Re: Active-Ergative langs (was Re: Ke'kh - degrees of volition)

> From: H. S. Teoh [mailto:hsteoh@quickfur.yi.org]
>
> On Wed, Sep 20, 2000 at 02:18:02PM -0400, The Gray Wizard wrote:
> [snip]
> > Syntactically accusative:
> >
> > John kissed Mary and left. (John kissed Mary and he left.)
> > John[A] kissed Mary[P] and (he[S]) left.
> [snip]
> > Syntactically ergative:
> >
> > John kissed Mary and left. (John kissed Mary and she left)
> > John[A] kissed Mary[P] and (she[S]) left.
> [snip]
> > John kissed-antip Mary and left (John kissed Mary and he left)
> > John[S] kissed-antip Mary[Obl] and (he[S]) left.
> [snip]
>
> This is very interesting! My conlang has a system of sentence adjoinment
> that handles these constructs in a different way: two sentences can be
> adjoined through a "linking word" -- usually a noun -- which is marked for
> *two* syntactic functions (noun cases), one for the first sentence, one
> for the second. (For examples, see my other post on "Look and thou shalt
> see".)
>
> The interesting thing about this is that in your first example sentence:
>       John kissed Mary and (he) left.
> my conlang would use "John" as the linking word, and would structure the
> sentence this way:
>       Mary(rcp) kiss(verb) <aux-cvy>John(org) left(verb)
> The auxilliary inflection <aux-cvy> marks the function of "John" in the
> *subsequent* sentence (John left); the normal inflection marks the
> function of "John" in the first sentence (John kissed Mary).
>
> For your second sentence:
>       John kissed Mary and (she) left.
> my conlang would use "Mary" as the linking word. Here, the sentence would
> be structured thus:
>       John(org) kiss(verb) <aux-cvy>Mary(rcp) left(verb)
> Here, the <aux-cvy> on the noun "Mary" marks its function on the second
> sentence "she left", while the normal case marking (rcp) marks "Mary" as
> the recipient of the verb (kiss) in the first sentence.
Interesting.  How would you handle the coordination of two transitive
clauses.

John kissed Mary and he pushed Ralph
John(org) kiss(verb) <aux-cvy>John(org) push(verb) Ralph(rcp)

John kissed Mary and she pushed Ralph
John(org) kiss(verb) <????>Mary(rcp) push(verb) Ralph(rcp)

Here Mary is the recipient in the first clause and the actor (org?) in the
second.  Is there a different linking inflection for this case?

> Furthermore, my conlang can do this with *every* noun case -- so you can
> use instrumental or locative nouns as "linking nouns" as well. So in
> effect, you can construct sentences like:
> 1)    He went to the countryside by the horse and (the horse) stopped at
>       the well.
> 2)    He went to the countryside by the horse and (he) stopped at the
>       well.
I like this capability and gave some though to implementing it in amman iar
which has ergative pivot constraints.  amman iar syntax, however, is so
heavily weighted toward the core arguments of the predicate that
constraining oblique arguments didn't seem to fit.

> As to whether my conlang is ergative or active or something else, I really
> don't know, because it doesn't analyse sentences as subject-predicate,
> but it uses an event(verb)-centric model. It's too long to attach to this
> current post, but if anyone's interested, I'll post it separately.
I don't think your analysis precludes ergative or active.  amman iar
sentences are typically analyzed as a predicate (verb) and its arguments.
Its a very verb-centric analysis.  In fact the language itself tends to be
very verb centric in that much of the syntax is driven by the relationship
between the verb and its arguments.  "Subject(S=A)" is not as meaningful a
concept in amman iar as is the pivot relation (S=P).  Its how the arguments
of the predicate (verb) are discriminated that counts.  From your examples
above, I would guess that the arguments of your verbs are discriminated
based on semantic roles and would appear to be active.

David

David E. Bell
The Gray Wizard
    dbell@graywizard.net
    www.graywizard.net

"Wisdom begins in wonder." - Socrates