Trask: History of Basque:
ondo "bottom, side; under" Basque
(h)ondar "remains; beach, sand" Basque
ondoren (ondo + gen.?)
"after" Basque
(ibaiondo "riverbank" Basque)
Warren Cowgill, Italic and Celtic Superlatives
and the Dialects of Indo-European:
húntrú "lower" Oscan
hondomu "from the lowest" Umbrian
hondra "infra" Umbrian
M. Löpelmann, Etymologisches Wörterbuch der baskischen Sprache:
mase "sieve" Basque
maso "club" Basque
maso "soft, tender, fresh" Basque
The American Heritage Dictionary of Indo-European Roots:
*bhes- "rub" Proto-IndoEuropean
1. Zero grade with unclear suffix
*(bh)s-amadho "sand" Proto-IndoEuropean
psamathos id. Greek
*sam(a)dam),
*sandam id. Proto-Germanic
sand English
2. Suffixed form
*(bh)s-abh- Proto-IndoEuropean
further suffixed
sabulum "sand" Latin
3. Suffixed form
*bhs-a:- Proto-IndoEuropean
pse:n "rub, scrape" Greek
pse:phos "ballot, pebble" Greek
TP:
Note the 'unclear suffix' in 1. The suffix in 2. looks
like a postposition too.
TP:
*bhs-e:na "sand" PIE? Substrate?
A. Sihler, New Comparative Grammar of Greek and Latin:
(h)are:na "sand" Latin
fase:na id. Sabine (Varro)
Pokorny has *bhes- in two versions,
"abreiben, zerreiben, ausstreuen" and
"hauchen, blasen"
In the first, he has
Albanian fs^in´, ps^in´, mes^in´ "kehre aus, dresche"
Doublets in p-, b-, m- is what one would expect in Basque, not
Albanian.
"sieve-under", if that's what the sand word means, is not exactly
IE. It makes more sense in Basque.
Pokorny does not include the last Sabine example. I think it's
interesting; it shows a loanword in *bhs-, *ps- appearing in Latin
as *pas-, as I guessed with Greek pse:r, Latin passer
http://groups.yahoo.com/group/cybalist/message/37414
Jens has proposed Latin /a/ as reflex of zero grade, which will
match here.
Given Gothic sparwa etc, with supposed s-mobile, should we consider
instead that sp- is a reflex of underlying ps-? Beekes notes in his
article on Pre-Greek
http://www.indo-european.nl/ied/pdf/pre-greek.pdf
that they have alternates eg ps-/pt-, ps-/sp-, p-/sp-
Now we know from Greek that -pt- weas once -py-. Linear B has signs
with initial -pt- which are best interpreted as -py-, which is good
also for syllabary-internal reasons, the language gets palatalised
consonants. So *pyar- (> *psar?) "sparrow"?
And BTW, do the abundant s-mobile (they can't all be causative
prefixes, as in 'melt' vs. 'smelt') in Germanic etc come from non-
palatalised/palatalised doublet roots *p-/*py-, *t-/*ty- etc, cf the
diminutive palatalised doublets in Basque and the 'pejorative j/ of
Continental Scandinavian?
Note that Beekes pre-Greek has alternation w/m, an alternation that
is fundamental in IE, but limited to a few roots and suffixes. That
would point to loaned elements in the fundamental machinery of PIE,
which means the corresponding construnctions in PIE (eg. 1pl in
verbs) must have been constructed a novo, when the loans appeared.
Torsten