Indo-Aryan migration vs Indigenous origin - scholarly debate

[Dominique.Thillaud]

Miguel Carrasquer Vidal answer me:

>
>>"the "-au" ending comes about only in dvivacanam" ?
>>No: nominative aSTau (RV, &c.) being a counter-example.
>>
>>        There was tentatives to explain aSTau by a dual, but without any
>>good result. The simpler was to consider *qet(w)r- (4), to cut the last "r"
>>(without understand it!), to suppose a zero degree with "qt" > "kt" and to
>>prefix by a dubious particle "O" (near) suggested by few Greek words and
>>Skr. A-gam- (shortened before two consonants?); finally: "two 4s brought
>>nearer"!!
>
>Leaving lurk-mode...  What's wrong with the much simpler one: Avestan
><aSti-> "width of four fingers" (IE *ok^t-is, *ok^t-os "4" => du.
>*ok^toH "8")?
>
>>        In fact, it seems we have a "vanishing" "w" at the end of the
>>Eurindian root, suggested by the ordinals Lat. octAuos, Gr. ogdo(w)os and
>>by Got. ahtau.
>>        The Indian variants aSTa, aSTA, aSTau find a parallel in Greek
>>where the two forms oktO- and okta- coexist early (Homer) in compounds;
>>okta- was explained by the analogy with hepta- and hexa- but nothing is
>>sure with this mysterious number and the "a" could be genuine (see the Osq.
>>form Uhtavis of the Lat. name Octavius).
>
>If the IE laryngeals h1, h2, h3 were analogous to the velar series k^,
>k, kw (i.e. they were phonetically palatal /x^/, plain /x/ and
>labialized /xw/), then reconstructing a dual form *ok^toh3 (/ok^toxw/)
>would explain all these forms (the Sanskrit dual with -u, Gothic
>ahtau, Greek *ogdowos "8th", and Latin Octa:vius < *oktoh2wi- <
>*ok^toh3i- [=*ok^toxwi-]).

        Undoubtly a simple and well-structured explanation. Alas, some
problems remain.
        Even writed *h3ekt-, *ok^t-os "4" don't fit well with *kwet- "4"
(the Greek tessares don't shows any prothesis in any dialect). The isolated
Avestan aSti- don't seems sufficient to postulate such a form.
        The analogy between h1, h2, h3 and k^, k, kw is oversimplifying
(and the last serie is dubious). It's true that a labial feature is
probable in h3 and used for a reconstruction *h3ekteh3 "8" (a voiced form
of h3 being useful to explain the Greek ordinal ogdowos) but that's not
sure that h3 is necessary to explain the direct animated thematic dual case
which could be -oh1(w) (see the common vedic dual forms in -A). I'm afraid
that all such choice are uniquely made to link aSti and aSTau and we can't
exclude the fact that, even if they are linked, aSti could be derived from
an aSTau falsely interpreted (such inverse derivations are well known in
linguistic).
        Not lurking but too few facts to decide ;^(
        Regards,
Dominique

Dominique THILLAUD
Universite' de Nice Sophia-Antipolis, France