Re: Active languages

Quoting Jörg Rhiemeier <joerg_rhiemeier@...>:

> Hallo!
>
> Thomas Wier wrote:
>
> > Patrick Littell wrote:
> > > They describe
> > > Choctaw as having neither a passive nor an antipassive, and Lezgian
> > > as having both. (Although Lezgian is sometimes described as ergative,
> > > it's active by their definition of active.) Lezgian does not appear to
> > > have a *morphological* passive or antipassive voice, though; it appears
> > > to simply allow the agent or patient to go unexpressed.
> >
> > Do they run any of the standard tests for a change in grammatical
> > relations?  I ask, because just stating that arguments can go
> > unexpressed is a pretty naive way to talk about voice systems.
>
> It is.  In Old Albic (my conlang), each (or both) of the arguments
> of a transitive clause can go unexpressed, with the verb not carrying
> the agreement markers:
>
> (1) Lennára               hena      lennas.
>     sing-PRES-3SG:P-3SG:A child-AGT song-OBJ
>     `A child sings a song.'
>
> (2) Lennara         hena.
>     sing-PRES-3SG:A child-AGT
>     `A child sings.'
>
> (3) Lenná           lennas.
>     sing-PRES-3SG:P song-OBJ
>     `A song is sung.'
>
> (4) Lenna.
>     sing-PRES
>     `There is singing.'
>
> Yet, I call these "zero-agent" and "zero-patient" constructions
> and avoid the terms "passive" and "antipassive" because I don't
> think these can be properly called "verb voices", lacking any
> kind of morphological marking on the verb (other than the
> absence of agreement markers) or changes to the cases of the
> nouns.
My Tairezazh does essentially the same. I've never thought of this as voices;
indeed, the original reason for introducing it was getting rid of passives.

A voice interpretation would be rather perverse, since the object of a
subjectless verb can't be the subject of a coordinated verb, but can be the
object of one.

                                                     Andreas

Reply