From: Tavi
Message: 69969
Date: 2012-08-12
>IMHO the French word would rather reflect Celtic (Gaulish) *kladiwo- (not **kladio-). However, the Italoid (aka Sorothaptic) protoform *k(a)lawijo- proposed by Coromines could explain the Celtic word through metathesis and reinforcement of yod (like in East Germanic): *klawijo- > *klajiwo- > *kladiwo-.
> I've been examining in detail what I told you Coromines had to say about
> gladius, and I now see I was completely misled. When he says the word
> may come directly from a Sorothaptic *klauiios he's not speaking of
> Celtic kladios (> Lat. gladius > Cat. gladi) but of a postulated origin
> *glavius for French glaive.
>
> That leaves out the question you're interested in: when does exactly the k- in CelticA direct Celtic borrowing (from e.g. Gaulish) can be excluded because we'd expect **cladivus instead of gladius. Thus we need some intermediate language where *kl- > gl- and *-w- > 0. From DGK's research, this would be Sabine.
> kladios becomes the g- in the Latin Celticism gladius.
>
> One riddle wrapped in another, as Churchill would say.Surely he knew about Matryoshka dolls. :-)
>