Re: HELP: Relative Clauses with Postpositions

Philippe wrote:

<< I would analyse the sentence somehow like:

plain-ACC Shinar-XXX-land-LOC 3PLHUM-find-PAST-3SGINA>>

If we want to get into the actual details, this is what it would look like:

saqo-lasan-a Sinal-ial iXol-y m@(-0) (0-)k-a-n-o-sqa-mist-uS-i.

This is the gloss, word by word (where (0) is a null morpheme):

Word 1: plain-suitable-acc.
Word 2: Sinar-gen.
Word 3: land-dat.
Word 4: in-positive
Word 5: 
3subj.3obj.-past-active-perfect-nonevidential-completive-search-affirmative-theme

The very last morpheme is a theme vowel, which is the last vowel of your 
basic verb, which takes the form CVCV.   It separates and attaches after all 
suffixes have been added.

So, some things I neglected to mention: My language doesn't actually have a 
locative case (surprise!   I totally forgot).   When you use a postposition, 
the actual meaning is conveyed by three things:

(1) The case of the postposition (there are three, plus person markers).
(2) The case of the object of the postposition (generally: 
accusative=allative; dative=essive; genitive=ablative).
(3) The motion marker on the motion verb (there is no motion verb here, so 
there isn't one, but it'd be either a cardinal direction [N, E, S, W, or some 
combination thereof], a prefix indicating movement away, a prefix indicating 
movement towards, or a prefix indicating nonspecific movement [used for sentences 
like "I walk while I'm inside the house"]).

So, I actually decided what to do, and I decided on the above--that is, 
rethinking the "noun-genitive" order, and, well, changing it.   I might change the 
name of the case now (I was influenced by Siglit, an Eskimo language).   Also, 
since there's no general locative, I needed to decide what postposition to 
use.   Rather than using the "at" preposition (which is the general locative, 
and can convey a number of functions), I went with "in", taking this reading as 
the one where Shinar is an actual place with borders, and so the plain is 
contained wholy within those borders (this would be the NATIONS are CONTAINERS 
metaphor).   Were I to read Shinar as not being a nation, but just a region, or a 
general place, I might use the general locative.

Anyway, thanks for the input!   I've said it once, and I'll say it again: 
This list is pretty much the best linguistic resource there is.

But wait, there's more...

I've just read the recent posts.   I'm still trying to digest.   Let me see 
if I can tease out Mu's statement here:

<<This works fine, doesn't it? The only problem I could imagine is that GEN
could also tag the argument of the verb like in plenty of languages. Imagine
that like in Estonian the partitive GEN tags the object (ACC) of the verb:>>

There is a type of verb in my language (which is called "Epiq", by the way, 
so I'll just start calling it that) that takes a genitive as its *subject*.   
These would be what would ordinarily be called "performative verbs", I believe. 
  Let's see if we can cause some problems:

"I promised you a pony."

This would be...

I-GEN. pony-INST. you-ACC. promised.

(I'm going to conveniently ignore the fact that my language lacks pronouns.)

So what if the subject was "the dog", and there was a man that owned the dog. 
  That would come out as:

man-GEN. dog-GEN. pony-INST. you-ACC. promised.

"The man's dog promised you a pony."

Ack!   Can this work?   It's certainly interesting.   When you hear the first 
word, you don't know whether the man is possessing something or is the 
subject of a class IV (performative) verb...   Then when you get to "dog", it can be 
either "the man's dog's...", or "the man's dog (performed something)".

Well, let me say, first of all, that I like this: I think it's neat.   Now, 
though, let me ask: Do you think this would cause to many problems for a 
speaker/listener, such that this strategy would get lost?

<<Quite apart from that--I don't understand why David's word order wouldn't
work:

land-LOC in shinar-GEN
in the land of shinar>>

This is starting to sound more attractive...   Especially since you'd never 
get the GEN.-GEN. problem, since word order is SOV and fixed.   Well, actually 
what you'd get is the first GEN. could *only* be interpreted as the subject of 
a performative verb.   This would clear up all the ambiguity.

Kou wrote:

<<rhaur lé sö nggíshabsöt cha dínsav Shinarsaub thfau zçírhéf

they-nom past-aux a plain-acc the place-loc Shinar-postpositional thfau* find>
>
The reason I was using the genitive is because I got the idea from Siglit, 
where the genitive is used for everything (and not called the genitive).   If I 
were to do it like this, it would be...

plain-acc. land-dat. in Shinar-dat. "called" [they found it].

Is that how it'd work?   If so, where is the second PP attaching to?   What's 
it modifying?   Seems like it'd be modifying "in the land", and not the land. 
  Or would it come out as:

plain-acc. land-dat. Shinar-dat. "called" in [they found it].

?

Okay, maybe I haven't made my official decision yet.   I'll mull it over 
tonight when I meet with my conlang group (we're designing an undergraduate 
language creation class for UCSD).

-David