Re: [tied] PIE Stop System

I'm sure we should accept exceptions in a one-vowel system. The classic
examples is Skt. s'vabhis 'with dogs' with vocalization of the /n/ in
*k^wnbhis, as opposed to sïumbhati 'beautifies' with vocalization of /w/
to [u] in *k^wnbhe- (with present-forming /n/). This can only be handled
by analogy, an analogy that triumphs over phonemic identity at that.
Although the example is from Sanskrit, the phenomenon pertains to all IE
and so must apply to PIE already.

The Sanskrit sandhi resistance ("pragrhya") of individual morphemes
demands the positing of a silent consonant (reflecting lost laryngeals).
That may be seen as a restriction on the automatic behaviour of the single
vowel if one so chooses; it looks like a matter of taste to me.

Jens

On Mon, 15 Sep 2003, P&G wrote:

>
> >> So you do find it possible that at a period there was only one
> phonemic
> >> vowel in PIE?
> >
> >Yes. The result would be as in Sanskrit where /a/ has precisely that
> >status.
>
> Not precisely that status.  Other phonemic vowels are necessary in
> Sanskrit
> in a few instances, e.g. the sequence of a dual followed by a word
> beginning
> with a vowel.
>     bandhu: a:sa:te ima:u
>     giri: a:rohatam
> Also (in Vedic) a locative from an i-stem.
> Also the pronoun ami: and a couple of Vedic pronouns.
>
> But perhaps we should allow exceptions even in a "one-vowel" analysis of
> PIE.
>
> Peter
>
>
> Yahoo! Groups Sponsor
>
> Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
>