ABSTRACTPhonological analysis in speech therapy clinics has traditionally been carried out by examining consonants in three word positions: initial, medial and final. More recent techniques, however, suggest a two‐pronged approach in which the occurrence of consonants is studied with regard to their position both in whole words and in the syllables making up these words. This study looks at the approach taken by three current procedures for phonological analysis to the traditional ‘word‐medial’ consonant or consonant sequence. The theoretical basis of one of these approaches is then examined, and phonological data from a normally developing 3‐year‐old is used to assess its descriptive validity. It is concluded that, for the child described, a sequence of two or more consonants in word‐medial position can be divided into syllable‐final and syllable‐initial consonants in a principled manner. When only a single consonant occurs in this position, the decision is more difficult and these consonants are described simply as intervocalic. It is suggested, however, that the correct approach to medial consonants and consonant sequences may vary from child to child, and that this has implications for the management of phonologically di