Re: Rinya cases

Daniel Andreasson wrote:

> Okay. I've got a serious case problem. My Rinya is
> ergative. As you know this means it marks semantic
> roles instead of syntactic functions. Now I've reached
> the point where I noticed that I could distinguish
> between actions that are volitional and those that are
> not by using different cases.
>
> In intransitive (itr) sentences this is not a problem.
> I either use the ergative case, which marks the agent and that
> the action is made on purpose), or the absolutive case
> (in most cases the subject of itr sentences is patient
> or experiencer) which marks that the action is done
> involuntarily.
>
> Now to the real problem. Transitive sentences with verbs
> of which the subject isn't the agent. To show that that
> the action is non-volitional the "subject" should be marked
> with the absolutive case. But then what happens to the
> "object"? I can't use the ergative for that.
>
> I'm thinking of introducing a new case to mark the
> Influenced Object/The Experienced, OBJECTIVE.
>
> Example:
> "Will looked at a mouse."
> Will is agent, he deliberately looked at the mouse.
> He was acting. Will = Erg. ; Mouse = Abs.
>
> "Will saw the mouse."
> Will is the experiencer, and mouse is the experienced.
> Will didn't act, he just happened to see the mouse.
> Will = ABS. ; Mouse = OBJ.
Hejsan, Daniel!  Hur gaar det med dig?

The solution you propose above seems reasonable to me.
As Christophe suggested, you could also use one of your
preexisting cases (dative or instrumental) to mark objects
in non-volitional transitive clauses, rather than inventing
a new case.  If so, then you'd have what I'd call a "covert
antipassive" construction (i.e. an antipassive construction
in which the form of the verb does not change).  Since
antipassivisation is associated with lack of volitionality
on the part of the agent in lots of ergative natlangs, this
would work out nicely, I think.

In my conlang Tokana, lack of volitionality on the
part of the agent is marked by having the agent noun phrase
in the instrumental case rather than the ergative case:

   Na      Tsion kahten      ikei
   the.Erg John  hit-the.Abs dog
   "John hit the dog" (intentionally)

   Inan     Tsionne   kahten      ikei
   the.Inst John-Inst hit-the.Abs dog
   "John hit the dog" (accidentally)

Note that the instrumental case is mandatory if the actor
is indefinite, or inanimate, or both:

   Ne      ikei kahte miohne
   the.Abs dog  hit   someone-Inst
   "Someone hit the dog"

   Ne      ikei kahtei       nakane    tiausi
   the.Abs dog  hit-the.Inst rock-Inst falling
   "The falling rock hit the dog"

A final note:  Your characterisation of ergative languages
as "languages which mark semantic roles instead of
grammatical functions" is not quite right.  In prototypical
ergative languages (like Basque, Dyirbal, and Tzotzil), the
ergative case is used for ALL transitive subjects, regardless
of their semantic role (agent, experiencer, source, theme,
etc.), while absolutive case is used for ALL intransitive
subjects and direct objects, again regardless of their
semantic role.  The kind of language you're talking about is
more properly called an "active" language.

Tokana, in which ergative case is used ONLY for definite
volitional agents, and in which ALL definite volitional agents
are marked with ergative case (regardless of the transitivity
of the verb), is an example of an active language.  One common
characteristic of such languages is the use of oblique
('non-core') cases to mark non-agent subjects.  For example,
Tokana uses the dative case to mark experiencer subjects of
transitive and intransitive perception/emotion verbs:

   Ineh    Tsione   kesta
   the.Dat John-Dat happy
   "John is happy"

   Ineh    Tsione   hielan      ikei
   the.Dat John-Dat see-the.Abs dog
   "John sees the dog"

From what I've seen, Rinya is more like a prototypical
ergative language, where the distribution of the ergative
and absolutive cases is determined primarily by valency
(whether the verb is trans. or intrans.), and only
secondarily by semantic factors such as volitionality.
Am I understanding things correctly?

Matt.