--- In cybalist@..., "Piotr Gasiorowski" <gpiotr@...> wrote:
> Willem R. Vermeer ("The rise of the North Russian dialect of Common
Slavic", _Studies in Slavic and General Linguistics_ 8 [1986]) argues
that the monophthongisation of the Proto-Slavic diphthongs had not
diffused to (Proto-)Krivichian before the progressive/second
regressive assimilation (regarded as one and the same change)
expanded to the Novgorod/Pskov area. Since *ai was still a diphthong
with a non-front first element, there was no input at all to the
"second regressive" part of the assimilation, hence its failure in
Krivichian. The "progressive" part resulted in a series of new
palatals that were restricted to stem-final positions and therefore
marginal to the system (Vermeer treats the Krivichian "c'okanje" as a
later innovation, not an archaism). As such, they were vulnerable to
analogical elimination. The one that normally survived was *c', which
was well entrenched in productive suffixes.
>
> Vermeer's analysis is attractive. The only inconsistency I can see
in it is the non-occurrence of the progressive assimilation after old
*ai and *ei in Krivichian. If they had preserved their final glide,
it ought to have triggered the palatalisation. One can propose (as
Vermeer does) that the glide had been lowered and the Proto-
Krivichian reflexes of the diphthongs at the time of the
palatalisation were something like *ae, *ee., which may be a little
ad hoc, though surely not implausible.
>
> I haven't got the original article to hand, but I have a PDF copy
of a more recen article by Vermeer (2000, from _Russian Linguistics_
24), which contains an exposition of his ideas. I am sending it to
you off-list.
>
> Piotr
>
Unfortunately I haven't got your e-mail. If you sent via Yahoo's
server, it probably just failed to deliver, as it usually does. In
case you would decide to retry, my e-mail is:
S.Tarasovas@...
Vermeer's position, as far as I can evaluate it from your summary,
doesn't look very attractive to me. Eliminating one problem (why the
second regressive assimilation failed) it introduces another (why the
Proto-Krivichian was late for the train as to the
monophthongization), and there are no extra evidence (apart from the
failed assimilation itself) for such a late monphthongization. While
the (Proto-)Slavic dialects were indeed a bit inconsistent as to the
output of the second wave of assimilations, I've never heard of any
observable inconsistency as to the monophthongization (with the only
exception of vowel+{*m,*n,*l,*r}-type diphthongoids, which are not
proper diphthongs), which would require rather early date for the
process, hardly after 500 AD (when North Krivichians reached the
Novgorod/Pskov area). Probably loans (both into and from Proto-
Slavic) could shed some light on that, but nothing relevant comes to
my mind at the moment.
And I see at least two more inconsistencies in his analysis:
1. North Krivichian _kIrky_ 'church' is left without explanation.
2. the process *ei > (lowering) *ee. > (raising) i(:) is not very
natural, and the process *ai > *ae > (in curcumflexed syllables) i(:)
is just implausible.
I would dare offer an alternative scenario, but first I would like to
summarize the results of the consonantal assimilations to the front
vowels and the palatal glide in Krivichian (E -- any front vowel
triggering the first regressive assimilation, I -- any (sequence of)
segment(s) triggering the progressive assimilation, [J] -- voiced
palatal fricative, [G] -- voiced velar fricative, [CJ] -- palatalized
[C]):
1.
*kE, *kj > [tc], *skE, *skj > [ckJ]
*gE, *gj > [J], *zgE, *zgj > [JgJ]
*xE, *xj > [c]
2.
*tj > [kJ], *stj > [ckJ]
*dj > [gJ], *zdj > [JgJ]
*sj > [xJ] (didn't merge with [c])
*zj > [GJ] (didn't merge with [J])
*sE > [c]
*zE > [J]
3.
*Ik > [k]/[tc]
*Ig > [g]/[J]
*Ix > [xJ] (actually we have only one example -- the stem _vIx-_
'all')
It must be noted that I failed to find any evidence of anything
aberrant as to the progressive assimilation in general (except _vIx-
_) and *Ig-assimilation specifically both in Zaliznyak's thesaurus to
the birch bark inscriptions (plus other early documents from the
Novgorod/Pskov areas) and in Nikolaev's papers on contemporary
Russian dialects formed on the North Krivichian substrate, so I have
no idea what makes Vermeer think that "The one that normally survived
was *c', which was well entrenched in productive suffixes.".
Now the scenario proper.
The progressive assimilation was a rather early process, and its
'stronger' phonetical environment points to that indirectly: it took
_close_ front vowel (*i or *I) to trigger it, (open-)mid *e was not
enough (and *i2 just didn't occur before velars).
Proto-Krivichian hadn't yet acquired [kJ] and [gJ] at that time.
Indeed, Finnish _kaatio_ 'thigh, trouser-leg' point to Proto-
Krivichian *[ga:tja:] (*tj -- still a cluster) or *[ga:tJa:], which
would in turn point to the following processes: *tj > *[tJ] > (*[c])
> [kJ], *dj > *[dJ] > (*[J-]) > [gJ].
Proto-Slavic clusters *skE, *skj (and *zgE, *zgj) probably first
assimilated their velars to the dental *s, yielding *[stJ] (and
*zdJ), before merging with the reflexes of *stj (and *zdj) in [ckJ]
(and [JgJ]) in North Krivichian.
Finding no support in existing phonems, phonetically palatalized *
[kJ] and *[gJ] either remain intact or jumped to the nearest existing
phonemes -- *tc and *(d)J (rather than *tj or *tJ and *dj or *dJ),
while *[xj] remained intact, because a supporting phoneme -- *xJ (<
*sj) -- already existed.
Later, by the time of the second regressive assimilation, *[tJ] and *
[dJ] had already yielded *[kJ] and *[gJ] (and *[xJ] remained), thus
velars, phonetically palatalized before *e^ and *i2, were
phonologized by a merger to already existing phonemes, forming a
unique feature of the North Krivichian phonemic inventory -- the
/kJ/,/gJ/,/xJ/ series, that triggered some interesting restuructrings
of the declensional system.
Sergei