On Thu, 19 Jun 2008 22:44:07 +0300, JR <[log in to unmask]> wrote:
>On Thu, Jun 19, 2008 at 9:04 PM, Jeffrey Jones <[log in to unmask]>
>> On Thu, 19 Jun 2008 14:20:12 +0300, JR <[log in to unmask]> wrote:
>> >
>> >On Thu, Jun 19, 2008 at 12:36 PM, Jeffrey Jones 
<[log in to unmask]>
>> >wrote:
>> >
>> >> On Thu, 19 Jun 2008 10:30:03 +0300, JR <[log in to unmask]> wrote:
>> >> >
>> >> >On Thu, Jun 19, 2008 at 4:38 AM, Jeffrey Jones
>> >> > <[log in to unmask]>
>> >> >wrote:
>> >> >
>> >> >> In Naisek, there are some grammatical voice prefixes, including a
>> >> >> causative and two kinds of passive. Another one adds a dative
>> >> >> subject, usually indicating a perceiver, to verbs which otherwise
>> >> >> have patientive subjects. ....
>> >> >>
>> >> >> Jeff
>> >> >
>> >> > Do I understand correctly that this is used to de-emphasize an
>> >> > already-existing patientive subject, and not to add a new argument?
>> >>
>> >> No, it adds a new argument, although the role of the dative may be
>> >> implicit.
>> >>
>> >> > Does the new dative trigger agreement in the verb (if there is
>> >> > agreement at all)?
>> >>
>> >> Yes.
>> >>
>> >> > Can it be used with any verb with a patientive subject? If not, how is
>> >> > its range limited?
>> >>
>> >> I think so. I forgot to say it's also used with modal auxiliaries.
>> >>
>> >> > When you say the dative subject "usually" indicates a perceiver, what
>> >> > are the other possibilities?
>> >>
>> >> Someone shaded or rained on, as in examples (1) and (2), and another
>> >> role I don't know the name for in (5).
>> >>
>> >> Note: ABS = patientive, ERG = agentive
>> >>
>> >> (1) Ho-paldu-bwe.
>> >>    ???-rain.PRS.IND-1NPD
>> >>    "It's raining on us."
>> >> Technically, paldax is impersonal, but there can still be a patientive
>> >> argument.
>> >> (2) Hi korilo su-bwe ho-naus-in.
>> >>    DEF-INA.S.ABS hazel.S.ABS COP.PRS.IND-1NPD ???-shade-VN
>> >>    "The hazel shades us."
>> >> (The COP copula + VN verbal noun construction = habitual aspect.)
>> >> (3) T-a-m matse t-i ho-jad-en-ti xiskw-e.
>> >>    3-ANI.S-GEN mother.S.DAT 3-INA.S.ABS ???-seem-FUT.IND-3SD
>> >>    terrible-SPO
>> >>    "It will seem terrible to her mother."
>> >> (4) Ho-laun-os-t-ki gaut-a daxme tep nu.
>> >>    ???-beautiful-CMP-VRB-1SD NUL-ANI.S.ABS woman.S.ABS than
>> >>    2S.ABS
>> >>    "There's no woman more beautiful to me than you."
>> >> (5) T-i juku ho-laip-ti-twe.
>> >>    3-INA.S.ABS very ???-easy-VRB-3PD
>> >>    "It's very easy for them."
>> >> (6) Johanna-i ho-bof-ti lo Tomas-a disp-ax-ta.
>> >>    Joan-DAT ???-modal.PRS.IND-3SD CPL Tom-ERG dance-SUB-3SE
>> >>    "Joan thinks Tom should dance."
>> >
>> >So far it looks like a circumstantial voice, like that of Malagasy ...
>> >though differing in the specifics. Can you give the "normal" versions of a
>> >few of these sentences, though, without using the construction in
>> >question?
>> >Say, numbers 1, 2, and 6. And what's the meaning of the modal in 6?
>> >
>> >Josh
>> I googled for circumstantial voice and it doesn't look like it qualifies.
>> (1a) Paldi.
>>    rain.PRS.IND.3SA
>>    "It's raining."
>> (2a) Hi korilo si naus-in.
>>    DEF-INA.S.ABS hazel.S.ABS COP.PRS.IND-3SA shade-VN
>>    "The hazel provides shade."
>> (6a) Tomas-a bof-ta disp-ax.
>>    Tom-ERG modal.PRS.IND-3SE dance-INF
>>    "Tom should dance."
> In the circumstantial voice (to answer Eldin here as well) an oblique
> argument is promoted to subject. That's the core of it.

I was going by what Wikipedia said (the only other references were all google 
books), which seemed to indicate that the circumstantial voice couldn't 
increase valency.

> What I really wanted to see when I asked for "normal" versions of those
> sentences was how experiencers were encoded in Naisek outside of this
> construction. You left them out altogether, but unless they're being elided
> here, this shows just as well that they are oblique, i.e., not core arguments
> required by the verb.

They're not expressable at all in the normal versions -- although they may be 
implicit participants in some cases -- so I understood that to mean they aren't 
even oblique arguments.

> The fact that these experiencers, when promoted to subject, appear with
> dative case, I would not attribute to the operation in question at all.
> According to your web page, dative case is used for experiencer-subjects
> (at least sometimes?) even in active voice, so I'd think this is just another
> application of the same principle of case assignment.

That's true; you could say that all the prefix does is make the experiencer 

>Of course your construction here is different from Malagasy not only in the
>case of the new subject, but in the types of oblique arguments that can be
>promoted in the first place. In Malagasy, it can be used for instruments,
>times, locations, beneficiaries, manners - perhaps any oblique, but I don't
>really know. In Naisek it's limited to experiencers. But the operation
>itself seems the same. You would just have to specify its range - you could
>even call it a Experiencer-Circumstantial voice if you want.
>Finally - instead of saying that the subject in the original structure must
>be a patient, would it be better to say simply that the *event* described
>must be some perceivable/experienceable state? Perhaps it amounts to the
>same thing most of the time, but from what I've seen, I wonder if this isn't
>the true determining factor. I'm especially troubled by (1), where there is
>no subject at all. Even if there could be one in another sentence with the
>same verb, there isn't one here (I assume the behavior is like that of the
>English 'rain', which can take a patient, but usually doesn't). Even aside
>from that though, it makes more sense to me that the possibility/behavior of
>an experiencer of an event would be dependent on some characteristic of the
>event as a whole, rather than the semantic role of any one argument.
>What do you think?

I think I'll have to look at a bunch more verbs (which I haven't invented yet). 
But it's definitely true that a verb that already has a dative case argument or 
an agentive case argument can't take this prefix.

>Eldin, in applicatives, an non-direct object, or oblique argument is
>promoted to direct object, as I understand. I don't see that ocurring here
>though. Not really sure about good web resources for the circumstantial -
>there's a little here, a little there. There's a page on wikipedia, but I
>think there's a mistake in it.

WRT mistake, what do you have in mind?

>(ObAFMC) Khafos has a circumstantial voice which is pretty close to
>Malagasy's, though it's used for indirect objects as well as any other
>oblique arguments. AFAIK, IOs are in Malagasy are promoted with the normal
>passive voice, just like DOs.