Re: [tied] a:/o: merger

On Tue, 02 Oct 2001 19:24:37 -0000, "Sergejus Tarasovas"
<S.Tarasovas@...> wrote:

>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!'.

The vocative is of course the first thing that comes to mind, but I
don't think a nom./voc. merger it can explain all. Hard yer always
goes to /o/ in Krivichian when unreduced?

>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.

Yes.

>Again, you assume /e^/ to
>be much like [AE], what is correct only for part of the dialects
>(other probably had [close e.]}.

I assume /e^/ to have been /íe/ (a falling diphthong), at least after
originally short /e/ had acquired a lengthened variant /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).

Where is that found? If perchance in a Latin text, it might be good
to remember that Latin -us had become /-Us/ or /-os/ (W. Romance
merger of short /u/ and long /o:/).