Re: The O-fix [was: The palatal sham :) ]

From: Piotr Gasiorowski
Message: 30917
Date: 2004-02-10

----- Original Message -----
From: "elmeras2000" <jer@...>
To: <cybalist@yahoogroups.com>
Sent: Monday, February 09, 2004 3:32 PM
Subject: The palatal sham :) (Re: [tied] Re: Albanian (1))


> I presented the analysis leading to the discovery of this element in
a congress paper, Akten der VIII. Fachtagung der Idg.Gs. (held in
Leiden 1987), which appeared in Innsbruck 1992. A fuller expose is
contained in my thesis from 1989, entitled Studien zur
Morphophonemik der indogermanischen Grundsprache (also Innsbruck).
Additional examples keep coming all the time, as do new information
about particular root structures. In the literature the matter is
generally reduced to "laryngeal loss after o-grade" which is now
commonly referred to as "Saussure's rule", because Saussure said he
saw no reflex of the laryngeals in a few words which all had o-
vocalism (footnote in a paper from 1904). The other properties of
the element seem not to have been addressed by anybody else. I have
found some echo on this list, but other than that, it does not
appear to have entered the scholarly debate. This is a true story.

I, for one, find the idea very interesting. I was planning to start a debate
about your "O-fix" when I'm done with other things. I've been making notes
for it for some time -- it seems to me that there are some unexplored vistas
of IE grammar where further support for your hypothesis could be found. I
haven't posted anything yet because the matter is fairly complicated and
tied up with my ideas concerning other phenomena that we both seem to be
interested in (especially the phonological aspects of composition and
reduplication), and the whole thing isn't quite ripe yet. What I'd like to
mention now, very briefly, is an unexpected o-grade in adjectives such as
Gk. poll�-. This is sometimes explained away as a historical product of
conjectural acrostatic *p�lh1u-/*p�lh1u-, which however leaves other
problems unexplained -- one of them is the schwebeablaut in the
well-attested comparative *pleh1-jos-/*pleh1-is-. My idea is that we should
start with derivatives of *pleh1-, such as *pl.h1-r�- ~ *pl.h1-n�-. We get
*pl.h1�- as the compositional variant (note that the most common *-u-
adjectives are notoriously found as first elements in compounds, which is
especially true of *h1s�-, but certainly also of *pl.h1�-). Another path of
derivation starts with *pl�h1-mn. (Gk. ple:ma, Lat. ple:mina:re), from which
we get the adjective *{O- + plh1mn- + -�-} --> *poln�s (note the
characteristic reduction of *h1) > poll�s. My idea in this case is that the
*o-infixation is a _phonetic_ process (actual metathesis), not a
_morphological_ one, which is why the *o is inserted after the initial
consonant (rather than replace the original root vowel), giving rise to a
schwebeablaut alternation. Now it's forms like Germanic *felu- 'much' (=
substantivised *pl.h1u-) or Skt. p�-par-ti (< *p�-pelh1-ti) that have to be
treated as secondary analogical creations, but they are neither numerous nor
particularly problematic.

This o-grade appears sporadically in *-u- adjectives as well, cf. Gk. pol�s
(a cross between *pal�- and poll�-?). Interestingly, in cases like *ohk^�-
we have _pre-laryngeal_ *o, which I take to be a sure indication of its
O-fixal origin (no matter how we account for its presence in a *-u- stem).

There's much more to share, but I'm not ready to show my hand yet.

PIotr