Inspired by the possibility that at least the accentuation
of 3pl. -oNtI' may be explained by Dybo's law, I've been
doing some rethinking about how verbal accentuation may have
developed from PIE to Balto-Slavic to Slavic.
The five basic Slavic verbal classes are:
I e-verbs
II ne-verbs
III je-verbs
IV i-verbs
V athematic verbs
Their IE roots are:
I e-verbs (mostly barytones, a few oxytones [tudáti-verbs]
II n/ne-verbs. Mobile in PIE, but in a weird way: Baltic
retained the n-infix (barytonic), Slavic the né-suffix
(oxytonic)
III je-verbs, mostly oxytones (a few barytones).
IV iteratives/causatives (-éje-), denominatives (-i-jé-) and
essives/fientives (-eh1(i)-). The first two categories were
"oxytonic" (i.e. non-barytonic). The i/ê-verbs are unclae.
Maybe they were mobile, with Slavic generalizing the
full-grade form of the suffix in the present (*-eih1-?),
Baltic the zero-grade form (*-h1i-).
V Athematic verbs. Mostly mobile, a few acrostatic
(barytones).
The transfer of athematic mobility to the thematic verbs
affected the barytones (essentially: the e-verbs), because
those verbs were capable of throwing the ictus forward in
the plural (which is the essence of mobility in the PIE
athematic verbs).
There are two possible scenarios for the early stages:
either Hirt's law worked before Pedersen's law (the transfer
of mobility to thematic verbs), or it worked after it.
First scenario:
PIE Hirt Pedersen
ath.mobile
'-mi ,, ,,
'-si ,, ,,
'-ti ,, ,,
-més '-mes ,,
-tés '-tes ,,
-énti '-(e)nti ,, = a.p. a
them.barytone:
'-o: ,, ,,
'-esi ,, ,,
'-eti ,, ,,
'-omes ,, -omés
'-etes ,, -etés
'-onti ,, -ónti = a.p. c
them.oxytone:
-j-ó: '-jo: ,,
-j-ési '-jesi ,,
-j-éti '-jeti ,,
-j-ómes '-jomes -jomés
-j-étes '-jetes -jetés
-j-ónti '-jonti -jónti = a.p. c
Second scenario:
PIE Pedersen Hirt
ath.mobile
'-mi ,, ,,
'-si ,, ,,
'-ti ,, ,,
-més ,, '-mes
-tés ,, '-tes
-énti ,, '-(e)nti = a.p. a
them.barytone:
'-o: ,, ,,
'-esi ,, ,,
'-eti ,, ,,
'-omes -omés ,,
'-etes -etés ,,
'-onti -ónti ['-onti] = a.p. c
them.oxytone:
-j-ó: ,, '-jo:
-j-ési ,, '-jesi
-j-éti ,, '-jeti
-j-ómes ,, '-jomes
-j-étes ,, '-jetes
-j-ónti ,, '-jonti = a.p. a
Clearly, the second scenario fits the facts best (je-, ne-
and i-verbs with a laryngeal in the root are a.p. a, not
a.p. c). It is remarkable that we can conclude that Hirt's
law had no effect whatsoever on thematic mobile verbs (it
didn't act on the 1/2 pl., because of the extra syllable,
and it's no big assumption that theme-stress on the only
form that it did work on, the 3pl., was analogically
restored soon enough). This explains the facts as diagnosed
in Rasmussen's 1986 paper on Hirt's law ("*de^~joN, dejetI'
"do", da~joN, dajetI' "give", zna~joN, znajetI' "know",
pa~soN, pasetI' "protect" (Hitt. pahs-) should have
triggered Hirt's law").
Meanwhile, the verbs unaffected by Hirt's law developed as
follows:
a.p. a => ath. acrostatics
a.p. b => thematic oxytones
a.p. c => ath. mobile verbs, thematic barytones
The ranks of a.p. a were replenished by athematic verbs and
thematic oxytones affected by Hirt's law.
Winter's law in Shintani's formulation only works in the
immediately pretonic position. This means that it would not
have affected any a.p. a verbs, nor any mobile _thematic_
verbs (same situation as with Hirt's law: only the 3pl.
would have been affected). Winter's law would have worked
in the plural of _athematic_ mobile verbs (with Ausgleich
possible either way), and generally in a.p. b verbs (always
thematic and theme-stressed). That would explain why a
thematic mobile verb like stérgo: ( = Grk. stérgo:) was
unaffected by Winter's law (Slav. *ster^goN, *stergtí,
*stergla') while athematic Lith. sérgmi was. We can then
also conclude that e.g. Slavic stri~goN, stri"gti (_with_
Winter's acute) goes back to an athematic mobile paradigm
*stréig-/*strig-', after Winter's law strei~g-/stri:g-',
with acute in the infinitive (we wouldn't find acutes in the
present anyway, because of Meillet's law).
So that little extra syllable in the thematic forms makes
quite a big difference...
=======================
Miguel Carrasquer Vidal
mcv@...