Re: Nostratic (was Re: Schwebeablaut (was Re: tolkien?))

Hallo!

On Mon, 15 Dec 2003 18:45:04 -0500,
Roger Mills <romilly@...> wrote:

> Jörg Rhiemeier wrote:
> > Well, one can investigate the origins of patterns observed in a
> > reconstructed protolanguage without trying to relate the family
> > in question to something else.
>
> Yes, good old internal reconstruction.  It can lead to good insights, even
> if they may not be provable to everyone's satisfaction.
Exactly.  My own internal reconstruction of PIE (which is still
patchy and fragmentary) makes some things make sense that seem
irregular in traditional PIE, but it surely contains controversial
matters.  For example, I reconstruct pre-PIE as an active-stative
language, with forms such as *bherth2em `I carry you'.
The main purpose of my PIE internal reconstruction is actually
to come up with a common ancestor of both PIE and Proto-Hesperic
(Hesperic is my main conlang family).

> > And regarding Nostratic, I am open-minded towards such hypotheses,
> > yet not content with the evidence provided.
>
> My view, too.
>
> > I think the most useful approach is to examine not six or more
> > families at a time, but to concentrate on two or three first,
> > and then look at similarities of morphology.  From my own studies
> > in this area, I consider it likely that Indo-European, Uralic,
> > Eskimo-Aleut and perhaps Etruscan are indeed related to each other.
>
> I tend to go along with that; many would add Afro-Asiatic, or at least
> Semitic...
Well, as Afro-Asiatic is pretty well accepted (though the reconstruction
of Proto-AA seems to be still lacking, judging from what I have found),
if Semitic is related with IE, than the whole Afro-Asiatic is.
It is of course not impossible that IE or some other languages
turn out to be parts of what we now call Afro-Asiatic, and may
be closer to Semitic than, say, to Chadic, but I consider this
unlikely.

>      The better versions maintain fairly strict standards, e.g. Family
> X has a 1st pers. morpheme /m/ and this seems to resemble the IE 1st pers.
> /m ~mi/ etc.  However, others try to claim Y's /n/ = IE /m/ because both are
> nasals, or Z's /w/ = IE /m/ because both are labials.......well..........
Yes, these are things that bug me about Nostratic as well.
We have these pronoun roots:

                 1st   2nd

Indo-European    *m-   *t-
Uralic           *m-   *t-
Eskimo-Aleut     *m-   *t-
Kartvelian       *m-   *c-
Altaic           *b-   *s-
Afro-Asiatic     *n-   *k-
Sumerian         *w-   *z-
Dravidian        *y-   *n-

The first three indeed match as well as they can.  Kartvelian may be
connectable to it when one assumes a (not implausible) change *t- > *c-.
Then things become hairy.  Altaic has the same places of articulation,
but different manners.  Afro-Asiatic is the converse: same manners,
different places.  Sumerian is like Altaic, only voiced (or whatever;
we know frustratingly little about Sumerian phonology, thanks to its
logographic script).  And the Dravidian pronouns show no similarity
to the other at all; they could just as well have arrived from Mars.

Given the fact that places of articulation are less likely to change
than manners, I would judge that the Altaic and Sumerian pronouns
are more likely to be related to the IE pronouns than the Afro-Asiatic
ones (we can forget about Dravidian).  The IE/Uralic similarity
is striking especially in light of the fact that both families
are neighbours and probably always have been.  Kartvelian is also
geographically close to IE, so a relationship seems quite likely.
Eskimo-Aleut is spoken quite far away from IE and Uralic, but we
know that its speakers came out of Siberia only about 5000 years
ago, and I have seen some really good evidence for a Uralic-Eskimo
connection (very similar phonologies, and verb conjugation paradigms
that are essentially the same in Proto-Uralic and Proto-Eskimo-Aleut).
But beyond that, things quickly become very uncertain.

> Paul Bennett wrote:
> > > The 2 (and a half) vowel e/o/0 ablaut (with i and u as vocalic
> consonants)
> > > suffices for me, although a six vowel a/i/u/a'/i'/u' ablaut is an
> > > interesting theory, that I had never seen before. Do you have any URLs
> to
> > > hand that explain it with examples, and describe why it is important?
>
> I'm not sure whether current handbooks go into the (somewhat out of fashion)
> status of *a, but you could look.
Handbooks on PIE usually don't concern themselves much with internal
reconstruction.  At least not the ones I have seen.

>       Failing that, join the yahoogroup
> Cybalist and cruise their extensive archive;
You need not be a member of Cybalist to read the message archive.
It is open to everyone:

http://groups.yahoo.com/group/cybalist/messages

I often look into it myself, and I am not on that list.

>       the most reliable and
> knowledgeable are Piotr Gasiorowski, Jens Rasmussen and Miguel Carrasquer
> Vidal.  Glen Gordon for comic relief.
Glen sometimes have some ideas to think about, but most of his stuff
seems to be built on poor scholarship, and he tends to flame those
who contradict him.

> At one point, Piotr had posted some extensive info on IE phonology and
> morphology in the Files section; I think it's still there.
On Tue, 16 Dec 2003 06:16:38 -0800,
paul-bennett <paul-bennett@...> wrote:

> On Mon, 15 Dec 2003 21:04:35  0100 =?iso-8859-1?Q?J=F6rg?= Rhiemeier
> <joerg_rhiemeier@...> wrote.
>
> >I think the most useful approach is to examine not six or more
> >families at a time, but to concentrate on two or three first,
> >and then look at similarities of morphology.  From my own studies
> >in this area, I consider it likely that Indo-European, Uralic,
> >Eskimo-Aleut and perhaps Etruscan are indeed related to each other.
>
> Yes. My own plan is to look at IE and Uralic, only, since AFAICT they're the
> most readily related. As yet, I don't even have enough information to be
> sure in my own mind of which PIE reconstruction I favour, although I tend to
> accept a fairly mainstream approach.
Yes, there are still many open questions on the PIE side,
and Uralic looks far worse still in that regard.

> Does your i/a/u ablaut make PIE easier to relate to Uralic? I know *nothing*
> of Uralic reconstruction.
I don't know, either, but some people try to reconstruct a 3-vowel
system for Pre-Proto-Uralic, before vowel harmony and all that
kicked in.  And what regards Uralic reconstruction, it seems that
there is still much work to do.  The only complete "reconstruction" of
Proto-Uralic I have seen was that by Decsy, and that doesn't hold
water because he not even used the comparative method (I guess
he is a Marrist)!  Instead, he considers each cognate set (or pseudo
cognate set, his method is incapable of distinguishing between the two)
*in isolation*, "reconstructing" a proto-form that yields the attested
forms with a minimum of changes.  Clearly, this is wrong, and yields
many false positives.  And what is worse, Decsy even contradicts
himself.  He claims that the speakers of Proto-Uralic didn't use
names (an anthropological impossibility!), yet he "reconstructs"
a PU word for `name'.

I have looked for Proto-Uralic material on the web many times,
but found little, so I am at my own devices to piece up my own
version from the little I have found, which is rather frustrating.
Apparently (this turned up over and over again in my Google searches),
many Uralicists now even dismiss the existence of Proto-Uralic,
but ascribe the observed similarities of the Uralic languages
to areal contact and lexical diffusion.  It seems that another family
goes the way of Altaic, which is also no longer accepted as a family
by many linguists.

> While I'm in the mood for asking, does the glottalic theory make PIE easier
> to relate to Uralic?
I don't know.  Bomhard and Kerns list more than 100 IE-Uralic cognate
pairs, and their sound correspondences take glottalic theory in
account (as opposed to Illich-Svitych and Dolgopolsky, who, I think,
have fewer IE-Uralic cognate pairs), but I seem to remember that
Bomhard and Kerns use Proto-Uralic forms from several different
sources (including Descy, see above!), and often rach down into
daughter languages on the Uralic and sometimes on the IE side.

In my own long-range studies, I concentrate on morphology.

Greetings,

Jörg.