Re: languages of pre-I.E. Europe and onwards

Hallo!

On Wed, 21 Jan 2009 15:36:42 -0800, Roger Mills wrote:

> Jörg Rhiemeier wrote:
> > Yes.  I have a list of non-IE Germanic words which I have
> > extracted from an etymological dictionary of German, and
> > another I once found on a web site which by now is gone;
>
> Would that be McCallister's list?  I have it.....
Yes.  Exactly that one.

On Wed, 21 Jan 2009 22:37:07 -0600, Eric Christopherson wrote:

> On Jan 21, 2009, at 2:21 PM, Jörg Rhiemeier wrote:
> [...]
> > Indo-European and Uralic look so similar that one can hardly
> > dismiss the impression that they are related to each other,
> > though "Proto-Indo-Uralic" (or whatever to call it), if it
> > ever existed, must have been spoken about 10,000 years ago,
> > if not earlier.  Also, the homelands of PIE and PU are quite
> > close to each other.  An IE-Uralic relation was suspected
> > already by the venerable Jakob Grimm.
>
> Could you write a little summary of what specifically is similar (or
> point to a web page or paper)? Similarly with Etruscan.
Quite a big chunk of Uralic inflectional morphology seems to
have cognates in Indo-European.  The most often cited elements
are the personal pronoun roots *m- (1st), *t- (2nd), *s- (3rd),
which are shared by both families.  Uralic has the plural markers
*-t and *-i, which are similar to the IE plural markers *-s and
*-i; if *-t > *-s in PIE, one would also expect *-k to turn into
a velar obstruent, and indeed the Uralic dual *-k seems to
correspond to IE dual *-h1.

Next, the case system.  The accusative is *-m in both families.
The Uralic genitive *-n may be cognate to the IE denominal
adjective suffix *-no- found in such adjectives as Latin
_marinus_ and _paganus_; an earlier athematic form of that
suffix may be found in PIE *dhghomon- 'human' (from *dhghom-
'earth'; to be understood in opposition to gods who dwell in
Heaven).  The Uralic ablative is reconstructed as either *-ta
or *-δa; possibly related to IE ablative in *-d.

Also, the verb conjugations are quite similar, clearly
related to the common pronoun roots in both languages.  The
Hungarian and Selkup 1st person intransitive *-k may have
something to do with the IE 1st person perfect (originally
stative) *-h2a.

See also http://en.wikipedia.org/wiki/Indo-Uralic_languages
for more examples, and comments thereon.

While the evidence adduced so far does not prove an IE-Uralic
relationship, I must say that IE and Uralic look eerily similar
for "unrelated" language families.  The fact that they originate
from areas only about 500 miles from each other makes a
relationship quite plausible.

The list of IE-Etruscan morphological similarities looks
a tad less impressive; I posted it a few years ago here:

http://listserv.brown.edu/archives/cgi-bin/wa?A2=ind0311C&L=CONLANG&P=R12370

On Thu, 22 Jan 2009 08:25:04 +0000, R A Brown wrote:

> Jörg Rhiemeier wrote:
> > Hallo!
> [snip]
> > Yes.  One cognate set is no cognate set.  You can always find
> > chance similarities.
>
> Like modern Greek _mati_ = 'eye' and Malay/Indonesian _mata_ = 'eye'
> both derived of course from Proto-Helleno-Malayan    :-D
A famous example.

> > In order to prove a relationship between
> > languages, you need at least about 100 cognate sets displaying
> > regular sound correspondences.
>
> ... and the _regular_ sound change bit is important. So many of the
> supposed correspondences are thrown together in a haphazard way.
Amen!  It is the *regularity* of the sound correspondences
that make the proof of relationship.  Sure, if the common
ancestor was spoken long ago, conditioned sound changes can
create a situation where the rules are difficult to unravel
and the correspondences look anything but regular; but with
many long-distance proposals, borrowing or chance resemblance
are more likely than common ancestry.

This is why (in my personal opinion) comparison of morphology
is more promising than open-class lexical comparison, though
it is bugged by the problem that the morphemes are short,
which increases the chance that they are similar by coincidence.
But if entire inflectional paradigms match and the same
allomorphies occur in a pair of languages, you are almost
certainly dealing with languages related by a common ancestor
(as is IMHO the case with IE and Uralic).

... brought to you by the Weeping Elf