How weird is Hittite? Not weird enough :)

[Vidhyanath Rao]

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

> "Vidhyanath Rao" <vidynath at math.ohio-state.edu> wrote:

> >Do you mean to imply that at some point in PIE to proto-Slavic, there was
> >a point at which present and past always had different stems, so that when
> >we see the same stem in present and aorist in Slavic, it must be an
> >innovation?

> No.  All IE languages that distinguish a present stem from an
> _aorist_ (not past) stem have I think some verbs where the
> distinction is not made.

If aorist is post-PIE, all inherited aorist forms must have been simple
pasts in PIE, right? I think that I don't understand the relative chronology
you have in mind.

> Verbs like vesti (ved-) can make  root-aorists, but also s-aorists
> (two kinds of them):

I think that the only question that matters is which are inherited and which
are new in Slavic. We find old forms supplementing newer forms in many
languages. Just because a form is found imbedded in a new paradigm does not
mean that it is also new.

> That's true, but there is more.  You forget that imperfective
> *presents* and perfective *pasts* also arose.

I am not sure what you mean here. I understood that your theory was that at
the point when Anatolian split off, the distinction was just present vs
past, that is to say, past was not limited to perfective. Root,
reduplicated, -neu and -ske/o and probably nasal presents existed at this
point. So their pasts were not limited to imperfective. If a new
imperfective present arose before Vedic, it can only have been bhereti type
thematics. But other presents abound in RV. What do we do with them,
especially their pasts which date back to the time before aspect?

There is another peculiarity here: Outside Greek, the imperfective past is
often formed not from the present stem but something else. To me it looks
like a new imperfective past (not a new imperfective that formed both
presents and pasts) arose, making pre-existing pasts into `perfectives',
while old (''Indo-Hittite'') present continued to exist (though
thematicization, either instead of or in addition to other formants kept
gaining ground).

>[deleted comments on -ske/o and -s etc.]
> Forms like Skt.
> gacchati / agacchat (*gwm-sk-e-ti, *e-gwm-sk-e-t), whatever their
> synchronic syntactic function or meaning, are historically
> iteratives, i.e. imperfective Aktionsart.  What's the problem?

The problem is that, to me, it is not obvious that gwmske/o- was an
iterative only and not directed durative. [This was basically Meillet's
suggestion].

I also object to ``iterative, i.e imperfective'' on general grounds:
`imperfective' and `perfective' should be limited to languages with a binary
contrast. If there are several different stem formations, but without binary
contrasts of complete/total vs incomplete/partial, and with telic stems
forming both true presents and simple pasts, the language does not have
aspect. [It might be on its way, but not completely there yet.]

[agacchat occurs once in Mandala 3, once in Mandala 8 and several times in
Mandalas 1 and 10. There are a couple of instances in the family books where
we have -Agacchat, where sandhi interferes.]

> One problem is that root-presents and root-aorists, which are
> both direct cousins of the Hittite simple past (mi-conjugation),
> get classified in different categories.

>(and, if I understand you correctly, Iranian) imperfect
> from the optative,

Past habitual in Old Persian and Avestan, not a general imperfective, and I
don't know anything about Middle Iranian. [Mahabharata and to a lesser
extent, Ramayana have optative endings used in past sense; but I have not
seen any special meaning attributed to them.]

> If what you're saying is that a Vedic/Skt. "imperfect" like
> <avahat> "he carried", without specific imperfective markers,
> goes back to a simple past tense as in Hittite or indeed Germanic
> [except for the augment, of course], I fully agree.  But the
> question is, what happened to the meaning of that form when
> s-aorists like <ava:ks.i:t> arose?  Did it become an
> imperfective/habitual past in pre-Vedic (e.g. proto-Indo-Iranian)
> like the Greek imperfect, or did it remain a simple aspectless
> narrative past?  And if the latter, what about marked
> "imperfects" like <agacchat>?

[avahat is marked (in contrast to root opt uhi:ta or imprv voLhvam), no? ]

The first question is precisely what I asked. To me, it seems to be simpler
to pick the latter rather than to assume an oscillation from past to
imperfective to past. agacchat is no problem: it was a telic and so its past
was not limited to imperfective. I see root extensions in -s as possible in
pre-IE and slowly becoming grammaticized, leading to sigmatic aorists. Vedic
sigmatic aorists (which continue to expand, becoming more common in
Brahmanas than in RV, while root `aorists' becming less common) were perhaps
``completives'' (a la Bybee et al, The evolution of grammar). The other
languages with sigmatic forms would have gone through such a stage, but the
more clearly marked sigmatic forms ousted the other pasts, or like in Greek,
expanded in frequency becoming perfective the same way that the prefixed
verbs did in Slavic. A similar thing happened in Pali, where the sigmatic
forms form the majority of the preterites, but with many more old
``imperfects'' mixed in than we see in Balto-Slavic.

The problem with my argument is with *avahat. If *uegh'et started as a
subjunctive, then *euegh'et must have had an conative meaning, with a root
form *(e)weght (not *(e)weghst) being the past. If this was the case, then
Greek expanded the imperfective meaning to all stems usable in the present
from this starting point, while IA eroded the past subjunctive meaning in
analogy with -sk-, -neu- and -n- stems which were limitd to telic functions
(not general imperfective). Of course, if thematic present and the
subjunctive developed from a common origin (Renou's ``thema eventualis''
(sp?) ) it would be easier to explain the Indic, while for Greek, we need to
assume expansion of a `completive' to perfective.