Becoming A Native Listener

Abstract

Xhe syllables, words, and sentences used in all human languages are formed from a set of speech sounds called phones. Only a subset of the phones is used in any particular language. Adults can easily perceive the dif ferences among the phones used to contrast meaning in their own language, but young infants go much farther: they are able to discriminate nearly every phonetic con trast on which they have been tested, including those they have never before heard. Our research has shown that this broad-based sensitivity declines by the time a baby is one year old. This phenome_ non provides a way to describe basic abilities in the young infant and ex plore the effects of experience on hu man speech perception. To put infants' abilities in per spective, adult speech perception must be understood. The phones that distinguish meaning in a partic ular language are called phonemes. There is considerable acoustic variability in the way each individual phoneme is realized in speech. For example, the phoneme Ibi is very different before the vowel led in beet from the way it is before the vowel /oo/ in boot. How do adults handle this variability? As first demon strated in a classic study by Liberman and his colleagues (1967), they treat these acoustically distinct instances of a single phoneme as equivalent. This equivalency is dem onstrated in the laboratory by presenting listeners with a series of pairs of computer-synthesized speech stimuli that differ by only one acoustic step along a physical continuum and asking them first to label and then to try to discriminate between the stimuli. Adult listeners are