Syntax Grammars
Ronald Ross
rross at expressmail.net
Wed Dec 8 01:01:10 EST 1999
clayton stirling bartholomew wrote:
> It would be entirely fair for someone to ask the question:
>
> But Clay, how would you improve on our current grammars?
>
> Here is a suggestion, just one of many possible approaches. To begin
> with we might completely reorganized around a functional model (NO
> connection with Systemic Functional Linguistics).
>
> Let's say we start out with an abstract language model that is a complex
> network of functional domains. By complex I mean that the relational
> connections between the functional domains are many to many, which gives
> us a network that looks a lot like Louw and Nida's semantic domain model
> for lexical semantics. But instead of lexical-semantic nodes in our
> network we have functional nodes which I will now try and illustrate.
>
> Let's take the simplest possible sentence. This sentence is made up of
> two entities (things) and one action. One of our entities is an "agent"
> and one entity is a "patient."
>
> So we have a structure:
>
> agent -> action -> patient
>
> Now our grammar would explain for us all the different ways our
> particular language, say Attic Greek can realize this structure. It will
> tell us all the different ways that an agent can be encoded, all the
> different ways an action can be encoded, and all the different ways a
> patient can be encoded. It will then tell us how the agent, action,
> patient encodings can be combined into an structure yielding an
> agent -> action -> patient sentence which would be acceptable to a
> native speaker of this language.
>
> The main difference between this and our current grammars is that it is
> not organized around the so called "parts of speech" but starts at the
> beginning with a functional model and applies it throughout the
> presentation of the material.
>
> With this approach the genitive case and an adjective would be found
> listed under the same functional domain when they both were being used
> to limit a substantive. The genitive case would also appear under other
> functional domains since it has a number of potential functions.
>
> The one big advantage to this model would be that students would not be
> tempted to think that the there are 36 (of 54 or 102) different kinds of
> genitives. This approach would completely disabuse the student of that
> awkward notion. The student would be encouraged to understand the
> genitive as just one of the possible ways to realize a particular
> function.
>
> Anyway,
>
> This is just a stab at a complex problem. Not intended to be a well
> thought out solution.
>
> Clay
I believe that this is exactly what is needed. This would allow for all kinds of
innovations (innovations in *Greek* grammars, that is), such as including under the
rubric of relative clauses all those structures the carry out the function of
relative clauses, even if they don't begin with a relative pronoun, such as
participial relative clauses ('the saying these things deciples') and
"prepositional" relative clauses ('the boy on the table'), etc. Personally, when I
look through even very recent grammars, I'm amazed to still see chapter titles like
"the syntax of the accusative" or "the syntax of the preposition". I couldn't agree
with you more!
Ronald Ross
UBS consultant
Department of Linguistics
University of Costa Rica
More information about the B-Greek
mailing list