path

From: dgkilday57
Message: 62323
Date: 2008-12-30

English <path>, with its Germanic cognates, was explained by Kluge as
a post-Grimm-shift borrowing from Scythian *paþ- 'way, path, road'
(with attested Iranian cognates), itself from the zero-grade of PIE
*pent- 'to tread, go'. One of the problems with this explanation is
the distribution of the Gmc. word, which is restricted to the western
part of the West Gmc. area. A late loanword from the East should not
have this distribution. T. (Geiger) Bynon observed that WGmc *paþ-
is found largely along the Rhine in the Germanic-Celtic contact area,
and so a Gaulish source of the word is plausible (TrPhS 64:67-87,
1966). Her choice for the Gaulish word is *pant-, cognate with Welsh
<pant> 'valley, dent'. This involves phonological difficulties, as
she admits. One expects *pant- to be Grimm-shifted into *fanþ-, and
loss of the nasal in western WGmc would require compensation,
yielding *fa:þ-. For the vocalism, she cites some onomastic
doublets, Gaulish <Carato-> beside <Caranto->, <Namatio-> beside
<Namanto->, etc., and suggests that Gaul. *pant- might have had a
doublet *pat-. (It is not clear to me how we know that the onomastic
doublets involve neither local orthographic custom nor compensative
lengthening within Gaulish dialects.) Getting her *pa(n)t- partially
Grimm-shifted into *pa(n)þ- requires Bynon to construct an elaborate
theory of loanword phonology between Celtic and Proto-Germanic which
dismisses other evidence about loanwords. In short, while Bynon's
suggestion that WGmc *paþ- is borrowed from Gaulish is highly
plausible, she tries much too hard to shoehorn *pant- into the role
of source.

To me it makes more sense to assume a Gaulish *bat- as the source.
Pokorny assigns some Insular Celtic words pertaining to death, Old
Irish <baîd> 'dies', etc., to PIE *gwa:- (i.e. *gweH2-) 'to go, come'
on the grounds that dying is a going forward from the realm of
mortals. In English, <pass> is used in a similar sense. This
Insular specialization of the word was not necessarily shared with
Gaulish. The Greek adjective <batós> '(easily) passed, passable', if
it comes from *gwm.to- like the Latin participle <ventum>, would have
*banto- as the expected Gaulish cognate. However, a parallel
adjective *gwH2to- from *gweH2- not *gwem- would yield Gaul. *bato-.
I propose that this form in the sense 'passable' was used in Gaulish,
typically as a substantive with a noun 'way, road' understood, and
borrowed as a noun by pre-Grimm-shift Germanic-speakers along the
lower Rhine, where it regularly became WGmc *paþa-, and remained
restricted to regional usage.

DGK