Last time, I suggested that AP(b) neuters like *pteróm
(peró) directly reflect the position of the PIE ictus (in
other words, AP(b) is here not due to Dybo's vel
Illich-Svitych's law). The fact that peró and friends
already occupied the o-stem neuter AP(b) spot at the time of
Dybo's (I-S's) law is the most likely explanation for why
the non-acute barytone neuters (dvòrU and friends) shifted
to o-stem AP(b) _masculines_.
Having now studied the verbs a little closer, I've come to
think that a similar phenomenon also occurred in the
accentual paradigm system of the verb.
I'm going to assume that mobility (or lack of) in the Slavic
verb, like mobility in the Slavic nouns and pronouns, is a
development of PIE mobility (or lack of).
If we look at the verbal categories as listed in LIV, we see
that the PIE accent types were approximately:
(I) mobile
1a. ghWén-ti ~ ghWn.-énti
1e. térh2u-ti ~ tr.h2-ú-nti
3a. bhe-bhóid-h2a ~ bhe-bhid-é:r
5a. wéid-s-ti ~ wid-s-énti
7a. mn-éh1-ti ~ mn-h1-énti
(2) barytonic
1b. dé:k^-ti ~ dék-n.t
1d. k^éi-e
1g. dhé-dhoh1-ti ~ dhé-dh&1-nti
1n. bhér-e-ti ~ bhér-o-nti
1r. spék-ye-ti ~ spék-yo-nti
1t. kléuH-de-ti ~ kléuH-do-nti
1u. pléh1-dhe-ti ~ pléh1-dho-nti
1v. plék^-te-ti ~ plék^-to-nti
2b. dhé:ig^h-s-t ~ dhéig^h-s-n.t
2c. wé-wkW-e-t ~ wé-wkW-o-nt
4b. swó:p-ye-ti ~ swó:p-yo-nti
6a. kWér-kWor-ti ~ kWér-kWr-n.ti
(3) mesotonic
1h. sti-stéh2-ti ~ *sti-sth2-énti
1i. si-sd-é-ti ~ si-sd-ónti
1k. li-né-kW-ti ~ li-n-kW-énti
1l. h3r.-néu-ti ~ h3r.-nw-énti
1m. tk^i-néH-ti ~ tk^i-nH-énti
1o. sup-é-ti ~ sup-ó-nti
1p. gWm.-ské-ti ~ gWm.-skó-nti
1q. g^n.h1-yé-ti ~ g^n.h1-yó-nti
1s. tr.p-éye-ti ~ tr.p-éyo-nti
4a. mon-éye-ti ~ mon-éyo-nti
5b. wi-wn.-sé-ti ~ wi-wn.-só-nti
8a. lip-h1yé-ti ~ lip-h1yó-nti
(4) oxytonic
1c. tuké
2a. gWém-t ~ gWm-ént
As is the case in the Slavic noun, the mobility of the
athematic types (e.g. 1a) was transferred to the normal
thematic type (1n), and both types (and closely related
ones) are now usually mobile (APc) in Slavic, if we
disregard exceptions due to Hirt's law.
However, the nasal presents (1k, 1l, 1m), yé-verbs (1q), a
part of the causative-iteratives in -éye- > -i:- (mostly the
iteratives) and an isolated ské-verb such as <iskati>
predominantly follow AP(b). I see no reason to think that
the ictus ever changed from PIE to PSlavic in verbs such as:
-noNti
*bhudh-né- bUnóNti
*dhus-né- dUxnóNti
*h2sus-né- sUxnóNti
*lip-né- lInóNti
*ml.kW-né- mlUknóNti
*muk-né- mUknóNti
*pis-né- pIxnóNti
*sup-né- sUnóNti
*tuk-né- tUknóNti
-ye-
*gWr.H-yé- z^IrjóN, z^rUtí
*lugh-yé- lUz^óN, lUgáti
*pis-yé- pis^óN, pIsáti
*h1m.-yé- jemljóN, jImáti
*h2mg^h-yé- veNz^óN, veNzáti
*ghWedh-yé- z^eNz^dóN, z^eNdáti
-ske-
*h2i-ské- jIskóN, jIskáti
-e-
*dhm.H-é- dUmóN, doNtí
*h1i-dh-é idóN, ití
*kes-é- c^esóN, c^esáti
*ghWn-é- z^enóN, gUnáti
*mn.H-é- mInóN, meNtí
*suk^-é- sUsóN, sUsáti
*h1m-é- jImóN, jeNtí
*kn-é- c^InóN, c^eNtí
*tuk-é tUkóN, tUkáti
As we also saw in the nouns, the presence of a class of
AP(b) verbs _before_ Dybo's law goes hand in hand with a
merger of categories in the common thematics: verbs in -e-
do not neatly distinguish between AP(b) and AP(c) [see
"Osnovy slavjanskoj akcentologii", pp. 36-37], although
traces of AP(b) e-verbs can be observed.
We can compare:
masculine o-stems:
barytone non-acute -> merge with mobile
neuter o-stems:
PIE oxytones have AP(b) before I-S's law
barytone non-acute -> shift to masculine
verbal e-stems:
barytone non-acute -> merge with mobile
verbal -je/ne-stems:
PIE mesostatics have AP(b) before I-S's law
?? any strange shifts ??
The verbs in -iti are special. The PIE accent was
mesostatic (mon-éye-ti, mon-éyo-nti). Slavic has split the
group in two: iteratives retain the mesostatic accent
(AP(b)), causatives have become mobile.
Within the AP(b) group, it is customary in Slavic
accentology to distinguish between an AP(b/1) and an AP(b/2)
group. The first (mainly the iteratives) are subject to
Stang's law, which retracts the ictus from a medial long
(circumflex?) syllable:
nosí'ti => nosjóN, nòsi:s^I, nòsi:tI
The other group (mainly denominatives and causatives that
for some reason didn't become mobile) is not subject to
Stang's law (or only in certain areas under certain
conditions):
loz^í'ti => loz^jóN, loz^í:s^I, loz^í:tI.
The denominatives of course do not derive from the PIE
iterative-causative in -éye-. Like the Latin i:-stems, they
originate in denominatives of i-stem nouns to wich the
suffix -yé- was added (-i-yé-), and this type spread
analogically (as a compund suffix -iyé-) to derive
denominative verbs from all kinds of stems.
Causative/iterative -éye- was reduced to -íye- (cf. i-stem
nom.pl. -Ije < *-eyes, trIje < *treyes "3"). Then, both
causative/iterative -íye- and denominative -iyé- were
reduced to -i:-, but the former had a falling tone, the
later had a rising tone. I think that explains their
different behaviours vis à vis Stang's law.
=======================
Miguel Carrasquer Vidal
mcv@...