>> I had not recognized that here s- < *k^-, and so failed to find
>> corresponding forms with *s-. But with the declension of *k^i-s a
>> question appeares. Is there any explanation of the *i/*e alternation
>> (that exists also in *i-s)?
Miguel:
> There are several. I believe that the anaphoric pronoun *is
> is suffixed to all (non-personal) pronouns.
I had offered that this alternation is an old one stemming from
IndoTyrrhenian (ancestor to IE and Tyrrhenian languages like
Etruscan). The alternation is the result of an earlier stage
with just *i becoming either *ei (stressed) or *e (unstressed).
This was a general sound rule that caused pure vowels like
*i and *u to become diphthongs when stressed and in an open
syllable.
The stem *ei continued on into MIE but when Syncope took
most of the unstressed vowels, it also took the vowels in
pronouns which happened to be naturally unstressed in a
spoken sentence anyway. Thus *ei- > Late IE *i-. Meanwhile
*e continued on in oblique cases like locatives, genitives,
etc. -- basically anything that wasn't nominative or
accusative. It remained unchanged right down to IE with
a slew of its own case suffixes.
As we can see, other pronominal stems have similar
alternations like *kWi-.
= gLeN