Bob wrote:

 Newmeyer in Language
>form and language function, p 50-1 (1998) observes that wh-phrases are
>"displaced from its subcategorized position and fronted" in at least
>four different kinds of constructions which differ functionally.
>
>        Questions
>                Who did you see?
>        Relative clauses
>                the woman who I saw
>        Free relatives
>                I'll buy what(ever) you are selling
>        Wh (pseudo) clefts
>                What John lost was his keys.
>
>Why should that be the case if "no language system exists independently
>of its uses"?  Despite the very different functional uses of these four
>constructions, Newmeyer argues they all have the same underlying
>abstract structural property.

Halliday deliberately set out to describe language in a useful way. He calls
what you refer to above "grammatical metaphor"; it is one very prevalent
phenomenon in expository text (particularly nominalizations). One of H's
early arguments was that, before you can ask if a description of language is
good, you have to ask, good for what?

>I am sorry that I did not make clear what I tried to show with the
>constraint on affirmative preposing.
>
>> I miss your point entirely on the second example.
>
>        1) Last night, John meant to insult his aunt.  And, insult her he did.
>        2) Last night, John meant to insult his aunt.  ?And, offend her he did.
>
>There is a constraint on what can be preposed.  It MUST have a
>morphological similarity with the proposition that is being affirmed.
>(1) is good: (2) is decidedly strange.  If USE explains language
>structure, why should such a formal constraint?

I disagree with your point. "2" sounds quite possible to me.

>Let us be clear what is at stake here for the teaching of grammar.  If
>Judy is right that "no language system exists independently of its
>uses," then all of the suggestions offered by Rei Noguchi to show
>students that they can recognize grammatical subjects and independent
>clauses are wrong.  The tag question test and the yes/no question test
>work because ALL clauses in English, regardless of the function
>(ideational, interpersonal, textual) those clauses perform, have the
>same underlying abstract structural property.

Bob, you clearly do not agree with functional grammar. I find it an endless
source of insight into language on multiple dimensions. It's not on my
agenda to convince you or anyone else. Regardless of what we come "wired"
with, a language system develops, phylogenetically and ontogenetically,
through its uses.

Judy



Judith Diamondstone  (732) 932-7496  Ext. 352
Graduate School of Education
Rutgers, the State University of New Jersey
10 Seminary Place
New Brunswick, NJ 08901-1183