ABSTRACT

The following chapter provides a broad survey of theories of speech perception that each attempt to explain how human listeners interpret acoustic sounds as a sequence of linguistic units such as words or even more basic speech sounds such as phonemes. These theories also have to explain several phenomena that have been observed in speech-perception research. For example, a given speech sound may have radically different acoustic characteristics when produced by different speakers or when produced in different contexts. In addition the same speech sound can be identified on the basis of multiple acoustic cues, any of which may also be used to identify other speech sounds in the same utterance. Therefore, the temporal extent of individual speech sounds can be widely overlapping in time. Theories of speech perception approach these problems in different ways. This chapter groups these theories on the basis of how the acoustic speech signal is mapped onto linguistic units. For example, many theories propose that the listener interprets sound with reference to the physical articulations that were made by the speaker. Others suggest that this is unnecessary and that auditory/acoustic properties can be mapped onto linguistic units directly. Intermediate theories also exist. Several other phenomena that support one theory or another are also described.