How weird is Hittite? Not weird enough :)

[Vidhyanath Rao]

I meant to respond to this, but had other things to attend
to. But I would like to resurrect this thread.

wrote, (eons ago)

>Exactly, the fact that the imperative has the secondary (short)
>endings shows that these were the unmarked, neutral ones.
>Present (Non-past) = neutral + ``here and now''.  The forms
>without the -i extension then become past forms (aorist or
>imperfect) by default.  But the distinction was already in
>Anatolian (-mi present vs. -m past).

According studies of contemporary language, present vs non-present
distinction is rare to non-existent (Comrie says the latter in his
``Tense''). Zero past does not seem to occur among languages with forms
restricted to past that is obligatory. It seems better to assume that SE
marked only person and number and could indicate past only when implied
by context or by inference. But then there must have been some way
(particles?) of overtly marking past when needed. Now, there is some
evidence for augment outside Gr.Arm.I-Ir (Rasmussen indicated some in a
post to the previous incarnation of this list). It is possible that the
augment disappeared as the past became overtly marked otherwise (we can
see this in progress in Pali). It may even be that the augment was
grammaticized independently in Gr. and I-Ir: I don't know if the
patterns of Myc., Homer and Avestan have been fully explained.

In fact the zero past vs marked non-past of Hittite seems quite unusual.
(i.e., Hiittite has its weirdness :-) I am not well read on the role of
sentence particles in Hittite. Do they have any role in this? [I
remember some work arguing that -kan marked perfectivity.]

>Some features that are unique to the "Indo-Greek" verbal system are:
>- the perfect as a separate "aspect", besides present (impfv.)
>and aorist (pfv.).

This is precisely what I objected to in the first place. There is no
morphologically marked aspect in Vedic or any stage of Sanskrit. The
so-called imperfect is the usual tense of narration and does not
indicate non-attainment of result or non-total event etc.  If PIE had
no perfective vs imperfective contrast and I-Ir did not either, then
why posit it for Gr-I-Ir?

The `perfect' is just a resultative. The pluperfect and moods of the
perfect do not carry any particular `aspectual' meaning.

To put it bluntly: The usual morphological classification of Vedic verb
forms found in grammar books has no syntactic justification, but is due
to 19 c. prejudices. It is a serious methodological error to base
syntactic comparisons on the mere names.

> In Hittite, the perfect is still simply the past tense of the stative
> (hi) conjugation.

There are examples of resultative > `present perfect' > (perfective)
past. But ``past state'' > resultative? IMHO, it would be better to
assume that Hittite extended its use of -i for present into the
`stative' by analogy, while the rest of IE extended the stative into
resultative with further evolution into (a kind of) past in individual
languages at different times. [Looking at some old messages, I found
that I have asked this before and you agreed that PIE `perfect' was
tenseless. In that case, Hittie -hi is an innovation.]

>- the imperfect as a simple past tense of the present ((augment
>+) present stem + secondary endings).

What does `simple past tense of present' mean? If it means aspectually
unmarked past, how does that indicate closer relationship with Greek?
If it means ``present (imperfective) in the past'', the claim is wrong.
And Armanian aorist has `eber' which is usually traced to `ebheret'.
Slavic aorist and Baltic preterit also have forms which seem to be
from present stem + secondary endings. so such a form is not
just Gr-I-Ir.

>- the subjunctive (conjunctive) as a thematic (of athematic
>verbs) or doubly thematic (of thematic verbs) formation, without
>additional markers.  The only parallel is I think Latin ero:, the
>future tense of "to be".

I thought that there were other examples of Latin futures that can be
traced to the root subjunctive and that there are a few Celtic traces as
well. That leaves only the present subjunctive. But in Vedic, we see the
present subjunctive (and moods generally) replace the root forms. So
these may have started as analogical creations.

>- The augment for past tenses.  Also found in Armenian (3rd.p.sg.
>of monosylabic verbs only).

See above.

>sigmatic aorist and future forms,

Opinion seems to be mixed about the sigmatic future of Sanskrit: the -s-
may be that of the sigmatic aorist or of the desiderative. General
opinion seems to prefer the latter. The only reason to even suggest the
former is the limitation of the -sya future to proximate future. But
then that very fact begs for an explanation if try to connect it to
other IE futures. Furthermore, why was the -sya future rare in RV with
its place seemly often taken by subjunctives? And when it did become
common, why was its domain restricted to proximate future, with
periphrastic future used for predictions?

> Italic, Celtic and Albanian ... but their forms are best described
> as s-preterites.

So what is the difference between s-preterite and s-aorist? But
be sure to make the case for Vedic aspect before appealing to
perfectivity.

>And if the Armenian imperfect is indeed derived from the optative,
>that's a remarkable Armenian-Tocharian (and Italo-Celtic?) isogloss.

Past habitual is indicated in Avestan and Old Persian by the optative
(sometimes augmented). I believe that Greek optative, in relative
clauses can be ~ past habitual (cf English ``would'').

If you want to change the picture of the IE verb because of the
difficulties in explaining the syntactic evolution of various languages,
you must explain the syntactic evolution of the Vedic verb as well,
before connecting it ore closely to the Greek verb. What is sauce for
the goose ...

-Nath