[tied] Re: Short and long vowels

From: elmeras2000
Message: 39319
Date: 2005-07-19

--- In cybalist@yahoogroups.com, "Richard Wordingham"
<richard.wordingham@...> wrote:
> --- In cybalist@yahoogroups.com, Jens Elmegaard Rasmussen
<jer@...>
> wrote:
>
> > Both scholars also showed, immediately assisted
> > by Rix, that some of the analogical explanations for schwa
reflected as
> > Greek /e/ and /o/ are morphologically impossible, and that there
> > consequently is no possibility of accounting for the Greek
laryngeal
> > reflexes by a PIE inventory containing less than three
laryngeals.
>
> I didn't think the number of PIE contrasts was the issue. I
> understood that Patrick's thesis was that the contrast lay not in
the
> laryngeal itself, but in the quality of the associated vowel. Thus
> *&1, *&2 and *&3 would phonetically include an oral vowel, and the
> distinction lie in that vowel rather than any associated consonant.

I understood yoour brave attempt to read some sense into his posting
that way of course; he apparently did not. Now, the examples I chose
did not have any vowels in the relevant segments in PIE: The PPP's
*wr.H1-tó-s, *k^r.H2-tó-s, *gWr.H3-tó-s had a syllabic /r/ followed
by fricatives, not by schwa. Since there are three different results
here too there must have been three different fricatives, unless one
posits three different varieties of "sonant r" in PIE. And in
parallel fashion then three different qualities of sonant l, sonant
n, and sonant m; add to this the necessity of three different
qualities of /i/ and /u/ to account for the cases where iH1/2/3 and
uH1/2/3 give Greek i:/ja:/jo: and u:/wa;/wo:. Add also the knowledge
that the laryngeals had not vanished since they count for segments
in the working of Hirt's Law in Balto-Slavic. It is a pity we do not
know the regular outcomes of *R.H1/2/3 in Anatolian. I do not know
what to make of Hittite daluki- 'long', dalukasti- 'length' from
*dl.H1gh-V-. The treatment of the syllabic resonants varies so much
from branch to branch that the IE point of departure is pretty
clear: it was simply a sonant sandwiched between consonants or
between a consonant and a word boundary. When the development of
secondary vowels shows three different vowel qualities in the later
life of these segments in Greek, that is proof definite of the
existence of three different spirants to impart these qualities.

>
> Patrick has consistently failed to address how, for example, *&1
would
> have differed from *e or *eH, but that is a failing of Patrick, not
> his thesis. A labelling of stages might help Patrick's argument.
> (What stage/branch, for example, does *k^era?-tó-s correspond
to?) I
> haven't yet seen any arguments demonstrating that the contrast was
not
> carried in a vowel, albeit ultra-short. The biggest problem I see
for
> Patrick's theory is in the apparent correlation of vowel colour and
> Hittite consonant.

What is *k^era?-tó-s a quotation from? The root was *k^erH2-; the IE
root is monosyllabic, and the accented *-tó- reduces the root to
zero-grade, yielding a monosyllable *k^rH2tós, whence PIE *k^r.H2tós
by addition of so much sonority that the /r/ counts for a syllable
(whatever that means in any real terms).

Patrick's "theory" has the same status as would a claim that the
consonant of the to-participle was not a /t/, but a uvular [q], with
the addition that the original t had only influenced the vowel, so
that what we should reconstruct for PIE would be q + "t-coloured o",
from where it would be quite possible to derived the later outputs
as -to- by assimilation of the q to the special timbre of the vowel.
I refuse to take this seriously.

Jens