Speech perception abilities are modified by linguistic experience to maximize sensitivity to acoustic contrasts that are important for one's linguistic community, while reducing sensitivity to other acoustic cues. Although some of these changes may be irreversible, in other cases adults may learn to perceive non-native speech sounds in a linguistically meaningful manner with limited perceptual training. The present study investigates the possibility of using a technique based on perceptual fading to train Canadian francophone adults to distinguish the voiced and voiceless "th" sounds of English: Ifj/, as in "the," versus 19/, as in "theta." Following a pretest to measure identification and discrimination performance with both natural and synthetic speech tokens, 10 subjects were trained using synthetic stimuli. Approximately 90 min ofthis training improved performance with both natural and synthetic tokens relative to that of untrained control subjects. The results suggest that there is a much higher degree of plasticity in these acoustic/linguistic categories than would be inferred from the normal performance of Canadian francophones who learn English as adults. The nature of the training technique is discussed in relation to other training paradigms. Linguistic experience produces major and very durable changes in the perception of some speech sounds. For example, English speakers place the phonetic boundary separating Ibl and Ipl at approximately +25 msec voiceonset time (VOT; see Lisker & Abramson, 1970; Williams, 1977, 1979), whereas Spanish speakers place this boundary at approximately 0 to-5 msec VOT (see Williams, 1977). Similarly, speakers of Canadian French place the voicedlvoicelesscategorical boundary at a shorter VOT value than do unilingual English or bilingual FrenchlEnglish speakers (see Carmazza, Yeni-Komshian, Zurif, & Carbone, 1973). One consequence of such subtle language-specific distinctions is that adults who learn a second language often have special difficulty with the perception and production of sounds that are distinct in the new language but allophonic in their native language. The example considered in this paper is one aspect of the voicedlvoiceless distinction between the En-This work was supported by grants from the Alberta Heritage Foundation for Medical Research, the National Health Research Development Program of Health and Welfare Canada, and the Natural Sciences and Engineering Research Council to D.GJ. and by an AHFMR Scholarship to D.E.M. We are grateful to Eleanor Rogers for providing testing facilities in Kingston and for helping to obtain subjects, to Carol McDermid for her assistance in providing subjects at Calgary, and to Fred Wightman and Terry Dolan for their hospitality at the Waisman Center, University of Wisconsin, where this work was completed while D.G.J. was a Visiting Fellow. Special thanks are due to Meg Cheesman, Terry Nearey, Curtis Ponton, and Mike Procter for advice and assistance throughout this project. Requests for reprints should be directed to

We trained unilingual adull Canadian francophone listeners to identify the English voiceless and voiced linguadcntal ("th") frictives, 101 and /&/, using synthetic exemplars of each phoneme. Identification training with feedback improved listeners' abilities to identify both natural and synthetic tokens. These results show that training with appropriately selected prototype stimuli can produce a linguistically meaningful improvement in a listener's ability to identify new, nonnative speech sounds -both natural and synthetic. However, it is not yet clear whether such training with a single prototype can improve performance as effectively as the fading technique used by Jamieson and Morosan (1986), since prototype training produced only a moderate improvement in listeners' identifications of synthetic stimuli containing brief frication. Differences between the techniques may reflect the need for listeners to experience the appropriate types of intraphonemic and intcrphonemic variability during training. Such variability may help to define the category prototype by desensitizing the subjects to differences between exemplars of the same category and by sharpening sensitivity to differences between categories.'The placement test used in this summer was The Ontario Test of English as a Second Language (Carlclon University anil the University of Ottawa). The tesl was piloted during the summer of 1986 and has not yet been published. 2 In the present experiment, the duration of frication varied within categories, representing a cue which could not be used to identify the category to which a lesl or training stimulus belonged. In some circumstances, fricative duration can provide a reliable category cue, however. For example. Cole and Cooper (1975) report that shortening the duration of naturally produced HI. Is/, and III stimuli increases the tendency for listeners to respond hi, Izl, and /dz/, respectively.

No abstract

Repp (1981) demonstrated that training with isolated cues can change how speech information is processed by native speakers. He used brief synthetic noises forming an Is/-Ifl continuum, followed by natural vowels containing formant transitions appropriate for either IfI or lsi. Prior to training, subjects' discrimination of fricative noise differences in syllabic context was poor and was influenced by the nature ofthe formant transitions. Discrimination training with isolated fricative noises did not improve subjects' discriminations of noise-vowel syllables. However, after training with both isolated fricative noises, and with noise-vowel syllables, discrimination of the noises in vowel context was improved for most subjects, and was substantially independent of the following vowel formant transition. 282

A perceptual fading technique [cf. Jamieson and Morosan, Percept. Psychophys. 40, 205–215 (1986)] was used to train two types of listeners to perceive new fricative contrasts. One group consisted of young adult francophones who had difficulty producing and perceiving the voiced/ voiceless “th” distinction in English. With this group, training using the fading technique improved identification accuracy both for the specific, synthetic targets used in training, and for multiple, natural tokens of these sounds, spoken by different individuals, both men and women. However, this learning was quite specific, as performance was poor when listeners were tested with the target sounds in new word positions, or when /d/ tokens were used in the distractor set. The second group of listeners were young children who consistently mispronounced at least one English fricative sound, and had been diagnosed as having a functional articulation disorder. A sizeable proportion of these children, all of whom had been selected on the basis of production difficulties, displayed atypical perceptual skills; for these children, training produced a measureable improvement in identification accuracy, and in some instances, perceptual training (without explicit production training) also improved the productions of the target sounds. [Work supported by NSERC AHFMR and HWC:NHRDP.]