I mean PIE (or Indo-Hittite, if you prefer) *e in unstressed
positions. Roughly,
["internal reconstruction" PIE]**(h1)es-�nti > ["comparative" PIE]*
(h1)@s�nti > Hittite asantsi (I prefer to use -ts- spellings on
the
list, since lay people generally tend to assume that z = [z]).
The initial weak vowel was dropped in most branches. I prefer this
analysis to the assumption of a change like *h1 > a, because it
accounts for many Hittite alternations that certainly don't involve
*h1. As for the details, I can't lay them out in a short posting, so
I'll prepare a longer account soon in order to share my pre-
publication views on reduction vowels in PIE with CyBaList members.
Piotr
--- In
cybalist@egroups.com, "Urban Lindqvist"
<urban.lindqvist@...> wrote:
> From: Piotr Gasiorowski
>
> [Piotr:] What is confusing about Hittite is that it also
has "strong" e corresponding to "weak" a! This is because originally
unstressed PIE *e could be reduced to a schwa-like vowel (*@) which
became Hittite a. [...]
> [Urban:] I don't follow. What kind of an *e do you have in mind? A
Pre(-Greco-Aryan)-PIE/Indo-Hittite unstressed *e, or an analogical *e
(*h1s-�nti --> *(h1)es-�nti > asanzi), or something else?
> By the way, are there any convincing examples against the
assumption *h1 > a /_[+dental] in Anatolian (or Hittite only)? This
would take care of adanzi and asanzi, and the rest would be
analogical.