LIMITS ON PHONETIC INTEGRATION IN DUPLEX PERCEPTION

Citation
Dh. Whalen et Am. Liberman, LIMITS ON PHONETIC INTEGRATION IN DUPLEX PERCEPTION, Perception & psychophysics, 58(6), 1996, pp. 857-870
Citations number
39
Categorie Soggetti
Psychology, Experimental",Psychology
Journal title
ISSN journal
00315117
Volume
58
Issue
6
Year of publication
1996
Pages
857 - 870
Database
ISI
SICI code
0031-5117(1996)58:6<857:LOPIID>2.0.ZU;2-N
Abstract
The telling fact about duplex perception is that listeners integrate i nto a unitary phonetic percept signals that are coherent from a phonet ic point of view, even though the signals are, on purely auditory grou nds, separate sources. Here we explore the limits on the integration o f a sinusoidal consonant cue (the F3 transition for [dal vs. [gal) wit h the resonances of the remainder of the syllable. Perceiving duplexly , listeners hear the whistle of the sinusoid, but also the [dal and [g al for which the sinusoid provides the critical information. In the fi rst experiment, phonetic integration was significantly reduced, but no t to zero, by a precursor that extended the transition cue forward in time so that it started 50 msec before the cue. The effect was the sam e above and below the duplexity threshold (the intensity of sinusoid i n the combined pattern at which the whistle was just barely audible). In the second experiment, integration was reduced once again by the pr ecursor, and also, but only below the duplexity threshold, by harmonic s of the cues that were simultaneous with it. The third experiment sho wed that the simultaneous harmonics reduced phonetic integration only by serving as distracters while also permitting the conclusion that th e precursor produced its effects by making the cue part of a coherent and competing auditory pattern, and so ''capturing'' it. The fourth ex periment supported this interpretation by showing that for some subjec ts the amount of capture was reduced when the capturing tone was itsel f captured by being made part of a tonal complex. The results support the assumption that the independent phonetic system will integrate acr oss disparate sources according to the cohesive power of that system a s measured against the evidence for separate sources.