Re: Macrolanguage (was: Forms for Montenegrin)

> It would also help a lot if the RA for ISO 639-3 expedited the addition of cnr to its tables since it is an anomalous situation to have, at any time, individual-language entries in 639-2 that are not in 639-3.

Indeed, BCP47 is designed around, and its proper operation, somewhat depends on this being an anomaly.


On December 24, 2017, at 9:59 AM, Peter Constable <[email protected]> wrote:

I think I borrowed the term "macrolanguage" from David Dalby, but the usage in ISO 639 is defined specifically for ISO 639. The macrolanguage relationships in ISO 639 may often appear to be classifications, but that is not what is meant. The term is defined in ISO 639-3 as follows:


"3.8

"macrolanguage code element

"a language code element with a scope representing multiple, closely-related individual languages that are deemed in some usage contexts to be a single language

"NOTE        Every macrolanguage code element has a normative correspondence to the individual language code elements representing the individual languages encompassed by the macrolanguage. This normative relationship between macrolanguage code elements and individual language code elements is documented in the code tables included in this part of ISO 639. The notion of macrolanguage is explained further in ?4.2.2."
The macrolanguage relationships are defined in 639-3, but are not defined in the current version of 639-2. So, as Doug notes, the macrolanguage info can't be added in the registry on the basis of ISO 639-2. And it can't be added on the basis of language classification since that's not what this field is in the LSTR:

"3.1.10.  Macrolanguage Field
"The field 'Macrolanguage' contains a primary language subtag (whose record appears in the registry).  This field indicates a language that encompasses this subtag's language according to assignments made by ISO 639-3."

It would be a bug in 639-3 data if it fails to add cnr to the items encompassed by hbs.

It would also help a lot if the RA for ISO 639-3 expedited the addition of cnr to its tables since it is an anomalous situation to have, at any time, individual-language entries in 639-2 that are not in 639-3.



Peter


From: Ietf-languages [mailto:[email protected]] On Behalf Of John Cowan
Sent: Friday, December 22, 2017 8:20 PM
To: Michael Everson <[email protected]>
Cc: ietflang IETF Languages Discussion <[email protected]>
Subject: Re: [Ietf-languages] Macrolanguage (was: Forms for Montenegrin)



On Fri, Dec 22, 2017 at 3:51 PM, Michael Everson <[email protected]<mailto:[email protected]>> wrote:
On 22 Dec 2017, at 20:27, Doug Ewell <[email protected]<mailto:[email protected]>> wrote:

> "Macrolanguage" is completely an ISO 639-3 concept.

The term is odd, but it's a familiar concept in linguistic classification.

It isn't, which is why Peter Constable (who is certainly familiar with linguistic classification) invented it.

You imply that Montenegrin's would either have no macrolanguage (but it is not a linguistic isolate) or that its macrolanguage would be anything other than Serbo-Croatian. But neither of those would be correct.

English does not have a macrolanguage, but it is not a language isolate.  Rather, it does not belong to a group of languages (as defined by linguistic standards) which are for some purposes treated as a single language with a single name, as is the case for Chinese, Arabic, Norwegian, Quechua, and Serbo-Croatian.  If on the other hand it were usual to call both English and various English-based pidgins and creoles simply "English", and only differentiate between them in specialized contexts, then English would indeed be part of a macrolanguage called "English (macrolanguage)".

--
John Cowan          http://vrici.lojban.org/~cowan<https://na01.safelinks.protection.outlook.com/?url=http:%2F%2Fvrici.lojban.org%2F~cowan&data=02%7C01%7Cpetercon%40microsoft.com%7Cb1df6591dec54af32b5208d549bc8544%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636495996404422629&sdata=nUSbmHeCOizsXp%2BqoUw%2B0K74r9mWba3g5f1RrctPAkk%3D&reserved=0>        [email protected]<mailto:[email protected]>
An observable characteristic is not necessarily a functional requirement.
        --John Hudson

_______________________________________________
Ietf-languages mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ietf-languages