At 16:34 10.03.1999 -0700, Rebecca S. Wheeler wrote:

>I guess I'm concerned about calling NP subjects, NP Direct objects, and all
>that stuff 'complements', even if it is traditionally done so. In that the
>term 'complement' has become more specifically relativized to the final
>term in "paint the barn teal" or "We made Jami president".

I for one am concerned about calling each and every structure we find in
the 'subject' slot of a sentence a noun phrase, regardless of what the
structure is like (single noun, noun group or noun phrase,that-clause,
what-sentence, chain, and what not).

It's like saying that there are plants outside my window, there is a plant
directly in front, there are plants on the ground, and several plants that
take away the sun. The purpose of science is to distinguish between things,
not to put them all in the same box.



>Back on the valence type approach, I'm accustomed to referring to a verb as
an n-ary argument structure -- that is, a verb takes 1, 2, 3 arguments. And
that these arguments (1 NP subject argument for intransitive verbs -- like
in "Dakota sleeps"; 1 NP subject and 1 NP object as in garden variety
transitives -- "Jami planted 400 tulips"; or 3 arguments as in "Jami
painted the door aubergine".

Ok, once Tesniere had created the valence model, others jumped on the
waggon (or maybe were already there). Anyway, it seems to me that the n-ary
model is a step backwards (apart from the fact that it misses the metaphor
that school students find quite helpful). The valence model describes a
three-fold influence of the verb on the structure of sentences, not just
one, namely
        1. number of complemental parts (1,2,3)
        2. types of parts (subject, etc)
        3. structures allowed in the various slots (e.g. ing-form sentence in
object slot, but not to-inf sentence)

The 'traditional' distinction between sentence parts should not be chucked
out without thinking twice, they have bearing on semantic and pragmatic
problems. Of course, if a linguistic model puts these problems in a
wastepaper basket called 'performance' ... But then it forgoes the
description of the lifeblood of language. A linguistic model that does not
deal with the content side and not with the usage side of language, is
irrelevant as far as the  school classroom is concerned, in particular the
foreign language classroom. (Don't get me wrong, I'm not saying it is
irrelevant as such. Whether a model is usable or not invariably depends on
the purpose it is to be used for. There is no best or most adequate model
in general, as the Chomsky people used to think. Any model is a tool for a
specific purpose and must be judged by that purpose.)

BTW, while Rebecca is concerned about the terms 'subject'..., I am
concerned about 'transitive' and 'intransitive' - the valence model does
away with these terms. Instead of having two main groups of verbs (object
or no object), in the valence model we have seven, which are of equal
importance - which is much nearer to the 'truth'.


--------------------------------------------------------------------------
Burkhard Leuschner -  Paedagogische Hochschule Schwaebisch Gmuend, Germany
E-mail: [log in to unmask]    [h]     Fax: +49 7383 2212
HTTP://WWW.PH-GMUEND.DE/PHG/PHONLINE/Englisch/index.htm