[Corpora-List] Is a complete grammar possible (beyond thecorpus itself)?

[Rob Freeman]

On 9/11/07, Oliver Mason <O.Mason at bham.ac.uk> wrote:
>
> On 11/09/2007, Rob Freeman <lists at chaoticlanguage.com> wrote:
> > The effect of "corpus linguistics" on the way we do syntax has been nil
> (for
> > lexicon it has been a revolution, but for syntax, nothing.)
> >
> > There is a vague idea we have to merge "lexical" and "syntactic" aspects
> of
> > text, but no-one has a clue how to do that.
>
> This is obviously not true.  In 1995 David Brazil published "A Grammar
> of Speech", which was based on authentic corpus data.
> Non-generativist description driven by a corpus.
>
> Then there is Pattern Grammar (Hunston and Francis 2000).  Also
> non-generativist, and also derived from corpus data.
>
> And finally, Linear Unit Grammar (Sinclair and Mauranen 2006).  They
> don't even use the traditional word-classes.
>

I agree people have made attempts, Oliver. I just don't think these work.
Most strike me as attempts to explain syntax _as_ lexicon. Which is brave,
but ultimately just as flawed as "all syntax" approach of the generativists.
The very small space given to syntax in cognitive style grammars strikes me
as falling into this category, "frames" etc.

Hunston and Francis appear to specifically use old ideas of grammatical
class like "noun" and "verb" to generalize their patterns. I've seen several
attempts like this, and it is clearly inadequate (though highlighting a
need, not to undervalue it.)

However, you've just drawn my attention to LUG, and I confess I know very
little about it. Is there an on-line reference I can look at, or can you
sketch how it works?

Likewise for David Brazil's "Grammar of Speech". How does he generalize
"linearly"?

-Rob
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20070911/d5309f6e/attachment.htm>
-------------- next part --------------
_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora