Re: [tied] Timing of ablaut

On Sun, 28 Sep 2003, Miguel Carrasquer wrote:

> On Sat, 27 Sep 2003 19:17:24 +0200 (CEST), Jens Elmegaard Rasmussen
> <jer@...> wrote:
>
> >> Wouldn't a form *sta:inV- also give circumflex?
> >>
> >With lengthened-grade PIE /a:/? No, that yields acute (Lith. íes^kau). I
> >am not sure whether it is relevant here (have lost your point a bit),
> >but long-diphthong roots form o-infix derivatives with acute in BSl.,
> >i.e. a root *steH2y- should form **stóyH2-no-, not *stóy-no-; the latter
> >can apparently only come from *stVyH-.
>
> The point I failed to make clearly was that ste^na/stains would be a
> thematized r/n-stem, and since the r/n-stem in question is *sta:yr. (the
> usual prototype of Greek stéa:r), the thematized version should be
> *sta:inos/*sta:ina:.  I don't think <íes^kau> is quite parallel in
> structure (but I'm not sure what the underlying structure of

íes^kau
is).

> In terms of the place of the laryngeal, *staHinos would be more
> analogous,
> I think, to the a:-stem dat/loc *-ah2(a)i, which has circumflex in B-S.

A preform *staHinos would change into *stayHnos in (i.e., before) PIE
already. Lith. íes^kau 'I seek' sure *is* parallel: the root is posited
by Pokorny as *ais-, the sk^-prs. seen in Skt. icchati is regularly
combined with an s-aorist, so lengthened grade is expected, the result
being *H2e:ys- which neatly explains the Baltic acute.

The dative *-aH2-ay is circumflex because it is disyllabic (the dative
desinence is immune to ablauting).

>
> >> I don't accept unmotivated /a/.
> >
> >I accept anything I find.
>
> Filtered through a theory, of course.

No, we make the theories on the basis of what we find - of anything we
find.

>
> >I am in no position to tell God what kind of
> >world to give us to analyze. I do try and reach for the simplest
> >understanding possible, but when things prove impossible I accept the
> >verdict of the material. Sometimes, however, a nice solution has only
> >been impossible *to me* because I got something wrong. The matter at
> >hand is particularly tenacious.
>
> Yes, there are /a/'s which I cannot understand (hopefully, which I cannot
> _yet_ understand).  But in this case, assuming *h2 in the root (given the
> obvious semantic ties to *stah2-, not hard to do) explains the /a/ in the
> most simple way for me.

If putting in /H2/ triggers a known rule to predict aspiration, and that
is not found, it is not unproblematic to accept it anyway.

> >> >
> >> >If this is from 'stand',
> >>
> >> It isn't.
> >
> >Well, you did write this (Sept. 24): "The connection with *steh2-, as
> >suggested in EIEC, seems appropriate for the semantics of the cognate
> >group." But okay, fine, it does not have to be that way.
>
> The connection I meant with with *steh2- is "Benvenistean", so to speak,
> like, say, *wer- and *wert-.

Plea of guilty again - that means the root forms are not as closely
related as normal in such calculations.

>
> > >[...]
> >> The fact is that we have *sti(:)a:, from which the verb *stya:-ye- is
> >> derived.
> >
> >I do not understand the legend "*sti:a:". If you have /H2/ preceding the
> >/y/ in the underlying form as you plainly do, the zero-grade *stH2i-
> >should have aspirated /th/.
>
> Yes, unless metathesized.

And by what rule would it become that?

>
> >> >Then I see no valid positive evidence for a distinction between /H3e/
> >> and /o/ in the way Brugmann's law works in Indo-Iranian.
> >>
> >> You would't.
> >
> >I would if you had adduced any, and that goes for the future also. I
> >still do not exclude that there may be such a difference, only I have
> >seen no valid evidence for it.
>
> I'm satisfied with *h3ep- for the moment.  It's only a single example,
> and
> it _may_ perhaps be *h3epH- (but surely needn't be), so more evidence
> would
> certainly be welcome.

You are ducking the issue: You *need* the root to be anit for the example
to be interesting in this context. Then will you not make sure your
evidence meets the only requirement of interest?

> >Now, on closer inspection I believe one must accept that the
> accentuation
> >of the form "styá:yate" given by Pokorny and Monier-Williams from where
> I
> >have quoted it, is not based on nothing, even if textual attestations
> are
> >limited to unaccented stya:yata:m. The classification of the verb as
> >belonging to "present-class I" is made by Pa:n.ini himself. Since the
> old
> >accent was obviously living in Panini's time, there is little point in
> >toying with the possibility that the stem /stya:ya-/ was accented on
> >/-yá-/ and not on the root. If that is credible - and I don't see how it
> >couldn't be - all analyses of the present stem as either a denominative,
> a
> >factitive or a stative are out of the question
>
> I hadn't considered the accent...
>
> Not _completely_ out of the question, however.  Macdonell says:
>
> "The causative accents the penultimate syllable of the stem, as
> krodh-áya-ti 'enrages'; the passive, the secondary intensive, and the
> denominative, accent the suffix -yá-; e.g. pan-yá-te 'is admired',
> rerih-yá-te 'licks repeatedly'; gopa:-yá-nti 'they protect'"
>
> However, there is a footnote to this:
>
> "A certain number of unmistakable denominatives, however, have the
> causative accent; e.g. mantrá-ya-ti 'takes counsel'(mántra)."

Now, it is of course easy to find an analogical pattern for -áya- to
replace -ayá-, but it is completely impossible to find one for -á:ya- to
replace -a:yá-. Therefore, I would assume that examples of -á:ya- are
genuine. Still, I don't know where to look for a full list of denominative
or de nominative-like stems in Vedic, so this will have to remain a gut
feeling. But I'm sure this is being looked into.

Maybe I should add the reason I have for accepting  vi- 'sheep' as
evidence at all. The only decisive piece of evidence I know for assessing
the question whether the root ends in /-w/ or /-wH/ is the form in
Germanic: Goth. awe{th}i 'flock of sheep', ON <ae:r>, acc. <a:> 'sheep'
from *awiz, *awi(N), not anything involving gg(v) from *-ww- < *-wH-. On
this basis, and on this basis alone it seems, *How-i-s > Skt. avi-s cannot
be suspected of reflecting a closed syllable. But it can still be
suspected of having *H2a- (or even *H2o- under a somewhat different
version of the ablaut-sensitive Brugmann), or of having generalized the
short vowel of cases with a closed syllable.


Jens