--- In
cybalist@yahoogroups.com, "Marco Moretti"
<marcomoretti69@...> wrote:
> However, it is wiser to consider Sumerian /girgir/ as derived form
an
> expressive form. Words for "to roll" are often of expressive
nature
> and of phonetic structure /*g/k(w)Vr/l/ in so many language that
very
> little can be inferred.
> I see little phonological and comparative substance here (using
Larry
> Trask's idiom).
There's a difference between finding words with a meaning similar
to 'to roll' of this form and actually having a word of this from as
_the_ word 'to roll'. Going through the English to Foreign
dictionaries at my disposal, I find, looking up 'to roll':
English _roll_, French _roulir_, Italian _rullare_
Latin _volvere_
Russian _vertet's'a_
Spanish _rodar_
Hebrew _gilge:l_ (which is what prompts your observation!)
Thai _muan_ (falling tone)
Only one of the six cognate sets falls in your category, so there
probably is something to explain. The only thing that weakens the
case for comparing non-Anatolian PIE *kWekWlos. *kWekWlos with
Semitic and Sumerian is the existence of another well-established
non-Anatolian PIE word set for 'wheel', namely *roth1ah2, *roth1os,
which are unexceptional deverbal nouns from *reth1 'roll'. Digging
in the archives, e.g. at
http://groups.yahoo.com/group/cybalist/message/9075 , Piotr makes
the point that there are 'branch-specific variations' in the
representative of *kWekWlos. He didn't go into details. Indo-
Iranian, Greek and Germanic all seem to me to reflect *kWekWlos,
while Baltic and Tocharian seem to reflect *kWokWlos (although I may
be missing something in the Latvian and Tocharian forms). Phrygian
_kikle:n_ seems to be the only really odd form.
Richard.