It was very nice of Rebecca to take up the point of the dialects in my
former message, and Millie's criteria are very well taken. Spontaneously, I
can see a difficulty in criterion (2) and another one in (4) through (6).

Several of these terms for dialects have the format <geography><basic
language>. "SchwŁzerdŁtch" (or how you spell it, the German of Switzerland),
"Tornedalsfinska" (the Finnish of the TorneŚ valley area, the border between
Sweden and Finland) are just two examples where the geographical part is
united with the language name. Is criterion (2) meant to be applied also on
this kind of names or only on those where the geographical part is a
separate word?

(4)-(6) tend to receive the same grades in such cases that have been brought
forward in this Plattdeutsch discussion, namely those instances where an
ethnic group has migrated into a foreign area but brought with it its
original language, that afterwards has kept a number of the characteristics
of the original tongue but developped on its own or in contact with the
other language(s) in the new environment. It has thus as a rule developped
in another direction than the mainstream. This is the acknowledged theory
behind the formation of the different Indo-european languages from a
supposed common tongue, as well as of the other families of languages, but
on a smaller and more recent scale it also is why you find those varieties
of low German under various names in other countries. Since these groups
tend to preserve their ethnic origin, they often strive to keep the oral
tradition alive, giving them YES for most of Millie's criteria (but,
admittedly, maybe not for (1)!).

I wonder if you could have a criterion distinguishing dialects occurring
within the dominating area of the main language from dialects found in
isolated communities. Of course, you would have to define "isolated
community" ...but the outcome could be that a high percentage of the
isolated ones are recognized as worthy of individual codes (if they are
enough well documented), while dialects within a central, dominating area
are kept under the main code. It does not help us all the way out, but I
think it could give us a neutral and realistic additional criterion that
goes well with the genesis of languages in general.     

This does not cover the issue of Rebecca's second message that just arrived.
It seems that some extension mechanism may save the already existing data
while giving better possibilities in the future, but from where does the
"val" of the example stem? I'd have no difficulties in adding an ISO3166
geographical code to an ISO639 code, as RFC1766 does, but the twin language
codes used in the "sgn-" case are, to me, justified only by the fact that
"sgn" does not mean a language in glyphs or sounds. Making that method
general is still to me a major change that seems to be hard to get accepted
as an amendment.



Sten Hedberg
Expert on cataloguing and standardisation
Uppsala University Library
POBox 510, S-751 20 UPPSALA Sweden
Voice +46-(0)18 471 3970
Fax  +46-(0)18 471 3941