| Abstract | Five monosyllabic words of Central Thai were recorded by ten native speakers & randomized into composite & individual tests. In the composite tests the utterances of all the speakers were randomized to prevent adaptation to any one speaker. In each of the individual tests, only a single speaker was used. Identifications were provided by 34 native speakers. The composite tests yielded a fair amount of confusion between the mid & low tones. This confusion was virtually eliminated in the individual tests. It is concluded that tones free of a linguistic context are better identified when the listener has access to the speaker's tone space. This effect is best shown by the mid-tone, which is likely to be confused with the low tone. The other 4 tones, which have much more FO movement & concomitant variation in amplitude, preserve their perceptual integrity more easily. |