ablative as ergative marker

[Scott DeLancey]

On Thu, 22 Oct 1998, Ralf-Stefan Georg wrote:

> My Thakali example may not be watertight (since INS is included),

and

> Theorising apart, where are clear examples of ABL/ERG (ideally excluding
> other functions, such as INS) ?

The ERG/INSTR/ABL syncretism is common in Tibeto-Burman (in a paper
in Linguistics of the Tibeto-Burman Area 8.1 (1984) I list Lotha Naga,
Newari, Geman Mishmi, Hani, and Kanashi as examples of languages where
they're identical, and quite a few more where they seem to share
material, or where comparative data suggests a shift from ABL -> ERG.
It seems to be the normal pattern in TB for ERG and INSTR to be the same
form, so there won't be many examples of ERG=ABL excluding INSTR use.
But then, I'm not sure I see why that's important.

Very interesting are cases like Classical Tibetan, where ERG/INSTR
and ABL aren't identical, but share part of their form:

		LOC na		ABL nas
		GEN kyi		ERG kyis

where one could argue that the common -s element is what defines
both the ERG and the ABL form.

> I hope I'll be forgiven for sharing indegested musings with the list,
but:
> could it be that the ABL function of this ERG marker is rather the end of
> the story than its beginning ? But maybe this would entail a
> grammaticalisation of a more concrete, spatial notion (source) out of a
> more abstract, gestaltless one (agent), something which may be hard to
> swallow, I'm aware of that (some people may even wish to rule this out
> completely ??).

I don't think we can rule this, or anything, out as completely
impossible--just about anything can happen once.  But I think it's very
unlikely as a recurring process, and whatever produces the widespread
ERG=ABL syncretism in Tibeto-Burman must be a fairly natural process,
not a one-time oddity.
     Anyway, there's enough other evidence for ablatives turning into
agent markers (e.g. the common use of ablative prepositions to mark
transitive agents in European and other languages, and the acquisitional
evidence in the Clark and Carpenter paper I mentioned in a previous
note) that there's no need to imagine any other explanation for this
kind of syncretism.

Scott DeLancey	
Department of Linguistics
University of Oregon
Eugene, OR 97403, USA

delancey at darkwing.uoregon.edu
http://www.uoregon.edu/~delancey/prohp.html