--- In
cybalist@yahoogroups.com, "mandicdavid" <davidmandic@...>
wrote:
>
> The 1sg. pres. ending -oH2 yielded acute -o: in Lithuanian (later
> short -u). In Slavic, it turned into -a: and was later expanded
with
> the secondary ending -m (cf. 1sg. aorist pekU < *pekwom, and also
> skr. bhára:mi < *bhero:+mi). The addition of -m probably followed
the
> change o: > a: in Slavic, because the o: yielded u: > y before
nasal
> stops in word-final position: kamy < *ka:mu: < *ak'mo:n. Also, -a:N
#
> wasn't affected by umlaut (A.sg. zemljoN; thus also 1.sg. pres.
bijoN
> etc.).
Actually, the addition of -m must have followed the raising of back
vowels before final nasals (or we would have gotten *-o:m > *-u:N > *-
uN > -U), and preceeded the shortening of long diphtongs (or we would
have gotten -o:m > -a(m)).
<Kamy> "stone" cannot come from *h2ák^mo:n, because that would have
yielded <kamU>. In fact, it _does_ yield <kam> in I forget which
Slavic dialect (Kashubian?). The Slavic form, like Lith. <akmuõ>,
comes from PIE *h2ák^mõ, a variant with the final resonant dropped
(only its falling tone remains).
There is another Slavic variant (Northern Russian?) <kama>, which may
come from PIE *h2ák^mo:, also with /n/ dropped, but with acute tone,
or from *h2ák^mo:n, with /n/ dropped in later (Balto-)Slavic times,
but still before the N-raising. Another such variant may underlie
<vodá>, where we would have expected *wo(n)duõ (= Lith. vanduõ)
to have yielded *v(o)dy or the full form *wo(n)do:r > *v(o)do(r) [I
won't mention the fact that we would have expected something other
than /o/ in the first syllable].
Apart from these two doubtful cases, all Slavic long diphthongs (i.e.
sequences V:R, where R is one of /m, n, l, r, w, y/ were shortened
(so were they, independently, in Lithuanian).
The acc.sg. of the a:-stems _was_ affected by the N-raising, but
since Slavic still had *a:, *o: and *u: at the time, the raising went
no further than *-a:N > *-o:N, which was then shortened to -aN > -oN.
To rehearse the possible developments from PIE:
circ. N-raising short. CS
*-am *-uN -U
*-om > *-am *-uN -U (o-stem Asg.)
*-um -U (u-stem Asg.)
*-a:m *-o:N *-aN -oN (a:-stem Asg.)
*-o:m *-u:N *-uN -U
*-õm *-u:m *-uN -U (o-stem Gpl.)
*-u:m *-uN -U
Only when /n/ and /y/ do not make a diphtong and /s/ follows, do we
get lengthening (or preservation of original length). We see this in
the Apl. and o-stem Ipl.:
N-rais. s-rais. lengthening CS
*-o:ns *-u:Nh -y (o-stem Apl.)
*-a:ns *-o:Nh *-u:Nh -y (a:-stem Apl.)
*-ins *-i:Nh -i (i- and C-stem Apl.)
*-uns *-u:Nh -y (u-stem Apl.)
*-ajs *-ujh *-u:jh -y (o-stem Ipl.)
A similar case is the nt-participle Nsg., where in part of Slavic we
have an early reduction of *-onts to *-ans, N/s-raising to *-uNh,
lengthening to *-u:Nh, and finally CS -y (e.g. OCS nesy). In Northern
Slavic, on the other hand, we have forms without such an early
reduction. There we have the development *-ants > *-aNts (for which
cf. e.g. 3pl. past *-ont > -oN), and only then (after the N- and s-
raising laws had stopped working) lengthening to *-a:Nh, yielding -a
(e.g. ORuss. nesa).