In general,
in the northern part of France,
from the Loire River Northward, (to the
Rhine)
that is to say, the place where P-celt Gauls are
massively attested
(Let us put aside this "Belgian" affair for a
while)
you often find the name of tribes under three forms
:
Nominative plural : tribe's name
Accusative singular neuter : province's
name
Ablative-locative plural : capital's
name
Example : Pictav-i (tribe)
Province : pictav-u > Poitou (territory
possessed by the Pictavi)
Capital : pictav-is > Poitiers (at the Pictavi's
city/market)
Most modern province capitals end with -s (coming
down from ablative-locative plural)
while the province often ends with a vowel. (from
the accusative neuter)
Bourges versus Berry, Angers versus Anjou
(etc)
It is not possible to make out if this traces back
to Gaulish or LAtin
because case markers are about the same in the two
languages.
In many cases, the Roman name has been replaced by
the tribe's name.
This is why I would consider probable that this
traces back to Gaulish :
it tends to erase Latin influence and show the
fairly strong survival potential of GAulish
under Roman occupation.
In some cases, in compound names, you found
genitive plural :
in that case, the tribe's name and the defined name
are separated by -or- or -eur- (from -orum)
this is a rare case.
Some examples have been created when Germanic
tribes began invaded the country
so this proves a full declension system with many
cases must have survived after the Roman Empire collapsed.
I don't know who first noticed this tribe +
province + capital structure
It is quite commonplace knowledge.
Reims with -s also fits in the mould.
And Paris as well.
Amien-S < Ambian-IS
Beauvai-S < Bellovac-IS
Soisson-S < Suesson-IS
Arra-S < Atrebat-IS
LAngre-S < Lingon-IS
As we say in French, I don't want to "make the
knife spin inside the wound"
Dear Torsten,
this is one more heavy blow on this "non Celtic
Belgian" hypothesis.