--- In phoNet@y..., "Piotr Gasiorowski" <gpiotr@i...> wrote:
Some Munda languages have borrowed long series of IA numerals, e.g.
Korwa has numerals of IA origin for the numbers 4-10 (includig
<nau> '9'), and I'm fairly certain that Bihor <la:>, which is a
member of a similar series, is indeed *nava in disguise. Experts in
Munda etymology should be asked first. Maybe they know where <lo>
comes from. Anyway, it is not an ancient pan-Indian numeral but a
recent borrowing from Indo-Aryan into Santali at best. They are not
_genetically_ related.

Why should borrowing from PIE into Munda be ruled out? [Assume for
the sake of argument that Munda speakers who were metal workers, were
present ca. 3300 BCE in Mehrgar.h].

A variant of numeral nine in Austro-asiatic is 'tom' (Kharia); this
is also found in Telugu, tommidi 'nine'. In many Dravidian languages,
there is no genetic word for 'nine'; it is explained as one less than
ten, e.g. Tamil, on-patu (pattu = ten); on is a reflex of u_na,
deficient (Skt.)

Here are some comments I have received about the l-n interchanges and
about nine in some Austro-asiatic languages.

"Pashto-English Dictionary" co-authored by M.H. Rahimi and M.S. Rohi,
Kabul, 1979 (and later reprints). Rahimi's entries include: _nmar_
and _lmar_ "sun" _lambar_ and _nambar_ "number" _luund_ "wet, moist"
which pairs with _nuund_ "damp" in Bellew's Pashto dictionary,
_lma'say_ and _nma'say_ (where _a'_ is a schwa) "grandson"
_lamCay_ "felt (rug)"; _namla_ "felt pad" (related to hindustani
_namda_ _nmuunJ_ "see _lmwwnJ_" for which the entry is missing; it
means "worship" and is related to Persian "namaaz" (I've used _J_ for
the _dz_ phoneme). _lg^a'Sta'l_ (where _g^_ is "ghain", the Iranian
phoneme sometimes written _gh_ and S is a retroflex S) "to roll" and
_ng^aSta'l_ [Wilma Heston, UPenn]

Regarding the Santali word /lo/ for "nine", Bodding says that it is a
Desi form, connected to Bengali /noe/ and thus of IE origin. I do not
have access here to the materials needed to verify this etymology,
but perhaps this is easy to do in Cennai. In Austro-Asiatic, the
Munda etymon for "nine" is South Munda: /tim/, and Kharia /tom/ with
good cognates on the Mon-Khmer side. That does not bring much
comfort, I am afraid. The Santali word /loa/ for a Ficus sp. is very
ancient in AA, with cognates as far as Khmer /lvia/; the lo- part of
the Santali word represents an older *lw- consonant cluster, and the -
o- part of /loa/ is not historically a true vowel but a semi-
consonant *-w-. [Gerard Diffloth, EFEO, Siem Reap, Cambodia]