On 2006-08-29 01:26, tgpedersen wrote:
> BTW, I reread the parts of Jens' 'Studien zur Morphophonemik
> der indogermanischen Grundsprache' where he introduces his
> R-infix which explains (part of) IE o-grade (I had gotten
> the impression from somewhere that he derived that infix
> ultimately from an R-prefix, but I can't find it in the text?).
>
> Any way, two things he doesn't explain (as far I could tell)
>
> 1) the o-grade of perfect sg.
That's a different phenomenon, connected with other kinds of
o-reduplication (see below). The o-grade of the perfect is specifically
accented, to begin with, whereas the o-grade resulting from infixation
is typically pretonic -- except in cases of evident accent shifts.
> 2) the semantics of that R-affix: what does/did it mean?
>
> So I had an idea:
> suppose that prefix was PPIE *a- (or *an- ?) and that it was
> identical to the verbal augment PPIE *a, PIE *e-, and
> that it changed the vowel of the root it was prefixed to
> by means of some type of 'progressive umlaut' (the e-grade
> of the augment in its classical sense would then be because
> it was only joined to the verbal root at a late time, after
> ablaut had run its course)?
It has already been pointed out on this list that the phonological
features of O-fixations are similar to those of second elements of
compounds. So little has remained of the first member, however, that its
full form can hardly be recovered through internal reconstruction within
IE. Nevertheless, Jens provides rather good arguments in favour of
regarding the infixed *O as some kind of sonorant whose vocalised reflex
is visible as PIE *o. Had it been a real full vowel originally, we would
expect accent retraction in all O-fixations.
> In other words
> perf
> PPIE 3sg *a-man a -> *am-an a -> *am-on e -> PIE *mon-e
> (cf double negation in Afrikaans 'nie <verb> ... nie')
> PPIE 3pl *ma-man-an -> *me-mnen- etc
>
> Note the two differnt syllabifications.
>
> The perf. sg. can't always have had reduplication; two full
> vowels in the stem is one too many.
A _reduplicated_ stem by definition provides more room. The perfect is
not the only IE reduplicated category with the o-grade. We also have the
intensive with full(er) reduplication, *CeR-CoRC-, best preserved in
Indo-Iranian. The original pattern must have been something like 3sg.
*gWH(e)n-gWHón-ti 'strikes and strikes' (with some reshaping reflected
as Ved. jáNgHanti, note the absence of palatalisation in the second
syllable), 3pl. *gWHén-gWHn-n.ti (i.e. *C(e)R-CóRC-/*CéR-CR.C-).
Thirdly, the athematic reduplicated aorist with the probable structure
*Ci-CóRC-/Cé-CR.C-. It's still disputable whether non-intensive
athematic reduplicated presents had the shape *Ci-CóRC- (like the
aorist) or *Ci-CéRC- in the singular. Both possibilities have been
argued; the plural was *Cé-CR.C- in either case. What the perfect has in
common with these formations is its disyllabic but athematic structure,
so the vowel contrast between the two syllables may be a manifestation
of a more general dissimilatory tendency (*CeRC-CoRC-, with later
simplifications, not unlike <ding-dong>, <criss-cross>, <bric-à-brac>,
<Zizou> etc. -- note that in such "ablaut reduplication" the typical
vowel pattern is front/back, also in languages unrelated to English or
French). This is also Jens's explanation, independent of his O-fix theory.
Piotr