--- In
cybalist@yahoogroups.com, "raucousd" <raucousd@...> wrote:
> I find it odd that anyone would say that the etymology of atar is
> unknown, and it's also curious that the Skt. words atharvan, more
> or less `fire priest' and Atharvaveda, (one of the vedas) are also
> said to have no etymology. I don't have Lubotsky, which you refer
> to often, but I'll try to find it.
Lubotsky's "Database query to Indo-Aryan inherited lexicon", based on
M. Mayrhofer's _Etymologisches Woerterbuch des Altindoarischen_
(Heidelberg 1986-1996), is online. The lexical entries given by
Mayrgofer have been in many cases revised by Lubotsky. The URL of
this resource's search page is:
http://tinyurl.com/644g4j
As to Old Indo-Aryan atharvan '(fire-)priest', it is most likely
unrelated to Old Iranian a:tar 'deified fire'. I see you here follow
Gemkrelidze and Ivanov, yet I must say that their hypothesis is
contradicted by more thorough comparative research on the history of
this word (Skt. atharvan). On p. 12ff. of his paper available online
at
http://www.humnet.ucla.edu/pies/pdfs/IESV/1/VVI_Horse.pdf
Ivanov claims that OIA atharvan is a possible borrowing from OIr.
a:thrauuan 'priest' (attested only in Young Avestan), which in his
view would derive from a:tar 'deified fire'; he further speculates
that Hurrian tari 'fire' may be a borrowing from the same root due to
the influence of Proto-Iranian fire cult on ancient Hurrian culture.
Of course, Ivanov must defend his (and Gamkrelidze's) hypothesis of a
PIE homeland in Armenia, a corollary of which is that Proto-Iranian
was spoken southeast of the Caucasus at a very early date; yet, some
combined linguistic evidence seems to point in another direction.
The form a:thra- instead of the expected athar- in the Y.Av. term may
be due to influence of Avestan a:tar (= a:thr) 'fire', yet the two
words don't seem to be cognate. On the contrary, OIA atharvan and
OIr. a:thrauuan seem to be true cognates, but the underlying Indo-
Iranian noun root, *athar-, has no IE cognates and seems to be a loan
word from an Indo-Iranian substrate that A. Lubotsky and G.-J.
Pinault identify with the language(s) of the Bactria-Margiana
Archeological Complex (BMAC).
The original meaning of *athar- appears to have been something
like 'magico-religious fluid' (thus Oldenberg), and the stem *-rwa-
or *-rb(H)a- is common to other alleged BMAC words borrowed into IIr.
as substrate words; thus:
BMAC substrate *atHarwan 'priest' > OIA atharvan, OIr. a:thrauuan
BMAC substrate *c'arwa 'name of a deity' > OIA s'arva 'name of a
god', OIr. sauruua 'name of a daeva'
BMAC substrate *g(H)andH(a)rw/b(H)a 'a mythical being' > OIA
gandharva, OIr. gan.d@...@ba
G.-J. Pinault, ("Une nouvelle connexion entre le substrat indo-
iranien et le tokharien commun", _Historische Sprachforschung_ 116
(2003), 175-189) has connected this hypothesized BMAC substrate word
*atHar- with Proto-Tocharian *etre 'hero' (Toch.A atär, B etre). This
lexical comparison is part of a wider set of triangulations among
OIA, OIr. and Toch. attested lexemes that allow Lubotsky, Pinault and
Witzel to argue for the existence of the said Central Asian
substratum of the BMAC as a source for many words in those three
ancient languages that appear related, but that have no convincing IE
etynology.
Regards,
Francesco