Thanks Steve for this interesting strand, especially the discussion about
syntax. I wonder if it would be helpful to distinguish two very different
meanings of the slogan 'autonomy of syntax' (which seems, to me at least,
to generate more heat than light).
There are at least two interpretations:
1. Representational Autonomy: syntax is an autonomous 'level' of structure,
distinct from phonology (obviously) and semantics (less obvious);
2. Processing Autonomy: syntactic structure can be determined (by a parsing
mechanism, human or not) without paying attention to meaning. (Presumably
nobody would argue that it could be determined without attention to
phonology!!)
Representational Autonomy is accepted by almost everyone, including most
functionalists (though not all); Cognitive Grammar is one of the few
theories that rejects it. (So I don't think it's helpful to speak as you do
of "the functionalist/autonomous divide".) In this respect CG is different
from other versions of the tradition called Cognitive Linguistics, which
includes not only Construction Grammar (e.g. Fillmore and Kay) but also my
own theory, Word Grammar. In both these theories syntax is very clearly
distinct from other levels. Minimalism takes a complicated position which
paradoxically is rather like CG (I like that - Chomsky and Langacker have
converged): syntax does nothing but bridge the gap between the two
interfaces LF (= semantics) and PF (= phonology) without defining a
distinct (autonomous) structure.
Processing Autonomy, so far as I know, isn't part of any theory of syntax.
Some theories simply ignore processing issues, but those that have anything
at all to say about processing accept that there's a great deal of feedback
from semantics and even context to syntactic choices.
Of course there are plenty of working systems, and theories of NLP, that
espouse Processing Autonomy in the name of statistical parsing: you decide
what the next word might be according to the transition probabilities set
up by the preceding n words. But typically these systems are tied to a
general rejection of syntactic theory. So again, paradoxically, the main
proponents of Processing Autonomy are those who reject Representational
Autonomy.
I gather from your message that at least some RT experts (e.g. Marjolein
Groefsema) reject Processing Autonomy, but that it's still contentious.
Speaking as a syntactician I think Word Grammar is fully compatible with
everything else in RT, and indeed that something like RT is precisely
what's needed in ambiguity-resolution in parsing; and I've often said so in
print.
Dick Hudson
> Interstingly, two of my SIL colleagues, in off-list communication
> (hence anonymity), presented very different views of RT vis-a-vis
> functionalist vs autonomous syntax. One colleague asked me to explain
> why I stated that, "Cognitive Grammar is obviously at odds with RT
> over the scope of grammar, as are various other functional
> approaches." This colleague continued, "When I teach courses in
> linguistic theory I usually start by sketching a continuum between
> "Autonomous theories of language" at one extreme, and "Inclusive
> theories of communication" at the other. Both CG and RT fall close to
> the latter extreme of the continuum (with CG even more extreme than
> RT)." Well, I'm booked to attend a course on CG in July, and I promise
> to keep an open mind. If I realise that I have been mistaken about CG
> I will gladly put the record straight, but if not I will try to sketch
> as precisely as I can the points of difference between RT and CG.
>
> Another SIL colleague sang the praises of Minimalism. He finds it
> extremely useful for practical descriptive analysis prior to getting
> into discourse and pragmatics. On my limited experience of MP I found
> this pretty incredible, but I respect him enough to believe that it is
> so. He also added, "The idea that different languages differ basically
> in parameter settings has given rise to a "new comparative syntax" with
> a substantial literature." Comparative linguistics is a passion of
> mine, so this gave me some encouragement. I'd like to know what this
> substantial literature consists of.
>
> Returning to the functionalist/autonomous divide, a robust challenge
> to autonomous syntax from a RT perspective is provided by Marjolein
> Groefsema's early work. (I hope Marjolein does not mind me ascribing
> early and, by implication, late periods to her work, a la
> Wittgenstein, so long before retirement age!) Thankyou to Begoqa
> Vicente for reminding me of this. The on-line anticipatory hypotheses
> that hearers use make autonomous syntax pretty well redundant. I was
> very impressed with Marjolein's work in this area, and adopted much of
> it in some of my own work in the late 90's. (It had many similarities
> with the work of Hauner, of which I have heard very little of late.) I
> don't know whether Marjolein has continued to develop this approach,
> although I am aware that it was not 'mainstream RT' in the 90's.
Richard (= Dick) Hudson
Phonetics and Linguistics, University College London,
Gower Street, London WC1E 6BT.
+44(0)20 7679 3152; fax +44(0)20 7383 4108;
http://www.phon.ucl.ac.uk/home/dick/home.htm
This archive was generated by hypermail 2b29 : Mon Jun 10 2002 - 10:32:07 GMT