How weird is Hittite? Not weird enough :)

[Vidhyanath Rao]

Miguel Carrasquer Vidal <mcv at wxs.nl> wrote:

> I don't know, maybe the -i originally marked something else
> (imperfective?).  For PIE, all we can recover is that it marked
> the present.  Compare Akkadian, where the unmarked form (iprus:
> -C1 C2 V C3) was the simple past, versus marked perfective
> [perfect] (ip-ta-ras: -C1 ta C2 V C3) and imperfective [durative]
> (ipar-r-as: -C1 a C2C2 V C3) forms.

I am completely ignorant of Akkadian. But it seems that the
durative/imperfective can also be used in the past. Then the
simple past may have been a perfective limited to the past
(according to Dahl, Tense and Aspect Systems, this is quite
common) and that being zero is found in a few places. But that is
not what you are proposing for PIE, if I understand you right.

[I have edited things out to save bandwidth. Hopefully those
interested can go to the archives if necessary.]

> The point is that neither the Armenian nor the Baltic and Slavic
> *imperfect* are simply made from the present stem + secondary
> endings.   Only Greek and Indo-Iranian make the imperfect that
> way.

What about the forms Szemerenyi quotes, Armenian eber, Slavic vede
and mino (with a cedilla under the o) as going back to forms made
from Indo-Greek present stem?

If you mean that these are aorist in Arm/Slavic, then aren't you
comparing apples and oranges here? If Vedic imperfect was not
imperfective, how can we compare it to Armenian, Slavic or Baltic
imperfects (the last of which is said to be past frequentative)?
This is what I meant by basing comparisons on names rather than
syntactical roles. [BTW, note that Vedic has a past habitual
(pura: [sma] saca:vahe) based on the present (not the present
>stem<) and asacat does not carry this meaning.]

Thus, my questions is, if present stem did not carry imperfective
meaning by itself in a stage that included these languages, why
should be posit it for a stage that includes Vedic when Vedic
`imperfect' is >the< tense of narration?

> Slavic does have some root *aorists*.  The Armenian aorist, apart
> from the 3rd.p.sg., cannot be derived from either root imperfects
> or aorists.

But the forms Szemerenyi cites are not root forms.

Turning now to the perfect: If resutative > past is known from
elsewhere, but past > resultative is unknown, then the languages
in which the reflex of the perfect is a (perfective) past have to
be considered as showing a later stage than those in which the
perfect is a resultative. Unless the direction of evolution can be
challenged, the use of perfect outside Indo-Greek has to be taken
to be secondary.
---