Is the perceptual magnet effect a matter of auditory sensitivity?

Abstract

Efforts to validate and extend Kuhl’s [P. K. Kuhl, Percept. Psychophys. 50, 93–107 (1991)] perceptual magnet effect for vowel prototypes in adult listeners has been somewhat elusive and particularly difficult to replicate in other than the /i/ vowel space. However, there is converging evidence that suggests the magnet effect, attributed to assimilation by a prototype, may more likely be due to differences in auditory sensitivity across the vowel space. Hawks’ [J. W. Hawks, J. Acoust. Soc. Am. 95, 1074–1084 (1994)] data on discrimination of formant changes for 17 synthetic vowels supported this notion, indicating that difference limens (DLs) based from an exemplar token for /i/ was among the largest (poorest) found, while tokens neighboring /i/ yielded the smallest DLs. If the perceptual magnet effect is a demonstration of differences in auditory sensitivity and Hawks’ data adequately reflects auditory sensitivity in the vowel space, then a likely location to find a magnet effect is within the space for /u/, which yielded the poorest DLs in Hawks’ study. The results of an experiment investigating this possibility through replication of Kuhl’s original protocol in the vowel space for /u/ will be presented.