Dynamic information in the identification and discrimination of vowels

The dynamic theory of vowel perception emphasizes the importance of the consonant transitions in CVC syllables to the identification of the vowel itself. However, the dynamic theory is relatively vague in terms of what perceptual processes may be involved and it says nothing about the role of auditory and/or phonetic memory for such 'dynamic events'. The present study created three synthetic [iota]-[epsilon] continua consisting of a [bVb] full-syllable continuum, a silent-center continuum (which had over 72% of the medial vowel replaced by silence) and an isolated vowel continuum to specifically examine (1) the use of auditory and phonetic memory for dynamic acoustic cues and (2) the ability of listeners to track the trajectories of very brief formant transitions. Experiment 1 demonstrated that there were no significant differences among the three continua in terms of listener identifications but that the silent-center tokens demonstrated significantly lower within-category discriminations, perhaps because of degraded auditory representation. Experiment 2 required listeners to make cross-continuum vowel discriminations and showed that some degree of acoustic similarity was important--listeners were particularly poor at discriminating between the silent-center and vowel-only tokens. This suggests that listeners were not able to make discriminations on the basis of abstract vowel labels alone. Experiment 3 demonstrated that listeners could make quite accurate vowel identifications even when as little as 1 pitch period of acoustic energy signaled the consonant transitions in the silent-center tokens. Experiment 4 showed that listeners' identifications of the silent-center stimuli were based on the formant changes in the transitions rather than their endpoint frequencies.