Re: Ergativity Question

Perhaps my problem is that I never thought of passivization as being
detransitivizing, but as reversing the roles or the word order. Let me
demonstrate the concept I'm trying to describe. Let's start with some
English sentences:

1) I ate.                (active, intransitive)
2) I ate corn.           (active, transitive)
3) Corn was eaten.       (passive, intransitive)
4) Corn was eaten by me. (passive, "transitive")

Granted, 4) is a bit akward in English, and requires the use of a
prepositional phrase, but its essentially the same as 2) but with a reversal
of word order and voice. Also, since I will be marking cases, I'm not going
to concern myself with changing the word order of these sentences.

In a nominative/accusative, I would mark it like this (using OBJ for the
object of the preposition)

1) I(NOM) ate(AV).
2) I(NOM) ate(AV) corn(ACC).
3) Corn(NOM) was eaten(PV).
4) Corn(NOM) was eaten(PV) by me(OBJ).

In ergative languages, for the last two we have:
3) Corn(ABS) was-eaten(PV).
4) Corn(ABS) was-eaten(PV) me(ERG).
(I'm not sure exactly how to mark the voice, though I consider it to be
passive-like)

I believe I've also read about something called antipassive, where 1)
becomes:
1) I(ABS) ate(APV).

But then we don't have any way to say 2) in the antipassive voice (AFAIK)
2*) I(ABS) ate(APV) corn(NOT-ERG!!!)

I guess this would be akin to the akwardness of case 4) in nom/acc langs. So
perhaps I could make a preposition that marks the patient, and do something
like this:
2) I(ABS) ate(APV) of corn(OBJ)
Come to think of it, I kinda like that. :-)

But the system I was originally proposing would merge NOM and ABS into a
single case -- for argument's sake lets call it SUB (for subject). Then we
would have:

1) I(SUB) ate(AV).
2) I(SUB) ate(AV) corn(ACC).
3) Corn(SUB) was(PV) eaten.
4) Corn(SUB) was(PV) eaten me(ERG).

In this case we would have an "active" or "antipassive" voice in 1) and 2),
which would use the "accusative" case to mark the second argument (patient).
But in 3) and 4), we have a "passive" voice whcih uses the "ergative" to
mark the second argument (agent). It just occured to me that this is similar
to (my understanding of) languages like Tagalog, where any role can become
the subject, and the verb changes form to denote the subjects role. In this
case, Active Voice means "subject is agent" and Passive Voice means "subject
is patient". But that's a whole 'nother kettle of corn. :-)

Thanks,
~Caleb

--- In conlang@yahoogroups.com, "Thomas R. Wier" <trwier@U...> wrote:
> From:    Caleb Hines <cph9fa@A...>
> > I'm pretty sure I understand the basics of what an ergative/absolutive
> > language is (as opposed to a nominative/accusative one), but I've got a
> > question. Would it make sense for a language to be ergative/absolutive
in
> > the passive voice, but nominative/accusative in the active voice? Are
there
> > any example natlangs or conlangs that do this? It's one possible system
I'm
> > considering for Akathanu (which is kinda funny, considering that
Akathanu
> > doesn't even really have verbs, except an extended form of the copula).
>
> It depends on what you mean by "passive".  In most systems that
> I'm familiar with, the "passive" is fundamentally characterized
> by demoting the agent somehow.  This also usually implies
> detransitivization, but not always.  Now, since ergativity is
> defined as treating the single argument of intransitive verbs the
> same as the patient of transitive verbs (by whatever criterion), it's
> not clear to me how you have ergative patterning with passives,
> but not other intransitives, and still remain "ergative". That is,
> if passivization is genuinely detransitivizing, it would
> effectively create two classes of intransitives.  On the one hand,
> you'd have regular intransitives, which are used when either an agent
> or patient argument is being focused on (i.e., active voice), and
> these intransitives' arguments are treated the same as the agent
> of  transitives. On the other hand, you've have special passivized
> intransitives, whose single argument would pattern like the patient
> of transitives.  And the problem with that is that it's a perfect
> example of a split-S system (aka active-stative), and thus not
> ergative at all.
>
> (Now, I'm wondering if you could get around this by having a passive
> that demotes the agent, but still requires both arguments.  I'm not
> sure how this would be different from an inverse system such as are
> found in some Algonquian languages, and thus also not ergative.)
>
> Another way of answering your question is that Bob Dixon's book
> _Ergativity_ (the 1994 version, not his earlier paper) does not
> mention any splits by voice -- only by tense, aspect, or mood
> (of those not conditioned solely by semantic properties of the
> arguments).  While this book is not the Bible of Ergativity, it
> is IMHO reasonable to call it at least the Talmud, or Hadith, of
> the phenomenon.
>
> ==========================================================================
> Thomas Wier        "I find it useful to meet my subjects personally,
> Dept. of Linguistics    because our secret police don't get it right
> University of Chicago   half the time." -- octogenarian Sheikh Zayed of
> 1010 E. 59th Street     Abu Dhabi, to a French reporter.
> Chicago, IL 60637