accusative and ergative languages

[Wolfgang Schulze]

Let me briefly comment upon some of the recent arguments:

"Patrick C. Ryan" wrote:

> A sequence like: "Me hurt. Tommy did it" is a virtual ergative so far as I
> would judge. I have heard children speak in this way.

Thjis surely is not ERG-like! It prestent tense, we ould hvae "me hurts
- Tommy does it" (or so) which shows that "me" still is in teh ACC case
(without AGR on the verb - "-s" has a "dummy agent" as a trigger. The
same is true with structures like German "mich friert" etc...

> (3)
> However, in Language A,
>         noun(B)+abs. verb
> will be interpreted as an activity is performed by an unspecified agent--

This is true for only some ERG systems. In such cases, the verb very
often gets a plural morphology to install AGR with a "hidden" agent. In
many other ERG systems, the ommission of the "agent" leads to new
instransitve stzrictures with the inference ABS > AGENT (cf. I-ERG
boy-ABS bring > boy-ABS come etc.).

Discussing the possible PASS background of ERG structures, JEns finally
asks:

> A truly innocent question for information: Are there other avenues that
> are _known_ to have led to the creation of an ergative than the one
> starting from an old passive?

Sure, there are plenty such avenues! Most of them have to do with the
Silverstein Hierarchy (or its expansion). The less a potiential agent in
(stronger) transitive scenarios is thought to bear inherent agentive
features, the more it becomes likely that this "light agent" is marked
by something that strengthens its agentive role. Some options are:
(alienable) genitives (which can be extended to "heavy agents" on an
inalienable basis), loctives (esp. with very light (or secondarily
lightened) agents or weather phenomena etc...), instrumentals that are
metaphorized to "agent" markers in the context of anthropomorphization),
true "agentive" markers that are grammaticalized from e.g. a deicitic
source, topic markers... Another possibility (though related to the
strategies mentioned so far) is the reanalysis of 'active' structures in
an ergative perspective (note that I do not want to suggest any 'active'
typology for IE here, see my ealier postings): the S-split would then be
harmonized on an S=0 level. All these strategies are based on the
semantic or functional grading of the agentive (in terms of the
manipulation of "lightness" and "heaviness"). Naturally, ERG techniques
can also evolve from a special 'treatment' of the objective: One of the
most prominent one is that of syntactic and/or pragmatic foregrounding
which means that O is syntactically referred to as an intransitive S.
Such a technique may be equivalent to passive strategies, however, this
is only ONE of the many possible inferences.
	The syntactic/lexicical interface is touched upon when causatives of
intransitives form the basis for newly established ERG features: Here,
the morphosyntax of the causer can be introduced in the paradigm of
other 'true' agentives via analogy.
	Finally, agreement strategies may play an important role in the game.
If, for instance, agreement is coupled with some kind of person
hierarchy the presense of any SAP in a clause may condition agreement
irresepctive their functional or semantic role. Hence, a scheme nSAP:A >
SAP:O would necessariliy produce an erg-like AGR pattern (in case AGR
becomes active), whereas SAP:A >nSAP:O would produce ACC-AGR. ERG AGR
patterns may also result from the reinterpretion of clausal layers, e.g.
the structure SV // AOV could be read as SV // A[OV] which means that O
becomes some kind of closer attribute to the (participle-like) verb...

Wolfgang

[Please note new phone number (office) :+89-2180 5343]
___________________________________
| Prof. Dr. Wolfgang Schulze
| Institut fuer Allgemeine und Indogermanische Sprachwissenschaft
| Ludwig-Maximilians-Universitaet Muenchen
| Geschwister-Scholl-Platz 1
| D-80539 Muenchen
| Tel:	+89-21802486 (secr.)
|      	+89-21805343 (office) NEW ! NEW !
| Fax:	+89-21805345
| Email: W.Schulze at mail.lrz-muenchen.de
| http://www.lrz-muenchen.de/~wschulze/
_____________________________________________________