This book proposes a new way of understanding the behavior of consonants and vowels in a broad cross-section of the world’s languages. A new model of subsegmental phonology within optimality theory that differs from standard autosegmental phonology both in its limited use of representational distinctions and in the form of the grammar to which the representations submit is introduced. The research focuses particularly on floating features and ghost segments, and demonstrates that the current understanding of segmental representation fails to characterize the full range of subsegmental phenomena found cross-linguistically. Zoll proposes instead an analysis in which the grammar derives the variety of surface phenomena from a single underlying representation. This work both enlarges the empirical foundation on which an adequate theory of segment structure must be based, and in developing such an account sheds new light on classic problems of subsegmental parsing.