This dissertation focuses on the emergence of auxiliary do in Middle English (1150- 1500). The development of auxiliary do has been the topic of a large body of studies, but no agreement is found among scholars on the circumstances that led to its rise. The present work combines a qualitative analysis of Old English data with a quantitative investigation of early Middle English poems, and offers a novel account that rests on the principles of Cognitive Linguistics. The development of auxiliary do differs from other instances of grammaticalisation in that it lacks the characteristics that are typically associated with grammaticalisation processes. The findings of my investigation indicate that auxiliary do is the product of the interaction of multiple construction sources with a series of system-internal and system-internal factors. The analysis of the Old English data suggests that some of the different uses of (ge)don (the Old English ancestor of do) anticipate several of the features displayed by auxiliary do in Middle English. I argue that the emergence of auxiliary do was influenced by causative do on the syntactic level, while pro-verb do contributed to its semantic development. The emergence of the auxiliary construction was supported by a number of synchronic factors that characterised the English language in early Middle English, as for instance the changes within the causative verb system and the presence of V - INF constructions in which V had auxiliary-like features. Furthermore, the quantitative investigation of the factors that underlie the use of auxiliary do in early Middle English poems shows that poetry played a role in the establishment and in the spread of the construction. The diachronic account proposed in this study challenges the assumption that auxiliary do developed from a single construction source and provides further evidence on the importance of multiplicity in accounting for certain cases of change.