--- In
cybalist@yahoogroups.com, "P&G" <petegray@...> wrote:
> Sanskrit (and one or two other languages) shows signs of
laryngeals in its
> earliest stages. In particular for your purposes, ablauting -o-
is treated
> differently from -o- < -h3e- /-eh3-. The ablauting -o- appears
long in an
> open syllable, whereas the laryngeal-coloured -o- does not. This
means that
> the two main sources of -o- in IE languages were still distinct in
Sanskrit.
> Wovel colouring may therefore be quite late.
Not so: Even if the o's of *H3ów-i-s and *dór-u sounded differently
at the time Brugmann's law operated (which I do not believe was the
case, but still, IF), it does not follow that vowel coloration was
post-PIE. *All* IE languages reflect full coloration of /e/ by /H2/
and /H3/, thee only thing left to quarrel about is whether the
result of H3-coloured */e/ was identical with /o/ from other
sources. However, the whole business is a mirage: The
unlengthened /a/ of Sanskrit ávis is no more of a special case than,
say, the unlengthened /a/ of pátis vs. Gk. pósis. In *pótis there is
no way the /o/ could be H3-triggered, so we *do* have Sanskrit
unlengthened reflexes of IE /o/ from other sources than /H3e/. What
source may that be? Ní h-ansae. It must be the unlengthened vowel of
the weak cases where the old /o/ was in a closed syllable, such as
the dative *pót-y-ey > Skt. pátye. Now, the weak cases of 'sheep'
also have closed syllables, dat. ávye. So, the nom. may very well
have been IE *Hów-i-s after all.
If the expert statements concerning Tocharian a:w assigning that
vocalism to an IE a-vocalism are correct, then the ablaut must have
been -o-/-e- preceded by */H2/, NOT */H3/, i.e. PIE nom. *H2ów-i-s,
dat. *H2áw-y-ey (from *H2éw-y-ey). That makes the vocalism of nom.
ávis completely identical with thhat of Skt. á:yu 'age' from *H2óy-u
which must have /H2/ because of the a-forms of Gk. aieí, Lat. aevus
which are based on *H2ey-w-. Now, in á:yu there *is* lengthening, so
this is not only a phonetic question, it is also a matter of
presence or absence of analogical influence.
> Quantative ablaut is hard to establish. It seems to have two
main sources:
> (a) lengthening in the nominative of animate nouns ending in -
r, -n, etc
> (b) lengthening due to laryngeals.
There were also lengthened vowel in roots, as Ved. stáuti 'invokes'
which must be *sté:w-ti. There is no known rule to produce length
here, so it probably is a direct device to express what is meant,
perhaps the durative aspect of the present stem? "Laryngeal length"
is not normally regarded a fact of ablaut, but that is of course a
matter of terminology which one is at liberty to change.
The events of qualitative ablaut e/o are of very different sources
and ages, and much of it has occurred in between - or may even be
identical with - events generally regarded as belonging to
quantitative ablaut. In fact, it is COMPLETELY misguided to
subdivide ablaut into just two kinds, quantitative and qualitative
and, on that basis, ask the question which one came first.
Jens