syntax and form-meaning pairs (=signs)

[harder at COCO.IHI.KU.DK]

This is not the first time that a discussion about syntax in relation to
cognitive linguistics creates problems of understanding, in spite of the
fact that neither functionalists nor cognitive linguists believe in
autonomous syntax, and both believe in the existence of grammar. In order to
make it clear that one can want to talk about syntax and still be dealing
with semantic phenomena I have suggested the term 'content syntax' for those
combinatorial relations that create larger meanings out of component
meanings (e.g. the head-modifier relation, etc). When one is talking about
content syntax, the issue is neither individual form-meaning pairs nor the
chimera of autonomous syntax.
        This is important in relation to a point made in George Lakoff's
second message. As I understand it, it seems to imply that if one agrees
that there syntax can be described in terms of form-meaning pairs, there is
no need to talk about the special properties of syntax. But the ability to
combine meaning fragments into larger wholes does, from an evolutionary as
well as a neurological point of view, seem to be rather a special skill.
Saying that it is distributed over the brain and that it depends on
connections rather than solely on a specific brain area does not appear to
capture its special nature very precisely. The best way to show the
superiority of a non-automous approach to syntax must be to show how the
special nature of the ability to create complex expressions can be captured
in a framework where the combimation of meanings is the essential part. But
this requires recognizing that the combinatory skill has properties that are
different from the ability to associate form and meaning in a holophrastic
sign. Mechanisms of combination do not disappear as a special problem in its
own right, even if it is non-autonomous and involves form-meaning pairing.
                                                        Peter Harder, U of
Copenhagen