Re: floor

--- In cybalist@yahoogroups.com, "Tavi" <oalexandre@...> wrote:

>
> --- In cybalist@yahoogroups.com, "Torsten" <tgpedersen@> wrote:
> >
> > Alternatively, we could follow Kuhn and propose the existence of not
> one but two substrate layers for Germanic in the NWBlock area:
> >
> > 1 a non-IE language, called the ar-/ur-language by Kuhn
> >
> > 2 an IE language, spoken but for a short time before Germanic took
> over
> >
> > and assign the Germanic-like NWBlock roots to the latter, eg Meid's
> > German flur, English floor, NWB placename Plore, OI lar "field"
> >
> Matasović reconstructs Proto-Celtic *fla:ro- 'floor', and he quotes
> Old Irish lár 'ground, surface, middle', although the semantic shift
> to 'field' is straightforward (cfr. Basque larre 'meadow', probably a
> Celtic loanword).

Why would Celtic *(f)la:rom, if borrowed into Basque, be reflected as <larre> rather than *laro?

DGK