Here's something I casually mentioned a while back without
making it explicit.
The Greek present thematic endings are a classical problem.
We have:
-o:, -eis, -ei
(imperfect: -on, -es, -e)
for expected:
-o:, *-esi > -ei, *-eti > -esi.
For Slavic, I had earlier suggested (based on the data from
the Novgorod birch-bark inscriptions), that the endings
*-eti > -etI and *-et > -e, -etU (c.q. *-onti > -oNtI and
*-ont > -oN, -oNtU), which both occur in Slavic, can be
derived from the indicative present (*-eti) and the
subjunctive (*-et). Jens suggested instead an injunctive
*-et (*-ont). If thematic presents derive from subjunctives
anyway (and the "double thematic" subjunctive is a
relatively recent formation), the two forms are historically
identical anyway.
In the case of Greek, the injunctive explanation fits the
facts better. In any case, what I would suggest is that
like in Balto-Slavic, pre-Greek used both indicative *-eti
and injunctive *-et as present tenses (originally perhaps
with a modal distinction that was soon watered down). In
the end, *-et completely replaced *-eti, and we would have
had:
phero: epheron
pheresi epheres
pheret epheret
After loss of -t:
phero: epheron
pheresi epheres
phere ephere
To distinguish the present from the injunctive proper, an -i
was added (after 2sg. pheresi) in the 3sg. present:
phero: epheron
pheresi epheres
pherei ephere
After loss of -s-:
phero: epheron
pherei epheres
pherei ephere
And finally addition of -s in the 2sg. (after preterite
-es):
phero: epheron
phereis epheres
pherei ephere
This way, the facts in Greek can be brought into agreement
with the facts in Balto-Slavic (where both *-et and *-eti
are reflected) and Old Irish (biur, bir, be(i)r must come
from *bhero:, *bhere(s)i, *bheret, cf. 2pl. *bherete >
berid, berith). The Latin merger of *-eti > -it/-et and
*-et > OLat. -id/-ed (in favour of *-eti) can perhaps be
ascribed to the same cause.
=======================
Miguel Carrasquer Vidal
mcv@...