isoltes

[Alexander Vovin]

On Thu, 27 Mar 1997, Mark Hale wrote:
 
> Alexander Vovin writes:
>
> > Austric (as consisting of Austronesian and
> > Austroasiatic, but not including Kadai) is accepted nowadays by all
> > known to me leading scholars in both Austroneasian and
> > Austroasiatic.
>
> No, it isn't. Or rather, maybe it is, but this only indicates that
> you don't know very many Austronesianists.
 
I can only return the compliment: if my memory is correct you were not
present at CAMAC 1993, where Diffloth, Blust, Reid, Starosta and others
were present, and where the issue was discussed at length in connection
with Reid's presentation, published in the last issue of Oceanic
Linguistics for 1994 (?#32). Of course, Benedict who was there too, denies
the PAN-PAA connection, but I believe the consensus reached was that AA-AN
is the only fruitful link to follow.
 
 An assertion that any such
> family (Austric) has been demonstrated to exist by the normal canons
> of the comparative method is absurd, given the published material
> on the subject.
 
Is not as absurd as you think. PAN and PAA share very striking common
morphological markers -- please see Reid's article I mentioned above. It
is rather quantitative than qualitative difficulty that gets in the way:
there are very few cognates discovered so far. But the few that are there
are quite impressive. Of course it is a long way before Austric reaches
the stage of elaboration of IE, but this is no reason to throw it in a
junk basket now and to condemn all research in this area.
 
 Indeed, the reconstruction of Austroasiatic itself
> is still very tentative -- if indeed it existed, which I personally
> seriously doubt. [I.e., I doubt there is any family which has the
> makeup of the currently declared makeup of Austroasiatic, though
> doubtless many of the languages within that putative group are
> related to one another.]
 
You are of course right in one aspect: there is no comprehensive AA
reconstruction yet. However, I believe that in the rest you are going too
far. It is true that some scholars doubt the relationship of Munda with
the rest of the family. But are you seriously suggesting e.g. that
Vietnamese
is not related to Mon, and none of them is related to Nikobarese or
Aslian?
 
>
> Can someone tell me why the question of "isolates" is of any
> linguistic interest whatsoever? If we all accept, as I assume we
> must, that the definition of "language" is a socio-political one,
> then the question reduces to "why do we sometimes find groups of
> speakers, sociopolitically defined to be speakers of 'one
> language', who can't be demonstrated, given the limited nature
> of our evidence, the small number of scholars who have invested
> a small amount of time (seen from a human history perspective)
> working on them, and the extremely limited number of plausibly
> reconstructed language families which we could possibly relate
> them to, to be related to another language?"  The answer is then,
> of course, contained in the question. We don't have the right
> kind of evidence (yet), for the reasons I've inserted into the
> question, and we may never have the evidence. Without such evidence,
> the languages will count as 'isolates'. But it isn't a property
> of the language -- it's a property of the evidentiary record.
> Sometimes it's good and allows reconstruction of an ancestor,
> sometimes it isn't and doesn't. Is this surprising or interesting
> in any way?
 
It remains unclear to me why we all must accept language as
"socio-political" in the first place.