On 29 Oct 2017, at 17:53, Mark Davis ☕️ <[log in to unmask]> wrote:
>
> The problem is mixing levels.
>
> The principled, formal criteria for 639 is to have codes that reflect only linguistic features: and bases distinctions only on mutual intelligibility.
I don’t know that I’d stipulate “only” here. Swedish, Norwegian, and Danish are mutually intelligible, though they are distinct. We do not distinguish them just because of spelling or political concerns, though (compare the two orthographies for Belarusian).
> ISO 639 itself does not need to be muddied by some arbitrary patchwork of extra political considerations that do not match the criteria for mutual intelligibility.
Again, intelligibility isn’t the only criterion for linguistic distinction.
> Why? Because there is IS already a standard (BCP47) that cleanly allows the separation of dialects based on political boundaries, leveraging both ISO 639 and ISO 3166.
Well, this depends. We can distinguish Valencian from Catalan by “cat” plus a country code (which does not necessarily imply any spelling or pronunciation differences), but when we are creating new subtags and I am asked to evaluate a proposal I apply linguistic criteria, as where the two Belarusian orthographies are distinguished.
> Some people do not realize just how broadly that standard is implemented: every modern OS for computers, smartphones, tablets, etc. supports it. Based on my experience, it is far, far more broadly implemented than just the ISO 639 codes by themselves. There are some old legacy systems that only support raw ISO 639 codes, but they will end up having to support BCP47 or be left behind.
Mark is right.
> The code for Montenegrin is thus already recognized: it is sr-ME.
Mark is right.
> Trying to add an additional ISO 639 code for Montenegrin would would not magically improve support for separate translations or handling of Montenegrin. It would do precisely the opposite, and impede the development of programs and content. If people really want to get Montenegrin supported more broadly, they should pour their efforts into more useful efforts.
This is no different than en-US and en-GB.
If I had a vote, I would vote against the adoption of an ISO 639 code for Montenegrin, because there has been no demonstration of its distinction _as_ a _language_.
This rationale is not difficult to understand, and should be communicated to the proposers. If they can dispute it on linguistic grounds, that’s one thing, but they would need to address the US/GB analogy. Explicitly.
Michael Everson
|