Wikipedia:Conlangs/Alternative proposal

Well, that was not what one would call a good start for our vote. Okay, never mind. Since we decided to revisit the discussion a little, I would like to propose an alternative questionnaire. Because frankly, I think the current questionnaire is needlessly chaotic, complex, unclear, and long. The odds are that the final result would be a set of criteria that are so complicated that no one will be able to control everything, and that's pretty much beside the point of this whole operation. Let's not forget why we started this: because anybody could call anything "non-notable" just on the basis of this own sentiments. What we want is to provide a set of objective criteria that can decide whether a conlang is notable enough for inclusion.

On the basis of the TOC, I have grouped the fourty question under a few headers:

• NUMBER OF SPEAKERS
  • 1 Has 10 speakers
  • 2 Has 30 speakers
  • 3 Has 100 speakers
  • 4 Has 300 speakers
  • 5 Has 1000 speakers
• COMPLETENESS
• Language
  • 21 Has a complete grammar
  • 24 Has a sufficiently large vocabulary
  • 28 The language can be used to define its own grammar
  • 29 The language can translate the Babel Text
• Corpus
  • 33 Has been used in at least 5 translation relays
  • 34 At least 10 original texts are published in the language
  • 35 At least 20 original texts published in the language
  • 36 Has a public corpus of at least 10,000 words
  • 37 Has a public corpus of at least 100,000 words
• UNIQUENESS
  • 27 Has an extreme grammar or vocabulary
  • 30 The language has a unique script
  • 32 Sets itself challenging artistic goals, and achieves them
  • 38 Demonstates plausible historical derivation from a natural language
  • 39 Demonstrates plausible historical derivation from another conlang
• PUBLICATION
  • 6 A book has been published about the language
  • 7 A book with an ISBN has been published about the language
  • 8 A book has been published in the language
  • 9 A book with an ISBN has been published in the language
  • 25 Has been used in professionally published media
• REPUTATION
  • 12 Among 100 most popular conlangs
  • 13 Among 200 most popular conlangs
  • 14 Been mistaken for a real language
  • 15 Caused controversy
  • 17 Inspired a notable conlang
  • 19 Has caused active discussion
  • 20 At least N independent discussions of the language
  • 26 Actively supported/developed
  • 40 Contributes to the notability of another conlang
• OTHER
  • 10 Was created before 1900
  • 11 Was created before 1950
  • 16 Has an ISO code
  • 18 Has a Wikipedia
  • 22 Was created by J.R.R. Tolkien
  • 23 Was created by Marc Okrand
  • 31 Is associated with a well developed conculture
• NOT MENTIONED IN THE QUESTIONNAIRE
  • Worked by m people for n years.
  • Differential criteria due to purpose
  • Has established notability outside of the conlanger community
  • Has this language influenced others?
  • What is the consensus of the conlang community about this particular language?
  • Does this language play a role in a broader work of fiction, like movies, books of fiction, games or Internet-based projects (in other words, not only in books with an ISBN number about or in the language itself)?
  • Has the language received any particular attention, for example in the media or the press?
  • Can the language be considered the exponent of a particular genre?

Some thoughts

Number of speakers

Two things. First of all, I think we should abolish all options higher than 100. As far as I know, only Esperanto, Interlingua, Ido and Klingon can safely be estimated on >100. I don't think even Glosa, Novial, Lingua Franca Nova, Volapük, Quenya or Toki Pona have more than 100 speakers. Putting the barrier higher than that would be nonsense. If you ask me, 20-30 speakers is more than enough to keep all kinds of unwanted IALs out.

Secondly, instead of saying "Has N speakers" we would perhaps better say "Has or had N speakers".

Apart from that, it is hopelessly difficult to find info about the actual number of speakers of a language. The proponents are undoubtedly best informed, but needless to say that they are not the most reliable source of information. Besides, how proficient must a person be to qualify as a speaker? And more importantly, how do we measure that?

All in all, the minor/major distinction could work here. Something like Minor = 10 and Major = 50 or 100.

Hear, hear! But of course, it might be decided that Glosa, Novial, Lingua Franca Nova, Volapük, Quenya and Toki Pona are unfit for wikipedia, them being silly wastes of nonsense drivel or whatnot (just copying from discussions and comments in various earlier VfDs here), meaning that this number must be set to a million or so so that only Esperanto survives. Btw, I think Esperanto is the only of them to have native (though bilingual speakers). There's currently a research-project about them at my local university. --Kaleissin 11:07:09, 2005-08-30 (UTC)

You are undoubtedly aware of an attempt at raising a boy bilingually in English and Klingon. I don't know about other languages, although I wouldn't be surprised if Interlingua and Ido had one or two native speakers as well. --IJzeren Jan 11:48, 30 August 2005 (UTC)

Strange BTW that Glosa has no article here, while it is featured in Belarusian, German, Esperanto, French, Interlingua, Norwegian and Russian... --IJzeren Jan 11:54, 30 August 2005 (UTC)

I agree on the rephrasing "has or had". And I agree that setting the bar higher than 100 speakers is silly; but since not everyone feels that way, I don't think we should omit such options from the poll/vote/whatever. That way we would open ourselves up to the accusation that this whole enterprise is biased toward conlangs and that we've deliberately announced this in a way to get only conlang-inclined people involved in the discussion and vote/poll/whatever. --Jim Henry | Talk 13:06, 30 August 2005 (UTC)

Well, not if we tell them that only four of five conlangs have more than 100 speakers. There's nothing wrong with adding some explanatory comment to each section (something similar exists already on the vo... erm, poll page). Putting the barrier higher than that is downright silly (but on the other hand, 100 or 1000 won't make a difference, since both Klingon and Ido are notable enough for other reasons). --IJzeren Jan 13:16, 30 August 2005 (UTC)

Except that by our reasoning, only those 4 or 5 conlangs deserve their own articles. At least, only those 4 or 5 on the grounds of the number of speakers. Speaking for myself, if the barrier's only 100, then it's a poor measure of notability, and I'm opposed to its use. Come to think of it, I think we should address the point Jim hinted at: is our method of determining consensus adequately proof of accusations that it slanted the results either towards inclusion or exclusion of conlangs? Jim and I both objected to the original poll on the grounds that its vote-counting method had a structural bias in it, for instance. The Literate Engineer 00:15, 5 September 2005 (UTC)

Alright, I have no objections against including a higher barrier (300, perhaps even 1000) as an option in the poll. I just wanted to make clear that practically, any barrier higher than 100 is pointless. I'll modify my proposal accordingly.

As for your second point: I agree with the point Jim brought in. Before we start the vote, it should be clear how the votes will be counted, and whose votes will be counted. I'm ready to discuss that before we start anything. But I also think there should be a moment when we stop talking about procedures; I have such a feeling that if we carry on much longer like this, everybody will lose interest and once again we'll end up with nothing at all. That would be a pity, not only because I genuinely believe it would be good to have some objective guidelines, but also for the word and time many of us have invested in this project. --IJzeren Jan 08:00, 5 September 2005 (UTC)

I think it's an apples and oranges thing again. There are conlangs that are meant to be used for communication. Some of them are actually used so. I agree that number of speakers is relevant for this subtype. There are conlangs that are not meant for communication. If they are sufficiently complete though, they can be used for communication and some of them are (eg. Quenya and Sindarin. Still, number of speakers is really irrelevant. The third type aren't languages at all, just sketches.... but sketches part of a greater work, like Marain is part of the Culture, ergo they get in on that criteria, merged with the main article or not. Speakers impossible ergo speakers irrelevant. --Kaleissin 08:12:35, 2005-09-05 (UTC)

Well, yes, of course. But don't forget that we all criteria are INCLUSIVE only. The opposite (that for example little or no speakers would diminish a language's notability) is not necessarily true. And, while I agree that the success/notability/whathaveyou of an artlang can not be measured by its number of speakers, you can hardly call it "notable because it has no speakers but it wasn't meant for that anyway". For artlangs, there are other criteria (but in my opinion, quality and/or completeness alone to not per so make a language notable enough for inclusion).

Sure, it's comparing apples and oranges. But that's our reality: there are numerous languages that have elements of auxlangs, artlang and/or engelangs at the same time; if classifying a language is a problem even for an expert, how do you expect the average wikipedian who has heard only of Esperanto and perhaps Klingon will handle it? Let's face it: most people here don't care for what purpose a language was made, they just want to know if an article is warranted or not. The purpose of my proposal is not only to be accurate or complete, but also to be applied easily. Now there's definitely something that could be said in favour of creating different sets of criteria for different conlang categories, but frankly, I don't think that will be necessary, since all categories are covered in the questionnaire anyway. --IJzeren Jan 09:10, 5 September 2005 (UTC)

I was writing for the benefit of the The Literate Engineer, not you. I don't know what knowledge the former has about conlangs and conlanging, and I find preaching to be easy. It might be "of course" to you but it isn't so to a lot of people. --Kaleissin 11:10:19, 2005-09-05 (UTC) (P.S. I'm starting to think definition-list markup for threads is a bad idea.)

Sorry, my fault! I've corrected it. --IJzeren Jan 11:24, 5 September 2005 (UTC)

I think it's important to treat the number-of-speakers as one of many potential reasons for notability. Thus, a bar of 100 does not exclude languages with fewer than 100 speakers, it just refuses languages with fewer than 100 speakers inclusion based solely on that criterion. WP:MUSIC uses the same or style criteria. Therefore, just because only a handful of conlangs would meet this criterion does not necessarily mean the barrier is too high. (IMO, 20-30 speakers wouldn't alone imply notability for me. In comparison, we certainly wouldn't accept a neologism known among 20-30 high school students.) Brighterorange 18:16, 1 September 2005 (UTC)

It takes less than a day to learn a neologism so that it actually sticks. It takes a child 15 years on average to truly master his/her native tongue. I don't have the number for languages learnt as an adult but often, adults seem not able to learn more languages at all. Apples and oranges, mon! --Kaleissin 20:13:43, 2005-09-01 (UTC)

Completeness

Personally, I don't think completeness makes a language notable per se, even if it has 10 million words! A language with that many words will be probably be notable for other reasons too (esteem, discussions etc.), so I don't think we need to oversubtilise this. The nine (!) current criteria make it needlessly complicated. I believe completeness is achieved by three things (no OR ... OR but AND ... AND):

• Has a complete grammar (how do we measure this?)
• Has a sufficiently large vocabulary (>= N words)
• Has a public corpus of at least N words; distinguishing between original texts and translations is not necessary IMO, and not always easy to accomplish anyway.

If it can be used for everyday communication, it must be complete, so the languages with more than one actual speaker already fulfills this. This criterium is more relevant for artlangs I think, to keep the sketches and mentions (like Marain) and Babel-text only, out. But of course, having a page about a language with links to a complete, free, well-written reference grammar might not be in the best interest of Wikipedia. --Kaleissin 11:14:00, 2005-08-30 (UTC)

For "has a complete gramar" I suggested the Lingua Questionnaire. The public corpus, I think, is as much an indication of how much the language is used as it is of how complete it is. And I agree that completeness per se does not make a language notable (or, more importantly, verifiable); but since not everyone agrees with us (Sai, for instance) we should not exclude such options from the poll at the outset. There's little or no danger of them getting a consensus in favor, especially now that a large number of non-conlangers have gotten involved in the discussion and will undoubtedly be voting in the poll (oops, "expressing their opinions"). --Jim Henry | Talk 13:13, 30 August 2005 (UTC)

Uniqueness

Here we have no less than five different options, to which I would like to add the option "Can the language be considered the exponent of a particular genre?" Well, in my own opinion all this says something about the interestingness of a language, but not about its notability. If a language is really unique, it will automatically generate attention and discussion.

I think we should vote about two things: can "uniqueness" be considered a major or a minor criterion, or none at all, and if so, which of these factors do we accept as factors that make a language unique?

Publication

Can't we just simplify this a little? This whole discussion about ISBN numbers and sales figures are far beyond the reach of the average wikipedian! If you ask me, the following would do:

• A book has been published about the language
• A book has been published in the language
• It has been used in professionally published media.

The next question would be: is this a major or a minor criterion for notability in general?

How do you define language here? Especially for media other than books, is a mention or a wordlist or some squiggles resembling writing, language? --Kaleissin 11:38:27, 2005-08-30 (UTC)

Well, in my personal opinion the lack of completeness of the language can be compensated by the importance of the medium. If a book of little importance features a conlang that is nothing but the name of the language and two words, then a wikipedia article is not warranted. But if the medium is incredibly super-important (like LOTR, or Harry Potter, or Star Trek), then things might turn out a little different. --IJzeren Jan 11:46, 30 August 2005 (UTC)

I suspect the person who originally suggested the "with an ISBN" restriction wanted to rule out vanity press or self-published items. But most recent self-published items have ISBNs -- they're easy to get -- and older professionally published books did not have ISBNs. Also, note that several people voting in the original poll wanted to make a distinction between books written by the conlang creator and books written by others. This relates to Robert West's criteria for verifiability and avoiding original research: can we reference some independent discussion of the language by people other than the language's creators? If so, maybe it doesn't matter what form such independent discussion takes, whether book, magazine article, or web page. --Jim Henry | Talk 15:18, 30 August 2005 (UTC)

As for the latter, I think that's sufficiently covered in the REPUTATION section. Excluding a book written by the author of the language is IMO nonsense; who else would have written the book? That has nothing to do with the no original research criterion for inclusion in the wikipedia (after all, we are talking about notability here). But my main point remains: let's keep it as simple as possible. If we refine every single criterion with numerous exceptions, ifs, buts and unlesses, we get a "notabilizer" (TM) that becomes very difficult to use. --IJzeren Jan 16:05, 30 August 2005 (UTC)

Reputation

• Has the language received any particular attention, for example in the media or the press?

Under this header I have grouped nine questions from the poll + four questions that were not listed. I think we have the following factors here:

• Popularity (among the top-100 and/or top-200)
• Controversy (including: been mistaken for a real language)
• Discussion (at least N independent discussions of the language; also: established notability outside of the conlanger community (how do we measure this?); and also: has the language received any particular attention, for example in the media or the press?)
• Influence on other conlangs (notable or not?)

A currently very frequently used tool for establishing notability of anything is Google. I think we should add this one too, simply because it's the most frequently used way.

When considering Google hits, we should distinguish between mentions and discussions, and set a higher bar for the former than the latter. Maybe at least 50 independent mentions or 3 independent discussions, or some such. --Jim Henry | Talk 15:22, 30 August 2005 (UTC)

Yes. And when counting ghits, I suggest we perform a simple google for the name of the language; when that name also has one or more different meanings, we should google for "CONLANG NAME" + "AUTHOR'S SURNAME"; if the search result is dominated by wikipedia and clones, we should google for "CONLANG NAME" (+ "AUTHOR'S SURNAME") "-wikipedia". --IJzeren Jan 15:55, 30 August 2005 (UTC)

At least N independent discussions

Note that several of the people voting Oppose on this criterion (and some of those supporting it) in Wikipedia:Conlangs/Straw poll#At least N independent discussions of the language object to the inclusion of "mailing list postings" in the definition. We probably need to split this up or redefine it. Maybe:

There exist at least N extended discussions of the language (books, magazine articles, web pages, mailing list postings...) by authors other than the creator(s) of the language. If you support, specify a value for N and the average value will be used if this criteria passes.

and

There exist at least N extended discussions of the language in edited sources (books, magazine articles, online peer-reviewed journals, moderated newsgroups, etc.) by authors other than the creator(s) of the language. If you support, specify a value for N and the average value will be used if this criteria passes.

--Jim Henry | Talk 19:58, 6 September 2005 (UTC)

Well, you certainly have a point. But do you really think it's worth the effort to split the question the way you propose? If you ask me, that's making things more complicated again. People always have the possibility to add a comment to their vote, and things like this will of course be taking into account. Well, okay, let me think this over, I'll try if I can come up with something... --IJzeren Jan 21:24, 6 September 2005 (UTC)

Personally, I like the first version. I think the "authors other than the conlang creator" is sufficient guard against gaming the system to get non-notable languages included. But I wanted to come up with an alternative that would satisfy the people who objected to it. I've posted messages to the talk pages of those who objected asking them to comment here. If it does satisfy their objections, then the modified version might get unanimous support (based on the votes so far on the straw poll page) -- which is probably worth some tinkering.

I'm a little surprised that no one objected to the "web pages" clause; it seems to me that, other things equal, a well thought-out mailing list posting by someone other than the conlang creator and a well thought-out review on the same person's personal website are of equal value in establishing verifiability, and both are slightly less valuable than reviews of similar quality in edited sources whether printed or online. --Jim Henry | Talk 22:01, 6 September 2005 (UTC)

Yeah, I guess you're right. Okay then, I'll adopt you proposal. --IJzeren Jan 07:43, 7 September 2005 (UTC)

Agreed, though I still have reservations that the first definition might allow crapflooding by enthusiasts, regardless of the merits or otherwise of the language. What I mean is, how do you judge whether the discussion has merit or not. Does a long discussion with a few very good threads and a lot of noise around it count for as much as a slightly more focused but on the whole shorter discussion? --Kgaughan 21:48, 7 September 2005 (UTC)

Other criteria

As for the remaining criteria, I think these can mostly stay as they are. The question: "Was created by J.R.R. Tolkien" and "... by Marc Okrand" should IMO be put as "Was created by someone in particular". This is the famous notability-by-proxy discussion. That way we have:

• Was created before a certain year (1900/1950)
• Has an ISO code
• Has a Wikipedia
• Was created by someone in particular
• Is associated with a well developed conculture

Two other suggested criteria, "Worked by m people for n years" and "Differential criteria due to purpose" are IMO too hard to implement.

The "someone in particular" should be better defined: for instance, "was created by someone who has a Wikipedia article about them, and said article is getting too long to include discussion of all of their works in the main biographical article, so it makes sense to spin off their minor conlangs as separate articles". Can you suggest a way to make this more concise? --Jim Henry | Talk 13:20, 30 August 2005 (UTC)

Not really, no. I'll happily adopt your formulation... --IJzeren Jan 13:22, 30 August 2005 (UTC)

Maybe if it were phrased something like, "Is a {insert better word than "meaningful" here} part of a notable person's body of work" it would better? Regardless of how it's phrased, I'm opposed to it on the GOOPTI principle. Same with the well-developed conculture one. Way I see it, if a conculture's well-developed, the conculture deserves an article. The conlang, though, is just part of that conculture, and 1)needs to be a part of that conculture's article so that the conculture article's less stubby, and 2)can't necessarily "stand on its own" independent of the rest of the conculture. The Literate Engineer 02:27, 1 September 2005 (UTC)

And now my proposal

Based on the points mentioned above, I have come to an alternative questionnaire. Since the vote has been stalled and everything is under discussion again, the votes will be discarded anyway, and we might as well revise the whole procedure. Here is my proposal, based on the current questionnaire, which I honestly believe is a little more organised, a little easier to fill out and much easier to apply.

Number of speakers (MAJOR/MINOR/OPPOSE)

What is the minimum amount of speakers required for fulfilling this option (pick one)

• Has/had 10 speakers
• Has/had 25 speakers
• Has/had 100 speakers
• Has/had 250 speakers
• Has/had 1000 speakers

Completeness (MAJOR/MINOR/OPPOSE)

Which of the following options must be fulfilled?

• Has a complete grammar (yes/no)
• Has a vocabulary of at least N words (N = 0 / 500 / 1,000 / 2,500 / 5,000 / 10,000 / 50,000)
• Has a public corpus of at least N words (N = 0 / 1,000 / 5,000 / 10,000 / 50,000 / 100,000)

Uniqueness (MAJOR/MINOR/OPPOSE)

Which of the following can make a language unique?

• Has an extreme grammar or vocabulary (yes/no)
• Has a unique script (yes/no)
• Sets itself challenging artistic goals, and achieves them (yes/no)
• Demonstrates plausible historical derivation from a natural language (yes/no)
• Demonstrates plausible historical derivation from another conlang (yes/no)
• Is the exponent of a particular genre (yes/no)

Publication (MAJOR/MINOR/OPPOSE)

Which of the following count as publication?

• A book has been published about the language (yes/no)
• A book has been published in the language (yes/no)
• It has been used in professionally published media (yes/no)

(Note: I think the three options are redundant here)

Reputation (MAJOR/MINOR/OPPOSE)

Which of the following can make this condition true?

• The language is in the Top-100/200 of Langmaker.com (100/200/no)
• Caused controversy (including: been mistaken for a real language) (yes/no)
• Produces at least N Googlehits
• At least N independent discussions of the language
• Received attention in the popular/scientific media (yes/no)
• Inspired other conlangs (only notable conlangs/any conlangs/not at all)
• Has established notability inside of the conlanger community (yes/no)
• Has established notability outside of the conlanger community (yes/no)

Year of creation (MAJOR/MINOR/OPPOSE)

• Was created before 1900/1950

Notability by proxy (MAJOR/MINOR/OPPOSE)

Which of the following can make this condition true?

• The language was created by a notable person (yes/no)
• The language is associated with a notable conculture (yes/no)

Next question

How many of the criteria need to be fulfilled for inclusion?

• 1 major or 2 minor
• 1 major or 3 minor
• 1 major or 4 minor
• otherwise, namely ...

Opinions? --IJzeren Jan 09:45, 30 August 2005 (UTC)

I'm almost ready to say 2 majors and 4-6 minors. As noted, only three or four conlangs have a number of speakers in the thousands, and really only Esperanto has native speakers. And I have a hard time thinking of anything other than thousands of speakers as being enough to establish notability all on its own. Although, the "extremity" thing in the uniqueness category, like for Toki Pona, almost has me convinced that it alone is enough. And of course, the circular reasoning of "Has established notability outside of the conlang community". That one, while I like having a requirement that it be notable outside a small, specialized field (which is why I proposed it), it's pretty useless to say that something's notable because it's... notable. The Literate Engineer 02:21, 1 September 2005 (UTC)

Well, I just added 250 and 1000 to the # of speakers criterion. If you believe that "anything other than thousands of speakers" should be enough to establish notability all on its own, you will have the possibility of voting accordingly. --IJzeren Jan 08:05, 5 September 2005 (UTC)

Voting procedure

Given the discussion about the way votes are counted, and the (IMO not entirely incorrect) suspiction of some that the current way of counting the votes is a little biased towards inclusion, I'd like to propose the following system, which basically is a variant on what Jim proposed (with the exception of his third point "if Minor comes third, Minor votes are redistributed ...", which IMO is indeed a little biased against minor criteria).

For each of the criteria, everybody has three voting options:

• SUPPORT (Major): Indicates notability, and merits inclusion.
• SUPPORT (Minor): Contributes to notability, but more than 2/3/4 of the minor criteria must be fulfilled to warrant inclusion.
• OPPOSE: Should not be used for determining conlang's notability; leads to false results, abusable, et multiple cetera.

First, Support and Oppose opinions will be counted; whichever has consensus will be considered community opinion. Then, if Support was chosen, the Oppose votes will be added to the MINOR votes.

In this way, I hope to address the issues raised by Jim Henry and The Literate Engineer on Wikipedia talk:Conlangs/Straw poll#Vote problems?. The reasoning: once a majority has spoken out for including something as a criterion, the Oppose votes should not be discarded after that; since one who opposes the idea would in all likeliness rather see it as a minor than as a major criterion, those votes are added there. On the other hand, it cannot be safely assumed that someone who votes Support (minor) would rather have it as a Major criterion than not at all.

For example, if the result of a criterion is 5 MAJOR, 2 MINOR, 5 OPPOSE, we conclude that a majority of 7 to 5 favours inclusion of the criterion, but that another majority of 7 to 5 does not want it as a MAJOR criterion. Hence, the conclusion is MINOR, even though MINOR in itself got less votes than the other two options.

A similar system could IMO also be applied to the subcriteria that have more options than just y/n (like, number of speakers, criteria with a variable N).

In addition, I agree with the current "rule", that "only registered user's opinions will be recognised; opinions of non-registered users will be ignored, even from IPs with many edits, because there is no way of knowing how many people an IP address represents."

Comments? --IJzeren Jan 08:44, 5 September 2005 (UTC)

That sounds good; it addresses some problems with my counter-proposal that Almafeta and others pointed out.

Re: only registered users; should this be "only registered users as of a certain date" ? --Jim Henry | Talk 20:13, 6 September 2005 (UTC)

Normally, that would be "users registered before the poll was started". That should be sufficient to prevent people from brushing up their sockpuppets! ;) --IJzeren Jan 21:18, 6 September 2005 (UTC)

Also: on VfD and similar votes/polls/whatever, generally a 2/3 majority is required to show clear consensus; a slight majority is not sufficient. Maybe we should do the same here. --Jim Henry | Talk 20:16, 6 September 2005 (UTC)

The problem is only that practically "no consensus" would equal "oppose" in that case. In other words, no criterion. I'd rather say that in case of a close vote, we simply mention it in our conclusion. --IJzeren Jan 21:18, 6 September 2005 (UTC)

Actually, now that I'm thinking of it, I'm having some second thoughts about this whole consensus thing, too. If you look at all those numerous conlang-related VFDs we have had in the past, you can see several recurring arguments used for and against. But the conclusion is always the same: consensus, i.e. a majority of at least 2/3, is required to delete an article. It's not the other way round, that a 2/3 majority is required for having it. In this poll, we are discussing inclusive criteria, primarily meant to be used in VFD discussions. If you follow the same logic as used in the VFDs, the oppose votes would need to form a 2/3 majority for a criterion to be not included.

If we would have done the reverse, i.e. asked for criteria to proof that the conlang is non-notable, things would have been complete different: a qualified majority would have been required for not using a criterion. In that case, we could have said: Language X is non-notable, because it has less than N speakers, it produces less than N Google hits, it has been mentioned in less than N independent discussions, it does not have an ISO code, it does not play a role in a significant work of fiction, etc.

--IJzeren Jan 14:18, 7 September 2005 (UTC)

Please consider dropping the voting, it's probably a waste of time

I am a little surprised that you spend so much energy discussing technicalities of voting. I'm afraid that you might be wasting your time. Let me bring up some interesting quotes from the Centralized discussion guidelines (since this discussion is linked from the Centralized discussions box on VfD.

• The point of these discussions is to establish consensus about a recurring theme that comes up on VFD or similar pages. (emphasis added)
• The point is to establish a consensual guideline through discussion, not to establish a rule through voting.
• Occasionally, however, it will be necessary to hold a vote to establish firmly where consensus lies, or to find out which of several options is preferred, e.g. for a template layout.
• If there is disagreement on what a standard, naming convention, or layout should be, then an approval vote should be held between the different options.

While Kim Bruning was obviously wrong in saying "voting" is never allowed, my interpretation of the above is that it mainly should be used to determine "formatting" and yes/no-choices (e.g. proposals for CSD criteria). Guidelines on inclusion of content should be primarily be developed through discussion, and if it isn't possible to reach consensus, there is no possibility to establish a guideline by staging a vote and "winning" by a narrow majority. What will happen in such a case is that things will be resolved on a case-by-case basis on VfD anyway (and the vote will have been a waste of everyone's time).

My point is that there simply is no consensus for inclusion of constructed languages based on merit ("Completeness", "Uniqueness" and, to some extent, "Reputation" above). This was raised in the discussion before the vote, e.g. here, by DenisMoskowitz and others, but I can't see that any real effort was made to discuss this through and reach a consensus conclusion. Instead, a (premature) vote was held. I then raised much the same point here and got some response from a few people - but now you seem to be preparing to hold another vote based on the same principles, albeit better structured.

Looking at the aborted "vote"/straw poll (which people still seem to be "voting" in), there seems to be significant opposition to almost every merit-based criteria. Taking into account that this page still seem to have attracted mainly people interested in conlangs, the opinions outside this group is likely to be even more negative towards such criteria. This theory is further strengthened by the recent VfD debate on the constructed language "Lara". While this example might not have fulfilled all your proposed criteria (I'm aware that IJzeren Jan voted to delete it), it might be relevant to point out that the rationales quoted by voters included general statements such as:

• "non-notable language not in common use"
• "original research"
• "Invented languages should be included if they are notable for academic (e.g., interesting and discussed example) or cultural (e.g., Klingon, etc.) reasons, even if those reasons are minor. No claim to notability exists here."
• "for a conlang to be noteworthy, it needs a LOT of notoriety outside its circle, in the mass media, etc. Klingon, Esperanto, and Tolkien's works fulfill those criteria, almost nothing else does."

I'm sparing you the outright mockery of conlangs from other voters. Let's just say that there is currently a massive majority for deletion, and that this majority seems to share the view that the notability of a constructed language has nothing to do with its completeness or linguistic inventiveness. Either you understand the rationales behind what seems to be the majority definition of notability, and try to reach criteria that comply with this view (preferrably by discussion rather than by voting) or the destiny of individual conlangs will still be decided on a one-by-one basis on VfD, regardless of what happens here.

Don't get me wrong. I have the greatest respect for the conlang hobby. I'm interested in linguistics myself. I haven't, and I will not, make fun of people who invent languages as a hobby. I just think most of them aren't more notable - i.e. more encyclopedic - than unpublished novels or demo recordings. Furthermore, I think this process went the wrong way rather early, and I just want everyone spending their time around here as constructively as possible. I don't think holding another vote based on the proposal above will be to do that.

It seems to me that many of the people intrested in constructed languages that have been active in this discussion have tried to devise a rudimentary peer review system of conlangs. I can understand this wish - and even the need - for this, but I am thorougly convinced that Wikipedia isn't the place for such a process, for various reasons that ultimately boil down to the fundamental principles of Wikipedia. (My advice would be, for those of you who are interested in this, to try to establish such a process in some other forum - e.g. the conlang wiki. With time, it might even develop into something that can influence notability judgements in Wikipedia.)

In order to be constructive, I'd like to try entirely different take on the issue of guidelines. My proposed guideline is short and refers to established principles for other topics:

Constructed languages can be regarded as works of science or works of art. Their notability should be judged by the standards generally applied to articles on such works; they should be the subject of significant peer-reviewed discussion and/or significant media attention. If a constructed language has (or has had) a significant number of active users, it can also be notable as a language in its own right. The standard requisites of Wikipedia:Verifiability apply to any such claim.

I think that this would be much simpler and far more useful than a set of rules entirely unique to conlangs. If you don't agree, I'd politely recommend you to discuss why, rather than to ignore this and simply hold another vote. / Alarm 21:52, 6 September 2005 (UTC)

The vote had been announced for August 28th or later, if the discussion was still ongoing. Since the discussion had pretty much died down by the 28th, with no significant new activity in a week or more, I suppose Almafeta thought it perfectly suitable to start the voting on the previously announced date. Then, of course, Mr. Bruning deleted the vote page and the resuling VfU controversy got a lot more people involved -- a good result, though I wish we had announced this more widely and gotten more people involved earlier.

I agree that completeness and merit alone do not make a conlang verifiable or notable. I expect that, with a larger cross-section of Wikipedians now involved, the criteria some people proposed which rely solely on completeness or artistic merit will have no chance of getting a consensus in their favor -- indeed they didn't have much chance even when the discussion was dominated by conlangers in the first week or two.

Your counter-proposal seems fairly good, but still a little too vague. I and a number of others think we should have some more quantifiable criteria to point at when people who dismiss the notability of conlanging in general nominate articles on notable conlangs for deletion. But maybe you're right and we have no chance of getting the proposed policy accepted.

I've tried to steer discussion about a peer-review process for conlangs over to the Conlang Wikicity, without much success so far. --Jim Henry | Talk 00:01, 7 September 2005 (UTC)

I'm with Jim here. Note the following things:

• Of course, I would also prefer to reach consensus by discussing things and not by voting. Things however have made it pretty clear that in the case of conlangs that's pretty much undoable.
• Quite a few things have been discussed thoroughly between the end of July and the beginning of September. Now the discussion has pretty much died out since a week or so.
• The reason of this vote/poll was in the first place to establish some quantifiable, verifiable ways to measure notability. Okay, it's often mentioned that notability is not a criterion, but in the meantime people use it constantly in VfDs. While your counterproposal surely makes sense, it doesn't solve the basic problem that significance remains a purely subjective matter.
• You will NEVER reach full consensus about including other conlangs than Esperanto. Even some of the "heavier" languages (Toki Pona, Verdurian) survived their VfDs rather for a lack of consensus than for a clear consensus for keeping them. Sadly, there are a lot of people around who think conlanging is merely making up gibberish à la "asdfdksfalj" and voilà, you have a conlang! Consequently, these people will ALWAYS vote for deletion, no matter the counterarguments.
• Speaking about counterarguments, how real are debates here anyway? On VfDs I constantly see people using arguments that simply aren't true; but even when someone proves them wrong, they rarely change their vote. And it stands.
• BTW, I notice that those who ardently seek to remove conlang-related stuff from Wikipedia, haven't spoken up at all yet. I personally find that a pity.
• I repeat: the list is NOT a set of proposals I and others like to be approved. It's merely an inventory of possible criteria, brought in during the discussion. To me personally, a thing like uniqueness (especially the famous "unique script" thing) has little or nothing to do with significance.
• The conclusion of all this is IMO: if we don't vote now, we will probably never vote. Unless someone comes up with significant new input for the debate, I'd like to proceed.

--IJzeren Jan 08:17, 7 September 2005 (UTC)

I still wonder what you think you are going to accomplish with this voting procedure. You are probably right that discussion had died down when the first vote was called, but I don't think that is always an argument for calling a vote. Hard work ought to be put into trying to reach a conclusion through discussion.

Note that a rough consensus, i.e. a large majority, is necessary for any guideline to be valid. "Winning" a vote with a small majority does not mean having reached a "consensus". One of few good reasons to call a vote is if there is significant and ardent opposition to a widely agreed proposal from what most people involved in the discussion believe to be a clear minority. In such a case a vote can be held to prove that a rough consensus exists. If, on the other hand, there is such a wide disagreement that a mainstream proposal can't be reached, it's quite possible to conclude in advance that the outcome of a vote won't be a rough consensus. And since no consensus about a guideline means that there is no guideline, there's no point in taking it to that point. What are you going to do with the result? You simply won't be able to use a 5-to-4 majority in a vote to convince anyone on VfD that an article ought to be kept or deleted - and even less so, a number of 5-to-4 majorities for different things being "minor" and "major" criteria, and others not.

A good guideline has to have wide recognition and respect. It needs to be easy to understand and convincingly phrased. Note that if I'm not mistaken, one of the most used guidelines on the inclusion/exclusion of content, WP:MUSIC, was developed without any voting at all.

Holding a representative vote on Wikipedia is a major operation. It should be announced in various relevant forums. Since it ideally involves a large number of people, it is rather energy-consuming. In my opinion, this creates an obligation not to stage a vote/poll unless it is absolutely necessary - and to avoid making it more complicated than necessary. In my opinion, the proposed new poll is way too complicated - in fact, more complicated than any other Wikipedia poll I've came across. It includes all imaginable criteria, including a lot of proposals that almost noone seems to seriously support any more. Having a "unique script" option to vote on will just strengthen the misconception that there is actually a bunch of conlangers actively supporting such a proposal - because the way things are normally done, it wouldn't have been taken to a vote otherwise.

I don't think your arguments for holding another vote are convincing. But if you decide to call a vote anyway, I'd recommend you to conduct it in a different manner. Rather than voting on a massive number of individual criteria, ask the proponents of different points of view to create comprehensive proposals that summarize their positions. Try to compromise in order to reduce their number to two or three, or at least as few as possible. Discuss them thoroughly, so you agree that their phrasing is unambiguous. Then ask voters to choose which one they support. This will make it all easier to understand. Asking all Wikipedians to choose between a number of different voting alternatives for every concievable criteria will just scare most non-conlangers away. It will also make the result of the voting impossible to interpret. / Alarm 19:44, 7 September 2005 (UTC)

I agree with Alarm. This is not the way to stage a vote. It's too complicated even for those who have a specific subject. I would prefer if it could be solved by true consensus rather than a vote. / Peter ^Isotalo 20:54, 7 September 2005 (UTC)

Look folks, don't get me wrong: if there is a way of building a true consensus about the inclusion of conlangs, I'd much prefer that to a vote. But tell me one thing: having read the discussion pages (basically, Wikipedia:Conlangs and all its subpages), can you see a single point of consensus there? Well, perhaps that a language with more than 1000 speakers is therefore notable by itself; but that qualifies only Esperanto and Interlingua, both languages that would qualify for a variety of other reasons anyway. The sad truth is that there are quite some people around here who hold some personal grudge against "conlangers and neologists" for reasons I completely fail to understand (obviously there is something in conlanging that offends certain people), who would happily delete ALL conlang articles if they only had the chance. And on the other side of the spectrum there are people who wouldn't mind including ALL conlangs here, arguing that WP is not paper and that notability is not a factor anyway. Ergo: virtually no criterion that might lead to the in- or exclusion of a conlang (except perhaps >1000 speakers and a bunch of books/movies) will be supported by a true consensus. But keep in mind what the original purpose of this whole discussion is! In conlang VFDs people constantly use the phrase "non-notable"; all we are trying to do here is giving some body to the whole concept of notability.

Now, I agree with you that a vote is not the best way of building consensus. But keep in mind that this poll is by no means meant to be the end of the discussion. On the contrary: I hope the poll will give some food to a renewed, and more specific, discussion about the successive criteria. The poll itself is only supposed to inventarise, to make it clear how people really think about all successive criteria. It is not meant to be Set In Stone after its completion. --IJzeren Jan 06:28, 8 September 2005 (UTC)

What kind of attempts has anyone made to read through proposals from various people and to summarize them? The current summary used for this poll is wholly inappropriate because of it's extreme level of detail and very absolute statements "X number of speakers", "X number of debates", "X number of books".

As far as I'm concerned, any truly notable conlang (and I'm not just talking about Esperanto-type notability) can demonstrate that notability in a way that should be obvious to non-fans or inherent deletionists like myself (even if I happen to be a general lingonerd). If it isn't obvious, I think people should fight harder to make it notable by using and cultivating and not lower the notability-definition of Wikipedia to something that could include almost anything.

Peter ^Isotalo 14:58, 8 September 2005 (UTC)

IJzeren Jan, I'll have to say it again: If you can't build a true consensus (meaning a significant majority) around a guideline, you won't have a valid guideline. As Wikipedia:Centralized discussion puts it: The point is to establish a consensual guideline through discussion, not to establish a rule through voting. There is no need use a Wikipedia-wide poll to "inventarise" every single imaginable criteria. If you, I, Peter, Jim and more than nine out of ten voters in the Lara deletion debate agree that "a unique script" or a vocabulary of 7,000 words does not in itself contribute to the notability of a conlang, then there is absolutely no need to include those criteria in a vote aiming to find criteria for inclusion with a consensus behind them. It is simply a waste of everyone's time.

I'd be happy to discuss criteria for notability, rather than voting on them, or discussing whether or not to vote. If you don't agree with the proposal in my first comment above, I would like to urge you to specify what is wrong with it and try to formulate a viable alternative - a simple and concise set of rules that you would like to use instead - rather than staging vote after vote. If such a guideline proposal is well-phrased and convincing it might actually help persuade sceptics that some conlangs are indeed notable. The proposed vote will do the exact opposite. / Alarm 19:46, 8 September 2005 (UTC)

Alarm is right that there is not much point in working on this vote unless there is a reasonable chance of coming up with a consensus. For that reason, I would suggest again that: (1) we require 2/3 majority support for a criterion to get in into the final policy, and (2) that we drop from the revised straw poll any criteria which, based on the results of the first poll, obviously have no consensus in favor. It would probably be a good idea to discuss the proposed criteria for a while longer now that more people have gotten involved, and make sure, before setting a vote date, that we've announced this discussion everywhere that people might reasonably expect such a thing to be announced. ...So, where should we announce this that we haven't already? And, besides the "unique script", which other criteria should we drop from the poll because it's already obvious they have no consensus? --Jim Henry | Talk 22:54, 8 September 2005 (UTC)

---

Alright, if we are going to use the Old straw poll as a source of preliminary information, it might be a good idea to inventarise the results:

Preliminary results of the Old Straw Poll

Note: the vote has been aborted in an early stage, and these results are NOT to be interpreted as community consensus. They are only supposed to give an impression of the direction consensus was going into.

criterion	votes (major/minor/oppose)	conclusion (majority)	conclusion (2/3 majority)
Has 10 speakers	5/4/10	oppose	oppose
Has 30 speakers	6/6/4	minor	minor
Has 100 speakers	10/2/3	major	major
Has 300 speakers	9/1/2	major	major
Has 1000 speakers	9/3/2	major	major
A book has been published about the language	8/4/6	minor	minor
A book with an ISBN has been published about the language	8/4/6	minor	minor
A book has been published in the language	7/4/4	minor	minor
A book with an ISBN has been published in the language	8/1/3	major	major
Was created before 1900	8/5/3	major	minor
Was created before 1950	3/3/6	-/-	oppose
Among 100 most popular conlangs	0/9/7	minor	oppose
Among 200 most popular conlangs	0/5/7	oppose	oppose
Been mistaken for a real language	6/1/6	minor	oppose
Caused controversy	8/2/3	major	major (minor)
Has an ISO code	6/2/3	major	major (minor)
Inspired a notable conlang	7/3/2	major	major (minor)
Has a Wikipedia	7/2/8	minor	oppose
Has caused active discussion	4/7/2	minor	minor
At least N independent discussions of the language (I'll leave N for what it is here)	8/2/2	major	major
Has a complete grammar	2/7/4	minor	minor
Was created by J.R.R. Tolkien	1/6/9	oppose	oppose
Was created by Marc Okrand	0/5/8	oppose	oppose
Has a sufficiently large vocabulary (N omitted for now)	3/6/6	minor	oppose
Has been used in professionally published media	6/5/2	minor	minor
Actively supported/developed	1/7/3	minor	minor
Has an extreme grammar or vocabulary	3/5/5	minor	oppose
The language can be used to define its own grammar	6/2/6	minor	oppose
The language can translate the Babel Text	0/5/6	oppose	oppose
The language has a unique script	0/3/9	oppose	oppose
Is associated with a well developed conculture	2/5/4	minor	oppose
Sets itself challenging artistic goals, and achieves them	1/5/7	oppose	oppose
Has been used in at least 5 translation relays	0/5/5	-/-	oppose
At least 10 original texts are published in the language	2/3/5	-/-	oppose
At least 20 original texts published in the language	3/4/3	minor	minor
Has a public corpus of at least 10,000 words	3/7/2	minor	minor
Has a public corpus of at least 100,000 words	10/3/2	major	major
Demonstates plausible historical derivation from a natural language	0/5/7	oppose	oppose
Demonstrates plausible historical derivation from another conlang	1/2/7	oppose	oppose
Contributes to the notability of another conlang	2/4/7	oppose	oppose

Based on the above, we can come to the following conclusions:

• As for the # of speakers, we agree that anything above 100 speakers counts as a MAJOR criterion, and anything between 30 and 100 as a MINOR criterion.
• Furthermore, there seems to be consensus about including the following as MAJOR criteria: "caused controversy", "has an ISO code", "inspired a notable conlang", "at least N independent discussions of the language by people other than the author", "a public corpus of at least 100,000 words".
• A book in or about the language counts at least as a MINOR criterion.
• The following criteria have a simple majority against them: "has 10 speakers", "among 200 most popular conlangs", "was created by J.R.R. Tolkien", "was created by Marc Okrand", "can translate the Babel text", "has a unique script", "set itself challenging artistic goals and achieves them", "has been used in at least 5 translation relays", "at least 10 original texts", "demonstrates plausible historical derivation", "contributes to the notability of another conlang". I suggest we omit these from further discussion.
• All the other criteria need further discussion.

If we decide to discuss only those elements that are supported by a 2/3 majority, this is what we are left with:

criterion	votes (major/minor/oppose)	conclusion (majority)	conclusion (2/3 majority)
Has 30 speakers	6/6/4	minor	minor
Has 100 speakers	10/2/3	major	major
Has 300 speakers	9/1/2	major	major
Has 1000 speakers	9/3/2	major	major
A book has been published about the language	8/4/6	minor	minor
A book with an ISBN has been published about the language	8/4/6	minor	minor
A book has been published in the language	7/4/4	minor	minor
A book with an ISBN has been published in the language	8/1/3	major	major
Was created before 1900	8/5/3	major	minor
Caused controversy	8/2/3	major	major (minor)
Has an ISO code	6/2/3	major	major (minor)
Inspired a notable conlang	7/3/2	major	major (minor)
Has caused active discussion	4/7/2	minor	minor
At least N independent discussions of the language (I'll leave N for what it is here)	8/2/2	major	major
Has a complete grammar	2/7/4	minor	minor
Has been used in professionally published media	6/5/2	minor	minor
Actively supported/developed	1/7/3	minor	minor
At least 20 original texts published in the language	3/4/3	minor	minor
Has a public corpus of at least 10,000 words	3/7/2	minor	minor
Has a public corpus of at least 100,000 words	10/3/2	major	major

--IJzeren Jan 09:11, 9 September 2005 (UTC)

Towards consensus instead of polls

Moved to Wikipedia:Conlangs/Towards consensus instead of polls

Well. Before any vote, straw poll or similar, we should first define in a way that leads to consensus which criteria should be used to decide if a constructed language is notable or not. Then a consensus on how to mesure the criteria, and finally a consensus (this last much probably vote driven) on which would be the mark.

Almafeta's original proposal of major and minor criteria has dominated the discusion... it seems we already has a concensus of using a major/minor criteria to judge notability. Also, proposed criteria are positive criteria (if a conlang does not fullfill all of them but one major, then is deemed notable).

Number of speakers of the language seems also a criterium to be important to positively stablish notability. (Not having speakers does not disqualify a language.) How to measure "speaker" has still no concensus. If it should be major, minor or the marks to determinate the major and minor limits do not have concensus yet.

Corpus in the language seems to be important, however I am not sure if we have a concensus yet. Less so how to measure corpus (printed works? ISBN printed works? translations?, etc.); nor the limits of the gauge.

Completeness of the language seems not to be a concentual result. (Probably some think that a good corpus should show enough completeness.)

People talking about the language, particularly people outside the creator(s) of the language in a series of verifiable sources, seems to be a criterium we most agree (a consensus?). We do not have agreed yet on how to measure this, but some elements are helping: books written on the language, discussion on the language in several internet phora, controversy (as stated by the press), etc. could be used.

There seems not to be any concensus about notability by proxy. Either minor Tolkien languages, languages belonging to alternative worlds like Star Treek, etc. If there would be, there should be no problem on how to mesure this.

Peer review... well, nobody seems to have support my idea of per peer review for artlangs, however Taliessin suggestions are pretty much what I had in mind.

So: my proposal

1. We should create discussions on different topics to establish consensus about rather those criteria serve to measure if a language is or is not notable, or not. Numbers should not be involved in these discutions. What exactly to measure, and how exactly to measure, sould not be important in this stage to determinate if we agree or not that the criteria is worth.
   1. Is number of speakers important?
   2. Is completeness important?
   3. Is a corpus important?
   4. Are outside discussions (including controversy) important?
   5. Is notability by proxy a valid point?
   6. Is per peer review a valid criterium?
2. It should be published what seems the concentual result of each discusion: "the following elements are important to judge if a conlang might be notable: ..."
3. Only on those criteriavjudged as important, discusion should continue on each alternative to measure the criteria. For example, if having a corpus is important:
   1. Having books published on the language.
   2. Having books with ISDN published on the language.
   3. Having a verifiably corpus of certain size.
   4. Having professional media shown the langauge.
4. It should be published what seems the concentual result of each discussion: "for measuring the corpus a book with an ISDN should be enough".
5. When the criteria needs a gauge mark, then a poll should be openned to stablish that mark.
6. Marks should be published.

All these published results should be edited in policy format in Wikipedia:Conlangs (with all these discussions moved to propper places. Then, everytime somebody seems that a non-notable conlang is being published, can well go and review the policy, and the policy can be refered to in any VfD (or however they are now called).

This policy will not overrule policies about Original research, Vanity, etc. which can also be invocked in an VfD.

— Carlos Th (talk) 21:24, 9 September 2005 (UTC)

Hmmm. We probably need a step 0: are conlangs a valid topic for w:en? — Carlos Th (talk) 21:27, 9 September 2005 (UTC)