> What I would hate to see is that we approve a project and then become a laughing stock because it turns out to be an unserious thing in one way or another.

 I don't think it's an unserious thing, and you surely don't have to worry that the language isn't real. It's as real as its grammar, its dictionary and all the texts written in it. Kotava doesn't claim to be some undiscovered natural language, so you can't call it a hoax either — unlike Siberian, which was a real language with an extensive dictionary and quite a few texts in it, too, yet the Siberian Wikipedia was deleted after the language turned out not to be what it claimed to be. The Klingon Wikipedia was deleted, as far as I can tell, because the language simply was too incomplete for a workable encyclopaedia. In the case of Kotava, neither of these problems are likely to arise.

The only thing that is not verifiably true is that Kotava is indeed a serious proposal for an international auxiliary language. The fact that it claimed 40 speakers right after its first publication on the Internet, this in spite of a total lack of publicity, could indicate that it is rather a hobby project of a dedicated group of friends (not that this would necessarily disqualify it, mind).

> Is there some proof here of any of: "independently proved number of speakers, use as an auxiliary language outside of online communities created solely for the purpose, usage outside of Wikimedia, publication of works in the language for general sale", as suggested by the admirable Language Proposal Policy?

Numbers of speakers are always trouble when it comes to constructed languages. The only figures we can establish objectively are the number of members of a mailing list, a forum, a Facebook group, etc. We can dig a bit deeper and establish how many people in these groups are really active users of the language, but that's it.

Kotava has a Facebook group with currently 67 members, but with little activity. Compared to other constructed languages, that is an extremely small number. For example, the Esperanto group has 21628 members, Interslavic 7788 members, Toki Pona 4146, Klingon 2155, Interlingua 1641, Lojban 1641, Lingua Franca Nova 550, Volapük 401, Ithkuil 353, Slovio 286, Folkspraak 260, Interlingue 215, and Novial 92). Obviously, not every member is a speaker and not every speaker is a member. Based on my own experience and research, I'd say that in general no more than 5–10% of these members can be considered regular, proficient users and another 10–20% are beginners and interested bystanders.

Kotava also has a forum (http://www.kotava.org/phpBB3/index.php) with 194 members and 7260 messages. Although the vast majority of all communication is in French, there is also a reasonable amount in Kotava. It is hard to tell how many people actually write in it, but I'd say at least ten. However, there has been remarkably little activitity on that forum these days (only 8 short messages in 2019), which makes me wonder if this language is actually used for anything else but making massive amounts of translations.

Perhaps I am being over-suspicious here, but the simple truth is that auxlang proponents have a vested interest in exaggerating its significance. Let me just remind you of Slovio, which claimed to have thousands and thousands of speakers, but in reality has never been used by more than 10–15 people. The author tried to build the impression of a whole user community by writing under dozens of different names, but as soon as he abandoned the project, all activity died out instantly. While I don't believe Kotava is a similar case, we can't know it for sure.

> "Kotava indeed has an impressive text corpus." Is there a convenient list somewhere?

Mainly here: http://www.europalingua.eu/wikikrenteem/?title=Emudexo

> I would suggest you consider looking at https://fr.wikipedia.org/wiki/Kotava. It also has sourcing problems, to be sure. But this conlang, unlike many, comes out of the francophone world. So I'm not sure the lack of awareness in the anglophone world is the most appropriate measure of the language.

That argument would be valid if we'd try to establish notability by googling for "Kotava language" or somesuch, but in the case of Kotava, there is no need for that at all. What I do notice, however, is that Kotava generates practically no results in news items, books and scholarly publications in any language, French included. All we have is three publications by Alan Reed Libert, none of which discusses Kotava in depth; two sentences in Turenne; and one mere sentence in a Brazilian magazine. These are exactly the same sources that we already had six years ago, in other words: even the fact that Kotava has had an ISO 639-3 code for over ten years now does not seem to have made it any more notable than it already was. Numerous books and articles about constructed languages that have been published in the meantime do not even mention it. Even Google gives very few hits, and most of those refer to a surname or a Belarussian village).

Besides, the French Wikipedia article is by no means better than the English article. It's 100% based on primary sources and all information basically comes from one single website. In other words, original research. That's the biggest problem with Kotava: the amount of information that can be verified in independent, reliable sources is totally insufficient to base a decent article on.

One thing that makes me kind of suspicious is the discrepancy between the rather bold claims made by Kotava supporters and this total lack of verifiable evidence. If Kotava really has 40 fluent speakers (BTW, the same figure as 12 years ago), 200–300 less advanced speakers and 1000–2000 people who started learning it at some point, how on Earth is it possible that (except for the above) not a single authority in the field, not even some local newspaper, has paid attention to it? Mind, these numbers are exceptionally high for a constructed language! Since the language apparently has managed to fly under the radar for over 40 years, I'd really like to know where all these people come from, and why they are totally invisible.

Questions like this are answered with defensive answers. Like this one: we are not interested in promulgating our language, we prefer to use it instead. A strange argument, since promulgation is essential for an International Auxiliary Language to function as such, and not true either, given the promotional nature of the original Wikipedia entries about Kotava and the behaviour of their author(s) in deletion discussions. Or this one: we are often neglected, because we are opposed to the supremacy of English (also strange, considering that 90% of all internal discussion is in French). Thus, the whole plea for notability is solely based on a large text corpus, a successful request for an ISO 639-3 code (which surely would have failed if it had been made a few years later) and complaints at the address of Wikipedia censors, jealous Esperantists and English supremacists.

Furthermore, what exactly do they NEED a Wikipedia for? Since Kotava is an a priori language, content in it is of no use to anyone outside its own small incrowd of two or three dozens of people, in other words: the same people who write it. This group already has a well-functioning infrastructure, including a Wikipedia clone and a Wiktionary clone, so what's the point of duplicating all this info? An additional Wikipedia project cannot serve any other purpose than promoting the language, which (as demonstrated by the examples of Lojban, Novial and Volapük) is not going to happen anyway.

The basic question is, therefore: should a language that doesn't fulfil the basic criteria of notability and verifiability on Wikipedia.en nevertheless be allowed to have its own Wikipedia?

I'm not saying it shouldn't. The language is certainly interesting, and I truly admire all the work done by those who generate content in Kotava. As I wrote earlier, the main criterion for establishing a new Wikipedia project should IMO be viability, and a Kotava Wikipedia is definitely viable. But one thing remains: merely writing/translating lots of text in a conlang does not make it notable.

Best regards

Jan van Steenbergen