FDG and language production 189language with a so-called ‘free word order’? By this we understand that the
order of constituents is much more strongly influenced by their cognitive
status as New, Given, Inferred, etc. For such languages the temporal order in
which the concepts become available may be much more directly reflected in
word order, with the declarative grammar serving to ensure above all that the
semantic relations between the constituents remain transparent to the ad-
dressee. In such languages we might expect the Focus to be expressed first,
in line with its cognitive priority; an example of this situation is discussed in
depth by Fortescue (this volume). Yet even in language with relatively ‘free’
word order, there is often one syntactic position, frequently not P1, which is
reserved for the Focused constituent. In Turkish (Van Schaaik 2001: 45) this
is the immediately preverbal position; in Hungarian (De Groot 1989: 105) it
is the position between Topic and Verb.
The non-initial positioning of Focus in actual utterances, as against its
cognitive priority, may also be understandable from a processing viewpoint.
The cognitive identification of a Focus concept, which is typically new in-
formation, and the selection of an appropriate lexicalization of that concept,
both require time. By contrast, associated information that is already acti-
vated and thus highly accessible (i.e. Topics and Settings) is immediately
available for expression. The speaker who expresses accessible information
first allows herself more time for the identification and lexicalization of the
Focus. Hannay (1991) has shown that even for a syntactically rather rigid
language like English, the relative positioning of Topic and Focus is depend-
ent upon what he calls the ‘message mode’: in an urgent situation, the
Reaction mode encourages initial placement of the Focus, with optional
back-up from a following Topic; where there is a highly accessible Topic,
the Topic mode induces a postponement of the Focus to a later syntactic po-
sition. It may indeed be advisable to add a specification of the message mode
to the representation of utterances in the interactional component.
One consequence to be drawn from the preceding discussion is that the
expression component is no longer, as it is in orthodox FG, the only place
where the order of constituents is determined. Its function now emerges as
that of balancing the competing demands of the interactional component,
which ‘wishes’ to see each component of its emerging message expressed
as soon as possible, of the representational component (as in traditional
FG), which requires an unequivocal reflection of its demands, and of the
expression component itself, which has its own language-specific rules and
regulations. These are the ‘simultaneous equations’ to which Levelt (1999:
95) refers.