arthur.reutenauer at normalesup.org
Fri Mar 11 16:45:32 CET 2016
On Fri, Mar 11, 2016 at 01:54:39PM +0000, Andrew Dunning wrote:
> This would also broadly parallel Greek’s distinction between el/grc (I assume that the latter would be a subtag if redone today).
I rather doubt it. They clearly are different languages, just like
Latin is different from any modern Romance language. The only
difference is that they both have "Greek" in their (English) names.
ISO 639 does not even need to concern itself with defining a cut-off
date, since it's a standard for names of languages, hence the ancient
one is mapped to [grc], and the modern one to [el] or [ell]; and what
language a document uses is for the person tagging it to decide. That
said, ISO 639-2 does actually give a recommendation, which I've always found
amusing: 1453 (see http://www.loc.gov/standards/iso639-2/php/English_list.php).
That is of course the year of the Ottoman conquest of Constantinople.
>>> 1. Use of i for /i, j/ and u for /u, v/ throughout.
>>> 2. Use of i for /i, j/ and u for /u/ and v for /v/.
>>> 3. Use of i for /i/ and j for /j/ and u for /u/ and v for /v/.
> Would these work better as script subtags rather than language variants?
That seems too specialised to be useful, but I'm not the one to judge.
> This isn’t just a Latin thing, and also applies to premodern variants of Romance languages.
Or even modern English, on an anecdotal basis; for example I work next
to this building, built in 1925: https://commons.wikimedia.org/wiki/File:Bush_House_NW.JPG
One could argue that the use of 'V' for 'U' is decorative, of course.
More information about the Ietf-languages