proposed ISO 639 change for "arn"

Doug Ewell doug at
Wed Jan 9 21:17:58 CET 2013

I've written to ISO 639-3/RA to ask, IF any change is to be made for
Mapudungun to get rid of the "offensive" code element 'arn', that they
simply retire the code element and replace it with another one.

There is no problem for BCP 47 if the RA does this. Every year, ISO
639-3 retires many code elements—sometimes dozens—and replaces them
with others. Often this is because of a code element split, where the
exact denotation has changed, but that doesn't factor into the BCP 47
response. The BCP 47 response is simple: add the new subtag(s) and
deprecate the old subtag, with an appropriate Preferred-Value or
Comments field. Problem solved, for BCP 47 anyway. (Windows and other
systems that already have 'arn' embedded probably will and should go on
using it, just as Java still uses 'iw' for Hebrew instead of 'he'.)

Proposed "solutions" such as macrolanguages, collection code elements,
merging of other languages into Mapudungun under a new individual code
element, or other "special cases" should NOT be considered unless that
solution would have been adopted anyway, based on prevailing knowledge
of the languages, and WITHOUT taking into account the sentiment to move
away from 'arn'. There is no reason to cleverly "code around" the
situation to avoid retiring 'arn', at least not any BCP 47 reason.
Ethnologue says that Huilliche is "barely intelligible with" Mapudungun,
which is about the worst basis for creating a macrolanguage that I can

Unfortunately, the fact that we are talking about a code element in
639-2 as well as in 639-3 means the final authority belongs to the JAC,
which probably means we won't see a decision in the very near future.
But this is the decision I support.

Doug Ewell | Thornton, CO, USA | @DougEwell ­

More information about the Ietf-languages mailing list