Process and pattern interpretations 167- A minimalistic functionalism?
The question remaining, then, is whether the two perspectives compared in
the previous section are compatible. I would claim that they are actually
not, either as a unitary model embracing both, or as reflecting two separate
modules on the same ‘ontological’ plane. The relationship is – and must
remain – one of complementarity. This is in part because the variable-plus-
restrictor (‘predicate calculus’) format of the pattern grammar has no direct
process correlate – at least not one involving experientially interpretable
meaning. I have proposed equating the processes involved with White-
headian prehensions. These are situated within real contexts, in a world of
entities rather than variables and of complex embedded communicational
purposes rather than discrete digital choices of ‘function’. More impor-
tantly, ‘emics’ cannot be matched directly to ‘etics’. Purposes are the
domain of the latter; the former define the various means at hand to express
them. Given the typological distinction that this chapter has illustrated be-
tween languages with a tighter or looser linkage of lexicon and grammar
(and between languages whose grammar reflects greater or lesser dis-
course-pragmatic vs. lexical control of grammar), one surely cannot expect
there to be a universally valid division of labour between a grammar mod-
ule and a discourse module. What we need is not so much a module or a
grammar of discourse, but a set of (rational) principles allowing us to link
specific communicational intentions to the means provided by the abstract
categories of grammar (cf. Itkonen 1983: 177).
Perhaps Hengeveld’s proposed reformulation of the basic architecture
of FG can be understood in just these terms. Presumably the idiosyncracy
of Nootka, from this perspective, has to do with the relative transparency or
fluidity of the expression level in this language, which matches interper-
sonal level choices much more directly (e.g. ascriptive act decisions as to
what is to be the main predicate) than in more familiar European lan-
guages. However the FG model is extended, reformulated or reinterpreted
in the future, we can at least require of it that it should be able to account
for languages of the Nootkan type in which morphosyntactic indication of
function is rather minimal and where inference from context is constantly
required.^20 If FG is to accommodate signed languages within its overall
framework the same could be said – the pragmatically determined fluidity
with which these articulate the flow of communication into discrete infor-
mation packets is indeed reminiscent of Nootka.
What in effect I have been arguing for in this chapter is a kind of func-
tionalist minimalism, parallel to the Minimalist Program of the generativists