Matt wrote :

> In some ergative languages (I'm thinking especially of Australian languages)
> the ergative and the instrumental are homophonous, and can be considered
> a single case form.  Participants marked by this case are interpreted as
> agents if animate, and instruments if inanimate:
>         John-ERG  knife-ERG  chicken-ABS  killed
>         "John killed the chicken with the knife"
> Tokana, it seems, is like PL in that the ergative case is reserved for
> volitional animates.  Non-volitional and/or inanimate participants are
> marked with the instrumental case:
>         Na       Tsion     mukteh      hitol
>         the-Erg  John-Erg  closed-the  door-Abs
>         "John closed the door (on purpose)"
>         Inan      Tsionne    mukteh      hitol
>         the-Inst  John-Inst  closed-the  door-Abs
>         "John closed the door (accidentally)"
>         Itan      suhune     mukteh      hitol
>         the-Inst  wind-Inst  closed-the  door-Abs
>         "The wind closed the door"

Christophe's language is almost like that. Funny that he re-makes nat- and conlangs he didn't learn.
My languages also work like that : cases equate voices and derive from the verbs 'to be' (=equative), 'to have as inalienable feature' (=attributive), 'to use' (=instrumental), 'to make' (=causative), 'to suffer' (=patientive) with a tag making them *inalienable attributes* of the predicate :

I hammer a nail with a stone :
me-ERG stone-INSTR nail-PAT hammer.

I flatten field :
me-CAUS field-EQUA flat-thing.

I strengthen you
me-CAUS you-ATTRIB  strength.

I clothe you with a coat
me-CAUS you-ATTRIB (coat-INSTR) coat.


See the original message at