OP stative verb ablaut?

[Koontz John E]

On Fri, 13 Feb 2004, Rory M Larson wrote:
> Here, I'm going to argue a little.  It seems that you are taking
> verbs from three different grammatical classes, combining them
> into one which you call "experiencer verbs", and holding them as
> exclusive of stative verbs.

I assume the three classes are dhiNge' (morphologically a simple root),
git?e (gi DAT + root), and the wa-forms (wa-root).  I do agree that these
are three different morphosyntaxes.  And I do agree that they all inflect
as statives in the traditional sense of 'one patient a stative makes',
modulo the peruliarities of inserting the patient pronominals into these
three different morphosyntaxes (aNdhiNge, dhidhiNge, ...; iNt?e, dhit?e,
...; aNwaNkhega, wadhikhega, ...).  However, the traditional analysis
mixes morphological critria (how many inflections of what kind) and
syntactic criteria (each and only each inflectional slot corresponds to
one argument, so once you know the inflections of the verb, you know the
argument structure of the verb).  It's this last assumption that I believe
turns out to be too facile, and so, in fact, the morphological statives
(the one-patient verbs) turn out to conceal two synctactic classes of verb
(the statives - as traditionally assumed - and the experiencer verbs - as
we've been discussing them).  Note that I'm not sure that all Siouan
languages have lots of experiencer verbs.  They seem to be pretty common
in Mississippi Valley.

> My understanding of "stative verbs" has always been that they are a set
> of words approximately equivalent to English adjectives, except that
> they conjugate as verbs, using the object, or patient, pronouns only.
> This would be a functional/morphological class.  By that definition,
> dhiNge' would be a stative, albeit a bit unusual in intrinsically
> referencing a non-existent 3rd person object in addition to an optional
> patient subject.

Exactly - the traditional morphological definition works, but the syntax
doesn't.  DhiNge' is approximately equivalent to an English adjective.
It's approximately equivalent to an English transitive verb.  And git?e is
approximately equivalent to something that doens't even exist as a working
verb type in English, leading to glosses like 'for one's own to die'.
It's true that uz^edha and wakhe'ga are equivalent in some degree to
English adjectives, but the first is also equivalent to 'to have (a) tired
...' or 'to be tired in/through/of one's ...'.  That's where the syntax
breaks down there.

As I pointed out in a preceding letter, if an experiencer verb
obligatorily incorporates it's patient, either as an incorporated noun or
was wa-, it seems to me on further reflection that it becomes a derived
stative.  The morphosyntax includes an extra patient, but the syntax no
longer does.  So, on that logic I'd have to concede that naNppe=...hiN 'be
hungry' and was^u's^e 'be brave' are, in fact, statives.

> This is just a nuance of the allowed argument list of the verb, much as
> 'give' differs from 'steal' in English.

Pondering this, 'give' is ditransitive (also transitive and intransitive),
while 'steal' is transitive (also intransitive) (but does admit some
peripheral arguments, too).  Maybe 'steal' vs. 'rob' would be a better
example of subtile differences?  But I'd say that dhiNge vs.  ttaNga is
actually more like a 'give' vs. 'steal' - a significant difference in the
number of arguments admitted.

> I think this argument depends on what we believe that wa- primordially
> represents.  Your view seems to be that it is originally a patient
> marker.

Yes, and there I'm afraid I still tend to prefer that analysis, though
plainly there are two or three different views on wa- among Siouanists.

> ..., it does seem to be the case that wa- can be added fairly freely to
> many unquestionably stative verbs, and the conjugation follows the
> expected (stative type) wa- paradigm.  Examples the speakers approved
> included:
>
>     ski'ge              heavy (objective description)
>     oNski'ge            I am heavy.
>     dhiski'ge           Thou art heavy.
>     waski'ge            We are heavy.
>
>     waski'ge            heavy (person is stout)
>     oNwoN'skige         I am heavy.
>     wadhi'skige         Thou art heavy.
>     wawa'skige          We are heavy.

This is nice example.  I'd argue again that wa- in the second set refers
to the body or parts of it.  So, both verbs are stative, but the different
semantics of the second derive from its experiencer morphosyntax.

I'm not too surprised to find a verb (or here, really, just the
root) alternating between stative and
experiencer with no marking, because this is a common pattern, e.g.,
English has verbs like 'give' and 'steal' mentioned above that can be used
without derivation with different patterns of argument:

I gave him a sandwich.  (ditrans)
I gave suitable gifts.  (trans)
I gave at the office.   (intrans)

I stole him a cookie.   (ditrans)
I stole a cookie.       (trans)
I stole constantly.     (intrans)

And there are other patterns, like:

It rolled down hill.    (intrans)
He rolled it down hill. (trans)

> My feeling is that prefixed wa- is to statives about what
> postfixed -s^toN is to active verbs in the sense of indicating
> characteristic as opposed to specific.

Dorsey has one case where he glosses this 'chronic'.  My kids would love
that.  However, I am happier with wa- referring chronically, albeit
sometimes obscurely, to patients.