Some final a.p. c verbs:
The infinitives dertí, z^ertí, pertí, stertí, volstí (if not
a borrowing from Goth.) and stergtí come from sequences
involving *-VRH-. As such they did not trigger Hirt's law,
as is evident from their oxitonicity. The presents dIroN,
z^IroN (z^eroN), pIroN, stIroN, voldoN, stergoN, being
thematic, also did not trigger Hirt's law (syllabically
*dr.-Hé, *wol&-dé-, etc.).
More problematic is why they did not trigger [-Dybo] as is
the case with all other a.p. c acute-root infinitives. The
fact that only these *-VRH-C- verbs fail to retract the
accent makes it likely that the phenomenon is connected to
the metathesis / polnoglasie suffered by these sequences
(and their circumflex *-VR-C- counterparts) in Late Slavic.
a.p. a verbs:
Both Kortlandt and Rasmussen state explicitly that Hirt's
law retracts the accent if the _immediately preceding_
syllable contains a (non-vocalized) laryngeal. Given that
both rarely agree on anything involving Slavic accentology,
this must be true. Or is it?
There's a small problem with a.p. a infinitives (as well as
l-participles and s-aorists). There are those that involve
a non-laryngeal or polysyllabic suffix: -iti and -ovati.
Strict application of Hirt's law would mean that (using Late
Common Slavic forms for convenience's sake) *stavití and
*z^alovatí => *z^alováti would not end up with the correct
placement of the ictus. Verbal formants that do contain a
laryngeal (-a-ti, -ê-ti, -noN-ti [I'm assuming this comes
from secondarily nasalized PIE *-nah2-, which I'm not sure
is the standard etymology]) would come out as e.g. dvigáti,
vidê'ti and dvignóNti, as you can only Hirt once [and there
wasn't anything anyway to trigger Hirt's law in the root of
say vidêti].
We can of course simply assume that in these cases we had
analogy after the barytone present system.
The introduction of the [-Dybo]-law fixes this problem (and
the formulation of Hirt's law by Kortlandt and Rasmussen
*is* correct): after Hirt's law, we had AP(b) forms:
*dvigáti, *dvignóNti, *vidê'ti, [*stavití], *z^alováti.
(in the inf. of the iterative i-stems, the stress was
retracted by the circumflex metatony rule [a.p. b nosi::tí
=> nosí:ti], but ê/i-verbs, denominatives and causatives(?)
may still have had an oxytone infinitive system before the
working of -Dybo. The present system of course always had
stress on the -i:-, be it acute or circumflex).
When acute-root AP(b) words became AP(a) by the reverse of
Dybo's law, the stress was retracted to the root [c.q. the
acute part of the root, as in plújoN but pljIváti, lelê'joN
etc.] (that formulation can even take care of *z^alováti >
z^álovati and *stavití > stáviti).
Given what was said at the beginning, I may still need
analogical retraction in the infinitive of e.g. vI"rgoN,
ver"gti.
Some other time, I'll discuss some interesting non-suffixed
a.p. a verbs.
=======================
Miguel Carrasquer Vidal
mcv@...