Phonetic Categorisation

Abstract

In phonetic categorisation, listeners hear a range of speech sounds forming a continuum of ambiguous sounds between two unambiguous endpoints (e.g. a voicing continuum between [b] and [p]). Listeners are required to identify the sounds as one or other of the two endpoints. The task has been used widely in phonetics and psycholinguistics to study, for example, categorical perception, selective adaptation, speaking rate normalisation and trading relations. Influences of the lexical status of the endpoints, of word frequency, of sentential context and of phonotactic permissibility have been explored by placing continua in different contexts. Phonetic categorisation has therefore been used in the study of many aspects of speech recognition. A broad selection of references (including several reviews) is given, and the task's many uses and its strengths and weaknesses are summarised.