From: Tavi
Message: 71024
Date: 2013-03-04
>Gamkrelidze-Ivanov and Nikolayev agree in deriving vitrum from *k´wei-t- 'light, white' (cfr. Lithuanian s^vìtra- 'sandpaper'), with *k´w- > Latin w- as in *k´wep- > vapor.
> > > Also some of De Vaan's etymologies, e.g. Latin vitrum
> > > 'glass; woad' from IE *wed-ro- 'water-like' are rather
> > > inventive, to say the least.
> >
> > It can't be all that inventive, since Sihler (223.5)
> > accepts it. He also says that 'for the semantics there
> > are a number of parallels'.
> >
> > > The problem is De Vaan's systematically tries to derive
> > > everything from the reconstructed "PIE" using "regular"
> > > sound correspondences, regardless of other considerations.
> >
> > Broadly speaking, that's a feature, not a bug. In
> > particular, when such a derivation is possible without
> > unreasonable contortions, it necessarily has primacy. This
> > isn't to say that it can't be displaced if a better
> > derivation is found, but the bar for any alternative is
> > pretty high.
>
> The real problem with de Vaan is his willingness to use slippery soundlaws and those (including some of Schrijver's) erected on a very slim etymological basis. The purported soundlaw *-dr- > Lat. -tr- has almost nothing but <taeter> against <taedet> behind it, and is contradicted by <quadri/u-> (which Sihler acknowledges but, true to character, does not explain).
>