I think attempting to quantify the word “uncommon” in this tag isn’t very productive, as was noted in the question itself.
For what its worth, the English language Wikipedia lists 312 languages for which Wikipedias have been written (see en.Wikipedia.org, search term “List of Wikipedias”). In that list there is listed for each of the Wikipedas the number of editors/authors who have written at least one article in that Wikipedia during the past month i.e., that is a usage data statistic that can be used as a reference to gauge as to how many SE-LL members might use an “uncommon languages” tag.
I for one believe the tag name “uncommon languages” is descriptive and accurate for what I think it implies. However, in answer to the question, my suggestions are:
- the tag title should remain “uncommon languages” and,
- the current definition should be changed to a longer and more descriptive one.
That being said, long definitions are also common in the SE-LL tag list. A longer definition for this tag (one I would prefer to see) could be something along the lines of:
Queries about one or more languages that are
a) rarely spoken or written by non-native speakers, e.g. Bugotu, Tok Pisin,
b) languages that have unusual spoken or written characteristics, such as lack of vowels in written words, e.g., Arabic, or
c) languages which are no longer used for communicating with others, e.g. Latin, Anglo-Saxon.
Note that in my suggested definition I make a distinction between “languages no longer used for communication” and the so-called “dead” languages, some of which (e.g., Latin) have a large geographic distribution of those who are learning to read and write the Latin language, but are rarely using it for communicating with others.
I noted that the stumbling block to re-defining the tag is the attempt to quantify the meaning of the word “uncommon” when the quality of the definition is the keyword to keep in mind. Some rare or unusual languages, such as the many different languages spoken in the islands of the South Pacific may have only a few native users of a language, or a relatively large number of users, such as Maori and Tahitian.
Geographic distribution is important, yes. But actually it is a lesser factor in determining whether or not a language is “uncommon”. Hawaiian, as spoken by native speakers, is a distinct language having a small number of native speakers, with a geographic distribution of those speakers limited to just one of the fifty U.S. states and U.S. territories. That being said, however, Hawaiian is included as a member of the Polynesian language family, which has a far broader distribution geographically and has more than two million native speakers and writers. Quite similarly, “Plattdeitsch” (Mennonite German) has a limited distribution geographically, but it is also included in the geographic distribution of the Germanic languages family, in which English is also member; that language group obviously has a geographic distribution of much more than two million native speakers. The point of this is that I believe trying to statistically quantifying a language is a dead end street.
The current definition in the “uncommon” tag lists ‘dead’ languages as falling under the “uncommon” tag; SE-LL already has a “dead languages” tag, so that part of the “uncommon” definition is redundant and can be eliminated.