Re: Agents

[doug@netcom.com (Doug Merritt)]

DLS9@aol.com (Guido) said:
> A question arises as to why langs don't (or only rarely?) put
>them into three separate categories. Efficiency? [ In regard to
>     transitive subject, intransitive subject, and transitive object ]

Grammatical role marking is necessary to disambiguate sentences.
There are (at least) three mechanisms that can achieve this:

	- word order
	- adpositions (e.g. prepositions)
	- case inflection

The three categories in question do in fact get adequately disambiguated
in any language, when you take all these mechanisms into account, so
there *are* three categories, not just two. The fact that only two
are marked might be called efficiency, I suppose. But it's important
to note that this is not purely a matter of case inflection, since that's
only one of the three mechanisms.

>And could
>ergative-language-speakers even be said to think of their intr subjs as
>agents at all? That was what I was groping at earlier.

All speakers of all languages think, at least at times, in terms of
cause and effect, and in terms of agents of cause. Linguistically
there is a strong separate between grammatical roles (where the
role of agent might not exist) and semantic roles (where the role of
agent does always exist).

A lot of confusion can arise if you don't keep grammatical roles
mentally separated from semantic roles. Naturally they can be strongly
related, but in principle they can be unrelated.

Consider: "I turned on the oven. An hour later, the cake was baked."
The latter sentence is grammatically passive, but semantically there
is an obvious agent.
	Doug