Whilst many Middle English romances were derived from Anglo-Norman texts in the form of chanson de geste, there is a long established view that there was no such thing as an Anglo-Norman chanson de geste. The Anglo-Norman versions of Horn and Boeve de Haumtone, both described by Dominica Legge as ‘romances in chanson de geste form’, are usually discussed with the other so-called ‘ancestral’ romances such as Gui de Warewic, Waldef, Havelok and Fouke Fitz Warin. Anglo-Norman redactions of continental chansons de geste have been labelled ‘chansons de geste’ but these are generally seen as ‘mere adaptations’ and given little attention as independent texts even when (as in the case of the Anglo-Norman Fierenbras) the reworking has been considerable.

This discrepancy raises questions regarding the definition of medieval genres. Are our genre labels simply modern critical constructs imposed on medieval texts? Critics have been asking this for nearly half a century. Does it matter if we ascribe different generic labels to these texts? The basic premise behind genre classification is the same as that behind other intertextual analysis: a clarification of what Northrop Frye called ‘traditions and affinities’. The perceived traditions and affinities will determine the horizon of expectations of the audience and an analysis of how the poets manipulate these expectations will bring us to a different appreciation of the texts.