Re: [tied] How to prepare **udon soup (was: PIE rhotacism)

On Mon, 19 Nov 2001 05:37:14, "Glen Gordon" <glengordon01@...>
wrote:

>First of all, Piotr devilishly uses *uden to downplay the likelihood
>of the early sound rule as if it were absolutely certain that this
>form dates back to the time of the sound change which is, in my
>examination, likely to have occured an entire millenium before
>common IE. Since a lot can happen in a millenium, I think there
>is ample room indeed for analogical changes to have occured after
>*-n became *-r.

OK.

>It's also unfair to whip out a single form without fully examining
>how it relates to the rest of the paradigm and the history of that
>paradigm.

In fairness to Piotr, he did no such thing.

[...]

> nom-acc/loc/voc sg. *wát:er
> gen.sg *wet:en-áse
> gen.pl *wet:en-áne

I'm sure you must have your reasons for reconstructing *n in the
gen.pl...

>Here, the above forms regularly change to *wodr, *wednos and
>*wednom without problems. We also see that *wet:ar-xe (*wedo:r)
>was coined at this time, showing the original vocalism
>alternations. The locative form wasn't any different from the
>nominative at this point in time aside from the lack of the
>postpositional *se of the nominative amongst animate forms (eg:
>NOM *kewane se // LOC *kewane bei). Also note that the weak case
>forms regularly show *wet:en-.
>
>With the loss of final vowels came a new mobile accent pattern. The
>paradigm changed as follows:
>
> nom-acc/loc/voc sg. *wát:er
> gen.sg *wet:en-ás
> gen.pl *wet:en-án
>
>Then, all other unstressed vowels (aside from strengthened vowels)
>began to drop and vocalic resonants started to form...
>
> nom-acc/loc/voc sg. *wát:r
> gen.sg *wet:n-ás
> gen.pl *wet:n-án
>
>It was at _this_ point, one sunny day in 5100 BCE, that a
>unique locative form began to develop independently from the
>nominative forms because the locative came to be associated with
>the weak cases. Naturally, this new form then was nothing other than the
>endingless weak case stem *wet:n- with "weak case accent". That
>is to say, it was given accent on the final syllable just like in
>the other weak case forms seen above. Of course, with the original
>vocalism of the final syllable lost, a dummy vowel *e was inserted
>therein. This gave:
>
> nom-acc/voc sg. *wát:r
> loc.sg. *wet:én
> gen.sg *wet:n-ás
> gen.pl *wet:n-án

I'm not too crazy about the "dummy vowel", but let's say that might
possibly explain a form like **wedén. Unfortunately, that form is not
really attested as such (although it might underlie something like
Hitt. weteni, if that's not simply /wetni/). What we have is in fact
*udén (Skt. <udán>).

My account is as follows: the "water" word shows traces of two
different paradigms, a static one with NA *wódr, and a collective with
NA *udó:r (*udór-h2).

We can reconstruct, after the "svarita lengthening rule":

NA **wá:da:n **wadá:n
G **wa:dána:s **wada:nás
AbI **wa:dána:t **wada:nát
DL **wa:dána **wada:ná

After zero-grade, r > n, a: > o, a > e ...:

NA **wódor **udór
G **wedénos **udenés
AbI **wedénot **udenét
DL **wedén **udené

After Rasmussen's "initial accent rule":

NA *wód(o)r *udór+h2 > *udó:r
G *wédnos *udéns
AbI *wédnot *udént [*] ~ *udnéh1
DL *wédn+i > *wédni *udén+i > *udéni

And with paradigmatic split of Abl/Ins, Dat/Loc:

NA *wódr *udó:r
G *wédnos *udén(o)s
Ab *wédnot *udénot
AbI *wédneh1 *udéneh1, *udnéh1
D *wédnei *udénei
L *wédni *udéni

[*] Unexpected Abl/Ins *udént results from strict application of the
rule, but may throw some extra light on the genesis of Greek <hudo:r>,
<hudatos>.

=======================
Miguel Carrasquer Vidal
mcv@...