Vowel Identification and Phonetic Contexts

Confusions in the perceptual identification of vowel samples taken out of various phonetic environments have been examined. Japanese words were uttered by a female speaker. The central portions of vowels within the words were electronically gated out for a 50‐msec duration, including tapered ends. For each of five Japanese vowels, fifteen samples out of different environments were prepared, and 75 samples in total were arranged in randomized orders for forced‐judgment‐identification tests. The result showed a considerable amount of confusion for some vowels uttered in certain phonetic environments. Some appreciable effects upon the score were also observed when the subjects had been trained in different ways immediately preceding the listening tests. The result, in general, conforms with explanations in terms of articulatory movements, previously reported formant data, and some other factors that would influence the subjects' judgments.