Evidence from the Romance languages has played a central role in linguistic debates regarding the relation between structure and change. One of the issues that has attracted most attention has been the connection between periphrasis and inflexion: Latin had rich systems of verbal and nominal inflexion that interacted in various ways with existing periphrastic patterns and structures that incorporated the ingredients of incipient Romance periphrases. These structures are of interest not only to those whose principal focus lies in exploring and explaining the development of the Romance languages, but also to those whose concern is with the broader theoretical implications for our understanding of language structure, variation, and change. This chapter sets out the issues, both theoretical — what models are most appropriate to describe and explain the Romance data? — and empirical — what sources of data are available and what issues arise in exploiting them? We begin with the criteria that serve to distinguish analytic patterns in general before going on to examine the specific properties of periphrastic constructions. From there we consider the interface between periphrasis and inflexion and the different ways they are modelled in current theoretical work. From synchrony we move to diachrony and the role of grammaticalization in the genesis of new periphrases and inflexions, and interaction on the one hand between internally and externally motivated change and on the other between reconstructed and attested patterns. We conclude with some general reflections on the relation between synchrony and diachrony.