Yet another areal feature in SA languages?

[Vidhyanath Rao]

A coincidence, reading Bybee et al ``The evolution of grammar'' and
Dahl, ``Tesne and aspect systems'' while trying to understand Vedic
verbal syntax, led me to notice that a certain somewhat unfamiliar
verbal category seems to be widespread in South Asia and that this may
be relevant to the difference in Vedic between the so-called imperfect
and the (sigmatic) aorist.

This category is that ``completive'' (terminology of Bybee et al).
According to Bybee et al, the common characteristics of the completive
are that it emphasizes the object(s) were totally affected or consumed,
or that the subjects or the objects are universally quantified; it has
an emphatic/emotive/surprise value or is used for `hot news' items.
However the first two are said to be atypical for forms which seem to be
connected to auxillaries meaning `finish'. What is interesting is that
several languages with such a construction have a distinct past used for
narration or a separate perfective. The completive is apparently not the
usual form used in narration.

The languages in the sample used by Bybee et al (a stratified sample
based on language grouping, with concessions to availability of
reference grammars) with such a form are fairly widespread (but
seemingly no European languages). But South Asia and nearby areas seem
to be particularly well represented. Bybee et al report such forms from
Chepang (Sino-Tibetean, Nepal, formed with je?), Kui (Dravidian, formed
with de), Maithili (IA, formed with cuk/ja/ga) and Car (Nicobarese).
This category seems to be common in Austroasiatic and Sino-Tibetean,
being reported in Koho, Palaung, Loatian, Haka, Lahu and Cantonese, in
addition to Chepang and Car mentioned above.

Dahl gives the Japanese ``-te shimaru'' construction and the (modern)
Tamil ``-viDu-'' construction as the examples of his ``conclusive''. The
Japanese form seems to have the same properties as the completive of
Bybee et al. The Tamil form seems to be a bit different but closer to
``completives'' than anything else. [The uses are discussed in detail in
a paper by E. Annamalai in ``International J. Dravidian Ling.'' vol 11
pp.103--126. He also attempts to categorize this in terms of
aspectology, using Comrie ``Aspect'' as a guide, but admittedly
unsatisfactorily.]

Here are my questions: Is such a category really widespread among South
Asian languages? What about historically older stages? In particular,
Sangam Tamil does not seem to use viDu as much as modern Tamil. But
there is ``iDu'' which is (somewhat) common and seems to be similar;
also the colloquial form ``vandiTTaa~'' looks like it can be from
``vandu+iTTaan''. Is iDu used an `auxiliary' in Sangam Tamil similar
to modern Tamil viDu? Pali does not seem to have anything like this. On
the other hand there seem to be some striking similarities with Vedic
(sigmatic) aorist. How about Prakrits? Do they have any periphrastic
constructions not discussed in the usual grammars that function like
``completives''?

In a separate post, I will describe the Tamil -viDu construction to make
this more understandable without reading the references mentioned. But I
wanted to get the questions into a post that is not too long.

-Nath