I think for the MARC list of codes, we will want to change the caption
"Sorbian languages" to "Sorbian (Other)."  This will signal to MARC
users that there may be separate codes for individual languages of the

Milicent Wewerka
Library of Congress

>>> Havard Hjulstad <[log in to unmask]> 07/16/03 04:41PM >>>
The issue raised by Peter (below) is very important (although actually
the "real" issue of the current thread; which is the reason why I have
changed the subject line).

With the new developments in language coding we need to re-think
the "group identifiers", and especially the "rest group identifiers". I
proposing that "group identifiers" be developed in a separate part of
For some implementations it is important that an item is, e.g.
"English" and
therefore "Germanic" and therefore "Indo-European". Other
need a "set classification" rather than a "hierarchical
classification", and
classes like "other Germanic languages" may make sense. But the
implementation obviously needs to know "Germanic languages other than
ones?". I am proposing a "principles part" of 639 to standardize such

As for Sorbian, there is currently no proposal to do anything with
"wen". It
may eventually move to the "group identifiers" part.


Havard Hjulstad    mailto:[log in to unmask]
  Solfallsveien 31
  NO-1430  As, Norway
  tel: +47 64963684  &  +47 64944233
  mob: +47 90145563

> -----Original Message-----
> From: ISO 639 Joint Advisory Committee [mailto:[log in to unmask]]On
> Of Peter Constable
> Sent: 16. juli 2003 22:06
> To: [log in to unmask]
> Subject: Re: ISO 639 JAC ballot 15-2003 and 16-2003 - Upper Sorbian
> Lower Sorbian - Preliminary results
> > Some of the comments have been relating to the issue of retaining
> as a
> > group identifier for Sorbian (or as a rest group identifier for
> > Sorbian languages"). Evidence has been presented in favour of
> the
> > group identifier, and no proposal will be made to deprecate it.
> The only issue I see in retaining the group identifier is clarifying
> denotation. There is a current anomoly in the the text of the
> standard kind
> of suggests that the denotation of a collective excludes any
> varieties that
> have their own identifier. That would mean that "wen" has a null
> denotation
> (unless we say it still includes historic varieties). Also, I have
> out in the past that there is a problem with collective categories
> have an "other xxx languages" denotation in that (in the general
> case) when
> a member variety of the collection is assigned its own identifier,
then by
> these rules (the collection excludes varieties with their own ID)
> data in that member variety that was tagged using the collective ID
is now
> incorrectly tagged. If the collection was inclusive, the existing
> would be correctly but sub-optimally tagged; by having the category
> exclusive, the existing data becomes incorrectly tagged. (And
> probably nothing in most systems to advise anyone that there is now
> incorrectly tagged data.)
> - Peter
> ------------------------------------------------------------------
> ---------
> Peter Constable
> Non-Roman Script Initiative, SIL International
> 7500 W. Camp Wisdom Rd., Dallas, TX 75236, USA
> Tel: +1 972 708 7485