> From: J Matthew Pearson
> Ed Heil wrote:
> > The only really weird thing is, I wouldn't expect that an ergative
> > language would use the ergative/absolutive distinction with a
> linking verb
> > like "to be."  It seems to me like that would only be with
> actions.  Maybe
> > someone who knows more about ergativity can enlighten..
> I know a bit about ergativity, and it seems weird to me too.  In
> my conlang Tokana, which is sort of ergative
> (cf. the recent thread on active languages), the linking verb,
> equivalent to "be" in English, takes two
> absolutive arguments.

Yes, this appears unusual to me as well.  In amman iar, a split ergative
language, the S-function argument takes the absolutive case as would be
expected in an ergative system, but the predicate nominal of a linking verb
takes the comitative case and uses the linking verb 'atho' which literally
means 'to attach, join, connect'.

David is a man
meldan adhanas athas
meldan -0   adhan -as  atho    -as
David  :ABS man   :COM connect :PRES
David       man   with is connected

David E. Bell
The Gray Wizard

Wisdom begins in wonder!