This paper addresses Browman and Goldstein’s claim that articulatory phonology is as adequate for capturing lexical contrasts and phonological patterns as it is for modeling articulator movements in speech production. It points out several respects in which articulatory phonology fails to express well-established phonological generalizations, such as the discrete nature of lexical contrasts and the hierarchical organization of features. It suggests than in some of these, at least, articulatory phonology could be revised in the direction of closer integration with phonological models without sacrificing its ability to account for phonetic data. Other problems may go deeper, reflecting Browman and Goldstein’s emphasis upon motor rather than cognitive aspects of phonological organization. However, they have offered strong support for the view that phonological and phonetic representations may be essentially congruent in structure.