Momentary-Durative

[Jens Elmegaard Rasmussen]

On Mon, 21 Jun 1999, petegray wrote:

> I (Peter) wrote >

>>> ... at first a tenseless verb was used with endings -m, -s, -t
>>> and so on.

> Jens said:

>> ... what does "at first" mean here?

> I meant only "prior to the time when tense was marked morphologically."   If
> you wish to take issue with me, you must also take issue with Friedrich
> Mueller (1857), Thurneysen (1885), the detailed proof by Kiparsky in
> Watkins' *verb* 45;  Bruggman, Kieckers, Burrow, Martinet, Kurylowicz,
> Erhart, Wright, Brandenstein, Szemerenyi, Beekes, Baldi, and others.
> [...]  There are other voices, speaking against this view, but they are
> few: Herbig (1896), Hattori (1970), Manczak (1980).    Your posting also
> recognised the existence of these tenseless forms in -m -s and so on.

> Jens said:

>>  If the tense markers such as the primary *-i
>> and the augment *e- were once independent words, how can we know they are
>> younger than the person markers?

> The formation is younger, not necessarily the elements.

Bingo, I guess we agree - and so do most of the above-mentioned
authorities. Watkins of course has even pointed out _what_ the augment was
before it came to mark past tense in an obligatory way in some (most?
all non-Anatolian?) of the daughter languages, equating the *e- as he did
with the Luvian sentence opener a-. Since the Hittite opener is nu- and
Old Irish has an empty preverb no- (nu-) in the imperfect doing much the
same job as the augment, there is good reason in calculating with two
temporal particles meaning 'now' and 'then'.
   But I do not think this calls for a time when the tenseless (and
moodless) injunctive was _alone_ - although it does not exclude that
either. But at any rate we seem to agree that the primary forms and the
augmented preterits must have existed as two-word juxtapositions before
they were univerbated. It looks much like the story of the determinate
adjective of Baltic and Slavic which is quite parallel and would be
accepted as inherited, were it not for very the different sandhi rules
that reveal that the univerbation had not yet occurred in Proto-Balto-
Slavic; still, the collocation is inherited as a fixed series of two
separate words.

> (I said):

>>> thematic -
>>> or especially sigmatic - aorists may be later than athematic and asigmatic.

> Jens replied:

>> Thematic yes, ... But
>> there is no indication that the sigmatic aorist is an innovation;

> Perhaps I misunderstand you.   It seems to me that a formation <root> +
> <suffix> + <ending> is necessarily derivative, and that the primary form is
> <root> + < ending>.

But is that not like saying that a given noun is older than the plural of
the same noun? Does that make much sense? Can't a speaker of a language
make derivatives from stems as soon as they arise? (I grant of course that
he could not do it before.)

> Jens went on:

>> It must
>> once have had a special function, and I suppose it was inchoative (the
>> aorist corresponding to the sk^-presents).

> The -sk^- presents are not normally incohative in IE, except in Latin.   We
> should not read back into PIE the situation we find in the languages with
> which we are most familiar.   Hittite uses -sk- for an iterative/durative;
> Tocharian for a causative.   Possibly the iterative /durative is more
> original, as there are traces of it in Homer as well as Tocharian.

I guess the widespread iterative value of sk-formations has started with
those that were reduplicated. The Lithuanian st-presents are also
inchoative.

> Likewise your connection of -s- aorists with -sk^- presents is not regular
> anywhere.  Many (if not most?) in Latin have -v- perfects, suggesting the
> root was a vowel or a laryngeal (e.g. creH-sco, gnoH-sco).  Some have
> reduplicated forms (disco didici).  LIkewise, aorists in other IE langs do
> not show the connection you suggest.

No, not regularly, I know, but often enough and in archaic-looking
examples enough to make it, in my estimate, an archaism you cannot
disregard.

[...]

Cheers,
Jens