--- In
cybalist@yahoogroups.com, "Sergejus Tarasovas"
<S.Tarasovas@...> wrote:
> I think one should take into account the fact that Lithuanian
doesn't
> have the *ji-stem while it does have the *ju-stem. Slavic words (I
am
> talking of non-Novgorod/Pskov Slavic here) entered Lithuanian at
the
> stage when the Nsg of Slavic *-jo-stems was already *-jI (or even
> still pre-Umlaut *-jü, but that depends on where one places the
> Umlaut in the relative chronology of Slavic)
In my view it is possible to be more explicit here, but it is
important to realize that Slavic umlaut is a slippery subject that
easily gets out of hand.
The principle is simple: Slavic umlaut is the fronting of back vowels
under the influence of preceding palatal consonants. Practice is
complex because it is one of those cases where it is of vital
importance to distinguish between the allophonic and the phonemic
phase of a change.
When umlaut fronted long and short *a (later a and o respectively) it
caused them to merge with long and short *e (later e^ and e), but
long and short *u (later y and U) remained distinct from long and
short *i (later i and I) because they were also rounded, so they were
realized as [ü].
Umlauted *u was still distinct from *i in the same position when the
progressive palatalization took place, otherwise we can't explain the
absence of the latter in *igo 'yoke'.
Now "we" know that the progressive palatalization was later than the
monophthongization of diphthongs, otherwise the reflex of *oi after
velars affected by the progressive palatalization would not be e^
(but **i).
"We" also know that the progressive palatalization was earlier than
the merger of the reflex of *ei with long *i because the former does
not cause palatalization, only the latter.
In this context it is important to realize that early Slavic *ei and
*ou were not immediately monophthongized to attested i and u
(contrary to what most handbooks say or imply), but by way of an
intermediate stage with long mid monophthongs, appr. long [e:] and
[o:]. This stage, which is phonetically more plausible, is actually
attested in a number of Slavic loans into Latvian (not Lithuanian!)
and Finnic, it is, or rather should be, completely unproblematic.
So on this reasoning the second palatalization of velars in Slavic
(progressive plus regressive) took place shortly after the
monophthongization of diphthongs but before the raising of the
outcome of the monophthongization of *ei and *ou at a time when long
and short *u were still rounded and the ancestors of Finnic and
Latvian were adopting their earliest layer of Slavic loans.
Put differently: the earliest generations of pre-Latvian Baltic-
speakers must have heard Asg (and presumably also Nsg) *[kanjü] where
attested Slavic has konjI. It may pehaps not be too fanciful that
this gave rise to the model by which Slavic jo-stems were adopted as
Baltic ju-stems.
Does this make sense? Can this be happening? W.