>=========
>Let's admit *pe:r and *ke:r existed.
>
>Why should it not be a synchronic
>(anatolian) lenghthening of
>*pern > *per:
of *perr, rather.
============
I mistyped :
I meant pe:r
as the synchronic result of *pern
Arnaud
=======
>*kerd > *ke:r
>Because Anatolian does not accept clusters
>and "prefers" long vowels.
It has nothing to do with Anatolian. We have Greek ké:r,
Armenian sirt < *k^e:rd, Vedic ha:rdi, OPrussian seyr, all
showing a long vowel.
Miguel
================
You are changing the topic.
It's no use giving the Central PIE examples
They do not prove anything.
I'm not convinced that the existence of inherited e:
is proved in Anatolian languages.
Now, what you have provided are examples :
1. with Logograms
SHA-er and E-er to be read *ke:r and *pe:r
This is more an interpretation than a reading.
2. supposing *ke:r and *pe:r exist
they can be explained synchronically
as *pe:r < *pern and ke:r < *kern
Next, I think it's highly strange
that /e/ exists only when it's long.
in Hittite.
This is a strange feature.
and probably a mis-reading.
Arnaud
==============
>What is the documentation for
>the other cases where you have *e: ?
My question is about Anatolian data ?
MAybe you have a clear example
without logogram and without final clusters.
Arnaud
=======================