Re: Agreement Idea

On May 29, 2007, at 2:57 PM, David J. Peterson wrote:

> I just finished grading one of my student's finals, and, as a result,
> came up with a bizarre agreement idea.
>
> Let's say a language had ten or so noun classes, SOV word order,
> and the verb, instead of agreeing with the subject, or the object,
> or anything like that, simply agreed with whatever the closest
> argument was:
>
> Class I: ka-
> Class II: me-
> Class III: ri-
> Class IV: tu-
>
> Oh, and let's say /-i/ dative, /-u/ accusative, /-e/ instrumental,
> /-o/ past tense:
>
> (1) kaven meluri ritapu tunese tuxoro.
> /I-man II-woman-DAT III-flower-ACC IV-tongs-INS IV-give-PAST/
> "The man gave the woman a flower with tongs."
>
> And, of course, word order would have to be variable for this
> to be interesting:
>
> (2) kaven meluri tunese ritapu rixoro.
> (3) kaven tunese ritapu meluri mexoro.
> (4) tunese ritapu meluri kaven kaxoro.
>
> And it'd be even more interesting with relative clauses:
>
> (5) kaven meluri se ridum rijalo tunese tuxoro.
> /I-man II-woman-DAT REL. III-frog III-see-PAST IV-flower-INS IV-
> give-PAST/
But <tunese> is IV-tongs-INS, not IV-flower-INS; it should be <ritapu>.

> "The man gave the woman the frog saw a flower."
Wow. I had a hard time deciphering even the gloss of that!

>
> Which could then become...
>
> (6) kaven tunese meluri se ridum rijalo rixoro.
Or rather <kaven ritapu meluri se ridum rijalo rixoro>.

>
> Then at this point, I suppose you could associate whatever
> pragmatic meaning (or even grammatical meaning) you wanted
> to the use of agreement.  Perhaps the only way a verb could be
> fronted was if it was fronted with the NP it agrees with...
Interesting.

>
> All right, back to grading.