ABSTRACT
In this chapter, the authors return to the theoretical issues raised in the Introduction of R&R. Our goal here is to elucidate these as far as possible, in the light of the analyses of the various cases of gram-maticalisation analysed in the chapter. The crucial property of functional categories is that they are only fully defined in relation to the syntactic system; in all other respects they are defective. The authors list all the empirical cases discussed in Roberts & Roussou, by providing a schematic representation of the relevant structural changes. The authors then identify the parametric change and, where possible, the cause of the change as well. The study of grammaticalisation raises the familiar tension between descriptive and explanatory adequacy in the diachronic domain. A descriptively adequate account of this class of changes results in defining pathways of change.
