Further tests of the ‘‘perceptual magnet effect’’ in the perception of [i]: Identification and change/no-change discrimination

Abstract

The present study investigated the existence of a "perceptual magnet" effect [Kuhl, Percept. Psychophys. 50, 93-107 (1991)] in a speech perception experiment. Sixteen subjects experienced in phonetics, transcribed the [i] stimuli used by Kuhl (1991) or identified the stimuli as [i] or not-[i] in experiment 1. In experiments 2 and 3, 50 adults who were not trained in phonetics, participated in an identification task requiring them to decide whether the vowel sounds they heard were the "[i] in 'peep'" or "not the [i] in 'peep'. "They also completed a "change/no change" discrimination task with standards consisting of the "prototype" [i] and other nonprototype stimuli at 67, 75, and 120 mels away from the prototype. Results from the first study showed that listeners did not identify the 120-mel token previously used by Kuhl (1991) as the nonprototype standard stimulus in discrimination experiments as [i]. Sensitivity results from experiments 2 and 3, as measured with d-prime partially supported the "perceptual magnet" hypothesis as suggested by Kuhl (1991) whereas miss results did not. Findings indicated that d-prime, a nonbiased measure of subjects' discrimination ability, better represented the perceptual magnet effect than misses alone. Thus results suggested that subjects were better able to discriminate stimuli surrounding a less "prototypical" [i] vowel as demonstrated by greater sensitivity, faster reaction times for hits and fewer number of misses. However, current results may also be explained by increased auditory resolution rather than prototype-based, "magnet" effects. It was concluded that whereas the perceptual magnet effect may exist for particular vowel categories, more data are needed to better understand the role of prototypes in speech perception.