Re: Basque mendi 'mountain'

--- In cybalist@yahoogroups.com, "bmscotttg" <bm.brian@...> wrote:

>
> The Irish word means both 'hill' and 'open space'. It can mean
> 'racecourse', 'assembly', 'meeting place in general', 'square',
> 'market place'. If at some point the emphasis in Insular Celtic
> came to be on the use as a place of assembly rather then the
> elevation, a shift to 'valley' would not be all that remarkable.
>

I think 'open space' is secondary to 'valley', as in e.g. Romance *ru:ga > Portuguese rua, French rue 'street', Basque (Biscayan) arruga 'square, marketplace', cognate to Hispanic Latin arrugia 'mine channel' (Pliny) > Spanish arroyo 'stream'. However, we've also got Spanish arruga 'wrinkle, crease', possibly a semantic specialization.

Given that the substrate  root *pant-/*pent- is found in Hispanic and Italian toponymy meaning either 'mountain' or 'ravine' (e.g. South Italian pentuma), my guess is it was borrowed into P-Celtic *pant- and Q-Celtic (Goidelic) *kWant-. Of course, the time of borrowing must postdate Common Celtic, which it already had *bando-/*bendo- 'peak top'.

> > Incredible!
>
> Rather less so than some of your stretches.
>

You're exaggerating as usual.

> > Also interesting is Old Irish imm-cella 'surrounds', from
> > Goidelic *kWal-na 'to go around', a doublet of Celtic
> > *Fal-na 'to approach, to drive' (Old Irish ad-ella).
>
> No, *kWelh1- and *pelh2- are clearly different.
>
> > Only in the traditional model. But macro-comparativists such as Bomhard link them.
>
> I doubt very much that Bomhard links them at the PIE level.
>

It depends on what you consider as PIE, as I must insist not all the traditional "PIE roots" have the same different chronologies nor belong to the same paleo-dialect.

So in diachronical terms *kW- must predate *p-. The reduction of labiovelars must be a very ancient isogloss, as it already took place in Proto-Altaic and pre-Germanic (remember our former discussion about the 'bear' word).