--- In
cybalist@yahoogroups.com, "tgpedersen" <tgpedersen@...> wrote:
>
> --- In cybalist@yahoogroups.com, "tgpedersen" <tgpedersen@> wrote:
> >
> >
> > > present stem:
> > > *XADV- > *XandV- > (syllabification) *Xan-dV- > *Xen-dV-
> > >
> > > perfect stem:
> > > *XAD > *Xand > (syllabification) *Xand > *Xãd > *Xõd > *Xod
> > >
> > > which would explain both the present infix and ablaut, which is
> > > nice. Obviously, both would have to have been reinterpreted as
> > > morphological processes and generalised to other stems, but
such
> has
> > > happened before
> > >
> >
> > Come to think of it, if open vs. closed syllable determines e-
> grade vs.
> > o-grade, that would explain the appearance of o-grade in
> causatives:
> > *mon-eye/o- etc (still assuming that the pure stem is a nominal
> form,
> > participle or the like (and perfective?), of the verb).
>
> Comment: because *mon was then an indendent word, and the syllable
> boundary followed the word boundary.
>
> >
>
> And, while I'm at it: I can't very well claim that prenasalised
> consonants caused ablaut before them without claiming nasals did
> too.
Perhaps blaming the whole ablaut phenomenon on generalisations from
stems in PPIE -a- and ending in prenasalised stops and nasals is a
bit wobbly. I should establish a larger "critical mass" before I
cry "generalisations".
According to Stang, the n-infix alternates in Baltic with the -st-
suffix. Tne n-infix occurs in roots ending in stops, resonants and
semivowels. I'll try to go for at least that.
Wrt. stops, I'll still appeal to generalisation from the
prenasalised, ie. voiced unaspirated, ones.
Resonants: I have argued for nasals above.
PPIE *-alV- > *-a-lV- > PIE *-e-lV; PPIE *-al\ > PIE *-ol
sounds reasonable if PPIE had "thin" /l/ in open, "thick" /l\/ in
closed syllables (cf. English)
PPIE *-alV- > *-a-rV- > PIE *-e-rV; PPIE *-ar. > PIE *-or
is OK if PPIE had regular /r/ in open, retroflex /r./ in closed
syllables
Semivowels: Traditionally, that means reconstructed ablaut series
eI/oI/i and eU/oU/u, ie. a combination of the ablaut vowel e/o/zero
with semivowel/wovel y/i and w/u. That means the whole ablaut
machinery, with the morphological processes involved, is based on a
single vowel, PPIE /a/, which in order to accomodate the necessary
inflection paradigms must occur in all roots, leading to the unlikely
single-vowel model of PIE. I see that differently. To me, the series
eI/oI/i and eU/oU/u are the result of phonological processes
operating on roots in -i- and -u-, similar to the ones in -a-
PPIE -a- -> PIE -e-, -o-, -
PPIE -i- -> PIE -ei-, -ei-, -i-
PPIE -u- -> PIE -ou-, -ou-, -u-
This is the situation in Old Latin, which has /ei/ and /ou/, but
no /eu/ or /oi/.
But in eg. Germanic (under Semitic influence, as Venneman believes?)
this was systematised into
PIE -e-, -o-, -
PIE -ei-, -oi-, -i-
PIE -eu-, -ou-, -u-
BTW In Danish the phoneme short [a] has two allophones which I'll
write /ä/ (a little more open than English /a/ in 'hat') and /a/
(standard a). The rules are: /a/ in closed syllables before labial or
velar, /ä/ otherwise (including open syllables). Obviously I could
rely on other features than nasality to make /o/'s out of /a/'s in
PIE reconstruction, if I should choose to do so.
Torsten