Wiktionary talk:Votes/pl-2011-04/CFI: Removing usage in a well-known work

Various problems

• The proposed version doesn't include the word "or", so makes it sound like we only accept words that are in clearly widespread use and are used in three durably archived works.
• A two-element list is a bit silly, especially when the first element is so short. The sentence could easily be written inline, as “‘Attested’ means verified, either through clearly widespread use, or through usage in permanently recorded media, conveying meaning, in at least three independent instances spanning at least a year.”
• "Clearly widespread use" is not really the same as "verified" or "attested". Previously it was listed there because there were so many list items that it made sense to structure it that way, but if we're only going to accept one kind of actual attestation, then we should say so, and then add that we do not actually require attestation for terms that are clearly in widespread use. (While we're at it, we can make explicit that "attested" has a technical sense, which isn't quite the same as our sense.)
• It seems a bit like "cheating" to make a proposal, find that most people don't agree with it, and that some people have other, related ideas, and then still go ahead with a vote that simply reflects the original proposal. The point of requiring discussion is supposed to be that the vote structure should take that discussion into account!

—Ruakh_TALK 14:52, 26 April 2011 (UTC)

Keep the 'rationale' section IMO. --Mglovesfun (talk) 15:05, 26 April 2011 (UTC)

Regarding the last point you make, I suspect this vote will go down in flames for the reasons you've highlighted. --Mglovesfun (talk) 15:11, 26 April 2011 (UTC)

By "go down in flames", do you mean that it will unequivocally fail, maybe as much as your nomination of Daniel. for desysopping? --Dan Polansky 16:52, 26 April 2011 (UTC)

Responses:

• 1. The proposal does not intend to remove the "or"; it intends to keep the bullet list as an or-clause AKA disjunctive clause. You seem to be skilled with the change markups, so can you add "or" to the proposal the way it makes most sense? Later: I have added the "or" where it seems to belong.
• 2. A two-element list is okay with me; this seems like a cosmetic issue.
• 3. I have not comment on the third point; this seems like a cosmetic issue.
• 4. I do not see that making a vote for a proposal with which some people disagree is cheating. Months have passed since the discussion that I started without anyone pursuing alternative ideas. I found the counterarguments unconvincing anyway. I do not really see why I should feel guilty about creating a vote that proposes what seems to me a reasonable course of action. Anyone is free to pursue alternative ideas, create discussions about them, and create a vote. If this vote fails, that is okay, even though I would like it to succeed. I do not understand why some people are so scared of votes that fail.
  • I do not think that I should create the vote in such a way as to take the discussion into account, as long as I disagree with the arguments presented in the discussion against the voted proposal.
• 5. (new point) Rationale is a good idea IMHO. Now we know that msh210 and you (Ruakh) dislike rationales. I like rationales; I have missed them in votes not created by me. Instead of starting a discussion on the suitability of rationales, you have removed a rationale from a vote that I have created. Now what should I do, start an edit war?

--Dan Polansky 16:41, 26 April 2011 (UTC)

Re: #1: Thanks.

Re: #2: Yes, this is a cosmetic issue; it is a matter of presentation rather than of substance.

Re: #3: This is more than a cosmetic issue. It would not cause me to vote "oppose", but I do not think that is a reason to ignore it.

Re: #4: I am not so bothered by votes that fail, but I would prefer that any given vote be the best vote that it can be. I would rather have a proposal that everyone is 90% O.K. with than a proposal that 70% of us 100% love and 30% of us 100% hate. Voting is a necessary evil, but I think your approach makes it more evil than is necessary. :-P

Re: #5: I do actually like rationales, when they're really rationales: when it's not otherwise clear why the change is being proposed. For example, if my reason for proposing a change is not "I prefer the new version" but rather "the new version is actually how we've already been doing it, for years now", then I will document that in the vote. But in this case your rationale was actually just arguing for the change, and even went so far as to forestall a reader's potential objection (by cheating: you knew, from previous discussion, that readers already had such objections). If your rationale had stopped after the first sentence, I'd at least have been O.K. with it, but as it was, I felt that it went unacceptably far.

—Ruakh_TALK 04:29, 27 April 2011 (UTC)

Re: #3: I admit that I have not noticed the issue when creating the vote. But I have now also realized that the issue that clearly widespread use is not really an attestation is there already with the current text of CFI. I do not see that the proposed change introduces the issue. The proposed change may make the issue easier to see for reducing the number of items in the list to focus on, but, again, it does not introduce it.

Re: #4: Voting is not necessarily an evil; this meme of "voting is evil" is doing much harm to Wiktionary. If you have a better vote in mind, feel free to create it; I will applaud.

Re: #5: Thank you for the explanation. Your edit summary was "remove rationale; that's what the linked-to discussions are for", so it looked like you agree with msh210 that there should be no rationales rather than thinking that my rationale was poor. By rationale I understand an answer to the question "why is the proposed change a good thing and why is it not a bad thing". The vote is not there to confirm an already clear community consensus. A rationale along the lines of "the new version is actually how we've already been doing it, for years now" is an argument by tradition, which is a rather poor form of argument, to be used as last resort in the absence of substantive reasons, IMHO anyway. It did not feel like cheating to deal with possible objections in a rationale; that is what a rationale should do, IMHO. Yes, I am taking advantage of a previous discussion, and showing that I have a response to some objections raised. Again, I do not see why this is cheating. I do not see that I intentionally create in people a false impression of how things are, other than presenting arguments that some will find wrong. People differ in their judgment of which argument is right and wrong, and I have presented no argument of which I would be convinced that it is wrong. The rationale has mere 77 words, including a sentence that points the reader to Beer parlour, so is really short. --Dan Polansky 17:10, 27 April 2011 (UTC)

Re: #3: I didn't say that the proposed change introduces the issue. On the contrary, I think I made clear that the proposed change merely aggravates an existing issue.

Re: #4: Voting is evil insofar as it seeks to substitute "supermajority" for "consensus". This substitution does harm.

Re: #5: That's not what rationale means, at least to me. To me, a "rationale" explains why something was done; it does not attempt to justify it. It is informative, not persuasive. My example was not an argument by tradition, firstly because it wasn't an argument, it was a rationale, and secondly because the underlying premise isn't that we should do something now because we've done it in the past, the underlying premise is that our policy pages should accurately reflect our practices.

—Ruakh_TALK 18:43, 27 April 2011 (UTC)

Re: #3: Point taken. I do not think the change really makes the existing issue very much worse, though; it merely makes it more visible for the reader of the text.

Re: #4: I do not know what "consensus" refers to other than the support of a large majority, much larger than mere 50%. Before I came to Wiktionary, I thought that it meant unanimity, but that is not what it is supposed to mean here. From what I have seen, consensus means in Wiktionary that two to four people support, no one visibly opposes, and most editors do not know that the thing is being discussed. Sometimes, consensus means that there is a bold editor asserting "there is a consensus". I have seen no proof or at least a sketch of a proof that decision making by consensus (whatever it means, I don't know) is morally good, while decision making using supermajoritarian votes is morally bad.

Re: #5: Wiktionary gives two relevant definitions for "rationale" (and one definition irrelevant to this discussion), whose accuracy I do not know, though: (a) "an explanation of the basis or fundamental reasons for something", (b) "a justification or rationalization for something". I do not see much difference between them; both match my understanding of rationale. I do not see argument as contrasting to rationale, but I will stand corrected, especially if you can back your claims by something, even though I do not know what that something could be. If you would be more okay with the term "justification" or "argument", I can use it instead. On the other note, the idea that our policy pages should accurately reflect our practices (so our practices should drive our policy pages) seems to be in contradiction with your practice of aligning our practice with what the policy pages say (so letting policy pages drive our practices). Insisting that a thing can be voted upon only after it became common practice seems inadvisable. I have a sense, though, that I am misrepresenting you in some way, so please correct me where I get you wrong. --Dan Polansky 20:29, 27 April 2011 (UTC)

Re: "If you would be more okay with the term 'justification' or 'argument', I can use it instead": My issue is not so much with possible misuse of the term "rationale", but with using the vote description as a platform to advocate a change. Rationales are good; stump speeches are not, no matter what you call them. (I'm writing as though there's a bright line between them, but in practice, of course, there's a lot of fuzziness. I wouldn't recommend a Draconian approach. But I felt that your "rationale" strayed well beyond the fuzzy area.)

Re: "On the other note, the idea that our policy pages should accurately reflect our practices (so our practices should drive our policy pages) seems to be in contradiction with your practice of aligning our practice with what the policy pages say (so letting policy pages drive our practices)": There is no contradiction. I think that our policy pages and our practices should match. Where they don't, one or the other needs to be changed. (This is how things work in real life as well, at least in the U.S.: often a given law will "fall off the radar" and go ignored for years, until something happens that suddenly focuses attention on it. When that happens, sometimes it starts being enforced, and sometimes it gets repealed.)

—Ruakh_TALK 21:20, 27 April 2011 (UTC)

Re: Justification: I as a voter often ask: where is the justification or argument that the proposer is making for the proposal? I do not need that the argument or justification be present on the vote page, but I want that the creator of the vote writes a polished argument and justification that deals with possible objections, and makes it available to me, whether via a link or otherwise. Regardless of which terminology we use, I want to see a sort of thing that you do not want to see. You say that the sort of thing is not good, while I say that I as a voter feel the need to have it. You have not demonstrated in any way that they are not good. I have not demonstrated that they are good either; I merely attest to my felt need to have them. --Dan Polansky 08:36, 28 April 2011 (UTC)

A clarifying addition: I have classed two of your points as cosmetic issues. I do not think they are invalid, and maybe they are not all that cosmetic. But they do not change the substance of the proposal. The two issues should not really be sufficient reasons for opposing the vote; an opposer should above all want to keep the usage-in-a-well-known-work attestation option in CFI. --Dan Polansky 17:13, 26 April 2011 (UTC)

Rationale

This is the rationale that I provide as the creator of the vote: This bullet point allows inclusion of terms that no one uses, such as Joyce's "bababadalgharagh...". Admittedly, the bullet point could theoretically simplify attestation of poorly attested languages such as Ancient Greek, but it is not tailored for the purpose, so it both overincludes (by including the above mentioned Joyce term) and underincludes (by failing to recognize single citations from less known works from these languages). To see more counterarguments, check the Beer parlour discussion linked to below.

--Dan Polansky 16:16, 26 April 2011 (UTC)

My revert

I have reverted three changes with which I disagree (diff), so here comes an explanation. Above all, the section "attestation" should only define what an attestation is in Wiktionary. If it defines the term in an unusual way, so be it. If it says that a term that is in widespread use is considered attested, that is okay. The section should not say that attestation is sometimes not required; that is what another section should say. The vote should make as few changes as possible, so the voter can only ponder the one change and not the whole pack. Furthermore, I do not think that two bullets are a bad thing. They make way for another bullet that may come in future, a bullet phrased like "Usage in one durably archived work for less well attested languages". This vote is about the second bullet; it is not about other perceived problems inherited from the past CFI. --Dan Polansky 16:14, 28 April 2011 (UTC)

Fair enough. It'll fail anyway, so I guess the details don't matter. —Ruakh_TALK 23:44, 28 April 2011 (UTC)

My most specific reason for opposing

Since we only require three citations in any work, allowing just one citation from a well-known work seems comparable. If I were to propose a change, it would be used in a 'very well known work'. --Mglovesfun (talk) 10:27, 29 April 2011 (UTC)

It is not really on par (or as you said, comparable). Three independent citations are a low number that attest to the word's being actually used. A single citation from a single author, no matter how notable the author and his work, only attest to author's ability of coining new wonna-be words. Only some words so coined make it into actual use. There is nothing wrong with author's coining new protologisms--someone has to do it--but the business of a dictionary is to document such words as are actually used, of which a single author hardly ever attests. The protologism "bababadalgharaghtakamminarronnkonnbronntonnerronntuonnthunntrovarrhounawnskawntoohoohoordenenthurnuk" exemplifies what I am referring to. As an aside, I hope you don't mind my responding to your post: if you do, just let me know. --Dan Polansky 10:49, 29 April 2011 (UTC)

I will try to put it differently. It seems that you are saying that citations from very well known works are of more worth as citations of use of a word than citations from less known works, including less known novels with ISBN, and most scientific papers. I do not see why this is the case at all. --Dan Polansky 10:58, 29 April 2011 (UTC)

Well it's certainly possible for a word to get three citations without being really part of the language. I could possibly support this if we were to say, a minimum of eight durably archived citations, and maybe saying a minimum of four durably archived citations with one of them being in a very well-known work. So, although I do object to the idea behind this vote a bit, my biggest objection is that this has to be accompanied with other changes to WT:CFI. Another reason, is that a lot people read well-known works, and if they want to know the meaning of a word, where can they look it up? Currently, they can look them up here, if we delete such words they'll have to use other dictionaries. Furthermore some such words aren't 'protologisms', they're just not attested in other works, or the works are hard to find. FWIW pescion that I created a month ago probably doesn't meet CFI in any way; it is attested once as the earliest spelling of what is now French poisson, but I'd be very confident (though not 100%) that there aren't two other citations out there. Sure, do reply as much as you like (I say that to anyone). --Mglovesfun (talk) 11:22, 29 April 2011 (UTC)

If a person reads a well-known work, finds a word that occurs only in that work, and needs to look this wonna-be word up in a dictionary, then the author did something wrong. If the reader of the word could not figure out what the word means, how could the authors of a dictionanry do it? They do not have a crystal ball, and the only quotations from which they can infer the meaning of the wonna-be word are the quotations in that single work. If the wanna-be word is of the form <adjective>-ness where the adjective is already understood by people, the dictionary author knows how to define that wonna-be word, but so does the reader. Authors of well-know works are not more reliable source of wordhood than authors of less known works, AFAICT; some of them are rather creative and unrestrained in their coinages.

Does "pescion" pertain to that discussed bullet point in any way? The discussed bullet point does not save the word, does it? There could be a desirable bullet point that saves "pescion", but the discussed one is not it. --Dan Polansky 11:41, 29 April 2011 (UTC)

Re pescion, no it's not relevant, I agree. --Mglovesfun (talk) 12:03, 29 April 2011 (UTC)

Re: "... they'll have to use other dictionaries." I would like to see some words that are included by the single-known-work rule and that other dictionaries have. "bababadalgharaghtakamminarronnkonnbronntonnerronntuonnthunntrovarrhounawnskawntoohoohoordenenthurnuk" is not such a word, judging from OneLook and Google books. --Dan Polansky 12:12, 29 April 2011 (UTC)

I could see that as a reason to keep the entries as well as one to delete them; we'd be exploiting a gap in the market that isn't cureently well dealt-with. --Mglovesfun (talk) 12:14, 29 April 2011 (UTC)

Sure, true enough. If there are good reasons for including this string, our inclusion of it creates an edge over other dictionaries. Our inclusion of protologism in the mainspace would further increase our edge over them, as would the inclusion of all attestable words from fictional universes, regardless of whether they are used to refer outside of that universe or not. If you ask me, Tolkien's "mûmak" is much more of a word that the Joyce's thunderclap. People are storing the word "mûmak" in their minds (it can be fairly easily remembered), and they are uttering it when talking about Tolkien's world to each other. --Dan Polansky 12:22, 29 April 2011 (UTC)

I disagree that any word only attested once is a protologism. You've picked some good examples above and on the Beer Parlour, but I don't feel like words used in Beowulf or The Song of Roland (as specific examples) that aren't attested anywhere else were simply 'made up' 'for a a laugh'. --Mglovesfun (talk) 12:27, 29 April 2011 (UTC)

But that's a different subject, the one of how to deal with attestation of words of poorly attested languages and poorly attested phases of languages. I think a bullet point along this could deal with this fairly well, containing some ambiguities nonetheless: "Usage in a single durably archived work for a less well attested language". In this bullet, there is no requirement on the work's being well known, but there is the requirement that the language or its phase must be poorly attested. --Dan Polansky 12:32, 29 April 2011 (UTC)

Then I don't accept that either; impasse. --Mglovesfun (talk) 12:35, 29 April 2011 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ You don't accept that alternative bullet point? Why? --Dan Polansky 12:37, 29 April 2011 (UTC)

Speaking only for myself — I don't see that alternative bullet point as relevant, because it's not in the vote. This discussion is rehashing discussions that we've already had and that you've completely disregarded. —Ruakh_TALK 14:55, 29 April 2011 (UTC)

Basically yes; they're two separate issues, well-known works and poorly attested languages. Since your (Dan Polansky's) point above isn't in this vote, I don't see how I can take it into account for this vote. Which brings me back to my point that if we want to remove this line of CFI, there need to be other changes accompanying it, and since those changes aren't in this vote I don't intend to support it. --Mglovesfun (talk) 16:20, 29 April 2011 (UTC)

It was you who has brought this separate issue into the discussion by mentioning Beowulf and The Song of Roland as such works that each word they contain is worthy of inclusion even if it is not used in any other work. To me, this is the only argument of yours that bears any weight; the rest is wholly unconvincing to me. --Dan Polansky 08:17, 2 May 2011 (UTC)

@Ruakh: I do not see what makes you say that I have "completely disregarded" all the previous discussion that was lead in Wiktionary about the well-known-works rule. I have linked to two discussions that I could find from the vote, and I have read both discussions. Actually, from what I can see in the two linked discussions, you have stated only one thing that looks like a reason, namely this: "I like that clause and have no desire to eliminate it, ...". You have as yet stated not a single specific reason for opposing this vote, other than the figurative and thus untransparent "throwing the baby out with the bathwater". It makes me wonder. --Dan Polansky 08:23, 2 May 2011 (UTC)

You misrepresent my words. I do not say that you have completely disregarded all the previous discussion; I say that you have completely disregarded the discussion that this discussion is rehashing. The number of reasons that I gave is irrelevant, since I explicitly said that I'm opposing for the all reasons that were given in those discussions (not just the reasons that I gave). And "throwing the baby out with the bathwater" is a common expression with a well-known meaning; see [[throw the baby out with the bathwater]]. —Ruakh_TALK 12:18, 5 May 2011 (UTC)

I am sorry for having misrepresented your "This discussion is rehashing discussions that we've already had and that you've completely disregarded." I have read all discussions, so I have not disregarded any of them. Anyway, saying that you oppose for all reasons without stating a single one seems evasive to me, and wholly inexplicit. You just throw in "throwing the baby out with the bathwater" without saying what the thrown baby is other than "all reasons"; other people follow. I am not saying that you are obliged to state your reasons, but in my book, you have in effect rejected to state your reasons. I do not deny that the expression with the baby is common; I am saying that it requires another layer of deciphering for no good. --Dan Polansky 07:37, 6 May 2011 (UTC)

Inclusion of Joyce

Personally, I don't see any reason we shouldn't cover every word in Joyce to the best of our ability to.--Prosfilaes 17:43, 29 April 2011 (UTC)

We have no ability to define words used in Finnegans Wake, for example. The foundation for definitions is citation evidence. Without more than one citation, we have no basis for guessing what a word is supposed to mean. Ƿidsiþ 07:41, 5 May 2011 (UTC)

Many nonce words are completely obvious in context, if not explicitly defined there. For Finnegans Wake, there's a large amount of discussion and theorizing that we can refer to. And if all else fails, simply the statement that we don't know what this means can be useful to the student, because then they know that You are not expected to understand this.--Prosfilaes 18:03, 5 May 2011 (UTC)

Then why should not we cover every word in less known works than Joyce's, such as every word that appears in any novel published in 19th century? --Dan Polansky 08:17, 2 May 2011 (UTC)

It's a lot of effect for little payoff, and it's hard to tell the typo from the nonce word in many of those poorly typeset novels.--Prosfilaes 18:03, 5 May 2011 (UTC)

Effort is not really a concern of CFI, if that is what you mean instead of "effect". Various contributors have various interests, leading them to contribute material that others find unworthy of their time.

Typos are not dealt with in the attestations section of CFI. Most typos meet this bullet point: "Usage in permanently recorded media, conveying meaning, in at least three independent instances spanning at least a year". They have to be recognized as typos by other means than attestation criteria. A typo usually looks similar in some way to its non-typo variant, but has much lower frequency than the non-typo. There is something on typos in CFI#Spellings. In any case, I would like to see a typo that is hard to recognize from a nonce word. --Dan Polansky 07:45, 6 May 2011 (UTC)