日時： 平成26年1月22日(水) 14:50～16:30
場所： 東京大学本郷キャンパス 文学部3号館6階 言語学演習室
題目： Micro- and Macro-phonetic Processes of Spontaneous Speech
要旨： The phonetic composition of a spontaneous utterance is generally not represented as a string of ‘words’ that are set in their ‘prototypical forms’. Such prototypical forms are usually extrapolations derived from an underlying string which could correspond to words in a ‘mental lexicon’. The ultimate representation of the spoken utterance is the physical sound signal. When the signal is mapped finally to a meaningful utterance through different layers such as phonology, morphophonology, phonetics needs to account for various processes at micro and macro levels. Micro level phonetic processes include sub-segmental gestures such as reduction and approximation that result from ‘undershoot’. Macro level processes involve ‘suprasegmental’ envelope that includes both ‘phonological features’ such as intonation, stress as well as paraphonological features such as emotion, attitude etc. of the speaker. Phonetics as applied to speech technology has to show how the raw speech signal could be correlated with the underlying meaning-carrying representation of the utterance through different higher layers of abstraction.