1. On-line Bangani Resources (courtesy of Piotr):
http://www-personal.umich.edu/~pehook/bangani.html
http://www.sai.uni-heidelberg.de/MIND/zoller/Bangani.html
http://iias.leidenuniv.nl/host/himalaya/individ/enigma.html
2. Status of Centum Layer:
Richard: Why are the Centum elements of Bangani described as a
substrate? Would not a superstrate be more likely given the size and
nature of the vocabulary?
Piotr: No, we are talking about a handful of Centum words in what is
otherwise an inconspicuous Indo-Aryan language.
Richard: I saw the tally quoted as 59 words, which seemed a very good
haul, far more than English has from Brythonic unless you start
counting place names.
3. Ancient Macedonian as Bangani Centum Source:
Piotr objects to the idea of an Ancient Macedonian source because he
thinks Alexander's armies would have been more likely to scatter
Greek loans in their wake and adds that some of the Bangani words are
definitely non-Greek (e.g. kOtrO 'fight' and gOsti 'guest of honour').
I was thinking more of a bunch of Macedonian troops that were
abandoned in the raid on the Punjab. Perhaps that's too romantic an
idea!
--- In cybalist@..., Piotr Gasiorowski <piotr.gasiorowski@...>
wrote:
> <kOtrO> is supposedly related to Celtic *catu- and Germanic
*xaþu-
(OE heaþu) 'battle', with an *-r- extension visible in German
Hader 'quarrel', Balto-Slavic (OCS kotora 'fight', Lith.
kata~ryti 'beat') and Sanskrit (<s'atru-> 'enemy'). The Balto-Slavic
words may be Centum loans themselves, and the ultimate reconstruction
would then be *k^at-u- ~ *k^at-(o)r- (Pokorny's *k^at- 'fight,
struggle). This works for the Bangani substrate.
How well does Greek kótos (2nd declension) 'grudge, spite' fit in
with the /kOtrO/ words? I can't see how to reconcile the Celtic and
Greek vowels without upsetting the Sanskrit correspondence. (There's
too much wrong with *k^h3tw-.) The Greek sense seemed fitting for
the Bangani sense of infighting, though suffixing the 'r' (which I
had been taking to be the adjective suffix, as in the Latin, dialect-
affected, pair <rufus> and <ruber>, and common enough in Greek and
Sanskrit to need mention in the rules for forming the comparative)
seems to require a lot of derivation
An accepted difference between normal Greek and Ancient Macedonian is
that PIE *bH > Greek /pH/, but Ancient Macedonian /b/, and
http://indoeuro.bizland.com/tree/balk/macedonian.html shows a
Greek /tH/ corresponding to Ancient Macedonian /d/. I therefore
conclude that PIE *gH > Ancient Macedonian /g/, so PIE *ghostis >
Ancient Macedonian *gosti- > Bangani /gOsti/ looks very reasonable.
Is the problem with this idea that Greek at most shows this root in
xenos < *ghsenos?
Actually
http://indoeuro.bizland.com/tree/balk/macedonian.html
doesn't persuade me that Ancient Macedonian was an 'aberrant Greek
dialect'. (And the Slavophiles didn't persuade me that it was
Slavonic.) And if it isn't, suggesting Ancient Macedonian as the
source is, alas, ignotum per ignotum. (At least it's not per
ignotius.)
Richard.