--- In cybalist@..., Miguel Carrasquer Vidal <mcv@...> wrote:
> On the other hand, I've written computer programs far more
> ugly and complex than what this amounts to (and they worked...).
But the language is a shared object. I like to create quirky programs
as well, but what if you should share code with those are to extend
you project. At least all the hype about Java, later CORBA and XML
was about not being ugly, rather clear and sharable. After all, I
don't think deep structures in my ancestor's heads were of such
disorder :)
> Unfortunately, I have misplaced your exposé on Krivichian. Before I
> can formulate a hypothesis, I need more information: does this -e
> palatalize a preceding consonant, and if so, how? What is its
> counterpart in the jo-stems?
Reasonable questions.
First of all, o-stems N. sg. masc. -e is registered for nouns,
pronouns (<te> 'that', <vIxe> 'all' [with failed 3rd pal.], <keto>
'who', <same> 'self'), adjectives and participles (nearly without any
exceptions). Neutra, other -U cases (Acc. sg., G. pl., D. pl, L. pl.
etc), as well as N. sg. of the u-stems and participles like <davU>
'having given' and pronoun <jazU> 'I' doesn't show that anomaly.
This -e does show behaviour (eg., in Havlik-law environments),
characteristic for an unreduced vowel.
This -e doesn't palatalize a preceding consonant in terms of the 1st
palatalization, cf. <zamUke> 'lock', <lixe> 'bad', <Ujene^gU> 'nom.
pr.'. Nevertheless, it caused semi-palatalization ("sm'agc^enije") of
the preceding consonant, just like normal [e] in after-the-1st-
palatalization environment.
Please also note that the later morphological levelling eliminated
the effect of the 1st paltalization from V. sg. in Krivichian:
<bratUke> 'brother!', <druz^Ike> 'friend!'.
'Soft' o-stems ( < *-jo-stems or -C'o-stems with a consonant in the
3rd palatalization environments) mostly have -i (<*-jI) and -I (< -
C'I) (but sporadically -e!).
If this is of any relevance, the 2nd palatalization failed in
Krivichian, cf. <ke^le> 'unbroken' (/e^/ was most likely rendered as
[AE], not [ie], like in SORuss), <xe^rI> 'gray cloth' etc. #rd
palatalization was even more irregular than in SORuss.
------------------------------------------------
P. S. Some notes to consider when promoting phonetical explanations
of Proto-Slavic auslaut phenomena.
1. Your vowel-shifting rules assume */o/ and */o:/ were just [o] and
[o:] during the whole period the auslaut processes were taking place.
But at least for some stages of Proto-Slavic [(labialized?) back a]
and [open a-like o] should be considered. Again, you assume /e^/ to
be much like [AE], what is correct only for part of the dialects
(other probably had [close e.]}.
2. Direct evidence from non-Slavic and early Slavic sources rather
points to -*o as an intermediary stage for *-os > *-U (cf. <Samo>, a
7thc. ruler).
Sergei