Differential Sensitivity to Synthetic Speech Sounds within and between Phoneme Categories
Katherine Safford Harris, Alvin M. Liberman, Howard S. Hoffman, Belver C. Griffith · The Journal of the Acoustical Society of America · 1956
The use of synthesizers provides an opportunity to vary speech-like sounds in small steps along a single acoustic continuum. We find that at some points on such a continuum large changes in the acoustic stimulus have no effect on phoneme identification, while at other points small changes cause the listener's identification to shift abruptly from one phoneme to another. Casual observation led us to suspect that there might be related discontinuities in the discriminability of these sounds—that is, that discrimination would be less sharp, other things equal, between sounds in the same phoneme category than between sounds which lie on opposite sides of a phoneme boundary. This effect might be related to the difficulties that linguists often experience in hearing certain differences among the sounds of an exotic language. Discrimination and identification functions were obtained for a series of stimuli which differed in the second-formant transition. In one part of the experiment these stimuli were presented singly to listeners for identification as b, d, or g; in another part, the discriminability of the stimuli was measured by an ABX technique. It was found that discrimination was, indeed, better in the vicinity of phoneme boundaries than it was near the middle of a category. The obtained discrimination function was quite close to a function predictable from the identification judgments on the extreme assumption that the listeners were able to discriminate the sounds only to the extent that they could differentially identify them as b, d, and g. [This work was supported in part by the Carnegie Corporation of New York, and in part by the Department of Defense in connection with Contract DA49-170-SC-1642.]