Re: Re[3]: [tied] Latin -idus as from dH- too

From: fournet.arnaud
Message: 55325
Date: 2008-03-16

----- Original Message -----
From: Brian M. Scott

> Yes
> I agree with you, Marius,
> Like some supposedly brilliant
> theory about Pre-PIE vowels :
> - which accounts for central PIE,
> - which can't be proved with Anatolian

Wrong: see

So Luwian may be the *only*
place to look for long e:
supposedly in pir < pe:r

> and conflicts with LAtin data.
> ye:kur with long e: > latin iecur (short !!)
> The discussion aborted when
> it became clear the idea does not hold water.

Looked more to me as if it stopped when Miguel decided that
he was wasting his time.



There remains a certain number
of clarifications and objections
to address :

1. What is the result of *per
in Lydian, Luwi
if any contrast with *pe:r existed ?

No answer has been provided.
Anatolian looks neutral, rather than
providing clear support.

2. Why has Latin a *short* e in iecur ?
We should expect **iacur or **ie:cur
if Miguel's theory was right ?

This is a counter-example
refuting the proposed model.

3. Why are supposed instances all
words that end with -r# ?
ker(d), gwher, H2ster,

I will add :
4. What about the case of n.gwen ?
Greek ade:n, Latin inguen

The discussion aborted
precisely when it was about to start
for real.

I'm not ready to buy long e: in PIE
on the flimsy basis of a couple of words
that all end in -r#,
the ABC of phonology teaches that
this may just be a contextual dvlpment.

The case of n.gwen shows the e:
may just be a dvlpment after sonants
like r l n

I consider that Latin iecur with short *e
is a *major* stumbling block
this word has a stop y_k after -e-
and it does not fit in the proposed model.

I also consider that the long e: issue
cannot be separated from the long o:

I wrote :

>You are in fact confirming my feeling
>that /e:/ is an innovation of Central PIE.
>Some Greek words have long o:
>klo:ps "stealer"
>tho:ps "flatterer"
>tro:ks "worm"
>sko:r "s*t"
>Obviously /o:/ is an innovation

if non laryngeal /o:/ is an innovation,
why should I believe that long /e:/ is
inherited ?

All well considered,
I now believe that the model proposed
by Miguel is most probably wrong.
Especially if Points 1234 remain unanswered.