Читать книгу The Handbook of Speech Perception - Группа авторов - Страница 77

In linguistic theory, features are binary; that is, they are either present (+) or absent (−). For example, the stop consonant [d] would be marked as [−nasal] to contrast it from the nasal consonant [n] which is marked [+nasal].The preceding section on “Feature dimensions” in speech perception and lexical access is consistent with this view. However, it has long been known that membership in phonetic categories is not all‐or‐none. Rather, there is a range of values associated with parameters or features of speech with some values appearing to be more representative than others. On the face of it, it would suggest then that, in speech perception, features may be graded. What is critical is whether such graded representations influence not only speech‐perception processes but also lexical access processes. We now turn to this question.

As we shall see, experimental evidence challenges the view that features are binary and rather supports the view that features are graded. This has consequences not only for speech‐perception processes but also for the mapping from phonetic category representations to lexical representations and for ultimately accessing the lexical semantic network.