A New Architecture for Functional Grammar (Functional Grammar Series)

48 Matthew P. Anstey

b. The grumpy Mr Smith went home.

interpersonal: Decl (RFoc: ‘Mr Smith’: ‘grumpy’)

representational: (dsx: Mr Smith: grumpy)

expression: (NP: /ðə grmpi mstə smθ/)

c. A grumpy Mr Smith went home.

interpersonal: Decl (R: ‘Mr Smith’: ‘grumpy’)

representational: (dsx: Mr Smith: grumpy)

expression: (NP: /ə grmpi mstə smθ/)

This representation has several notable features. Firstly, the expression
layer indicates the referential unity of the three utterances: in each the
Speaker refers to Mr Smith. Secondly, the interpersonal layer can represent
grammaticalized pragmatic differences if need be. Thus for sentence (b) we
could argue that the speaker intends both to refer and to draw attention to
what he is referring to, written formally as RFoc. Thirdly, the semantic, or
intrinsic definiteness of ‘Mr Smith’ is shown in all three utterances in the
representational layer, avoiding the dilemma that A grumpy Mr Smith...
poses for FG1-3. Fourthly, the variation in syntactic structure of the three
utterances is visible in the expression layer. So for the sake of argument,
suppose that linguists agreed that (c) was simply a variation of (a) contain-
ing no semantic or pragmatic differences. In this hypothetical case the
language provides the Speaker a choice between multiple syntactic realiza-
tions for the same pre-syntactic input.
The point of this brief analysis is obviously to demonstrate that FG 4
provides a framework to tackle such linguistic data, a framework I would
suggest is eminently superior to FG1-3. The linguistic product emerging
from the complex interplay between pragmatics, semantics, and syntax can
now be more adequately described. More importantly, coupled with vari-
ous linguistic, philosophical, and psycholinguistic arguments, such data, it
is hoped, can be convincingly explained.
In other words, Hengeveld’s contribution in FG 4 is to introduce a tripar-
tite formal notation corresponding to the tripartite functional hierarchy of
influence: pragmatics > semantics > syntax becomes instantiated as inter-
personal layer > representational layer > expression layer. It is my view
that this change alone constitutes the most important development for
Functional Grammar, for without it the theory can not unravel the Gordian
knot that has arisen from the well-intended but ultimately self-defeating
conflation of pragmatic, semantic, and syntactic information onto the
monolinear URs of FG1-3.