A reply to ‘ ‘ Innate phonetic boundaries revisited ’ ’ [ J . Acoust .

Before launching into a response to the letter by Aslin, Werker, and Morgan ~2002!, I would like to make clear my belief that the work on infant speech perception has been well executed, and has been critical to the development of general theoretical perspectives on speech perception. I respect and appreciate the contributions of these authors, and consider my own work with children to be a natural extension of their and others’ work on infant speech perception. At the same time, the questions starting to be asked by speech researchers who are not focused solely on infant perception reflect a dramatic change in theoretical perspective, one that is not matched in magnitude by changes in perspective in the infant speech perception work. Many investigators in other areas of speech perception are no longer concerned with trying to identify the acoustic correlates of phonetic categories, or with asking at what age human listeners are able to classify those correlates, or with asking if nonhuman animals can classify those correlates. Nor are other investigators much concerned with distinctions between ‘‘auditory’’ and ‘‘phonetic’’ perception, a recurring theme in the letter of Aslin et al. A general paradigm shift is underway in studies of speech perception toward a focus on how listeners perceptually organize the many properties of the speech signal. For example, Robert Remez and his colleagues ~e.g., Remez et al., 1994! and Robert Shannon and his colleagues ~e.g., Shannon et al., 1995! have convincingly demonstrated that mature listeners can organize signals that are void of the detailed properties that we usually consider acoustic correlates ~or ‘‘cues’’! of phonetic categories in such a way that they still manage to arrive at a phonetically structured message. Investigators studying cross-linguistic speech perception are grappling with fundamental notions of what the objects of speech perception are, and what that means for our understanding of how the perceptual organization of the speech signal is shaped by experiences with a native language ~Strange, 1995!. What we consider to be the initial state of the human capacity for speech perception affects the further development of theories related to these topics, as well as many others. The purpose of the brief article serving as the target of Aslin et al.’s letter was to state explicitly that the evidence is weak to support the position that infants are born able to discriminate all the phonemes of the world’s languages. The authors say it best themselves on page 1259 of their letter, ‘‘Thus, at best, the infant literature supports a weak view of