180 J. Lachlan Mackenzie
(c) or one can claim that the processor embodies the grammar, i.e. that
grammar is itself procedural.
I believe that there are certain developments in current FG that suggest
that greater psychological adequacy is achievable through a combination of
Jackendoff’s positions 2 and 3. More specifically, I will explore the possi-
bility that Hengeveld’s (this volume) interactional and expression levels
can be modelled as procedural, with the intermediate (and in some cases
by-passed) representational level functioning as a declarative module that
is consulted by the processing mechanisms.
- FG as a declarative grammar
A central tenet of the functionalist stance is that linguistic form results
from a complex of choices. For example, there is in certain languages a
choice whether or not to apply the Subject function to a first argument,
which is reflected in English in the active vs. passive voice (Dik 1997a:
248–250). The choice would appear to be the speaker’s. However, this
raises an apparent problem for FG, and other functionalist approaches that
stress ‘choice’: psycholinguistic findings strongly suggest that most of
these choices are simply not accessible to the language-user’s awareness.
As Levelt (1989: 21) puts it, “A speaker doesn’t have to ponder the issue of
whether to make the recipient of GIVE an indirect object (as in John gave
Mary the book) or an oblique object (as in John gave the book to Mary)”.
This is known as ‘cognitive impenetrability’: many aspects of language
processing are automatized, and not accessible to choice.
The conclusion must be that, when FG invokes psycholinguistic expla-
nations, it is trying to make understandable why the automatic processes of
language are as they are. This applies as much to the hearer’s processing of
utterances as to the speaker’s production of them. To take an example, the
explanation offered for LIPOC is that “[i]t is easier [sc. for the hearer,
JLM] to perceive, process, and store complex information when this infor-
mation is presented in chunks of increasing internal complexity” (Dik
1978: 212). We must beware, however, of assuming that the speaker is ac-
tually choosing to alleviate the hearer’s interpretive task: language users do
not consciously apply linguistic principles in their effort to express them-
selves. Surely the explanation offered by Dik must be understood as
historical and selectionist: of all possible orderings of information, the one
that has sedimented into automatic processing is that which, everything