Functional Discourse Grammar and language
production
J. Lachlan Mackenzie
- Introduction
One of the central requirements placed on a Functional Grammar is that
“... such a grammar must also aim at psychological adequacy, in the sense
that it must relate as closely as possible to psychological models of linguis-
tic competence and linguistic behaviour’ (Dik 1997a: 13). The FG model is
presented in a quasi-productive mode (Dik 1997a: 57), but at the same time
sees itself as essentially generative – indeed the primary purpose of the
computer implementation of FG (Dik 1992) was to test and enhance its
generativity. Recognizing that to confuse generation with production is a
category error, FG has insisted that it is not a model of language produc-
tion. Yet the demands of psychological adequacy have encouraged
speculation on the relation between the model, in its various manifesta-
tions, and the ever more sophisticated models of language production
currently available. The purpose of this chapter is to examine the potential
of Functional Discourse Grammar for a further rapprochement between FG
and a psycholinguistic consensus on language production.
Jackendoff (1997: 7–8) has argued that there are three possible positions
on the relation between grammar and the processes of speech production
and perception:
(a) one can deny any relationship, insulating grammar from psycholin-
guistic findings (the traditional generativist position);
(b) one can maintain that processing mechanisms can ‘consult’ or ‘in-
voke’ a declarative grammar;