I beg your pardon. I thought most of my questions were relevant to the 
request regarding Korean. I am trying to understand on grounds a "group 
node" is to be proposed for Part 5, especially one that corresponds to an 
individual language present in Part 2 (or only in Part 3, for that 
matter). On the grounds that it is an isolate, or an otherwise independent 
branch in a large family? I can understand that (and see that it has 
implicit support in the "definition" of language family on page 12 of the 
standard document). My question about the criteria that should met in 
order to qualify, and my question about the intended comprehensiveness of 
Part 5 are still directly relevant to the question of Korean (and a lot of 
other potential "group nodes" that are missing).

Please remind us (me) of the principle as it was discussed on this list.


Håvard Hjulstad <[log in to unmask]> 
Sent by: ISO 639 Joint Advisory Committee <[log in to unmask]>
2008-11-12 06:34 AM
Please respond to
ISO 639 Joint Advisory Committee <[log in to unmask]>

[log in to unmask]

Re: ISO 639-5 item needed - Korean (family)

As to the concern in your second paragraph: I am very sorry that I haven’t 
been able to follow up 639 issues as closely as I would have wished to 
during the last several months. I am now in the process of catching up, 
doing quite a lot of it on my free time. Please let us not discuss this 
under the current thread. (To avoid drowning: please just one issue per 
The main issue here seems to have two aspects: (1) Should 639-5 have 
“group nodes” at positions in the hierarchy where there is just one (and 
cannot theoretically be more than one) sub-item. (2) If so, is the 
proposal below ok.
I don’t have strong opinions, but some experts did express very clearly 
during the development of 639-5 that it is indeed desirable to have such 
“group nodes”. As project editor I just implemented what was a clear 
consensus among project participants: (a) All “top nodes” need to be 
“group nodes”; (b) “important” nodes further down (e.g. hyx) should be 
“group nodes”. The current proposal just intends to fill a gap (an error 
on the part of the project editor).
The principle was discussed during the development of 639-5. A change of 
the principle would be a matter of a revision of 639-5.
Best regards,
Håvard Hjulstad
  Standard Norge / Standards Norway
  [log in to unmask]
Fra: ISO 639 Joint Advisory Committee [mailto:[log in to unmask]] På vegne av 
Joan Spanne
Sendt: 12. november 2008 00:28
Til: [log in to unmask]
Emne: Re: ISO 639-5 item needed - Korean (family)


I have in the past asked specifically for a clarification of the meaning 
of “hyx – Armenian (family) – arménien (famille)” and “jpx – Japanese 
(family) – japonais (famille)” in contrast to the individual language in 
each case. Until that is given, I am not in favor of adding further code 
elements that are not clearly distinguished from individual language 
elements in Parts 2 and 3. 

I have copied my message of 6 Sept.  2007 which includes these and a 
number of other questions to which you have made no response. There are 
also other outstanding maintenance issues with regard to Part 5, most 
notably that Papuan languages [paa] should be above Trans-New Guinea 
languages hierarchically, not below it. 


----- Forwarded by Joan Spanne/IntlAdmin/WCT on 2008-11-11 05:09 PM ----- 

Sent by: Joan Spanne 
2007-09-06 02:54 PM 

ISO 639 Joint Advisory Committee <[log in to unmask]> 
[log in to unmask] 
Re: ISO 639-5 bigger issues

Håvard and all, 

I said in my short message that I had some other issues to raise. These 
are my more substantive questions: 

I think there is an error with one code element: 
Carib [car] is an individual language, named Galibi Carib, in both 639-2 
and 639-3. Carib languages probably warrant a code element (but without 
criteria, I could not tell), but it has to be different from [car] (that 
probably lingered from the DIS table, since the decision on Galibi Carib 
was after the first discussions of Part 5)

Larger issues: 
First and foremost: what is the basis for a group being included in the 
code set of Part 5? What criteria should be met in order for a group 
(whether genetic language family or other) to warrant a code element? It 
appears the list is not attempting comprehensiveness in including groups 
that encompass all individual languages, even at the highest level of the 
hierarchy (130 or more of the languages in Part 3 may not map to any group 
code element in Part 5--geographic or genetic--depending on the genetic 
classification to which the language might be assigned). 

What relationships may exist between code elements within Part 5? Is it a 
rule of the standard that no code element overlaps (in terms of the 
individual languages, and perhaps also proto-languages, which would come 
within its denotation) with another code element except in a strictly 
hierarchical relationship? Or may the denotations of code elements 
overlap, allowing for different geographic boundary sets or different 
genetic trees? 

The standard makes almost no reference to Part 3 (apart from listing its 
existence) What is the relationship between code elements of Part 3 (and 
individual language code elements of Part 2) and code elements of Part 5? 
Should a Part 3 code element t map to a single code element (the lowest 
appropriate in a hierarchy) or could there be mappings to more than one in 
Part 5? (This would be related to the previous question.) 

A specfic aspect of the previous quesion is: what is the distinction 
between a macrolanguage code element of Part 3 (by definition an 
individual language code element in Part 2) and a code element in Part 5 
that has apparently identical meaning, but is labelled a Family--or not 
labeled at all (e.g. Albanian -3 [sqi] , -5 [sqj]; Quechuan -3 [que], -5 

What is the purpose of a language family code element for a "family" of 
one language that is part of a higher level family? The exception 
situation is noted in the document, but no motivation for this is given. 
This would pertain to Armenian (family) [hyx], though there are historical 
varieties also coded for Armenian; and Egyptian [eqx]. (What is the point 
even for a "family" of one that is an isolate?) 

And adding to Milicent's list of questions (e.g., the Bantu question) of 
the meaning of specific code element (making use of Ethnlogue 15th ed. 
classification data), what is intended, for: 
Chinese (family) [zhn at present], does this include Dungan [dgn], or is 
it the same as [zho] Chinese (macrolanguage) (a problem as metioned 
Ijo languages [ijo], does this include Defaka [afn], an "Ijoid" language? 
Japanese (family) [jpx], does this include the 11 Ryukyuan languages that 
Ethnologue classes as "Japanese languages"? If so, it might be clearer to 
refer to them, than to Japanese [jpn] as an individual language in Parts 
1, 2, and 3 and call them "Japanese languages". Otherwise, what else is in 
the Japanese family? 
Is Mongolian [xgn] to be interpreted as the Altaic sub-family "Mongolian 
(languages)" (14 languages) or "Mongolian proper" (2 languages, part of 
Mongolian family)? If the former, then certainly "Mongolian languages" is 
Is Formosan languages geographic or genetic? There are 8 Formosan (or 
Northwest or East Formosan) languages, but 12 other languages native to 
Taiwan, in other very small Austronesian families, all of them families 
that are only in Taiwan.
There may be more with ambiguous meanings. 

I had a long discussion with Gary Simons on these matters. I will send his 
write-up of that in a separate email. 


Håvard Hjulstad <[log in to unmask]> 
Sent by: ISO 639 Joint Advisory Committee <[log in to unmask]> 
2008-11-11 02:28 PM 

Please respond to
ISO 639 Joint Advisory Committee <[log in to unmask]>

[log in to unmask] 

ISO 639-5 item needed - Korean (family)

Dear ISO 639 RAs-JAC members, 
At the meeting of ISO/TC 37/SC 2/WG 1 in Moscow in August 2008 
representatives from the Democratic People’s Republic of Korea pointed out 
that an ISO 639-5 identifier is needed for “Korean (family)”, 
corresponding to, e.g., “hyx – Armenian (family) – arménien (famille)” and 
“jpx – Japanese (family) – japonais (famille)”. The Korean language is an 
individual language (ko / kor), which is the sole member of the Korean 
language family. The input from the Korean delegates was accepted by the 
WG, and it was explained that the process of including an identifier for 
“Korean (family)” would be brought forward to the 639 RAs-JAC. 
Finding a suitable identifier starting with k  is next to impossible. 
Identifier: wko 
English name: Korean (family) 
French name: coréen (famille) 
Any comments or discussion? 
Best regards, 
Håvard Hjulstad 
  Standard Norge / Standards Norway
 Postboks 242, NO-1326 Lysaker
 besøksadresse / visiting address: Strandveien 18 
  tel: (+47) 67838600  |  faks / fax: (+47) 67838601 
  direkte tel / direct tel: (+47) 67838645 
  [log in to unmask]