Algonquian Parallel? Muskogean Parallel?

[Koontz John E]

On Tue, 1 Oct 2002, Pamela Munro wrote:
> You're not hallucinating, John. Muskogean languages certainly have "a
> class of experiencer verbs", by which I assume you might mean dative
> subject verbs (if people will allow me that phrasing).

I've been trying to avoid that in the Siouan context because (in OP,
anyway) there are two forms of experiencer subject verbs, ones that use
the simple stative series of pronominals, like dhiNge 'to lack' (aNdhiNge,
dhidhiNge, dhiNge), and ones that are formed on dative stems and use the
dativized patient series, like git?e 'one's own to die' (iNt?e, dhit?e,
git?e).

> So we have, like Siouan, a typical "active" system with (as I call them)
>
> -- class I subject intransitive verbs (similar to e.g. Lakhota wa- 'I'),
> primarily active;
> -- class II subject intransitive verbs (similar to e.g. Lakhota ma-
> 'I'), primarily non-active;
> -- class I subject, class II object transitive verbs.
>
> We also have class III "datives", which can be either objects or
> subjects. As objects, they can be the only object (with verbs like
> Chickasaw i-hollo 'love' [I am writing nasal vowels as underlined; if
> this does not come across in your email let me know and I can send
> this to you another way],

No trace of underlining for me.  I assume this is iNhollo, then?

> which takes a I subject, III object [with the dative prefix im-]) or
> can be a second object added to an ordinary transitive, as with
> im-pilachi 'send to'. As subjects, they are typically intransitive
> (e.g. in-takho'bi 'be lazy').

Thus it seems that there are both I, II transitives and I, III
transitives, as also in Mississippi Valley Siouan, where the latter are
like transitive dative stems (and datives stems are usually transitive, of
course).

Cf. naN?aN' 'hear'  (anaN'?aN 'I hear it', aNnaN'?aN 'he hears me', ...)
    gi'naN?aN 'hear for' (e'naN?aN 'I hear his', iN'naN?aN 'he hears
        mine', ...)

> However, we also have occasional transitive II and III subject verbs,
> such as banna 'want' (II subject) and im-alhkaniya 'forget' (III
> subject). These, like the 'lack' verbs that have been the subject of
> recent discussion, take a subject that may be non-third person, but must
> have a third person object:  ... (exx.)

And git?e' is an example of a III subject verb.

Of course, it's a bit of a stretch, though not an unnatural one, to speak
of I vs. II vs. III in a Siouan context, since it's more like Pronouns {I,
II} x Stems {non-dative, dative}, where I is agent and II patient.  While
it's true that "dative" involves raising a less direct object to character
of concord-governing primary object, an effect on the interpretation of
the II (patient) pronominal, the morphophonemic effect on the vowels of
the verb having a dative marker in the mix applies to both the I and II
pronouns.  There are really four series of pronominals I, II,
I-in-the-presence-of-dative, and II-in-the-presence-of-dative.  In OP some
dative forms don't even use the modified series - the effect of the dative
marker can be absorbed by the locative or just not occur.

> ... There are numerous other syntactic subject tests, too, [apart from
> case marking of NPs] all of which agree on what the subject is here.)

I wonder if any of these would apply in Assiniboine for Shannon?  Is there
a reference one could consult?

> Further: In Chickasaw only one object can agree. So 'send to' for
> example cannot have a non-third person patient, even though the simple
> transitive pilachi 'send' can. (In some languages, such as Choctaw, this
> is not the case, and you can get three agreeing arguments on a verb.)
> Moreover, some Choctaw speakers allow a non-third person patient for
> 'want', so you can have two II markers on the same verb. No Chickasaw
> speaker I've worked with allows this, though.

This variety of treatments of secondary objects sounds like it might be
relevant in the Siouan context, too, since things seem to be different in
Dakotan and Dhegiha and maybe even in Omaha-Ponca vs. Osage and Kaw.

JEK