The absence of I amn't for the first person singular present tense negative form is taken to indicate that there is a gap in the paradigm. Recent accounts take a morphosyntactic approach and phonology is largely ignored. Such accounts typically focus on contemporary forms of Standard English. This paper, in contrast, compares nineteenth-century and contemporary West Yorkshire (WY) aux+n't forms and pursues a largely phonological solution. The paper sets out to demonstrate that WY has never had a *amn't gap and that changes over the past century shed light on the *amn't gap problem. Contemporary WY is known to exhibit a phenomenon called secondary contraction, whereby shouldn't [ʃʊdʔ], for example, may be pronounced [ʃunʔ]. I argue that secondary contraction is responsible for the creation of homophones for amn't and aren't: [aːnt]/[aːt]. I will consider the possibility that certain aux+n't forms have become lexicalised and that this has triggered secondary contraction as a phonological repair strategy. With the analysis of WY data as a backdrop, the paper then pursues the possibility that lexicalisation may have occurred, at a much earlier date, in precursors of Standard British English (SBE). Indeed, it seems plausible that homophony for amn't and aren't may have led to prescription against new realisations of amn't. The paper will show that grammatically amn't has evolved in exactly the same way as other auxn't forms, and it is only commentators who have treated it differently. If this is so, the *amn't gap in SBE is man-made rather than grammatical in nature.