This book seeks to point out the difficulties which remain in phonetic theory because of a legacy from prephonemic days when phonetics and phonemics were one. Sounds were shown to have been chosen for description because of their use in speech rather than because of their articulatory or acoustic nature; many sounds were ignored because they were not produced in a certain way. After illustrating this problem with a discussion of marginal sounds, nonspeech sounds, units of sound, structural function, and classification criteria, the book presents a system for phonetic classification which should allow, with slight modification or minor additional categories, for the analysis of all sounds--not only a group determined by non-articulatory means. The productive as well as the controlling mechanisms of speech are also considered. (Author/VM)