Identification of vowels spoken in isolation versus vowels spoken in consonantal context

Strange et al. [J. Acoust. Soc. Am. 60, 213--224 (1976)] report that naive listeners misidentified approximately three times as many tokens of vowels spoken in isolation as the corresponding vowels spoken in consonantal context. On the basis of these findings, they argue that the superior intelligibility of vowels in consonantal context is due to coarticulatory effects: "acoustic information distributed over the temporal course of the syllable is utilized regularly by the listener to identify vowels" (p. 213). The present study tested naive listeners' identification of eleven American English vowels spoken in isolation and in consonantal context with an experimental design comparable to that used by Strange et al. Here, however, listening tests were administered under high quality listening conditions, speakers and listeners were closely matched for regional dialect, and problems with the response alternatives for the vowels were minimized by having listeners identify the isolated vowels and /tVt/ syllables by rhyming them with English words. The results of the tests indicated that isolated vowels could be identified quite well; listeners misidentified only 2% of the vowels when tokens were blocked by speaker and only 8% when tokens were randomized across speakers. Further the tests did not reveal any overall difference in identifiability between the isolated vowels and vowels in consonantal context.