Re: THEORY: Ergative, word order, and predicates

On 9/25/06, Eric Christopherson <rakko@...> wrote:
> I seem to recall that typologically speaking, it is common for head-
> final languages to be SOV (or otherwise verb-last).  Does this
> correlation also apply for ergative languages?  I'm thinking that in
> an accusative language, SOV translates to nominative noun-accusative
> noun-verb, and since the nominative is less marked, that means it's
> less marked-more marked-verb.  Thus, when speaking of an ergative
> language, in which the absolutive is the less marked case, if I
> follow the same markedness-based ordering, I end up with the order
> absolutive noun-ergative noun-verb, which in turn is OSV.  I hope
> that all makes sense; if not, my main question is: is it more
> "typical" for a head-final ergative language to be SOV, or OSV?
>
SOV.  OSV as a basic constituent order is vanishingly rare.  But
anyway, the markedness or lack thereof of case markings doesn't really
have anything to do with the word order of a language.

The following might ease some confusion.  There are three (main) ways
ergativity will show up in a language:

Your language could be ergative as regards case marking: you could
mark intransitive subjects with the same case as transitive objects.
(Not uncommon.)
Your language could be ergative as regards agreement: the verb could
use the transitive object agreement markers for intransitive subjects.
 (Less common.)
Your language could be ergative as regards word order: intransitive
subjects could be put in the same position in the sentence as
transitive objects.  (This last one rarely shows up, since it only
makes a difference in word-medial languages.  You get SVO in
transitives and VS in intransitives, or OVS in transitives and SV in
intransitives.)

These three don't all need to agree; it's quite common to have
mismatches.  Frequently your case marking will be ergative but your
agreement will be accusative.  (And since most ergative languages
aren't verb-medial, the third question doesn't come up.)

Another thing to clear up some possible confusion:

Ergativity is just an answer to the question: "How do I mark the sole
argument of an intransitive sentence?"  That's pretty much it.  Some
languages mark it like a transitive subject, and some like a
transitive object, and others do something else.  That's pretty much
it.  When it comes to transitive sentences, everything's the same.
Subjects are still subjects in ergative languages, and objects are
still objects; the only thing that's different is how the speaker
marks (and the listener picks out) the sole argument of an
intransitive.

> Also, while thinking about this, I started wondering, is the
> definition of a predicate the same in an ergative language as in an
> accusative one, i.e. the verb with its object?  Or would a predicate
> in an ergative language be the verb along with its subject?
The former.  Ergativity doesn't mess with this sort of thing; subjects
are still subjects.

Instead of thinking of ergative languages as ones that have "mixed up"
subjects and objects, it's much better to think of them as languages
that are (for transitives) just like English.  The difference is just
in intransitives, and the choice they make isn't mixed up... they've
just made a different generalization.

Take these two sentences: "Jonas ran" and "Jonas died".  In the first,
Jonas is an agent, but in the second, he's a patient.  A language has
to decide: "How do we mark 'Jonas'?"

Accusative languages say "I don't care that they're different; mark
'em both as agents.  (That is, mark 'em the way we do the agent in a
transitive.)"

Ergative languages say "I don't care that they're different; mark 'em
both as patients.  (That is, mark 'em the way we do the patient in a
transitive.)"

The languages that care -- those that mark the intransitive agents
like transitive agents and intransitive patients like transitive
patients -- we call "Split-S" or "Fluid-S" or "Active/Stative" or
"Active", depending on the details and our terminology.

> Bonus
> question: if a predicate is still composed of the verb and its
> object, does that mean that an intransitive predicate consists of the
> verb and its subject (since the intransitive subject patterns with
> the transitive object)?
>
No, stuff like this isn't affected by ergativity.  It's partly true
that "the intransitive subject patterns with the transitive object",
but only for certain things, and it differs between languages whether
it's the case patterns or agreement patterns, etc.  Some patterns,
like case marking, are immediately visible, and we make a big deal out
of them, but in most ways, the subject/object relations aren't
affected.

So we might ask -- and it's been asked here before: "Take the
incorporation of an object into a verb.  In an ergative langauge, do
you incorporated the subject?"  The answer is no; ergativity is
irrelevant to this.  Because ergativity doesn't make subjects into
objects or objects into subjects or anything like that.

We could also ask "Say it's an intransitive sentence.  Since
intransitive subjects are like transitive objects, and since
transitive objects can be incorporated, can intransitive subjects be
incorporated?"  The answer is the same in an ergative language as in
an accusative language: "Sometimes, it's complicated, ask me later."

Anyway, hope this helps.

-- Pat

Reply