I had wanted to say something about the group of a.p. b
o-grade verbs (kolti, polti, porti, borti), but I couldn't
figure out why they were a.p. b in the present system (as
against a.p. a in the infinitive system).
I was now looking at verbs with zero grade root ending in a
sonant, in particular z^Ir"ti, z^ir(j)óN ("to sacrifice",
*gWerH-) and tIr"ti, tIrjóN ("to rub", *terh1-). The
infinitives are a.p. a, as expected (*gWr.H-téi and
*tr.h1-téi, with Hirt's law *gWI'rtei, *tI'rtei), but the
present system again suggests a non-acute root, without
retraction of the accent in *gWr.H-jó:, *tr.h1-jó:. The
phenomenon is not merely Slavic, because the Lithuanian
forms show the same: gìrti vs. giriù and tìrti vs. tiriù.
Apparently, in sequences *-RHj-, the laryngeal just
disappears, at least if the *-j- belongs to the verbal
suffix *-je/jo- (which then behaves as if it was
*-ie-/*-io-, with vocalic onset). That seems a more elegant
solution than assuming that *-je- is a secondary
substitution of plain *-e- in all Baltic and Slavic forms.
In particular, it allows for a more satisfying solution of
a.p. c verbs such as z^IroN (z^ertí), stIroN (stertí), pIroN
(pertí), dIroN (dertí), skvIroN (skvertí?), mIroN (mertí).
LIV classifies most of these (except mIroN, *mer-) as
"tudáti"-presents, which I'm not enthusiastic about.
Tudáti-presents should not give mobility, and the fact that
e-verbs are almost automatically mobile in Slavic speaks
against the presence of any significant number of
tudáti-presents in Proto-Slavic. If on the other hand, all
of them were, like mIroN/mertí, original jé-verbs which lost
the palatalization after /r/, there is no problem, provided
the sequence -RHj- becomes -Rj-. We then have e.g.
*pr.H-jó:, a normal jé-verb, which remains a.p. b when the
laryngeal is dropped (*pIr-jó:), and which automatically
becomes mobile after the loss of /j/, as is semi-regular
after /r/: *pIr-jó:, *pIr-jétI > *pI'ro:, pIretI'.
The other zero-grade mobile verbs as listed for example on
p. 203 of Dybo's Slavjanskaja akcentologija (Table 89) end
in -n: pInoN (peNti), tInoN (teNti), klInoN (kleNti). The
analogical model here is tInoN < *tm.-né- (root *temh1-), a
real né-verb, after which the others presumably also adopted
né-conjugation (e.g. *pn.-né- from *(s)penh1-, which was
earlier perhaps athematic *pénh1-ti ~ *pn.h1-énti =
pn.nénti). Given Lith. tìnti, tinù; pìnti, pinù this was
Balto-Slavic. In Slavic the forms were eventually
re-interpreted again as e-stems (pIn-oN, tIn-oN) and thus
became mobile (similarly, noN-verbs ending in a vowel
[minoNti etc.] also became mobile).
mInóN (meN"ti, *menH-, Lith. minù, mínti) and c^InóN
(c^eNtí, *ken-) for some reason escaped reanalysis as
e-stems and remained a.p. b, as pseudo-né-stems.
That leaves only dUmóN (doN"ti, *dhmeH-), z^ImóN (z^eNtí,
*gem-) and jImóN (jeNtí, *h1em-) as potential tudáti-verbs
among the verbs listed by Dybo in the aforementioned table,
but there are other possibilities (Lith. dùmti, dumiù is a
jé-verb, as is Slav. h1em-jé- > jemlj- [but not Lith. imù
im~ti]).
=======================
Miguel Carrasquer Vidal
mcv@...