k?u and related argument problems.

[Pamela Munro]

Sorry to be obscure!

In English, with a verb like "eat" we can use two nouns, e.g. "John" and
"bread": we can say "John ate bread". But with a verb like "dine", we
can't: we can say "John dined", but not "*John dined bread". (Yes, you
can put "bread" in the sentence, but you have to do it in a
prepositional phrase, "John dined on bread".) That's what I mean by
"adding random nouns", which may have been an odd phrase, but I'm just
trying to avoid theory-specific terminology. You can't freely add
"bread" to a sentence with "dine", even though semantically it makes
sense that when one dines food is involved.

Thus, in Lakhota, some verbs are associated with one ordinary noun
phrase (not part of a postpositional phrase), some with two, and some,
like k'u 'give', with three. I can't give you a Lakhota example of a
verb that seems semantically as though it should take more associated
nouns than it does (my Lakhota isn't ready enough for that), but I can
give you a Chickasaw one.

In Chickasaw, owwatta means 'hunt'. But it's intransitive -- it is not
possible to add a noun phrase specifying what was hunted (like 'deer')
no matter how much you want to. (Thus, it's like English 'dine'.)

Thus, I consider the number of non-oblique noun phrases (neither in pre-
or postpositional phrases or marked with oblique case markers) a verb
can be associated with is its number of arguments. I feel that this
shows that the number of arguments a verb has is not necessarily
associated with the inflectional pattern it takes. I pointed out that in
English we'd certainly say a verb like 'hunt' or 'eat' is transitive
(though both can also be used intransitively, as David pointed out),
even though the English verbs inflect only for their subject, never for
their object. As David noted, that is a language-specific fact. So, in
contrast, in Lakhota (and also Chickasaw) a verb may have three
arguments, but it can be inflected to show the person and number of only
two of these.

Anyway, that's what I was trying to say. (I don't think that whether one
accepts the PAH really matters for this, but I'll leave that matter to
others!)

Pam

R. Rankin wrote:

> Excuse a question from someone who has always done more phonology and
> morphology than syntax.  Pam wrote something in one of postings to the
> effect that "you can't just add nouns in Dakotan sentences."  (The
> quote is inexact because I don't have the message here on my home
> computer -- sorry.)  This was in reference to the argument structure
> of k?u.
>
> I guess I'm having a little trouble picturing what this means and what
> its implications are.  Could Pam or anyone give me an example of this
> restriction, especially compared to some language (English would be
> fine) in which you *can* add such forbidden nouns, whatever they might
> be (but wouldn't be able to in Dakotan)?
>
> Sorry to be so thick, but I'm missing something here. (And I'm also
> thinking about how this whole discussion might play out if you accept
> the 'pronominal argument hypothesis' for Siouan.  Then all the
> nominals would be something like adjuncts, and I'm curious to know
> what the restrictions on such adjuncts might be.)
>
> Bob
>

--
Pamela Munro,
Professor, Linguistics, UCLA
UCLA Box 951543
Los Angeles, CA 90095-1543
http://www.linguistics.ucla.edu/people/munro/munro.htm