Descriptions are given for citation tones and disyllabic lexical tone sandhi. The results show that, according to Chao's (1928) synchronic typology of Wu, Lishui does not have a typically Wu phonology. Segmentally, Lishui lacks the diagnostic three-way distinction between syllable-initial stops and affricates, distinguishing only +/-aspirated obstruents (with the -aspirated obstruents showing free variation in voicing). Like the Northern Min dialects to its south, its tone sandhi is right-focussed: word-final tones are realised mostly the same as in isolation (although there are some changes), but non-word-final tones have different realisations determined by their input tone and the following tone. Lishui has seven citation tones: mid dipping [324], high and low falling-level [433], [211], high falling [52], low convex [342], and high [5] and low [23] stopped tones. The normalised tonal acoustics show that citation realisations are consistent between speakers. Although the realisation of tones on non-word-final syllables is also shown by normalisation to be mostly consistent between speakers, considerable between-speaker variation is demonstrated in the incidence of these realisations. One of the interesting findings is that the number of realisations on non-word-final syllables, although representing neutralisation of citation contrasts, is often greater than found in citation tones. Moreover, there are very few instances of phonetically motivated sandhi changes, and the changes are usually better expressed in terms of abstract categories which relate to Middle Chinese tonal categories. It is shown that, as a result, a derivational account of the non-word-final tonal realisations is messy. The between-speaker variation in tone sandhi suggests that Lishui tone sandhi is changing quickly. Show more