Re: [Ietf-languages] Latin Sub tags

[Doug Ewell <[e-mail]>]

Hugh Paterson III wrote:

> For example, if new Latin tags/codes were submitted to iso639-3, and
> they were accepted, the registrar has already indicated that the
> normal process is to deprecate the current tag/code [lat].  WHAT IF,
> that current tag were to be assumed by the ISO639-5 registrar? In a
> sense by proposing new Latin language tags for old-classical-neo one
> would not be negating the current tag as the current tag has in
> essence been operating as a macro-language tag (in a hypothetical
> situation where these other Latin tags are accounted as languages).
> The assumption of deprecation seems erroneous to me. But maybe I’m not
> understanding an important part of the information standard management
> process.

Any time someone brings up macrolanguages and/or collection codes, it raises red flags for me, for precisely this reason.

ISO 639-5, or what has at least been traditionally ISO 639-5, has NOTHING to do with macrolanguages.

Macrolanguages are very specifically an ISO 639-3 concept. The term refers to an identifier for multiple languages which, in certain contexts, are considered to be a single language. “Chinese” and “Arabic” are classic examples; in some contexts, each of these is considered a single language, and in others, each refers to multiple languages (e.g. Mandarin, Gan, Hakka, Wu...), each of which is clearly identified in 639-3 as an “encompassed language.”

Macrolanguages are highly relevant to BCP 47: they are the basis for the BCP 47 concept of extended language subtags, and each individual language identified in 639-3 as “encompassed” has a BCP 47 primary language subtag with a Macrolanguage field in the Registry.

https://iso639-3.sil.org/about/scope#Macrolanguages

Collection codes are the domain of (traditionally) ISO 639-5, and are NOT the same thing. These are simply bags of languages that are related in some way, either by family (“Bantu languages”) or geography (“Austronesian languages”) or by some other criterion. ISO 639-5 does not provide any list of the individual languages that belong to such a collection.

These are NOT sets of languages that are sometimes considered to be a single language. Nobody would say, for instance, “This document is written in ‘Bantu languages’” or “I need a ‘Bantu languages’ spell-checker.” As such, they usually have little relevance to BCP 47 language tagging situations.

There are some “family” collections in 639-5, such as “Chinese (family),” but that is not the same entity as the 639-3 macrolanguage [zho] (which is realized as 'zh' in BCP 47 via the equivalent 639-1 code element). The membership might be different, although that is hard to tell because there is no list of the ISO 639-3 collection.

https://iso639-3.sil.org/about/scope#Collections%20of%20languages

PLEASE do not mix up these two concepts.

> In the past we have seen tags removed/deprecated and then reactivated.
> So why can’t a tag move from single language to a macro-language where
> the macro-language is an entity/concept through time? Likewise why can
> a code move from representing a single language to representing a
> group of languages? Especially if that is how the tag has been
> functioning already?

In January 2010, ISO 639-3/RA reclassified the individual language [lav] Latvian as a macrolanguage, encompassing the two new individual languages [lvs] Standard Latvian and [ltg] Latgalian. In the same year, [bnc] Bontok was also reclassified as a macrolanguage, encompassing three new individual languages. (There may have been other such conversions more recently; Latvian was the one that popped into my head.) Note that there is no intermediate stage of deprecating anything.

So yes, it is possible and has been done. And it seems reasonable that, because the name “Latin” and the ISO 639 code elements and BCP 47 subtag for “Latin” have already been used, and are still in use, for all stages of Latin, this could be a candidate for this sort of conversion, if the new language code elements you suggest are approved. Again, the RA will consider each linguistic situation on its own, not simply because it was done in the past for this language or not done for that language.

--
Doug Ewell, CC, ALB | Lakewood, CO, US | ewellic.org