Re: gl > gr attested in Romance?

Henrik Theiling skrev:
 > Hi!
 >
 > Is the sound shift /gl/ > /gr/ and/or /kl/ > /kr/ attested
 > in Romance?
 >
 > Or at least, would you think it's feasible like the /bl/ >
 > /br/ in pt. 'branco' < BLANCU(M)?

I don't know if it is attested, but it seems plausuble
enough. IIRC there are/were not many words with GL in Latin
to begin with.

Is, BTW _branco_ an isolated example?  I think not.

 > I am thinking about 'grorie' < GLORIA(M) in a new toy
 > romlang I started yesterday.

Why not? Although in this particular word I'd expect the
presence of an _r_ later in the word to inhibit the change,
or if the lang totally eschews _Cl_ clusters to yield
_grolie_, since remote dissimilation between liquids is
quite common in Latin and Romance. Meyer-Lübke's
Einführung lists a number of instances, but I don't have it
by me ATM. AFMOC Rhodrese regularly has D > r / V_V, -- e.g.
FIDES > _fair_ -- except when there is already an R in the
word, in which case D > l, e.g. PRUDENTEM > _prulent_.
Another exception is where D comes to stand before a
sonorant: CREDERE > _craidre_. It also draws some original
Rs with it, e.g. MERCURI DIE > _melcordi_. I still don't
know what to do with CRUDELEM, but since CRUDUS > _crul_
probably is in the language _crulair_ is thinkable.

 > Also, what about /Nn/ > /Ngr/ like /mn/ > /mbr/ (sp.
 > 'nombre' < NOM(I)NE(M))?
 >
 > Tere, I'm thinking of 'ringre' < RE:GNU(M).

The only snag is that [Nn] for _gn_ is AFAIK unattested in
Romance, which IIANM has /J/ or /nn/ > /n/, in all the
relevand words, e.g. Old French _renne_. But wait, Rumanian
has LIGNU > _lemn_, and I guess all of /J/, /n(n)/, /mn/ may
be derivable from [Nn] if one wants to.

FWIW Rhodrese also has Cn > Cr -- which BTW is attested in
the form _Rhodray_ itself, which also shows that D > r is
posterior to syncope!