Voting on: Removing the item "use in a well-known work, or" from WT:CFI, specifically WT:ATTEST, placing ", or" at the end of the item "clearly widespread use".
Support. As noted on the talk page, the last time we had a vote like this, the "well-known work" criterium was used to include many words from extinct and poorly attested languages, and I opposed its removal. Now, words from extinct and poorly attested have their own special allowances, and the only entries which rely on the "well-known work" criterium are one-off nonces whose very nature means we are usually unable to offer verifiable definitions for them. I agree with DCDuring's comment on the talk page that we should not have main-namespace entries for such things. We can create an appendix of nonces found in well-known works, and use {{only in}} to direct anyone who actually looks up Joyce's thunder word (etc) to it, the way we already use {{only in}} to direct people to appendices of phobias and of dictionary-only words. (And just like the use of {{only in}} to direct users to those appendices is not regulated by CFI, I agree with the view expressed on the talk page that this use does not need to be spelt out in CFI. If anything, spelling out certain uses of {{only in}} would be instruction creep and would imply that other uses were not allowed, when in fact {{only in}} seems to be designed to be flexible.) - -sche(discuss)19:26, 23 March 2014 (UTC)[reply]
Oppose. Approval of this proposal will result in the loss of recognition for nonce words found in well-known works. I could agree to having nonce words listed in Appendix: namespace and linking to them from mainspace via {{only in}}, but the current proposal makes no provision for that. (I would rather have abstained, since voting is the worst possible form of decision-making and we should really avoid it as much as possible, but around here abstaining from a vote means having your opinion ignored.) —Aɴɢʀ (talk) 19:43, 23 March 2014 (UTC)[reply]
This vote is solely on "Removing the item 'use in a well-known work, or' from WT:CFI, specifically WT:ATTEST, placing ', or' at the end of the item 'clearly widespread use'." Your comments above reflect what you would like to happen if this proposal is approved, but that's not what the proposal itself says. —Aɴɢʀ (talk) 20:21, 23 March 2014 (UTC)[reply]
I think this vote itself already constitutes an undesirable level of instruction creep and bureaucratic masturbation. I hate votes and wish we wouldn't use them in our decision-making. —Aɴɢʀ (talk) 20:40, 23 March 2014 (UTC)[reply]
In the past, Dan and sometimes others have opposed deciding things (most recently e.g. the repurposing of {{m}}) by discussion alone, and requested votes. Even I, who prefer to decide most things via discussion, think changes to our enumerated CFI are so major that they should be done by vote, like this. So we're at something of an impasse: it's hard (if not impossible) to make decisions when different sets of users can't even agree on which methods/venues of decision-making are acceptable. - -sche(discuss)21:09, 23 March 2014 (UTC)[reply]
Re: "since voting is the worst possible form of decision-making". Prove it. Or even better, acknowledge it to be a blatant nonsense; "not the best possible form of decision making"--maybe, "the worst possible form of decision-making"--ridiculous nonsense, even a criminal one. --Dan Polansky (talk) 09:26, 29 March 2014 (UTC)[reply]
This vote and virtually every other vote we've had at Wiktionary proves it. What's blatant nonsense is calling for a "vote" on every little thing we want to do, as is calling ideas you disagree with "criminal". Pretty sure disagreeing with Dan Polansky isn't against the law in my country. —Aɴɢʀ (talk) 10:24, 29 March 2014 (UTC)[reply]
Mentioning votes, even bad votes, does not constitute a proof that "voting is the worst possible form of decision-making". Your idea of what a proof is seems seriously lacking. --Dan Polansky (talk) 10:38, 29 March 2014 (UTC)[reply]
Oppose. In the discussion of this subject, User:DCDuring commented: "There are lots of terms whose meaning I'd like to know, but which actually have no meaning in a linguistic sense". If such words are not found in Wiktionary, the reader may just as likely conclude that we are missing the word, rather than excluding the word. If we have an entry that explains the absence of meaning, then we are at least answering that question. Please note that I'm not wholly averse to a redirect from that entry to an appendix, but the proposal does not, by its terms, require the implementation of such an appendix, and I just don't know that it makes any more sense to do it that way. bd2412T21:38, 23 March 2014 (UTC)[reply]
Oppose. I support appendicising nonce words, but, as Ruakh said in the previous vote, this solution throws out the baby with the bathwater. — Ungoliant(falai)02:17, 25 March 2014 (UTC)[reply]
My problem with this vote is not the lack of specification of what happens to nonce words if it passes, it’s the actual removal of the “usage in a well-known work” clause. — Ungoliant(falai)02:46, 25 March 2014 (UTC)[reply]
Then what do you mean by "I support appendicising nonce words"? Do you want nonce words to be present in the main namespace and in an appendix? (That would be a perfectly reasonable position ... it's just not one that I had considered that anyone might have, until now.) If you want nonce words to be present only in an appendix (not in the main namespace), then it would seem you should support this proposal: extinct and poorly documented languages are now have their own CFI — what words would be excluded if the "well-known work" provision were removed, other than nonce words? - -sche(discuss)03:03, 25 March 2014 (UTC)[reply]
No, I don’t support keeping the nonce words in the MNS if they are moved. There are two other babies I see in the bathwater:
regional terms: the author of a well-known work may be only person to use a term limited to his dialect;
ease of citing: even if a term has a lot of uses, citing from a well-known work takes one-third of the time.
If we want to appendicise nonces, we have to know which words are nonces and which aren't, right? And the way to know if something is a nonce or not is to find out if multiple uses of it exist. Well, if they do, the term is not a nonce — and if it were listed at RFV, it'd be passed under the "three uses" clause. (I'm counting words that have only been used twice as "nonces".) If multiple uses don't exist, then the term is a nonce, and either gets kept (if we keep the "well-known work" clause) or appendicised (if we get rid of the "well-known work" clause). Your thought about how it would be easier to cite a well-known work than to cite three works is an interesting thought, but can you point to any RFV discussion in which it actually happened that participants agreed that a word was used in 3+ works but that they could only be bothered to cite one of them — can you point to any word that was passed under the "well-known work" clause even though it could have passed under the "three uses" clause? Keeping the "well-known work" clause for such an uncommon and easily overcomeable situation as that seems less like keeping the bathwater because the baby hasn't quite finished its bath, and more like abandoning the baby in the bathtub because drying babies off is slightly difficult.
The presumption that any term that occurs only once in a language's total corpus is limited to the context in which it occurs seems like a solid presumption for a descriptivist dictionary to make. How would you refute that presumption? In other words, how would you demonstrate that any nonce was not region/dialect-specific? You couldn't point to speakers of other dialects using the terms, because if you could, the term wouldn't be a nonce. - -sche(discuss)04:46, 25 March 2014 (UTC)[reply]
What? No. I’ll cut the metaphor and state with (hopefully) more clarity what I think should happen: nonce words should be moved to the Appendix namespace. These appendices should be linked to from the main namespace using {{only in}}. Getting rid of the well-known work clause would allow us to remove the nonce words from the main namespace, but it would also get rid of that clause’s useful aspects, so I oppose getting rid of it. — Ungoliant(falai)10:51, 29 March 2014 (UTC)[reply]
Do you have an idea how to fix the negative consequence, other than removing the item while expressly stating that nonces are allowed in an appendix? --Dan Polansky (talk) 11:13, 29 March 2014 (UTC)[reply]
What definition of nonce would you use and how would you distinguish single-attested nonces from single-attested non-nonces using evidence? --Dan Polansky (talk) 11:23, 29 March 2014 (UTC)[reply]
So you would use this definition of "nonce": "A word invented for the occasion." Now how do you distinguish single-attested nonces from single-attested non-nonces using evidence? --Dan Polansky (talk) 11:58, 29 March 2014 (UTC)[reply]
“Research may clarify the term’s condition; you can find resources saying that bababa...thurnuk is Joyce’s coinage, for example.” — Ungoliant(falai)13:22, 29 March 2014 (UTC)[reply]
Wiktionary:Beer parlour archive/2011/January#CFI: Removing usage in a well-known work, January 2011
Support per Wiktionary talk:Votes/pl-2014-03/CFI: Removing usage in a well-known work 2#Rationale. --Dan Polansky (talk) 11:47, 22 March 2014 (UTC)
Oppose. Approval of this proposal will result in the loss of recognition for nonce words found in well-known works. I could agree to having nonce words listed in Appendix: namespace and linking to them from mainspace via 