The central tenet of functionalism is that language is first and foremost a tool for communication between human beings, and that this fact has profound influences on the ways in which language, and individual languages, have developed. Given this, the central aim should be to explicate how people communicate with one another via language. This means that a truly functional theory must be a theory of language, rather than only a theory of grammar, and that it must model the speaker/hearer, and so be about both the patterns revealed by linguistic anlaysis and the processes by which these patterns are put to use in actual communication.

Work so far done within FDG already contains within it the seeds of the proposed expansion. Bakker and Siewierska’s (2004, Bakker 2005) proposals for a speaker model of FDG should be looked at in more detail, with particular reference to possible support from psycholinguistics and cognitive psychology.

Changes would also be needed with respect to discoursal adequacy. Firstly, the dynamic aspect of discourse, according to which speakers formulate their moves and acts in real time, would need to be foregrounded even more than at present. Secondly, in order to account for higher level discourse structuring, it would be necessary to abandon the criterion that only phenomena with an overt reflex in the grammar should be included in our investigations, this being the reason why no unit above the move is currently postulated, and why purely inferential aspects of meaning find no reflection in FDG itself. Thirdly, the relationship between discourse and context would need to be developed in some detail.

Finally, building on the work of Butler (2008a, 2008b) and Connolly (2004, 2007, 2008), we need to develop more detailed accounts of conceptualisation and construal within the conceptual component, and the contextual component needs to be broadened to deal with a wider range of relationships between grammar and context.