On Fri, 14 Sep 2001 00:07:07 +0200, Miguel Carrasquer Vidal
<
mcv@...> wrote:
>On Thu, 13 Sep 2001 20:20:18, "Glen Gordon" <glengordon01@...>
>wrote:
>
>>I think we agree that syllabic *r must surely be *[&r]
>
>No, I can't agree at all.
I hit the "Send" button too soon. I was going to say:
No, I can't agree at all. Syllabic *r was just that, syllabic /r/.
The proof is in Sanskrit, where /r./ was assigned a separate symbol in
the precursors of the Devanagari script (Brahmi and Kharoshthi). Had
/r./ been /&r/ or /r&/, its vocalic element would have merged with /&/
from "syllabic" H1, H2 and H3 ("schwa indogermanicum") and would have
gone to /ir/ or /ri/ already before Vedic Sanskrit. The developments
were in fact:
pre-Vedic Vedic post-Vedic
/&/ -> /i/ /i/ /i/
/r./ /r./ /r./ -> /r&/ -> /ri/
The same in e.g. Germanic, where schwa went to /a/ [as everywhere else
but in Indo-Iranian]:
/&/ -> /a/
/r./ -> /ur/
And /r./ was still syllabic in the Common Slavic period, as is proven
by the fact that the reflexes show great dialectal variety: we have
true syllabic /r./ in Serbian/Croat, Slovenian, Czech and Slovak;
/er/, /or/ (< /Ir/, /Ur/) in Russian; /rI/, /rU/ in Church Slavonic;
/&r/ (CrC) and /r&/ (CrCC) in Bulgarian; and we have /ier/ (~ <ierz>)
and /ar/ in Polish [the latter form directly disproving any
possibility of reconstructing /Ur/ [alternating with /rU/] in Common
Slavonic, as Bräuer tries].