Lexical variation and rime -tone correlation in early tonal acquisition: a longitudinal study of Mandarin Chinese

Studies of the acquisition of tones in Chinese dialects have documented the early mastery of tones in terms of accuracy in children’s tonal production, suggesting that children produce the distinctive tones of their target language competently by age two or earlier (Li and Thompson 1977, J. Tse 1977, Clumeck 1980, Hsu 1996, Zhu and Dodd 2000). These studies have also revealed interesting patterns of order of tone acquisition (such as the level tone and the high falling tone occurring before the fall-rise tone in Mandarin), and of tonal substitution in child language (such as the substitution of a high rising tone for the fall-rise tone in Mandarin). However, the claim about early tonal acquisition deserves critical scrutiny even at the level of phonetic production, given that detailed longitudinal studies have revealed that some tones such as the low rising in Cantonese are acquired late (A. Tse 1993), and a clear lexical effect has been observed for the acquisition of Taiwanese tone (Tsay 2004). In line with other studies suggesting lexical diffusion in phonological development (Ferguson and Farwell 1975, Hsieh 1977), the fact that children can produce a tone accurately on selected morphemes may not mean s/he can produce the same tone accurately for most or all lexical items being acquired. Recent pitch studies of the babbling of infants acquiring Cantonese and Mandarin have revealed preferences for level pitch over falling and rising pitch contours, which were in turn favored over concave and convex contours (Chen 2005, Lee, Lee and Chen 2005). How these preferential pitch patterns in babbling evolve into the adult tonal system would need to be understood on the basis of acoustical analysis of infant pitch production, as well as analysis of adult input. The present study addresses the following issues: (a) What is the acquisition order of Mandarin tones, with respect to consistency and accuracy of tonal production? (b) To what extent does accuracy of tonal production depend on the lexical dimension? (c) Given that tone needs to be realized on the syllable, would there be any correlation between tonal complexity and the complexity of the rime of the syllable, in the spirit of Woo (1969) and Yip (1980,1989) (d) Do the tonal substitution patterns in child language reflect early preferences for pitch contour types?