The list of available languages (and their names), for use by MonolingualTextValue and friends, should combine three different sources:
Note that there two use cases, which need different lists of languages:
one for languages that is supported for the UI (Names.php, pretty much),
and one for input (labels, description, monolingual text, etc).
We need as starting list of language the languages defined by the ISO 639-3 norm. The best is to sort the list with the three-letter alphabetic codes and then to give the name of the languages in the language of the user. It is very difficult to choose the correct language when it is labelled in original alphabet like 中文（简体） or 中文（繁體）.
I rearranged the various tasks around this feature. This is now the tracking bug for a) supporting the same set of languages everywhere in the content and b) supporting languages on wikidata which are not yet present in any of the data sets we have access to.
Just find this list for variation of metalanguage: http://www.iana.org/assignments/language-subtag-registry/language-subtag-registry
Theoretically, if I understand correctly, all valid ISO 639-3 codes could be used. This is the famous ~7000 number. Getting a list of valid codes shouldn't be hard.
If it doesn't scale, I'd support having an easy procedure for adding codes for needed languages. Adding a code to ULS's langdb is technically very easy.
We kind of have this now:
private function getMonolingualTextLanguages() {
return $this->getWikibaseContentLanguages()->getContentLanguages( 'monolingualtext' );
}There is also an API for this.
For example, so I'll close this.