This article is an attempt to account for data, collected from twenty speakers of Tyneside English (TE), which seem to indicate that syllabification of the oral stops in certain positions in English is subject to dialectal variation, and that the issue of syllabification is closely linked to patterns of glottal reinforcement and glottal substitution in the TE variety. Models of syllable structure based on data from ‘standard’ accents such as Received Pronunciation (RP) therefore fail to account for patterns in nonstandard varieties. In what follows, I propose some revisions to the RP-based account provided in Giegerich (1999) which are necessary to account for the TE data and, furthermore, I emphasize the need for a more holistic approach to studies of variation, which attempts to synthesize structural and sociolinguistic principles.