Re: THEORY: Why more than two grammatical relations?

On Sun, 21 Oct 2007 04:21:41 -0400, J. 'Mach' Wust
<j_mach_wust@...> wrote:
>On Fri, 19 Oct 2007 13:38:15 -0400, Eldin Raigmore
><eldin_raigmore@...> wrote:
>>I do not know whether or not that is the case; but languages like Greek and
>>Latin and German and many other European languages have usually been
>>analyzed as having grammatical relations.
>Certainly, German is analyzed as having subjects and objects. It distinguishes
>four kind of object according to most analysis: Accusative object, dative
>object, genitive object and prepositional object.
Ordinarily the accusative and dative objects would be considered Grammatical
Relations.

In the usual genitive phrase, the possessor (or other genitive-case noun) is
not an object of any verb.  In clauses where some object of the verb must be
given the genitive case, this is usually an example of "quirky" or "non-
canonical" object-marking; of "quirky case".

For the most part prepositional objects are also not objects of the verb, but
rather of their preposition; they may well be arguments, but if so they'll usually
be considered oblique arguments rather than grammatical relations.

However,

>Prepositional objects are considered objects because the prepositions are not
>chosen according to their semantics, but are determined by the verb in much
>the same way objects are.
Wow! I didn't know that!

So, maybe the "prepositional objects" _are_ objects of the verb; in which case
they're probably in grammatical relations.

(Unless I've misunderstood you, and "prepositional objects" are just another
set of examples of "quirky case" non-canonical marking of direct or indirect
objects.)

>mach
Thanks, Mach.