Towards a lexical fuzzy logical model of perception: the time-course of audiovisual speech processing in word identification.