(Reposting and elaborating from a VfD 'keep' comment.)
Okay, I see two points here. First off, I can understand the "verifiability" clause... for most other items on Wikipedia.
HOWEVER, conlangs are in my opinion different in a way that makes this not just unnecessary, but exclusionary. Verifiability by definition is so that you know that the info is good, the thing being written about actually exists as described, etc. With a conlang or other work of art, it is 'self-defining'. If the author posts the conlang, ipso facto, it exists as described. Any talk of "verifying" it - e.g. by others talking about it, etc. - goes solely towards 'notability' (or "merit", as you prefer).
So that bring up my second point. I don't see reason to exclude things from Wikipedia for notability. If it's small and not interesting, then don't mention it in the main articles, or in the more exclusive "these conlangs are interesting/notable" lists. But there is no reason I can understand not to have an article about it for whatever it's worth, or to not include it on an all-inclusive list. So I only support "notability" for that 'alone': determining what to include in high-level / central articles. (And FWIW, I would support including as an article even sketch conlangs, if there's anything to say about them, so long as they don't cause namespace problems [e.g. colliding with something important] and aren't included on "notable conlangs" lists.)
Perhaps this PoV can help as a middle path. I won't touch the "how do you determine notability / merit" discussion at this point - seems to be going just fine by itself. I'm just going to suggest that it be applied solely to deciding whether something should be included in *major pages* (like Conlangs or Artlangs etc), lists of notable conlangs (duh), etc - not whether there should be a separate page on it. --Saizai 15:47, 17 August 2005 (UTC)
- I think you are missing the difference between an encyclopedic article and a listing in a collection of miscellany (see WP:NOT). Existence isn't the point: does the article say anything meaningful about the subject? If it does, it makes claims that are not self-verifying. For example, the creator can claim that it has a 5,000 word vocabulary, but that doesn't make it so, or true in a way that the reader might expect. Suppose that the language may has an idiosyncratic vocabulary that will not support basic conversation: absent such a critique, the article is not NPOV, which is nearly the only rock-bottom non-negotiable policy that Wikipedia has. If the article restricts itself to the name of the language and the creator, it hasn't said anything worth saying, and is a "permanent substub," which is also WP:NOT. Robert A West 22:08, 17 August 2005 (UTC)
- Sai, there are so many problems with your proposal that I don't know where to begin.
- First of all, I hope that you don't mean that the author of a conlang posting his description as a Wikipedia article should be considered sufficient proof of existence, and hence that such an article should be kept. That goes against WP:NOR and Wikipedia is not a publisher of original thought and possibly other principles as well.
- Second, I think you are confusing the map with the terrain, so to speak. A web page that describes the constructed language is not the constructed language itself. It may be the case that most of it exists in the tattered notebooks in the author's drawer, or that the web page has become significantly out-of-date.
- Third, the author may be mistaken in his characterisations of his own language. On at least one occasion, I have seen people make claims of the grammatical properties or genetic origins that, on examination of the grammar or example texts, clearly make no sense. But to analyse the grammar or sample texts ourselves (assuming the absence of third-party analyses or mentions) would amount to original research. And we can't do that. Hence, including with no secondary sources on a subject as obscure as this is opening the encyclopedia up to errors that are difficult to discover and correct.
- Fourth, it is also possible that the author is deliberately lying. If I were to publish, today, a web page that says that Quuxish has a vocabulary of 12 000 words, by your rules WP editors would have to take that at face value. But in fact I never created a language like that, the only word I made up is "Quuxish", and that was right now. The other 12 000 don't exist, not even in my head. arj 22:26, 17 August 2005 (UTC)
- I strongly agree with arj: the creator's web page by itself is not necessarily sufficient evidence of the properties of the language. If that's all the information about it that exists, it is not verifiable without original research; the first person besides the creator who writes extensively about it will probably take some position about how valid and consistent the creator's claims are, comment on how complete and expressive the language is/seems, etc., which amounts to original research. Such articles are welcome at the Conlang Wikicity, but they would be out of place here. Once one or more such independent reviews have been published, a Wikipedia article may be written that is NPOV and verifiable, referencing the claims of the creator, the comments of one or more reviewers or critics, etc. --Jim Henry | Talk 16:35, 18 August 2005 (UTC)
...
Responding to the above:
I understand the concern of "mistaking the map for the terrain". Let me modify what I said a bit.
As far as I am concerned, authors' notes and etc., while probably the most "primary source", are not really accessible. If you want to consider them such, then you would need to consider webpages *about* them to then be a de facto *secondary* source, a sort of review *of* the conlang. Which makes it pass some of the criteria suggested for "verifiability" and "notability". I think we'd agree that this is a bit silly.
So, yes, it is possible that a conlang as described may not be a full description of the conlang. However, for all intents and purposes, it nevertheless *is* the conlang - at least insofar as the writing of a design can be said to 'be' the design. As someone who looks at conlangs mainly from a theoretical perspective, I would hold to saying that - at least in the case of conlangs which are sufficiently and cogently described on their webpages (e.g. Toki Pona, Ithkuil, et al) - they are self-defining. If you describe it, it exists. I think it's an interesting question whether an insufficiently described language (e.g. Hepatpod B) is a "real" language - one might consider it a sort of "fictional" conlang (purely apart from the fact that it happens to be described in a work of fiction) - but it could just as well be turned into a real one with more elaboration. (In fact, that's what my threads re. non-linear writing on CONLANG were partially attempting to do.)
(Let me clause this: if it's internally inconsistent, then one could rightly make an argument that its primary source - the webpage about it - is flawed in terms of verifiability and reliability, and that would be an argument for exclusion or at least for a remark to that effect on the Wikipedia article about it.)
I agree with RAW and Arj's comments against Wikipedia being a primary source of itself. I am not advocating that it should be - I think I've been consistent in saying that we are still referencing material that is 'primarily' described elsewhere, e.g. on its homepage, and that that source should be reliable (i.e. online, non-broken, legible). If the authors' notes describe something different than what is described on the accessible website, well, I'd regard that as two different but related languages - the one online is the only accessible one, so that's what you go with.
This also goes for the comment about a Wikipedia article making claims that are not accurate of the accessible source material. If it says that language X has a 5kword vocab, and its webpage only lists 5 words, then this is false reporting - as I said, I don't consider "authors' notes" to be a relevant source. If it is a mere relex, or worse a one-to-one ("foobar" means "car"), then that should be noted, since that's a bit different than claiming a bona fide 20kword vocabulary.
What I *am* advocating is this, the first line of WP:NOT: 'This means that there is no practical limit to number of topics we can cover other than verifiability and the other points presented on this page.'
Again, the essence: I don't consider notability to be a reason not to have an article about it.
I think I'll add one more clarification that I think a couple of you mistook... I do not mean to say that a webpage talking *about* a conlang is itself the conlang. E.g. Arj's comment #4, echoed by Jim Henry - if I write the sentence "The language Foobar has a 5kword vocabulary and is a Romance/Mandarin/ASL creole", that is obviously not enough to make that language actually exist. We agree on that.
What I mean by self-defining is something that actually *does* describe the language - e.g. Ithkuil's webpage. It describes the language in full (AFAIK it's not the most current version and is missing some vocab, but close enough); the language is, ipso facto, that which the webpage describes. I'm not sure how better I can put that. I would contrast this to things like the Klingon language website, which I would consider a *secondary* source that talks about the language that is *primarily* described - defined, conjured into existence - in the Klingon Dictionary books. But I think in most cases, the website is the primary source, and should be treated as such.
Hopefully that clarifies my point a bit. --Sai Emrys 02:36, 19 August 2005 (UTC)
- The thing I've realized I dislike about disregarding notability is that it seems to me that notability's the elusively-defined criterion that keeps things from running afoul of section 1.4 of WP:NOT (Not a Propaganda Machine), in addition to helping keep articles from violating section 1.7 (Not an indiscriminate collection of information). Basically, I see the notability test as answering two questions: Does it matter? Why? So, does Ithkuil matter? Why? Does Klingon matter? Why? Do the Mona Lisa, the Dreyfuss Affair, existentialism, or rape culture matter? Why? (Or why not?) That, to me, is the notability test. And I'll admit, I believe things that don't matter don't deserve articles. I admit, of the 9 sections of WP:NOT that deal with articles, I think the first one's the least important. But still, if a conlang doesn't matter, it shouldn't be included. Same as anything else, right? The Literate Engineer 06:12, 22 August 2005 (UTC)
- My issue with this is that it becomes subjective. Sure, one shouldn't be deluged by things that "don't matter" when trying to do a high-level review, or when trying to just be told (e.g. through conlangs) what the notable events, languages, ideas, etc. are. I fully agree with that.
- But I fail to see any harm in having seperate articles detailing what *is* known about a conlang, assuming it is still out there in some accessible primary form (online, book, whatever). That they "don't matter" is not to me an argument for deleting any mention of them, especially when namespace and storage space are not at issue. --Sai Emrys 19:20, 24 August 2005 (UTC)