Tom wrote:
<<
Hello, anyone who feels like answering.
 >>

I have kind of a response, but not a specific one.  I had some
trouble trying to picture what you were describing.  Examples
would be helpful.  :)

Anyway, you mentioned applicatives, and whether are not they
can occur in ergative languages.  The answer is, of course, yes.
And I have a kind of example from Tukang Besi, though I don't
believe this language can accurately be classified as accusative,
ergative, trigger, or anything I've even ever really seen.  It's
quite an extraordinary language, and I hope I can explain it
accurately.

All of this is based on a talk that was given at the undergrad.
typology class I TA'd at UCSD by Mark Donohue (visiting from
the University of Singapore).  He's an amazing guy.  I'm going
to try to reconstruct what he said about Tukang Besi (a language
for which he wrote the grammar) based on his handout, which
can be downloaded at the following url:

http://ling.ucsd.edu/~djp/dlstuff/verbinitialhandout.pdf

Tukang Besi is a language of central Indonesia, and is at least
distantly related to Tagalog.  It has an agreement pattern and
case marking system that strikes me as really quite fascinating.
Below are some facts which would lead one to believe that
Tukang Besi is simply a nominative/accusative language:

(1)
(a) Ku'ita te ana (na iaku).  /1sg.-see ACC. child (NOM. I)/ "I saw a  
child."
(b) *Ku'ita na iaku te ana.  For same.  [word order = VOS]
(c) Kurato (na iaku) di kampo.  /1sg.-arrive (NOM. I) OBL. village/   
"I arrived at the village."

So, essentially, the order is VOS, the language is a prodrop language,
and oblique arguments follow the subject.

Now here's an alternative way of expressing (1a).

(2)
(a) Ku'ita'e (te iaku) na ana.  /1sg.-see-3sg. (ACC.? I) NOM.?  
child/  "I saw the child."
(b) Ku'ita'e na ana te iaku.  Variant word order is grammatical.

So now the case marking has switched, the subject is now obligatory
and the object non-obligatory, and the word order is flexible.
Importantly, though, this wasn't triggered by a valence-changing
affix, but by the presence of optional third person object agreement
on the verb (somewhat reminiscent of Georgian).

So far, the system can be summarized as follows:

(3)
(a) Monovalent verbs: Agrees with the subject via prefix, and the
lone argument is marked with /na/.
(b) Bivalent verbs: If the object is not marked on the verb, the
agent is marked with a prefix, the word order is VP(A), /na/
marks the agent, and /te/ the patient.
(c) Bivalent verbs: If the object is marked on the verb, the agent
is marked with a prefix and the patient with a suffix.  The word
order is either V(P)A or VA(P), /te/ marks the agent, and /na/
marks the patient.

A confusing, but regular system.  Now for a ditransitive sentence:

(4)
(a) Kuhu'uke te boku (te iaku) na ana.  /1sg.-give-3sg. ACC.? book  
(ACC.? I) NOM.? child/  "I gave the child a book."

Now the "nominative" case is marking the recipient, and the
"accusative" case is marking both the agent and patient (or theme),
with the subject yet again the pro-droppable argument.

The way Donohue characterized these case markers is as follows:

(5)
(a) di = a general oblique/adjunct marker
(b) te = marks core terms not marked by /na/
(c) na = marks one obligatory term in the clause (marks the P or IO  
if the verb has object marking; otherwise, marks S or A)

Interestingly, there are several passive markers which tend
not to be used (or are certainly not as common as the English
passive) which are rather specific:

(6)
(a) Noto'ita na ana.  "The child was seen."
(b) Note'ita na ana.  "The child happened to get seen."
(c) Nomo'ita na ana.  "The child was visible."
(d) Nopo'ita'ita na ana.  "The children looked at each other.

That latter is a reciprocal marker.

In addition to this, there's an applicative marker (and it seems
to be lexicalized which verb takes which applicative marker).
So you can get the following (using an intransitive verb):

(7)
(a) Norato na mori di kampo.  /3sg.-arrive NOM. student OBL.  
village/  "The students arrived at the village."  (Normal)
(b) Noratomi te kampo na mori.  "The students arrived at the  
village."  (Applicative)
(c) Noratomi'e na kampo te mori.  "The students arrived at the  
village."  (Applicative + Obj. Marking)
(d) Notoratomi na kampo.  "The village was arrived at." (Applicative  
+ Passive)

So essentially there are a bunch of different ways to say the
same thing.  Why?  Donohue said that the privileged nature of
the /na/ marker allowed it to be used pragmatically for particular
stylistic reasons.  He offers a couple stories (which I won't  
transcribe,
but just describe) as examples.  One of them involves a chicken
and a woman named Wa Sabusaburengki.  The story goes like this
(I'll mark what argument gets marked with what case in parentheses):

(8)
(a) Once upon a time, there was a (INST) lady, and her (te) name
was (te) Wa Sabusaburengki.
(b) (Te) Wa Sabusaburengki was going to decapitate a (te) chicken.
(c) Just as (pro-drop) she was about to decapitate that (na) chicken...
(d) ...(na) the chicken said...

So up until the second verb "decapitate", no verb has object agreement.
Object agreement is used on the second verb "decapitate", so that the
chicken, the patient, can be marked with /na/.  And why mark "the
chicken" with /na/?  Because up until that point, the story introduces
the woman and talks about what she's doing.  /Na/ is used to alert
the listener that a change in subject is coming up.  And then when
"the chicken" appears as the subject again, /na/ is used again.

This example was fairly simple.  Mark also gave a much longer
story that essentially tracks all its arguments throughout in one
way or another so that they can be referred to with /na/ whenever
they become the focus of attention.  This is done with agreement
marking on the verb, applicatives, and passives--all of it for
pragmatic purposes.

So, that's about a third of his handout.  The rest goes into the
history of the language, which is even more interesting (and
which Austronesian buffs will find familiar).  Anyway, this is
a really interesting system that I've wanted to share for awhile,
but I've been busy.  It probably would've been better to try
to explain it right when I heard about and it was fresh in my
mind, but at least I can offer the handout for you to look at on
your own.

-David
*******************************************************************
"sunly eleSkarez ygralleryf ydZZixelje je ox2mejze."
"No eternal reward will forgive us now for wasting the dawn."

-Jim Morrison

http://dedalvs.free.fr/