214 María de los Ángeles Gómez-González
constructions, apparently in conformity with evidence reported from lan-
guage processing, language acquisition, aphasia and categorization (Bates
and Goodman 1997).^3 As a corollary, there would be no need to posit un-
derlying structures or derivations, nor (expression) rules to link the
different components.
Different substantiations of this kind of non-hierarchical non-modular
framework are subscribed to by the exponents of WCF. Roughly, these ap-
proaches, which are, as has been remarked, in many respects analogous to
FDG, (a) defend a materialist model of perception with an empiricist func-
tionalist and universalist orientation, and (b) postulate real-time dynamic
programmes that discard the division between a core grammar and a pe-
riphery grammar, under the assumption that all constructions, whether
relatively recurrent or highly idiosyncratic, stand on an equal footing for
description, resorting to, among other things, (c) non-discrete categories or
prototype theories, (d) monostratal or ‘surface level’ syntactic descriptions,
and (e) making no sanctioned distinction between grammatical knowledge,
knowledge of language use and other sorts of knowledge under the as-
sumption that “knowledge of language is knowledge” (Goldberg 1995;
Langacker 1987, 1991, 2001b; Noonan 1999, Kay and Fillmore forthc.).
This latter stance (e) apparently flies in the face of Hengeveld’s outline
of FDG (this volume), in which Cognition and the Communicative Context
are presented at the sides of the schema as two separate, though connected
modules. This representation could lead to the assumption – counterin-
tuively in my view – that Cognition (i.e. (long-term) knowledge) and
communicative context (i.e. (short-term) linguistic co(n)textual information)
can be dissociated or can operate independently of each other. Instead, we
have every reason to suspect that in encoding and decoding strategies the
communicative context is filtered through and interpreted by the speaker’s
cognitive system: the latter encompasses the former as it were, both with a
certain amount of independence but nonetheless coordinated, the underlying
rationale being that processing proceeds simultaneously on multiple time
scales, and with respect to numerous parameters. Moreover, it seems that one-
way arrows running from and towards Cognition, the Communicative Con-
text and the three levels of grammar in different guises infuse the model
with a rigid one-way directionality that impoverishes the supposedly inter-
actionist (bi-directional or multidimensional) dynamicity that the model
strives to confer to the cognition-discourse-grammar relationship. In sum, a
modification of this design is here called for to better symbolize
Hengeveld’s claim that “the three levels interact with a cognitive component
and with a communicative component” (this volume: 3).