Greetings!
> Sorry! The link should be:
> http://www.snl.ch/dhs/externe/protect/sprached.html
> Peter S.
Besides the interesting topical discussion on language boundaries, the key
issue that we are dealing here with, is the one of SCALE. A key concept
that Geographers are grappling with for a very, very long time.
Certain processes are visible at one scale, but not at others. This issue
is also known as the modifyable areal unit problem (MAUP) or in the social
sciences, as ecological fallacy. One cannot assume that if a
(statistical) relationship exists at a finer spatial level of detail (e.g.
local level, such as communes in Switzerland), the same relationship will
be existent at a more generalized or coarser spatial level (e.g. Canton
level in Switzerland). MAUP goes both ways, scaling up or down.
The web pointers have one crucial item in common: they are at a
coarse spatial level of detail, or how cartographers would say, the are
small scale. If one selects a map at the communal level, thus higher level
of detail (e.g. commune boundaries are shown), the 'odd german speaking
spots' in the French speaking part of the Canton of Berne become clearly
visible:
-Mont-Tramelan (pop 125) german speakers 107 = 85.60%
-Chatelat (pop 128), german speakers 72 = 56.25%
-Rebevelier (pop 43), german speakers 25 = 58.14%
They all share a border with the French speaking Canton of Jura, by the
way.
(source Atlas of Switzerland, interactive CD ROM, 2000, data is from the
1990 Swiss census).
The Mennonite connection is interesting, as the above communes seem to
have a rather high proportion of catholics, relatively higher than most
neighbouring communes (81, 89, and 92 percent, same source as above).
Just a couple of thoughts that might be of interest,
sara
__________________________ dr. s a r a i r i n a f a b r i k a n t
u c santa barbara g e o g r a p h y
5805 ellison - santa barbara, ca 93106-4060
tel. +1 805 893-5305 | fax. +1 805 893-3146
sara@... | www.geog.ucsb.edu/~sara