On Mon, 28 Feb 2005 09:14:05 +0000, P&G
<G&
P@...> wrote:
>On this whole topic of the PIE perfect and past tense, there is an
>interesting article by Helena Kurzová, "From Indo-European to Latin", which
>suggests that PIE had a two-way distinction in nouns and verbs. Nouns were
>animate or inanimate, verbs active or "inactive". She suggests that these
>lexical verb classes were determined by meaning, and had different endings,
>the active verb showing consonantal aorist & present endings, the inactive
>verb showing vocalic or laryngeal endings. The inactive verb indicated
>statives and processes.
>
>The active verb later develops into imperfective (present tense) and
>perfective (aorist tense), while the inactive develops into verbs of state
>(perfect) and verbs of process (middle). Because the difference is lexical,
>we get root presents (which develop aorists by suffixation), root aorists
>(which develop presents by suffixation), root perfects (surviving as fossils
>in some IE languages),and root middles (the media tantum, many of which
>develop new active forms).
>
>She offers the Hittite middle as the best representative of the PIE inactive
>endings that developed into perfects in Greek & Latin. (es-ha-hari,
>es-ta-ri, es-a(-ri) ).
>
>She even suggests active roots were CeC, while inactive were CRC or CHC.
Speaking of root structure.
I was trying to investigate whether there was a special
connection between the s-aorist (especially the s-aorist
conjunctive), the desiderative (I just wrote "precative" in
another message: I meant "desiderative") and
*-sk^e-presents. They all share a common element *-s, so I
checked in LIV whether there's a basis for a connection. I
also remembered Jens' theory that there is a connection
between the s-aorist and *-sk^e.
At first sight, the latter would seem to be improbable:
*sk^e-presents are marked presents, which we would expect to
be made primarily from root-aorists. And s-aorists are the
productive way (in IE languages that have an s-aorist) to
form aorists from root-presents.
Still, I found some 15 cases where a primary s-aorist and a
ske-present are made from the same root (against 38 cases of
root aorist + ske, as expected).
I noticed a tendency for these cases to have a root
structure CeC, and especially CeH:
(stop) *weg^h-, *tres-, *tep-, *prek^-; (resonant) *men-,
*ghWer-; (laryngeal) *(s)neh2-, *peh2(i)-, *mneh2-, *yeh2-,
*g^neh3-.
Exceptions are:
*myeuh1-, *kerH-, *h2leu-, *h2eis-
Almost all roots of the structure *C(C)eH which make a
sk^e-present also make an s-aorist. The statistical
significance of CeC and CER roots seems less clear.
This reminded me of the fact that in Vedic, stems in -a: (<
*-eH) are special in that they have the ending -ur (< *-r.s)
in the _imperfect_, where it doesn't belong. The same is
also true for a few roots ending in -eC (e.g. tvis.-).
What can be the reason for this?
Following the reasoning above, characterized presents such
as *-sk^e should be made from root aorists. The root aorist
of roots ending in a laryngeal (or sometimes a single -C) in
Vedic has the paradigm:
-m
-s
-t
-má
-tá
-úr
(other verbs usually have 3pl. -án, from the imperfect)
The 3sg. ending was originally *-s (the aorist third person
marker), to go with 3pl. *-érs (in Vedic equated with *-r.s
from the s-aorist).
The aorist paradigms can be reconstructed as:
active stative
*-m *-h2e
*-s *-th2e
*-s *-s
*-érs *-r.s
For how the stative aorist became the s-aorist see my
previous messages. As to the root aorist, I had assumed
that there, 3sg. *-s had been replaced by *-t very early on
without leaving a trace, while 3pl. *-érs was replaced by
*-ént in most verbs, with only the category of -eH verbs
(etc.) being more conservative.
I now realize that some verbs of this root structure were
even more conservative than that: they even retained 3sg.
*-s!
So a form like *g^noh3-s-t essentially _is_ a root aorist,
and as such it's normal that it should make a characterized
present *g^nh3-sk^é- (although I suspect it's no coincidence
that it's specifically a *sk^e-present, so Jens is right
that there is a connection between the two). Likewise,
Jasanoff is right that there is a connection between the
hi-past/s-aorist and Hitt. ganeszi (*gnoh3-s-t(i)).
=======================
Miguel Carrasquer Vidal
mcv@...