ABSTRACT
This study investigates the errors made by 60 English-, Japanese-, and Korean-speaking learners of Chinese when producing the two contour lexical tones T2 (rising tone) and T4 (falling tone) in disyllabic words. It finds that T2 is produced at a greater rate of accuracy in word-initial positions, while T4 is produced at a greater rate of accuracy in word-final positions. Two intertonal effects shared across the three groups of speakers were also observed: 1) the accuracy rate of T2 is always greater when it is followed by tones with low onsets than when it is followed by tones with high onsets, and 2) the accuracy rate of T4 is always greater when it is followed by low tones than when it is followed by other tones. The study presents evidence suggesting that second language tones are partially constrained by anticipatory dissimilation.
