Active alignment beyond intransitive sentences

[Trebor]

I understand that languages with active alignment mark the subject of intransitive sentences in different ways depending on the degree of control it has over the action. So, a person who is walking will be marked as an agent and a person who is sleeping will be marked as a patient (in split-S systems), or a person who is running will be marked as an agent if doing so voluntarily and as a patient if doing so involuntarily (in fluid-S systems).

I also know that languages with accusative alignment may mark a subject without control over the action in a special way. But the examples provided in the relevant Wikipedia article come mostly from the Indo-European family, and comprehensive glosses are not provided for the Georgian data. I have had little success in locating the kind of information I am seeking elsewhere.

So, I would like to find out how split-S and fluid-S languages deal with subjects, direct objects, and indirect objects given their active alignment, in terms of:

- word order, particularly if case-marking and verb-argument agreement are absent;

- case-marking, no matter the amount of person/number inflection possible on the verb; and

- verb-argument agreement, especially if polypersonal marking is present.

Thank you for any information you can offer. Links to online articles and book recommendations are welcome.

[Trebor]

Well, I've managed to find a little information on this topic, at last (I wasn't using good keywords):

1) According to this paper (PDF), "Active-Stative Agreement in Choctaw and Lakota", and this page, "Sketch of Lakhota, Pt.II", this language has a small number of transitive verbs which are stative rather than active.

2) According to this summary of a linguistics address (PDF) (alas, only one page in length), "The active/stative split of verbs in Baure (Arawak)", this language has some such verbs as well, a phenomenon described as "highly unusual".

Is it really the case that split-S and fluid-S systems generally abandon their concern for matters agentive and patientive when multiple arguments are involved?

[vokzhen]

Is it really that surprising? There's a strong cross-linguistic correlation with transitive subjects being agents, and objects being patients. Active-stative languages can be thought to borrow these features into the intransitives, but they're still present in transitive constructions. That is, the agentive and patientive aren't abandoned in transitives, but are upheld strongly: the subjects are agents and the objects patients. And like other languages, they probably (though I have no data on hand) allow other constructions for subverting transitive subject agentivity, like ambitransitives, passives, and oblique-marking (the last I know is an argument for pre-PIE being active-stative, as is a high number of pairs like see/look and hear/listen, though they seem to be opposite what I'd expect with the transitive being patientive and the intransitive being agentive).

There might be morphological grounds for this as well, according to WALS active languages are primarily head-marking over case-marking (28 to 4), and the vast majority of those (23 of 26) mark both agent and patient roles (due to how the data's presented, there's no information on the possibility of zero-marking active languages). Allowing subjects to be patientive might simply create too complex a verbal system to be stable, or might never arise in the first place if e.g. using dative marking for nonagent subjects is a much easier solution.

Tangentially, from my understanding, perfect fluid-S languages are rare or nonexistent, with most/all having a subset of highly active intransitives, like your example run, that require agent marking, and/or a class of highly inactive statives like be.hungry or shiver that require patient marking.

[Trebor]

It does seem reasonable that languages wouldn't complicate matters by reusing intransitive patientive declensions/conjugations for the subject of a transitive verb, given that it will often be the agent anyway.

But I cannot seem to find much information about what alternative strategies are used in active languages to mark a non-agentive subject. If so many such tongues [edits: are head-marking], then surely linguists will have something to say about how they implement polypersonal agreement...

[gach]

It's a bit unusual but there are a more languages that have split marking for transitive agents and patients as well as intransitive subjects. This paper (PDF) discusses a few more examples where dative agreement is introduced and allowed for some or all types of core arguments of the verbs.

A universal characteristic of dative arguments is their low affectedness. Thus in transitive sentences they can create both agents of low agentivity (as with verbs like "to forget" and "to remember") and patients of low patientivity (as with verbs like "to hate" or "to help"). Maybe of interest are also double accusative verbs (with both A=ACC and P=ACC) that can also signal low transitivity. The above paper gives "to forget" as an example of such in Saweru and, despite not mentioning this agreement type for Choctaw, the verb "to want" is supposed to agree like this in it.

Then, if we have dative head marking for core arguments, should we also consider more familiar European dative subjects or other wider quirky case systems as split marking not that dissimilar to canonical split intransitivity?

[Trebor]

Thanks for that paper, Gach--it was interesting reading.

Then, if we have dative head marking for core arguments, should we also consider more familiar European dative subjects or other wider quirky case systems as split marking not that dissimilar to canonical split intransitivity?

Yes, such makes a lot of sense. Are there any active-stative languages with quirky subjects and polypersonal marking?

[sirdanilot]

This field within descriptive linguistics is a dark spot to me (with which I mean split-transitivity, ergativity, antipassives etc. etc.). Waaay too complicated, haha, so I hope I don't have to deal with this kind of thing when doing fieldwork. I understand the general gist of ergative languages but that's about where it stops for me.

There's a large book by Dixon on this thing (forgot the exact title), perhaps I should read it sometime...