Stimulus dominance in fused dichotic syllables: trouble for the category goodness hypothesis.

Abstract

Repp [J. Acoust. Soc. Am. 60, 456-469 (1976)] hypothesized that dichotic stimulus dominance relationships between fused speech sounds reflect the relative category goodness of the competing stimuli, i.e., their relative perceptual distances from the listener's linguistic category prototypes. In experiment I, 15 synthetic syllables from a /bae/-/dae/-/gae) continuum were dichotically fused with three selected stimuli from the same continuum and presented to listeners for identification. Stimuli in the vicinity of phonetic category boundaries were found to make weaker dichotic competitors than stimuli from well within a phonetic category, as predicted. However, some systematic deviations from the predictions and large individual differences were noted. In experiment II, synthetic syllables from a /ba/-/da/-/ga/ continuum were paired with either of the two endpoint stimuli (/ba/,/ga/) in dichotic or mixed presentation. The resulting fused hybrid stimuli were presented in identification and AXB discrimination tests. The hybrids were perceived quite categorically, and there was little difference between dichotic and mixed modes of presentation. These results did not replicate earlier data [Repp, Haskins SR-45/46, 123-139 (1976)] that had suggested that discrimination of dichotic hybrid stimuli might be based on a level of representation preceding phonetic categorization. The category goodness hypothesis, which implies such a level, was neither supported nor contradicted. In experiment III, the prediction was tested that increasing the aspiration amplitude of a /ta/ syllable, and thus its category goodness, would increase the perceptual dominance of this stimulus over a /da/ simultaneously presented to the other ear. This hypothesis was not supported by the data. Rather, the response pattern suggested that a listener's success in perceptually integrating the aspiration noise (of /ta/) with the fused vocalic portion (of /da/ and /ta/) into a single phonetic percept depended on the perceived spatial locations of these stimulus portions. Taken together, the results of the three experiments cast doubt on the category goodness hypothesis and suggest that dichotic stimulus dominance in fused speech sounds is determined by psychoacoustic factors, some as yet undefined.