Objective: The scientific study of the perception of spoken language has been an exciting, prolific, and productive area of research for more than 50 yr. We have learned much about infants' and adults' remarkable capacities for perceiving and understanding the sounds of their language, as evidenced by our increasingly sophisticated theories of acquisition, process, and representation. We present a selective, but we hope, representative review of the past half century of research on speech perception, paying particular attention to the historical and theoretical contexts within which this research was conducted. Our foci in this review fall on three principle topics: early work on the discrimination and categorization of speech sounds, more recent efforts to understand the processes and representations that subserve spoken word recognition, and research on how infants acquire the capacity to perceive their native language. Our intent is to provide the reader a sense of the progress our field has experienced over the last half century in understanding the human's extraordinary capacity for the perception of spoken language.
The Beginnings of Speech Perception Research: Some Core IssuesResearch on the perception of speech began in earnest during the 1950s. Not surprisingly, the agenda for the field was set by the kinds of issues that preoccupied language researchers of that era. In the field of linguistics, a major goal was to devise rigorous scientific procedures that would yield a correct structural description of a particular language when provided with a detailed corpus of utterances (Bloomfield, 1933;Harris, 1955). This approach, known as taxonomic linguistics, made certain assumptions about how such a description should proceed. The general view was that a language is hierarchically organized at a number of distinctive levels, so that an accurate description of language structure required descriptions of the organization at each level of the hierarchy. Moreover, because the aim was to achieve a scientific description, the structural analysis of a particular language began with what could be directly observed, namely, the acoustic waveforms of utterances. Because of this emphasis on what could be objectively observed, the description provided for a given level was not supposed to depend on the description of any of the higher levels. In principle, then, a description of a language would begin with an account of how acoustic properties map onto phonetic segments (the phonetic level), how the phonetic segments mapped to particular phonemes (the phonemic level), how the phonemes are combined to form morphemes in the language (morphemic level), and eventually to how the morphemes are combined to form sentences (the syntactic level). Thus, although speech perception research is very much an interdisciplinary field, drawing on concepts and methods from physics, engineering, linguistics, and psychology, many of the core problems that drove the early research efforts reflected the view that languages are hierarchical...