argument structure k'u etc.

[ROOD DAVID S]

Pam, I'm not quite sure that this is a very important discussion, even
though I started it (one of my pet peeves, for years, has been the claim
that 'give' takes indirect objects universally).  However, it's kind of
fun.
	The kind of verb you're asking for (only 3rd person singular
argument possible, not phonomenological) seems like it'd be hard to come
by in any language -- can you provide a possible example from English or
some other language you know?  All I can think of would be the modal-like
verbs like iyecheca 'necessary' or nachece 'maybe', which take
propositions as their "subjects". Or maybe an intransitive that cannot
have an animate argument, so "pi" would be impossible -- e.g. do plants
'grow' or 'die' with a different verb than do people?  Two minutes of
research in Ingham gives a verb "uya" for 'to grow (longer), as hair';
back checking with Buechel indicates that it's used for grain springing
up, and then the sentence example he gives is "tokiyatanhan uya hwo?"
which he glosses 'where does the wind come from?' -- so I'm very confused,
but maybe "uya" is a candidate.  Are there any speakers of Lakhota
out there who can comment on this word?  Does it ever take a
plural "subject"?  And if not, what does that tell us?  The morphology of
third singular won't reveal anything about whether or not that "subject"
is an argument.
	More important, perhaps, is some agreement about what we mean by
"argument".  I intended my "marked on the verb" definition to be
Lakhota-specific; I think you have to find criteria for grammatical
argument status one language at a time.  And as John said, (I think), my
example with three arguments for the causative works only because there
are two verb stems involved.  I have no objection to claims that the
logical structure of 'give' universally includes three entities, but I do
object to the hypothesis that the recipient is in some sense "secondary"
or "indirect" in all languages.   Given that _k'u_ can take only two
affixes at most, and that one is the giver and the other the recipient, I
still claim that the third "entity" involved is not part of the core
argument structure of this verb in this language.
	I run into similar problems when people claim that "eat" in "we
eat every afternoon at 4:00" has an "implied object" because you have to
eat something.  I think it is purely intransitive in that kind of context,
and has only one argument, logic or no logic.
	Maybe my structuralist upbringing is leaking through here (I
learned lingusitics first from Charles Hockett), but I need concrete
evidence for grammatical structures before I accept them, and Lakhota
"k'u" only shows two entities, even if other things are hanging around in
the vicinity.
	David

David S. Rood
Dept. of Linguistics
Univ. of Colorado
295 UCB
Boulder, CO 80309-0295
USA
rood at colorado.edu

On Fri, 1 Apr 2005, Pamela Munro wrote:

> Hm, maybe I have another argument against defining "argument" narrowly
> as "what can be marked on the verb". My Lakhota isn't ready enough to
> supply an example, but I would certainly guess that there are some
> intransitive verbs out there that can take only a singular nominal
> (third person) subject. (I want one that can take a nominal subject,
> not, say, a phenomenological verb that takes no subject, but I want one
> where said subject is incompatible with plurality, so we can't get -pi
> -- maybe I'm reaching, but I would guess there are such verbs.) In this
> case (if such a verb exists), we have a verb that I think everyone would
> want to say has one argument (the subject), but where nothing is marked
> as subject on the verb. In other words, it seems to me that ability to
> show marking on the verb need not be directly related to the number of
> arguments a verb takes. (I think that's a matter of what used to be
> called subcategorization.)
>
> Pam
>
> Koontz John E wrote:
>
> >On Fri, 1 Apr 2005, Pamela Munro wrote:
> >
> >
> >>Well, if you define "argument" as "something marked on the verb" (which
> >>is fine, and I thought might be the case), this seems fine -- but in
> >>fact, I think there are other ways to go. For instance, you really can't
> >>freely add random nouns to Lakhota sentences. I think k'u really is
> >>semantically and syntactically completed with three associated
> >>"participants" (to choose a different term, which I often use when
> >>talking with people who will be confused by "argument). I think each of
> >>the three is fully as much a part of the sentence as, say, the subject
> >>of an intransitive -- thus, I think this is a structural feature, not
> >>just a semantic one. (I thought this before, but wondered if you knew
> >>any obscure syntactic tests I was unaware of.)
> >>
> >>
> >
> >I agree with this.  I'm not really sure how the term argument should be
> >used, and I don't want to do violence to accepted usage for it.  However,
> >by restricting themselves historically to seeing only the arguments (or
> >participants) defined by the canonical property of indexing in the verb
> >Siouanists have been missing several important categories of verbs. If
> >non-indexed arguments are considered, then we don't have just actives,
> >statives, and transitives, we also have ditransitives like k?u and the
> >"experiencer subject" pattern(s).  (I just slipped a letter or so up and
> >said "dative subject" where I meant "experiencer subject").
> >
> >
> >
> >
>
> --
> Pamela Munro,
> Professor, Linguistics, UCLA
> UCLA Box 951543
> Los Angeles, CA 90095-1543
> http://www.linguistics.ucla.edu/people/munro/munro.htm
>
>