Re: Identity of the 'language of geminates'

From: tgpedersen
Message: 60992
Date: 2008-10-18

--- In cybalist@yahoogroups.com, "Brian M. Scott" <BMScott@...> wrote:
>
> At 3:45:12 AM on Friday, October 17, 2008, tgpedersen wrote:
>
> > --- In cybalist@yahoogroups.com, "Brian M. Scott"
> > <BMScott@> wrote:
>
> >> At 2:18:05 PM on Thursday, October 16, 2008, tgpedersen
> >> wrote:
>
> >>> --- In cybalist@yahoogroups.com, "indravayu" <sonno3@>
> >>> wrote:
>
> >>>>> Welsh perth is ultimately, it seems, from Lat.
> >>>>> pert(ic)a, the Roman surveying instrument par
> >>>>> excellence, recall the pertica militaris 'sectioned
> >>>>> land allotments as payment deeded (Celtic and
> >>>>> Germanic) mercenaries'.
>
> >>>> Perth (aalong with Gaulish Perta) is generally believed
> >>>> to be a derivative of Common Celtic *kWerkWo- "oak"
> >>>> (from PIE *perkWo-)
>
> >>> That leaves Jysk perte "beat", pirke:firke "squeeze
> >>> out", Da. pirke, dial. perke "to prod" (with un-Germanic
> >>> p-) unexplained.
>
> >> There's no obvious reason to connect them with <perth>
> >> 'wood, copse' in the first place.
>
> > I didn't get that; so there's no connection Lat.
> > pert(ic)a > Welsh perth?
>
> If MWel. <perth> 'bush, hedge' is from *perkWu-, then there
> may indeed be no connection:

And if it isn't there may, eg. if Vendryes is right below.

> I don't think that the
> etymology of <pertica> is known (though it apparently does
> have cognates in Umbrian <percam> 'rod' and Oscan
> <perekais>).

Yes, it certainly looks like a substrate loan in Italic.
Ernout-Meillet:
'pertica, -ae f. : perche; spécialement "perche à prendre des mesures"
p. mi:lita:ris, p. decempeda, qui servait à mesurer les lots concédés
aux soldats; puis ce "lot" lui-même (cf. fr."perche"): quodcumque
coloniae est assignatum, id uniuersum pertica appellatur, Front.,
Limit. Agr., p. 43 Goes.
Ancien (Pl.), usuel; sert de cognomen. Panroman, sauf roumain.
M.L.6433. Irl. pertic.
Dérivés: pertica:lis, -rius, -tus.
Cf. ombr. percam "uirgam" (?), osq. perek, "pertici:s" (mesure).
Pas de rapprochement sûr hors de l'italique et du celtique, où M.
Vendryes rapproche m.gall. erchyll "mât de navire", Rev.celt.48.
'

> For that matter, there's no necessary connection between the
> Danish words and <pertica>.

Nothing is 'necessary' in linguistics, as you should know. Things are
possible, and some things are more likely than others. The presence of
perte "beat" in Jysk (where because of p- it must be substrate, it
can't be Germanic) and kwarthæ "brim, edge" in ODa. is suspicious. You
could rule them out on geographical grounds, if you knew that there is
no chance they could have influenced the Brythonic word, but that
would be circular since that contradicts the hypothesis we are testing
and so is impermissible unless grounded in independent evidence.


Torsten