Subjects of Causatives

[vegfarandi]

Are there languages that mark the subject of causatives with an oblique case, while retaining the original nominative/ergative marking for the original subject? I'm toying with this in Imutan, but wondering if this is naturalistic.

[akam chinjir]

I don't know how it would work with causative marking on the verb, but if you allow ablative (or whatever) cause phrases, and then let those be topicalised, you might get pretty close to what you want.

(Something like I died → I died of hunger → Of hunger I died.)

...But I don't know how well that could really work when the cause is an agent. (I ate → I ate because of Charlie → Because of Charlie I ate doesn't really get you to Charlie fed me.)

[Xwtek]

Are there languages that mark the subject of causatives with an oblique case, while retaining the original nominative/ergative marking for the original subject? I'm toying with this in Imutan, but wondering if this is naturalistic.

It's called benefactive applicative

[zompist]

Are there languages that mark the subject of causatives with an oblique case, while retaining the original nominative/ergative marking for the original subject? I'm toying with this in Imutan, but wondering if this is naturalistic.

It's called benefactive applicative

What? A benefective isn't a causative; it's, er, a benefactive. That is, it says who benefits from an action, not who caused it.

A typical discussion of "benefactive applicative" I could find was this paper, which includes examples translated "Buy it for me", "He went on an errand for me", "He died for me".

[circeus]

Are there languages that mark the subject of causatives with an oblique case, while retaining the original nominative/ergative marking for the original subject? I'm toying with this in Imutan, but wondering if this is naturalistic.

It would not be naturalistic to have an oblique causer if the causer is a syntactic subject. That is only sensible as causative processes by definition add an additional core argument to the verb that becomes the subject.

Retaining the subject markings on the original subject is entirely possible (there are languages that do so), however, no language marks the causer in a causative with a nonsubject marking (because of the above definition). If you do, the construction may be semantically a causative, but syntactically what you get is really more akin to some sort of passive:

Michael ate → Michael was fed by John

[akam chinjir]

Retaining the subject markings on the original subject is entirely possible (there are languages that do so), however, no language marks the causer in a causative with a nonsubject marking (because of the above definition). If you do, the construction may be semantically a causative, but syntactically what you get is really more akin to some sort of passive:

No room for non-canonical subject-marking here?

...isn't at least some ergative case marking supposed to derive from oblique-marked agents in passive constructions getting reinterpreted as subjects? What's to stop the same sort of thing from happening with (other) causatives? (Granted, if there's good reason to think that's never happened, it would be good to have some idea of why.)

[circeus]

Retaining the subject markings on the original subject is entirely possible (there are languages that do so), however, no language marks the causer in a causative with a nonsubject marking (because of the above definition). If you do, the construction may be semantically a causative, but syntactically what you get is really more akin to some sort of passive:

No room for non-canonical subject-marking here?

...isn't at least some ergative case marking supposed to derive from oblique-marked agents in passive constructions getting reinterpreted as subjects? What's to stop the same sort of thing from happening with (other) causatives? (Granted, if there's good reason to think that's never happened, it would be good to have some idea of why.)

This is not a development I'm familiar with, but you're not providing any counterpoint since the original construction was still not a causative. Once it became the subject, that marking was the normal subject marking.

Again, the thing is that a causative is formally defined by the raising of a a causer to subject. By this definition, it is entirely impossible to have a "true" causative construction if the causer is not a subject. You can create a construction with a causative meaning, but it's not a causative construction (my terminology may be a little fuzzy here and I apologise for it. I hope it does not make the argument incomprehensible). WALS has two chapter discussing causatives, and in neither the idea of a non-subject causer is ever entertained in any way.

Dixon, R.M.W. 2000. “A typology of causatives: form, syntax and meaning”. In Changing Valency: Case Studies in Transitivity, Dixon, R.M.W. and Alexandra Y. Aikhenvald, eds. New York: Cambridge University Press. p.30–83.

The typology is summarized here. Note that causer are invariably marked "A" (i.e. with the language-appropriate mark of the subject of a transitive verb).

[gestaltist]

What you're describing sounds like a causative case to me. I don't think it's reconcilable with the causative voice since raising the causer to the role of the syntactic subject is in the definition of the causative voice. But I see no reason why you couldn't have a causative case used with the usual active voice (or even requiring some additional syntactic quirk):

He-NOM killed John-ACC Mark-CAUS. = "He killed John, and Mark made him do it."

[Pabappa]

What you're describing sounds like a causative case to me.
....
He-NOM killed John-ACC Mark-CAUS. = "He killed John, and Mark made him do it."

If it helps, my conlangs do this, and in most of them the causative case is identical with the locative. Haven't had any problems so far.

[akam chinjir]

This is not a development I'm familiar with, but you're not providing any counterpoint since the original construction was still not a causative.

The idea was that it would become a causative, once the topicalised because of X phrase got reinterpreted as a subject.

Once it became the subject, that marking was the normal subject marking.

Yeah, good point.

I guess I also don't really see how the original subject keeps its original subject marking.

[Salmoneus]

If vec was just talking about semantically causative situations (because, as pointed out, a syntactically causative construction is precisely one where the cause is NOT oblique), then English qualifies: Alice fed Bob at the instigation of Charles. Of course, we use a complex preposition, but there's no reason a language couldn't use a case.

[Imralu]

Isn't the whole point of a causative verb that it promotes the cause to the subject case and demotes the old subject, generally to the object?

To me, this question is a bit like asking "Can we have passive verbs with the agent marked in the subject case and the patient marked as an object?"

[Pabappa]

I have never understood causative verbs at all ... i kept quiet in this thread because I didnt have anything to say until the question about the causative noun case came up, and even that is only something i know through my conlang. but i get the impression causative verbs are just a cover term for different things that can be handled different ways in different languages. maybe some of the time the agent gets demoted to an absolutive case and the patient turns into something else too, but that would only come up if you had at least three noun cases you could use to distinguish them, or had a specific word order in use for causative verbs only.

The archaic Germanic causative verbs, at least, didnt change the case structure at all .... " I drink water" and "I drench the cloth" both used nominative for "i" and accusative for the object. the third noun, if present, would just have been in a dependent clause.

edit: wait, i see waht youre saying now ... "cloth" could be considered an agent because its the thing that's soaking up the water, and in this sentence the real patient is just not used in the sentence. still, i dont know, there might be other ways to do it.

[akam chinjir]

Isn't the whole point of a causative verb that it promotes the cause to the subject case and demotes the old subject, generally to the object?

But the question was about case-marking, not about whether you get a new subject and so on. I mean, in plenty of languages the experiencer argument of certain psychological verbs takes dative case-marking, it doesn't mean those verbs don't have subjects.

[kodé]

Isn't the whole point of a causative verb that it promotes the cause to the subject case and demotes the old subject, generally to the object?

But the question was about case-marking, not about whether you get a new subject and so on. I mean, in plenty of languages the experiencer argument of certain psychological verbs takes dative case-marking, it doesn't mean those verbs don't have subjects.

I was confused about this for a while. I think what was getting me was “mark the subject … with an oblique case”, since the way “subject” is typically used doesn’t jibe with having oblique case. I’m guessing veg meant a semantic causer, so that the case marking would be similar in spirit to an instrumental or to how the ablative is used in Latin sometimes, but with the argument being animate and volitional. I can’t think of a language that does something like this consistently, though—if you introduce a semantic causer, there’s almost always a change in argument structure and the causer usually gets promoted to subject, likely because it’s volitional and agentive.

Hmm… would the marking of agents in non-active voices (or however you wanna call it) in Philippine type alignment systems work? In the language I know best with this type of alignment, SaySiyat, you mark the agent of a patient focus/patient voice verb with the genitive and it’s arguably not an argument syntactically. So I guess I’d have to see what you get when you have the patient focus/voice form of a causativized verb.