Principles of Operation (was LANGUAGE SUBTAG REQUEST FORM Erzgebirgisch)

Doug Ewell dewell at
Sun Jan 27 08:03:12 CET 2008

CE Whitehead <cewcathar at hotmail dot com> wrote:

> I agree that de is best probably for the purposes at hand--but it 
> might be possible to have two prefixes, right?

NO.  This is absolutely not the sort of situation for which multiple 
prefixes are appropriate.

Variant subtags can have two or more prefixes when the variation in 
question applies to two or more languages or language varieties.  We 
already have examples of this in the Registry:

*  The variant subtag '1994' has five prefixes, corresponding to the
   Slovenian dialect Resian and four of its sub-dialects, showing that
   the orthography represented by '1994' can apply to any of these
   language varieties and highlighting the correct syntactic usage
   ("sl-biske-1994", not "sl-1994-biske").

*  The variant subtag 'baku1926' has 10 prefixes, corresponding to 10
   different languages, all of which were once written in the
   orthography represented by 'baku1926'.

This can extend to other types of variations besides orthographies. 
Although this example is meant for illustration only and IS NOT A 
SERIOUS PROPOSAL, it would be possible to imagine a variant subtag 
'cheech' to indicate the characteristic "cholo" spoken accent used by 
comic actor Cheech Marin, and furthermore it would be possible to 
associate this accent with both English and Spanish usage.  In this 
case, the subtag 'cheech' might have two prefixes, 'en' and 'es'.

The extreme case of this is a variant subtag with no prefixes at all, 
meaning that the variant is appropriate for use with (almost) any 
prefix.  The subtags 'fonipa' and 'fonupa' have no prefix, because 
virtually any spoken language can theoretically be written in IPA or 
UPA.  (Obviously there are reasonable, common-sense limits to "any 
prefix."  It would not, for instance, make any sense to write 
"de-1996-fonipa" or "sr-Cyrl-fonipa" because other subtags conflict with 
the notion of "written in IPA.")

What is NOT appropriate is to think of multiple prefixes as a vehicle 
for creating aliases for the same language variation.  Registering 
'erzgeb' with the two prefixes 'de' and 'sxu' would create an impossible 
situation: no matching algorithm would ever compare "de-erzgeb" and 
"sxu-erzgeb" as equal, any more than it would compare "az-baku1926" and 
"tt-baku1926" as equal.  The only way this tagging scenario would make 
any sense would be if "German Erzgebirgisch" and "Upper Saxon 
Erzgebirgisch" were somehow determined to be different linguistic 
varieties, which I don't think anyone has proposed.

No, what we need to do is figure out which ONE prefix is the best to use 
for the one dialect "Erzgebirgisch," and register that one.

Doug Ewell  *  Fullerton, California, USA  *  RFC 4645  *  UTN #14  ˆ

More information about the Ietf-languages mailing list