Universality of GR's

[Avery Andrews]

Joan Bresnan said:
 
> I think Chris Manning's point is very well put, and Avery's proposal
> is also cogent.  (But why call this recent work "lfg-like"?  

I guess I put it that way because some of the authors were trying to be
somewhat neutral w.r.t. LFG or HPSG, and at any rate they're all
proposing substantial innovations, which one can think about expressing
within more than one formal framework.

> For this reason, my own best guess about active languages remains that
> they have a relatively rigid mapping between roles and functions:
> Actor = Subject, Undergoer = (Primary) Object, in this lg type.

So how does this work out in the Philippine or Balinese-type
Austronesian languages, where the non-pivot Actor is like a traditional
object in certain respects, and quite different in others (Peter Austin's
Sassak stuff being a rather extreme example of this).

`Active' languages are also rather various in what phenomena they present;
Dixon (1994) says that they all wind up having some phenomena that unify
S_a and S_o, and that's certainly true for some of them (Crow (Siouan),
at any rate, where all S have number agreement with the verb in the
third person, while objects don't; but in Acehnese I don't recall any
such unifying phenemona that don't extend to all core arguments.  My guess
about Crow & Warlpiri is that they have the `pivot' (Chris' g-subject),
although why is not clear (it's not strongly associated with any pragmatic
functions, since there's no voice morphology to alter it's semantic role),
whereas there's no evidence for obligatory
g-subjects in Acehnese, although the optional preverbal position might be
an optional pivot position; this makes Acehnese similar to many other
Austronesian languages in having considerable freedom in which core argument
can be the pivot.

  Avery (hopefully not yet merely Avery-like, tho who knows what too much
   PERL does to the brain...