Discourse Structure, the Generalized Parallelism
Hypothesis and the Architecture of Functional
Grammar
Ahmed Moutaouakil
- Introduction
One of the most extensively discussed issues in the recent FG literature is
the way in which the current sentence-oriented model can be extended to
handle supra-sentential phenomena. In order to come to grips with this
problem, two kinds of solutions have been proposed: an ‘extrinsic’ solution
consisting in combining FG with other theories (e.g. Gulla 1997) and an
‘intrinsic’ solution intended to achieve the extension in question by inter-
nally transforming the FG apparatus itself. Within the latter line of thought,
two approaches can be distinguished: an expanding ‘upward layering’ ap-
proach (Hengeveld 1997; Moutaouakil 1998, among others) involving a
continuum of successively larger units, and a modular approach (Kroon
1997) according to which sentential and supra-sentential (discourse) phe-
nomena are to be dealt with in two separate modules.
Hengeveld (this volume) proposes a model of Functional Discourse
Grammar (FDG) which integrates these two approaches. My aim in this
study is to further argue for Hengeveld’s claim and show that the two ap-
proaches in question should indeed not be considered incompatible. In the
light of this assumption and in the perspective of a new architecture for FG,
I will suggest a general frame for describing and explaining discourse phe-
nomena which is both hierarchical and modular. This frame differs in
certain substantial as well as organizational aspects from Hengeveld’s
model of FDG but it goes in the same direction and strives to achieve the
same theoretical goal, simplification and unification of the theory of FG.
Elaborating on the idea put forward in Dik (1997b) that the hierarchical
structure and the functional relations postulated for the clause can be pro-