>>This first shift in Early IndoTyrrhenian certainly must have gone
>> >>something like *i to *ei (that is, */@y/) and *u to *eu (*/@w/).
>
>I don't think I've ever encountered such a sound shift. Can you give
>examples from any language?
What I am basically trying to say is that vowel length doesn't seem to have
ever been important in earlier stages of IE and beyond like you seem to
think. There was never a phonemic contrast between ProtoSteppe *i: versus
*i. However, the accent probably played a role to make things "appear" long
or short. I have to give examples to clarify...
If we have ProtoSteppe *ku, we should obtain IndoTyrrhenian *kWe if, and
only if, the word is unstressed. We find this word attested in IE as an
interrogative enclitic. If it were a "full" word (noun or verb) we would
obtain *keu. Both derive from ProtoSteppe *ku but the accent is different
for both cases. There is no need to write *ku: versus *ku in ProtoSteppe
because any vowel length difference would have been dependent on stress.
There is also no need to write the accent explicitly because the accent was
regular (placed on the initial syllable in IndoTyrrhenian and penultimate in
MidIE).
I think this is just an orthographical problem more than anything. Actually,
you're right sorta: *ku(:) > *keu (as verb/noun) but *ku > *kWe (as
enclitic). It's just that explicitly reconstructing long vowels in
ProtoSteppe or even Nostratic makes me woozy. Hope this clarifies my
position better.
Miguel:
>This is Szemer�nyi's proposal, and I believe it's incorrect >(*wo:kw-s,
>*die:w-s, etc.).
You didn't read properly. This is due to the _first_ lengthening. That is,
*wo:kWs derives from a very early form *wekW, a noun derived from the verb
*wekW-. As a monosyllabic noun, it was lengthened to *we:kW in Early IE
(MidIE *we:kW-se).
>This is Rasmussen's proposal, and I believe it's incorect (*-h2 isn't
>lost in the collective [n.pl.]).
It isn't lost finally following a consonant? Examples?
>So what about e.g. Arm. sirt (*k^e:rd + *-is)?
Armenian is weird. :) Hmm, well could it be a new animate formation or
derivative? What is *-is?
>>The qualitative alternation is quite clearly determined for the most
>> >>part by accentual alternations in the Tonal stage of Late IE.
>
>Do explain.
- First, in Mid IE, there were only two vowels, *a and *e (schwa).
- Suddenly, unaccented schwas were slaughtered
and didn't survive :(
- Then the accent changed from stress to tonal. Hooray!
- Some accent patterns changed as well. Boo!
- The remaining schwas fractured into *e (front) and *o (back).
- This change was dependent on the new accent _at the time_.
- *� > *e
- *e > *o
- Then the tonal accent did a few more twists and turns
to try and cover up its earlier evil deeds.
So... the alternation between normal grade and lengthened grade (*e versus
*e:) is the most ancient of all vocalic alternations and far predates the
next two alternations. The alternation between normal grade and zero grade
came next (*e versus NULL), followed by an alternation between *e and *o.
The normal/zero alternation is seen in the conjugation of *es- "to be"
because of the ancient differences in accent between singular and plural
stemming from the uneventful penultimate accent of MidIE.
When the accent changed to tonal along with some accent patterns, the
remaining *e's (all the accented schwas in MidIE) were split into accented
and unaccented schwa just like it was before. This time however, accented
schwa became a front *e as it tends to do in other languages. The unaccented
schwa was pushed back to *o, perhaps in order to fully distinguish it from
front *e. To further complicate things, it appears that *e changes to *o
partly because of surrounding labial phonemes like *m, *w or *xW.
That's all, folks.
- gLeN
_____________________________________________________________________________________
Get more from the Web. FREE MSN Explorer download :
http://explorer.msn.com