Daniel Andreasson wrote:

> Durie is the man on Acehnese. Who's written the article you've found?
> Just so I can stay clear of it. Don't want to read something with the
> name "universal" and "theory" in the title... ;-)

Maria Bittner and Ken Hale. Two generativists who don't work on European
languages very often. They get my respect. :)

(Ken Hale has done so much work on Warlpiri, that his children are fluent in
it.  White Americans fluent in a dying Australian language! Astounding!)

> > (But can you imagine a single theory that derives accusative, ergative,
> and
> > active languages! It was developed by Ken Hale, who IMNSHO is the most
> > brilliant linguist alive today. I wanted to study with him for graduate
> > school, but he retired the year before I started applying.)
> Yeah, I can actually. I think Dixon has a nice discussion of it
> in "Ergativity". I don't know if you can call that a theory in
> the sense you're used to, but still.

Hmm. Well, you're absolutely right. I did mean "theory" in a different
sense.  I meant it in the form of "one principle that can account for all
kinds of data, including stuff that isn't directly related to the topic it
was developed for".  They give one principle of case, without reference to
the different patterns (ie, accusative, ergative, active), and derive all
the patterns from that one principle.  They're idea, in layman's terms, is
that each noun must be within the sphere of influence of ("governed by") a
"case assigning" element (ie adposition or verb).  The differences in the
types come out when two elements are in the same area

> Hmm. Originally, incorporation of an object or not had to do with
> the definiteness of the object. I have _no_ idea what happens to
> indirect objects in transitive clauses. I'll have to think more about
> this. Glad you pointed this out, I had completely forgotten about the
> indirect objects. :-)

You could treat indirect objects like subjects.  Some people have proposed
that indirect objects are in fact some kind of secondary subject.  According
to these people, "X gives Y Z" can be semantically decomposed into "X causes
Y has Z". So "John gave Mary a gift" is "John caused Mary to have a gift."
There is interesting support for the idea. Compare the following examples:

*Who did the brother of give Mary a gift? (Pretty incomprehensible)
*Who did John give the sister of a gift? (Huh? What is that supposed to
Who did John give Mary a picture of? (Perfectly okay.)

So the indirect object patterns more with the subject than the direct

On the other hand, this goes against the idea that in head marking
languages, the indirect object takes over the role as the agreeing object.
(Unless you're a generativist, in which case that is expected because
agreement disregards grammatical roles almost entirely in favor of

I'm not sure how to implement this for Pimak, but I thought I might make the