This book describes the most comprehensive system yet developed for the automatic conversion of English text to intelligible and natural sounding synthetic speech. It offers detailed accounts of the various components any speech technologist needs to consider - algorithms for morphological analysis, letter-to-sound rules, syntactic analysis, lexical stress, timing and pitch, together with segmental synthesis. The MIT text-to-speech system, on which Jonathan Allen has collaborated with M. Sharon Hunnicutt, Dennis Klatt and other colleagues, has enhanced the standards for intelligibility, linguistic sophistication and methods of evaluation. This study will be an invaluable resource not only for professionals in the field, but for any reader with an informed interest in natural language processing.