a.p. a verbs
kra"doN, kra"sti
It's interesting to compare this verb with a.p. c kla~doN,
kla"sti. Both are supposed to be backformed on the
athematic imperative *-dhí (like idoN, êdoN, boNdoN), and
both have *h2 in their root, yet one is AP(a) and the other
AP(c). The assumption that the mobile present of klasti is
based on zero-grade *kl.h2-dhí doesn't help much, as Hirt's
law would normally apply there too (and why didn't the same
happen in the case of *kr.h2-dhí?). My conjecture, in the
case of kla"sti, that the oxytone suffix *-dhí blocked
Hirt's law (cf. *-ah2-kós > -a:kÚ, or *plah2-nós > Latv.
plâns) explains the lack of retraction, but then why didn't
the same apply to kra"sti? Well, perhaps Hirt's law
*didn't* work here (as suggested by Latvian krât "gather,
earn", which, however, is not based on the imperative), and
the verb was originally AP(b), like jIdoN "go" (< *i-dhí),
the ictus being pulled back later by -Dybo. That still
doesn't explain why klasti is mobile and krasti not, but
perhaps that's simply the tendency of medium-frequency
("normal") verbs to become mobile (given a classification
where krasti is +low-frequency, iti is +high-frequency and a
verb like klasti is marked for neither). [I still find it a
bit strange that "steal!" was in the same league as "go!",
"be!" and "put down!", but cf. the Latvian semantics].
ê"doN, ê"xati
Also based on the imperative (*jah2-dhí). The infinitive is
based on the IIa s-aorist (*je:h2-s- > *jêx-), which is
itself not attested (in OCS) (we have aor. I <jadU>, and IIb
<jadoxU>).
n-infix verbs:
seN"doN sê"sti
leN"goN legti'
sta"noN sta"ti
boN"doN by"ti
The roots (as found in the infinitive system) are *sed- (>
*se:d- by Winter's law), *legh- (> *leg-), *stah2- (>
*sta:-) and *bhuh2- ~ *bhuh2-dhí "be!" (> bu:d-). Unlike
Lithuanian, Slavic normally uses the suffix -né instead of
the n-infix. However, in these verbs of position (sit, lie,
stand, be), Slavic maintained (or rather, introduced) the
infix. The acute accent in leN"goN is presumably analogical
after the other verbs in this mini-class: the infinitive
system is mobile.
Verbs in -V:ti:
bri"ti, c^i"ti, s^i"ti, (bi"ti)
vy"ti, kry"ti, my"ti, ny"ti, ry"ti
Already discussed previously. Note that in my new
interpretation (the "-Dybo" law), ry"ti is regular even if
the laryngeal is *h2/*h3 (as suggested by Toch.
rwa:tar/rwa:tsi). Pres. *ruh2-jé-, inf. *ruh2-téi would not
have triggered Hirt's law (due to the previous action of
Francis/Normier), and would have remained AP(b) until the
Meillet/+Dybo/-Dybo reorganization, when the verb was
retracted to AP(a). Only _mobile_ paradigms with *ih2/3,
*uh2/3 are evidence for Francis/Normier in Slavic
(especially such clear cases as z^iti, byti, piti).
grê"ti, spe^"ti
du"ti, -u"ti, c^ju"ti
zna"ti
The verbs in -(j)u"ti (*-eu&-, *-ou&-) were necessarily also
unaffected by Hirt's law, and shifted to AP(a) by -Dybo
(being jé-verbs, they were originally AP(b)).
=======================
Miguel Carrasquer Vidal
mcv@...