ABSTRACT
Words are composed of phonemes, so the first stage of speech recognition is to identify the phonemes in the input. In essence, speech recognition is the reverse process, beginning with the perception of speech sounds and ending up with an idea in the listener’s mind. For the sake of simplicity, it is assumed that the mental lexicon is the same for both speech recognition and production. The function of speech is to convey our ideas to other people through the medium of sound. Speaker and listener need to cooperate with one another in order to achieve effective communication. Language is typically used in a social context and, social activities can only proceed effectively if there is cooperation between speaker and listener. Listeners use the phonemic information derived from the speech stream to identify words. Theories of word identification suggest that, in addition to this bottom-up processing, listeners also use their knowledge of words to help them in the identification process.
