Wikipedia talk:Manual of Style

Proposed revision: links within quotes

As much as possible, avoid linking from within quotes, which may clutter the quotation, violate the principle of leaving quotations unchanged, and mislead or confuse the reader.

Despite being called into question time and again, the above rule remains essentially the same as when it was added in Oct 2006 following a discussion. When I look at these conversations, it doesn't even look like the rule was ever built upon a clear, unanimous consensus in the first place, or has ever since been firmly supported, established or enforced. I think this obscures its overall validity as a WP:POLICY as well, as expressed in the last comment at this discussion by Pmanderson.

I also couldn't agree more with the last comment about this issue on this talk, by SMcCandlish: "it's more helpful to say 'do this' than just 'don't do that'." So here's my proposed revision, in which I attempt to elaborate on the current rule rather than to alter it:

Version 1

Wikilinks within quotations must be kept to a bare minimum. Only link a word or phrase that is unequivocally referring to a unique and specific entity or notion—that is, usually a proper name or technical jargon—except when it is universally recognized and therefore easily understood by most readers (see Wikipedia:Manual of Style/Linking § Overlinking). Excessive: Smith wrote, "The [[1990s]] [[German]] [[film]], [[Cinematography|shot]] on [[70 mm film|70 mm]], is [[Epic film|bigger]] and more [[Film budgeting|expensive]] than ''[[Ben-Hur (1959 film)|Ben-Hur]]''." Modest: Smith wrote, "The 1990s German film, shot on [[70 mm film|70 mm]], is bigger and more expensive than ''[[Ben-Hur (1959 film)|Ben-Hur]]''." If such a subject is mentioned somewhere else in the article, however, then link those instead of the one inside the quote (see also Wikipedia:Manual of Style/Linking § Repeated links). And even if not, consider adding or moving a mention of the subject to the surrounding encyclopedic passage. Confusing: "[[Paris, Texas|Paris]] still has the best", food critic John Smith said in 2007. (Reader must click on link to discover that the obvious interpretation of "Paris" is incorrect.) Acceptable: "[[Paris, Texas|Paris[, Texas]]] still has the best", food critic John Smith said in 2007. Better: "Paris still has the best", food critic John Smith said in 2007, referring to [[Paris, Texas]]. Most importantly, never give links to words that are remotely semantically ambiguous or use piped links to direct to articles whose subject is significantly broader or narrower than the displayed text (Easter egg links). This is to avoid leaving any room for original research or violation of text–source integrity. Bad: [[United States Declaration of Independence|Four score and seven years ago]] our [[Founding Fathers of the United States|fathers]] brought forth on [[North America|this continent]] a new nation, conceived in liberty, and dedicated to the proposition that [[all men are created equal]]. OK: Four score and seven years ago our fathers brought forth on this continent a new nation, conceived in liberty, and dedicated to the proposition that all men are created equal.

Clunky (?) examples aside, I think this generally sums up the points discussed in this talk before (as cited). What I would love to find out is if people think this is overall in the right direction or not, whether they might agree or disagree about some minutiae. Nardog (talk) 08:14, 4 September 2016 (UTC)

Notes

• It's about time we tackled this idiotic provision. Without mulling it carefully (bedtime!) the above is a great start; subject to the OP's permission of course, I added a note to one of the examples.

• Re "Only link a word or phrase that is unequivocally referring to a unique and specific entity or notion—that is, usually a proper name or technical jargon": Too strong, I think. There's no difference between linking within a quote and linking in a paraphrase of that quote: we need to be sure that what we're linking to reflects what the source was referring to; beyond that, if the link helps the reader understand the quote, that's no different from a link that helps the reader understand a paraphrase.
• Re "except when it is universally recognized and therefore easily understood by most readers": Isn't this just trying to say what WP:OVERLINK says, and (I say again) the guiding principles for linking don't need to be any different inside quotes than they are outside quotes, so they don't need to be restated here in different ways.
• Here's another example that might be useful:

The outer of the two rooms is of Alabama marble, with fluted columns and Ionic capitals.

EEng 08:45, 4 September 2016 (UTC)

Yes, I am certainly being conservative here. My approach as a starting point was basically to fist clarify, paraphrase, and elaborate upon the current rule to bring it closer to the reality of what editors do. Nardog (talk) 02:18, 5 September 2016 (UTC)

• I definitely support this change, whether or not the above word is perfect or eventually changed. (It's a #honour to have inspired something.) To be honest, I completely forgot this policy existed. I linked something in a quote the other day. That's just one example of how this policy is widely ignored and not supported by consensus. It definitely needs to be changed, whether or not it's changed to the above specifically. McLerristarr | Mclay1 08:49, 4 September 2016 (UTC)
• Support (with some copyedits). We've needed to rectify this for a long time. Pretty much no one follows the current rule; it's surely the #1 most-triggered WP:IAR, because there are very few circumstances where adding explanatory text outside the quote to provide link points for the words already used in the quote produces better material. (The "Paris, Texas" example is good, because "Paris" by itself is confusing, and per WP:REUSE we cannot depend on an explanatory link to be available.) It instead usually results in redundant and repetitive material that is frustrating for editors and brow-beating and intelligence-insulting for readers. All rules in MoS and other guidelines (and policies) are not followed some of the time by some editors, because people just show up and start editing without reading all these rules first (and we want it that way; these rules are primary for clean-up gnomes). However, we should not retain a bogus rule that is intentionally ignored as nonsense by virtually all editors who are well aware of the rule; it's a matter of WP:CREEP and WP:COMMONSENSE (and WP:POLICY, which tells us these pages exist to codify actual best practices, not try to force changes that no one actually practices).

  Copyedits: Do not use <tt>...</tt>. This element has not even existed in HTML for years (and if you're habitually using it, please stop - cleaning up after it is a maintenance headache). The correct element in this context is <code>...</code>, and it should wrap the entire example that represents wikicode, not just the part with linking brackets. "Only link a word or phrase that is unequivocally referring to a unique and specific entity or notion—that is, usually a proper name or technical jargon—except when it is universally recognized and therefore easily understood by most readers" doesn't flow right. Try "... or technical jargon. Do not link something that is universally recognized ...". Use {{Crossref}} around crossreferences. "If such a subject is mentioned somewhere else in the article, however, then link those instead of the one inside the quote" doesn't have plurality agreement, and we generally do not want people to link multiple instances, remember. Compress the verbiage; e.g., that entire string can be replaced with "If the term is used outside the quote in the article, link it there instead". Other parts of it can similarly be compressed. The whole segment that starts with "Most importantly, never give links to words that are remotely semantically ambiguous" suffers from this problem. We should also not introduce anything like "never" and "remotely", per WP:BEANS; they're just drama-generation tools (see other thread on this talk page about terrible idea for a rule against "inconsequential" changes). Guidelines do best with "do" and "do not" wording versus "always" and "never" (which seems to deny that WP:IAR exists), and subjective pronouncements like "remotely" and "inconsequential" and "most importantly" incite interpretational disputes. "This is to ..." wording is awkward; it's better to integrate policy rationales directly into the rule's sentence. That whole bit should probably be rewritten. We also don't need to provide the "OK" example that just shows no linking. I think part of our point here would be that well-known quotations are often best left without linking inserted into them. It makes more sense to just say so that to provide an "un-example" with no links in it, for no clear reason.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  13:10, 4 September 2016 (UTC)

Thanks for your suggestions. I'm not an experienced editor here on enwp, let alone its policies, so this is very helpful stuff. So please feel free to make changes below. Nardog (talk) 02:18, 5 September 2016 (UTC)

This is tangential and rather applies to the entire MOSes, but my problem with <code>...</code> is that it totally butchers the {{xt}}/{{!xt}} styling (at least with the default CSS). I found this template {{Plaincode}}, can we use this within {{xt}}/{{!xt}}'s? Nardog (talk) 02:43, 5 September 2016 (UTC)

We've not needed to do this, and have {{mxt}} (green) and {{!mxt}} (red) for this. The main problem is mix-and-matching styles on the same line; just use one:

{{em|Excessive}}: <code>{{!mxt|Smith wrote, "The <nowiki>[[1990s]] [[German]]</nowiki> ...}}</code>

gives:

Excessive: Smith wrote, "The [[1990s]] [[German]] ...

In this case, it's neither necessary nor desirable to show this markup, though. Just use plain {{xt}} and allow the links to link. This will illustrate the "sea of blue" effect in the quoted material.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  20:54, 6 September 2016 (UTC)

Version 2 (currently waiting for SM to edit in his suggested changes)

SMcCandlish, instead of everyone else wading through all your suggestions, why don't you just edit them into the version I've helpfully pasted in below, in a series of self-contained quanta? Start with the most obvious, unobjectionable ones, leaving the ones people may want to discuss or modify for the last. Then people can step through your edits, follow your reasoning in your edit summaries, and revert or modify for further discussion here. (Nardog, I hope you won't object to this approach.) I've already added the "Alabama marble" example, just so it doesn't get lost. EEng 18:37, 4 September 2016 (UTC)

Per suggestion below, I've made some edits [https://en.luquay.com/w/index.php?title=Wikipedia_talk:Manual_of_Style&diff=738826688&oldid=738825763to the V2 text. Mitch Ames (talk) 08:08, 11 September 2016 (UTC)

V2

Wikilinks within quotations must be kept to a minimum. Only link a word or phrase that is unequivocally referring to a unique and specific entity or notion—usually a proper name or technical jargon—and not a common term easily understood by most readers. (See Wikipedia:Manual of Style/Linking § Overlinking.) Example: The outer of the two rooms is of Alabama marble, with fluted columns and Ionic capitals. Excessive: Smith wrote, "The [[1990s]] [[German]] [[film]], [[Cinematography|shot]] on [[70 mm film|70 mm]], is [[Epic film|bigger]] and more [[Film budgeting|expensive]] than ''[[Ben-Hur (1959 film)|Ben-Hur]]''." Appropriate: Smith wrote, "The 1990s German film, shot on [[70 mm film|70 mm]], is bigger and more expensive than ''[[Ben-Hur (1959 film)|Ben-Hur]]''." Where possible, link a word or phrase elsewhere in the article instead of in the quotation. If it is appropriate to link a word or phrase that does not appear elsewhere in the article, consider mentioning and linking it in text near the quote. Confusing: "[[Paris, Texas|Paris]] still has the best", food critic John Smith said in 2007. (Reader must click on link to discover that the obvious interpretation of "Paris" is incorrect.) Acceptable: "{{bracket|[[Paris, Texas]]}} still has the best", food critic John Smith said in 2007. Better: "Paris still has the best", food critic John Smith said in 2007, referring to [[Paris, Texas]]. Most importantly, never give links to words that are remotely semantically ambiguous or use piped links to direct to articles whose subject is significantly broader or narrower than the displayed text (Easter egg links). This is to avoid leaving any room for original research or violation of text–source integrity. Bad: [[United States Declaration of Independence|Four score and seven years ago]] our [[Founding Fathers of the United States|fathers]] brought forth on [[North America|this continent]] a new nation, conceived in liberty, and dedicated to the proposition that [[all men are created equal]]. OK: Four score and seven years ago our fathers brought forth on this continent a new nation, conceived in liberty, and dedicated to the proposition that all men are created equal.

• OP here. I support this because the whole goal was to just suggest a starting point to revise the current rule, so that you, more experienced editors, can improve it. So please be WP:BOLD. Nardog (talk) 02:18, 5 September 2016 (UTC)

@Jayaguru-Shishya: notified me and others of this discussion and thinks it should be at Wikipedia:Manual of Style/Linking. I don't agree, since the impugned statement is on Wikipedia:Manual of Style whereas Wikipedia:Manual of Style/Linking has nothing about quotes. As regards the substantive issue; I am in favour of not having any "rule" on this and letting editors use their discretion. I am sure some readers find links within quotes to be ugly, but they are a minority who need not be accommodated. Sometimes wikilinking in a quote is convenient and transparent. It shouldn't violate WP:EASTEREGG, but that's not specific to quotes. jnestorius^(talk) 16:14, 5 September 2016 (UTC)

I don't know if Jayaguru-Shishya thinks it should be at Wikipedia:Manual of Style/Linking, because it DOES have something about quotes: one sentence in WP:LINKSTYLE. And I actually sort of agree a written rule about this is redundant, but that was already once dismissed (years ago, though). But wouldn't you agree a revision would be at least better than leaving the current rule (Nirvana fallacy)? I think a change is more easily attained when it's gradual than when it's drastic. Nardog (talk) 16:49, 5 September 2016 (UTC)

So you have already tried proposing this at Wikipedia:Manual of Style/Linking, but it didn't end up that well and now you are posting it here? Did I get it right? That'd be WP:FORUMSHOPing, I am afraid. Posting the same thing over and over in hope to "have better luck next time" isn't advisable either.

Anyway, this is the wrong place to discuss the changes concerning WP:MOSLINK. I bet the majority of the editors there aren't even aware of this discussion. Cheers! Jayaguru-Shishya (talk) 17:59, 5 September 2016 (UTC)

I'm afraid but no, you didn't get it right. I honestly don't even know how one could possibly get such an impression. I proposed this here first, then linked to this discussion on MOS:LINK where it virtually has the summarized version of WP:MOS#Linking (because I had the exact same thought as you that editors there might not notice) only to be reverted by you, so I started a discussion there on WT:MOSLINK as you suggested when reverting. I don't know how that's forumshopping or "posting the same thing over and over". Nardog (talk) 18:29, 5 September 2016 (UTC)

If it's going to affect the wording at the main MoS it's better to discuss it here. See below; an RfC at MOS:NUM is now leading to dispute here because the RfC didn't take into account the wording at this page only at MOS:NUM. In general, any non-trivial MoS-related discussion is probably better held here, because far more people watch and comment at this page than at something like WT:MOSLINK.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  03:13, 6 September 2016 (UTC)

Re: request to edit in changes – Okay, but I'll have to come back to it later. The "give me PoV-pushing quote boxes or give me death" stuff further up the page has soured my appetite for editing here today, and I'm focusing on article copyediting, category cleanup, and template work.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  03:13, 6 September 2016 (UTC)

I've no objection to discussion here about important changes to MOS offspring guidelines, as long as they're section-linked at the offspring talkpage. Is the green-background text above the existing, or has it been edited? I don't really want to encourage editors to use awkward square-bracket placements, although it's not exactly wrong to do it. "Paris[, Texas] still has the best" might be better as [[[Paris, Texas|Paris, Texas]]]. Also, we might consider suggesting that wherever it's convenient to link an item in the vicinity of a quotation rather than withing it, that might be preferable. Tony (talk) 03:47, 6 September 2016 (UTC)

If you mean the block beginning Wikilinks within quotations must be kept to a bare minimum with the vertical bar at the left (I ask because to me it looks more bluish than green), it's all new; the current guideline says simply As much as possible, avoid linking from within quotes, which may clutter the quotation, violate the principle of leaving quotations unchanged, and mislead or confuse the reader. I'm editing your suggestions into Version 2, subject of course to others' approval

• You'll notice, if you look at the rendering of your own post just now, that triple brackets don't work the way one wants -- you have to use {{bracket}}.
• I added Where possible, use linking elsewhere in the article to make linking inside the quotation unnecessary.

I have to say, though, that someone's suggestion that we might simply drop the current guideline is an attractive one. If trouble arises, then we can add guidance along the lines of what we've been discussing. EEng 04:08, 6 September 2016 (UTC)

There are other approaches than {{bracket}}; I use [[[Link here]]] because I memorized those HTML character entity codes years ago. You can also do [<nowiki />[[Link here]]]. In all cases, it's fiddly, so we don't want to advise it (no one will comply, and it will be easily broken).  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  20:37, 6 September 2016 (UTC)

I should have said, "You need to use {{bracket}} or something." If this basic proposal gets traction, then we can talk about what markup to recommend (since, when you think about it, inside a quote is major use case for a bracketed link). I'm glad we've got back the old Sandy McCandlish we know and love. EEng 21:31, 6 September 2016 (UTC)

I think the biggest argument against removing the rule entirely would be precisely because it has existed for so long. So instead of paraphrasing or elaborating on it, it may be just enough to say something along the lines of "With regard to linking, quoted texts are no different from any other types of texts. But since they are more susceptible to a violation of WP:NOR or WP:EGG, see to it that they adhere to MOS:LINK." Nardog (talk) 08:20, 7 September 2016 (UTC)

Thinking it over, I'm thinking we might keep the first and second examples above, drop the third, and drop the explanatory text in favor of something much shorter such as what Nardog just suggested. But to not confuse things too much, let's wait for SMcCandlish's edits to Version 2, then think about what to do next. EEng 00:16, 8 September 2016 (UTC)

You all can just integrate what I suggested if you want; I probably can't get to this until tonight or tomorrow.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  17:17, 8 September 2016 (UTC)

• I think we'll all be happy to wait for you. You know best what you have in mind. EEng 19:45, 8 September 2016 (UTC)

Some sugestions:

• bare minimum should be minimum. The word "bare" adds no semantic value.
• instead of

Where possible, use linking elsewhere in the article to make linking inside the quotation unnecessary.

use

Where possible, link a word or phrase elsewhere in the article instead of in the quotation.

• In

that is, usually a proper name or technical jargon

the juxtaposition of "that is" and "usually" seems wrong. I suggest just deleting the former, ie:

... entity or notion—usually a proper name or technical jargon—except ...

• The sentence

Only link a word or phrase that is unequivocally referring to a unique and specific entity or notion ... except when it is universally recognized and therefore easily understood by most readers

is potentially ambiguous. The "except" clause could be read as being an exception to "only" rather than "link". I suggest instead:

Only link a word or phrase that is unequivocally referring to a unique and specific entity or notion—usually a proper name or technical jargon—and not a common term easily understood by most readers (see Wikipedia:Manual of Style/Linking § Overlinking)

Mitch Ames (talk) 07:11, 11 September 2016 (UTC)

Version 2: comments

• "If such a subject is mentioned somewhere else in the article, however, then link those instead of the one inside the quote (see also Wikipedia:Manual of Style/Linking § Repeated links). And even if not, consider adding or moving a mention of the subject to the surrounding encyclopedic passage.."

  What is "such a subject"? What does "those" refer back to? And even if not what? Even without addressing those points, a better opening might be: "However, if such a subject is mentioned elsewhere in the article, link those ..."

• Is "most importantly" necessary? ("Never" is already intensive.) "Remotely semantically" is a little clumsy, and what does "semantically" add here? Either remove altogether or replace with "at all ambiguous? Tony (talk) 07:39, 11 September 2016 (UTC)

• Mitch Ames, Tony1: Will you please just edit your ideas directly into V2? A series of small edits, each with a clear edit summary, will allow others to follow your changes. Likely most if not all will be uncontroversi, but if someone dislikes something we can discuss it then. EEng 07:56, 11 September 2016 (UTC)

please just edit your ideas directly into V2 - Done. Mitch Ames (talk) 08:10, 11 September 2016 (UTC)

This edit to the V2 text might address the first of Tony1's points. Mitch Ames (talk) 08:26, 11 September 2016 (UTC)

More Version 2 comments

Version 2 now has changes made by Mitch Ames, which look good to me. Others should feel free to continue modifying it, or if there are concerns that can't be addressed by simple editing, comments can be added here. SMcCandlish, I'm sure we'd all like for you to edit in anything you still want to see changed. (If such edits are extensive, you might want to start a V3.) EEng 21:32, 11 September 2016 (UTC)

SMcCandlish, we're waiting. (And similarly at MOSDATE, too.) Can you please make time for this? EEng 08:33, 18 September 2016 (UTC)

Briefly, on linking from quotes. This should still be minimised as much as possible. The point of a quote is usually to transmit something in the voice of the person giving the quote (i.e. not in Wikipedia's voice). Linking distracts from this. If something is important enough to link from within a quote, then 99% of the time it will be mentioned elsewhere in the article, or should be mentioned elsewhere, and can be linked from there. In many cases, an explanatory footnote, with links, is better than linking from the quote itself. The reason this is rarely done appears to be because many editors are not familiar with how to generate actual footnotes containing text (as opposed to citations and references). Linking from within a quote is, IMO, a lazy way of explaining things to the reader. Taking the time to write the surrounding text so that things are explained without the need to link, is usually better. This is difficult to express in a guideline, though. Wikipedia:Nesting footnotes is useful to explain the 'reference within note' and similar approaches (hands up who uses the obscure methods 'Subnote within note' and 'Subnote within reference'?). Carcharoth (talk) 13:52, 22 September 2016 (UTC)

Is there anything you think should be changed in Version 2? EEng 14:15, 22 September 2016 (UTC)

It is difficult to engage with the examples, because they lack the context provided by the surrounding text which would in many cases suggest ways to avoid linking at all. Example three, I would add the year the quoted text was uttered/published and explain that 'Four score and seven years ago' is 87 years ago and that it is referring to the US Declaration of Independence in 1776. Or even simpler, when attributing the quote, just link to Gettysburg Address (1863). That should be sufficient for most readers. This is why the context of how you are giving the quote is important. Made-up or truncated examples (stripped of their context) don't really work here. Carcharoth (talk) 15:01, 22 September 2016 (UTC)

I would prefer simple text: "Be especially careful with wikilinks in quotations, ensuring that they are unambiguously appropriate to the text. As editors we do not impose additional meaning upon the words of others, except through the editorial voice, and only then when supported by reliable sources."

Why? Because these same problems can arise with wikilinks in article text. An incorrect or biased link from a quote is hardly more damaging than misleading contextual links. ...is widely considered [[Nazi|right wing]].... We can put words in someone's mouth but we can't put a link in their mouth.

All the best: Rich Farmbrough, 21:07, 22 September 2016 (UTC).

Version 3

This integrates Tony1's comments (I hope). Rich Farmbrough, I'm uncertain how to integrate the changes you're talking about above. Can you edit that into V3 for us? EEng 04:27, 23 September 2016 (UTC)

Wikilinks within quotations need special care, and for this reason are often worth avoiding altogether. Only link a word or phrase that is unequivocally referring to a unique and specific entity or notion—usually a proper name or technical jargon—and not a common term easily understood by most readers. (See Wikipedia:Manual of Style/Linking § Overlinking.)

Example:

Opulent Architecture said, "The outer of the two rooms is of Alabama marble, with fluted columns and Ionic capitals."

Excessive:	Smith wrote, "The [[1990s]] [[German]] [[film]], [[Cinematography|shot]] on [[70 mm film|70 mm]], is [[Epic film|bigger]] and more [[Film budgeting|expensive]] than ''[[Ben-Hur (1959 film)|Ben-Hur]]''."
Appropriate:	Smith wrote, "The 1990s German film, shot on [[70 mm film|70 mm]], is bigger and more expensive than ''[[Ben-Hur (1959 film)|Ben-Hur]]''."

Where appropriate, put a link in nearby text instead of in the quotation; consider adding suitable text where none exists.

Confusing:	"[[Paris, Texas|Paris]] still has the best", food critic John Smith said in 2007. (Reader must click on link to discover that the obvious interpretation of "Paris" is incorrect.)
Acceptable:	"{{bracket|[[Paris, Texas]]}} still has the best", food critic John Smith said in 2007.
Better:	"Paris still has the best", food critic John Smith said in 2007, referring to [[Paris, Texas]].

Don't link to words that are ambiguous or use piped links to direct to articles whose subject is significantly broader or narrower than the displayed text (Easter egg links). This is to avoid original research or violation of text–source integrity.

Bad:	[[United States Declaration of Independence|Four score and seven years ago]] our [[Founding Fathers of the United States|fathers]] brought forth on [[North America|this continent]] a new nation, conceived in liberty, and dedicated to the proposition that [[all men are created equal]].
OK:	Four score and seven years ago our fathers brought forth on this continent a new nation, conceived in liberty, and dedicated to the proposition that all men are created equal.

I have made some revisions to the above. I am inclined to remove the advice that applies to all links in running text, such as overlinking. All the best: Rich Farmbrough, 12:33, 24 September 2016 (UTC).

I've made minor correction to the punctuation in Rich Farmbrough's new text. Mitch Ames (talk) 05:56, 25 September 2016 (UTC)

This link [1] shows the changes from V2 to V3. I'm fine with them. Comments? EEng 06:14, 25 September 2016 (UTC)

I disagree with this change:

Only link a word or phrase that is create links that are unequivocally referring to a unique and specific entity or notion...

In the original wording the word/phrase in the quote must be unambiguously referring to a specific entity, implying that the link target must be that specific entity. The newer wording says that the target must be a specific entity but does not say that the quoted word/phrase must be unambiguous, allowing the editor to infer the meaning of an ambiguous quote - which we ought not allow. Mitch Ames (talk) 06:41, 25 September 2016 (UTC)

Hmmmm. You may have a point. EEng 06:58, 25 September 2016 (UTC)

There being no further comment, I've edited V3 to undo the change Mitch Ames dislikes. Pinging everyone who's participated so far so see how close V3 now brings us to agreement: Mitch Ames, SMcCandlish, Tony1, Rich Farmbrough, Nardog, Mclay1, jnestorius, Jayaguru-Shishya, Carcharoth. EEng 08:40, 28 September 2016 (UTC)

I support V3. Nardog (talk) 10:30, 28 September 2016 (UTC)

Well, I support V3 in that I consider it better than the current one, but I don't disagree with any of the comments below. Nardog (talk) 14:44, 30 September 2016 (UTC)

Oppose I don't see anything in the V3 "bad" examples that is specific to quotations. Overlinking is overlinking wherever it occurs; ditto for eastereggs. There are three cases:

1. A snippet where the wikilinking is acceptable whether or not the snippet is a direct quotation
2. A snippet where the wikilinking is unacceptable whether or not the snippet is a direct quotation
3. A snippet where the wikilinking is acceptable if the snippet is ordinary running text but unacceptable if the identical wording is a direct quotation

Examples of type #1 and #2 are useless in helping editors see the point and scope of a quotation-specific policy. So advocates of any policy need to focus on finding examples of type 3. jnestorius^(talk) 10:37, 28 September 2016 (UTC)

I don't like it. I generally disagree with the whole "avoid it" thing. The point of linking in any case is so that readers can gain an understanding of the context. This is no different in quotes. The only thing that needs to be avoided is linking to something that the person quoted didn't mean. However, there will be many, many cases where it is obvious or explicit what the person meant. It may even have been explained in another part of the quote that is not given in the article. So I don't agree at all that links "are often worth avoiding altogether". If someone's talking about the Beatles, we know they're talking about the Beatles. There's no need to avoid that outside of the rules that apply to normal linking. There's no reason readers will be any more or less aware of terms used within quotes. I would just delete "and for this reason are often worth avoiding altogether". "Wikilinks within quotations need special care" says enough. This bit – "Don't link to words that are ambiguous" – is either unclear or wrong. If the word (I think term is better) is ambiguous in that it requires interpretation that could be incorrect, then, yes, a link should be avoided for the all the stated reasons. However, if a term is ambiguous in that a reader may misunderstand it, but it's actually certain what the person was referring to, then that's a very good to reason to include a link. If a person says (and hypothetically no previous context was given), "I really hate spam", then "spam" needs to be linked to explain whether the person means the food or junk emails. Without that explanation, the quote is pointless. Yes, you could put an explanation in [brackets], but there may be situations where it's better to link. McLerristarr | Mclay1 14:14, 28 September 2016 (UTC)

• I find Jnestorius' and Mclay1's points appealing. I wonder if we may be heading toward something very minimal, like a simple statement like "Wikilinks within quotations need special care..." and maybe just a bit more than that, and maybe or maybe not an example. But let's take this a bit at a time. How about if we start by changing

Wikilinks within quotations need special care, and for this reason are often worth avoiding altogether. Only link a word or phrase that is unequivocally referring to a unique and specific entity or notion—usually a proper name or technical jargon—and not a common term easily understood by most readers.

to read

Wikilinks within quotations need special care. Only link a word or phrase that unequivocally refers to a unique and specific entity or notion‍—‌usually a proper name or technical jargon.

I don't think we need to xref to MOS/Linking § Overlinking because that applies to all links. Can we all agree on that change, for starters? EEng 18:20, 28 September 2016 (UTC)

• I agree with jnestorius, or if there must be a version of the rule, with Mclay1. Direct quotes are not so frequent that overlinking or easter-egg linking won't be easily enough spotted. Harfarhs (talk) 20:16, 28 September 2016 (UTC)
• I agree with the recent comments that examples which are clearly just overlinking, whether in a quote or not, and examples that are clearly appropriate links, whether inside quotes or not, aren't helpful. Of the three example types, only the third is useful, the example of linking that would be okay in regular prose but unhelpful inside a quoted passage. I agree with the "keep is simple" trend; just using that sort of example and suggesting that quote linking should be handled with more care is probably sufficient, at least for now. If a specific, identifiable pattern of overlinking in quotes arises, it can be addressed more narrowly later. I think our goal here should be to acknowledge that whether there was ever a consensus to completely avoid linking inside quotations (dubious), there no longer is one, just concerns about "mangling" quotations with over- or inappropriate (e.g. PoV-pushing) links. Since we already have material on avoiding both of those generally, it would be instruction creep to re-explain that here specifically about quotations in any detail. Just saying not to overlink or use misleading links is probably sufficient, since we can link to the existing broad advice against doing that.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  10:23, 30 September 2016 (UTC)

A bold alternative

I'm losing hope that we'll be able to agree on particular examples, at least for now. How about if we start by simply replacing the current

As much as possible, avoid linking from within quotes, which may clutter the quotation, violate the principle of leaving quotations unchanged, and mislead or confuse the reader.

to

(Version EE) Wikilinks within quotations need special care. Only link a word or phrase that unequivocally refers to a unique and specific entity or notion‍—‌usually a proper name or technical jargon.

Can we agree to just do that for now, and then someday, if it seems needed, reopen a discussion about adding some examples? (We can tinker with the exact text, of course, but what I'm looking for here is agreement on the basic idea that for now we're just going to do something simple similar to what I just gave, not all the examples and stuff.) EEng 05:12, 1 October 2016 (UTC)

• Not an objection... just tossing in a thought (feel free to ignore if it complicates reaching agreement). In my experience, when a quote contains a word or phrase that could be linked... that word or phrase will almost always also appear elsewhere in the article - ie in the running text of the article. When this occurs, I think it is better to place the link in the running text, rather than in the quote. Blueboar (talk) 12:17, 1 October 2016 (UTC)

Yes, that's one of the points made in the more elaborate guideline that's been stalled, so when that discussion's revived that will be in the mix. EEng 02:45, 3 October 2016 (UTC)

Blueboar, would you be able to support the text proposed in this subthread? EEng 05:41, 3 October 2016 (UTC)

• I support this inasmuch as I think the point is, as I mentioned when I started this proposal, that the current rule has been called into question numerous times, and an overwhelming proportion of people seem to think that it at least needs to be replaced (the weird thing is, some of the criticisms raised against the proposed versions are also applicable to the current version). If so, we can just replace it with a minimal, least controversial version to start with, and revise it if necessary whether boldly or via discussion. Nardog (talk) 15:56, 2 October 2016 (UTC)
• Pinging SMcCandlish, Mclay1, Mitch Ames, jnestorius, Tony1, Rich Farmbrough, Harfarhs, Jayaguru-Shishya, Carcharoth. EEng 02:45, 3 October 2016 (UTC)
• Agree with Blueboar's point. Also, the points to not clutter quotations and not mislead or confuse the reader shouldn't be lost. I agree with the goal of replacing the current poor advice with better advice, with or without examples.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  05:33, 3 October 2016 (UTC)

SMcCandlish, if you can find it in your heart, can you support the simple text I proposed at the beginning of this subthread, so we can make at least some definite progress after all this work? EEng 05:40, 3 October 2016 (UTC)

Not a matter of heart, just a matter of that wording not, as Mclay1 points out, including any rationale why quotes should be linked with care. The more I read it, the more I don't think it addresses the nature of the beast. I'm not sure "a unique or specific entity or notion" is really the heart of the matter. We could make it simpler yet, e.g. with: "Avoid linking from within quotations in a way that may mislead or confuse the reader, or when the same link can be provided earlier in the article or immediately after the quotation." (This would also address Blueboar's point.) I don't have an outright objection to working in the other material, e.g. with "Most often, an appropriate link in a quotation will be to a specific entity or notion." ("Unique" and "specific" are redundant.) We could also address the most common forms of misleading/confusing the reader, e.g. with: "In particular, avoid linking anachronistically to an article on a term that has changed meaning since the quotation was originally published, or to a general topic when a more specific subtopic was intended by the original speaker or writer. Piped links can be used to direct readers to specific sections of general articles."  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  07:26, 3 October 2016 (UTC)

• Support. Blueboar may be right, but I don't think it needs stating in the guidelines. However, I think it does need to briefly state WHY links within quotes need special care, i.e. to avoid POV interpretation. The current (the one currently in the MOS) explanation is not good, and I disagree with the concept of clutter (blue or black text – who cares?). McLerristarr | Mclay1 06:04, 3 October 2016 (UTC)
• If we consider that the cluttering part is already covered by MOS:LINK and its concerns about overlinking, then we're still left with including concerns about PoV/misleading/confusing; you say "Support", but simultaneously point out the need to say why handling with care is required, which the proposed language doesn't do.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  07:26, 3 October 2016 (UTC)
• Support (in case that isn't obvious). EEng 06:37, 3 October 2016 (UTC)

---

OK, per Sandy McCandlish, can we all get behind

(Version SM) Wikilinks within quotations need special care. Avoid linking from within quotations in a way that may mislead or confuse the reader, or when the same link can be provided earlier in the article or immediately after the quotation.

--as a first step? EEng 08:00, 3 October 2016 (UTC)

I don't object... however, I think we are really dealing with two separate issues: 1) whether to place a link within a quotation in the first place (vs elsewhere in the text)... and 2) when we do link within a quotation, taking care that the target is appropriate to the quote.

So... I would flip the suggested wording around... to something like: Best practice is to avoid Wikilinks within quotations. See if the desired link can be placed in the main text of the article (either earlier in the article text, or soon after the quotation). If this is not possible, use care to avoid linking to targets that may mislead or confuse the reader. Blueboar (talk) 11:47, 3 October 2016 (UTC)

I would support Blueboar's wording, or something very similar. Carcharoth (talk) 11:56, 3 October 2016 (UTC)

Disagree, I'm afraid. I think "Best practice is to avoid.." is too close to what we already have, and will end up with the same problems. I would support "Wikilinks within quotations need.." - it seems to me to deal effectively with the two issues. Harfarhs (talk) 12:59, 3 October 2016 (UTC)

(Version JN) How about Take special care with Wikilinks within quotations, to avoid linking to targets that may mislead or confuse the reader. See if the desired link can instead be placed in the main text of the article, either before or soon after the quotation. ? jnestorius^(talk) 13:43, 3 October 2016 (UTC)

That works as well. For those who don't want to see wikilinks in quotations discouraged, the problem here is that those who are cautious with how they link don't need reminding. It is those who add links liberally (thinking that a link is better, maybe, than nothing at all) often do so without realising that it can be a problem. It is those people that need this guidance (not that they will have read it anyway, but they will be pointed at it). Carcharoth (talk) 14:34, 3 October 2016 (UTC)

I'd be so happy just to get rid of the current no-links-in-quotes rule I'll get behind almost anything. EEng 15:47, 3 October 2016 (UTC)

In cases where the quotation is taken from an online source, there could be a perception that a link was present in the original cited text. Is there a way to distinguish between links added by Wikipedia editors and links from the original source? isaacl (talk) 16:02, 3 October 2016 (UTC)

That's a very apposite question - almost the reverse situation to the one for which this thread was opened. It's likely that we will not be able to maintain those links embedded in the original text, which may well link to other articles/headlines in the original publication. This would breach guidelines on use of embedded extermal links? Are we meant to not use, such passages at all? Or just to ignore the original links? Martinevans123 (talk) 16:12, 3 October 2016 (UTC)

Most sources of the kind that contain their own links (e.g. blogs) aren't RS for most purposes anyway, so I think this is an unusual case. ME123 is right that we can't maintain them anyway. They should just be silently omitted. No one's going to think links we add were in the original. In some weird case where a link in the source is important, we should say,

Smith wrote in his blog, "I believe in freedom." (Jones pointed out that Smith linked the word America to a Libertarian Party website.)

But please, please, can we first focus on OKing what's in the text on the table (Version JN, above, at this point), instead of what maybe could be added? EEng 17:13, 3 October 2016 (UTC)

A lot of newspaper sites and news analysis sites are now including hypertext links. I raised this specifically because I wanted to gauge opinion on the implications of adding a link to a quotation: would readers consider any included hypertext links to also be a direct quote from the source material? If so, then there is an issue with Wikipedia editors adding links to a quote. isaacl (talk) 15:50, 7 October 2016 (UTC)

Then I suggest you to open a new thread. This could not be the right place to gauge opinion on that. Nardog (talk) 04:30, 9 October 2016 (UTC)

To put it another way, I'm not sure I am in favour of adding links within quotes taken from written sources, as this alters the original quote in an imperceptible manner. Accordingly I have raised the issue in this discussion thread on the proposal. isaacl (talk) 05:49, 9 October 2016 (UTC)

• How about Take special care with Wikilinks in quotations: avoid linking to targets that may not represent the meaning within the quote. See if the desired link can instead be placed in the main text of the article, either before or soon after the quotation.

(I think we can cause as much damage by bad context links, as I mentioned above, but nonetheless I prefer this wording.)

All the best: Rich Farmbrough, 20:19, 3 October 2016 (UTC).

I think Rich Farmbrough's version is an improvement - it addresses something that has been bothering me but that I've not been able to put into words until now. In particular the earlier versions' Avoid linking ... that may mislead or confuse the reader... is a bit vague as to why it may mislead/confuse. RF's avoid linking to targets that may not represent the meaning within the quote is more specific to what (I think) we intend. However I'd go a bit further in that respect and propose:

(Version MA) Take special care with Wikilinks in quotations: link only to targets that correspond to the meaning intended by the quote's author, and only when the author's meaning is explicitly known (rather than assumed). Ideally, link from text outside of the quote instead.

Mitch Ames (talk) 13:10, 4 October 2016 (UTC)

Suggestion withdrawn - I am going to withdraw my suggestion that we include something about linking outside the quote... that is really a separate issue from what is being discussed, and mentioning it in conjunction with the whole "appropriate target/meaning of the author" issue gives the impression that linking to a target that the author of the quote did not intend is somehow OK... if placed outside the quote. That is wrong. Choosing an appropriate target for a link is important, regardless of whether the link is placed within the quote or elsewhere in the text. Blueboar (talk) 15:40, 4 October 2016 (UTC)

With your permission I'm striking your now-withdrawn text to focus attention on the prior text on the table, Version JN, with an eye toward seeing if there's any serious objection to it. Again, I'm hoping we can focus on the acceptability of what's there, and not worry for now about what might be nice to add. EEng 18:22, 4 October 2016 (UTC) Later: Ooops, undoing my strikeout, because somehow I thought Blueboar's comment was from Mitch Ames. This discussion is getting too confusing, and my clumsiness didn't help. Blueboar, at this point I can't tell what suggestion you were withdrawing (energy... fading... can't.... focus........). EEng 03:27, 5 October 2016 (UTC)

Repeated in next subthread, below

Caution Fidelity Minimality Attribution SM Wikilinks within quotations need special care. Avoid linking from within quotations in a way that may mislead or confuse the reader, or when the same link can be provided earlier in the article or immediately after the quotation. JN Take special care with Wikilinks within quotations, to avoid linking to targets that may mislead or confuse the reader. See if the desired link can instead be placed in the main text of the article, either before or soon after the quotation. RF Take special care with Wikilinks in quotations: avoid linking to targets that may not represent the meaning within the quote. MA Take special care with Wikilinks in quotations: link only to targets that correspond to the meaning intended by the quote's author, and only when the author's meaning is explicitly known (rather than assumed). Ideally, link from text outside of the quote instead. MA2 Take special care when adding links to quotations: link only to targets that correspond to the meaning intended by the quote's author, and only when the author's meaning is verifiable. Ideally, link from text outside of the quote instead. If a quote includes any link, add an editorial note, [link in original] or [link added], as appropriate. EE2 Add links to quotations only where the targets clearly correspond to the quotation's meaning. Where possible, link from text outside of the quotation instead – either before it or soon after.

Note that in Version MA (and JN and RF, although obviously I prefer MA), the second sentence ("link from outside the quote instead") could be dropped if necessary, without affecting the first sentence. The first sentence is the fundamental principle and would stand alone. Mitch Ames (talk) 11:43, 5 October 2016 (UTC)

Support I guess I prefer Rich Farmbrough's wording, but I can certainly get behind this one for the sake of pro tem consensus :) Harfarhs (talk) 19:22, 4 October 2016 (UTC)

Harfarhs, since you made your comment during the time I'd mistakenly struck out MA's proposed text, can you clarify which version you're supporting? EEng 03:23, 5 October 2016 (UTC)

Any reason SMcCandlish's suggestion of including the "why" can't be implemented? (Would it be too much clutter?). Tony (talk) 04:23, 5 October 2016 (UTC)

Support version JN or version SM - either would work fine for me. Harfarhs (talk) 17:32, 6 October 2016 (UTC)

• We've got four similar proposals on the table, and I really like to not see this stall because we can't decide what direction to go in. To try to narrow things down, I'd like to start like this: Mitch Ames, I think your "only when the author's meaning is explicitly known (rather than assumed)" is too strong -- when are we going to have an explicit indication of the author's meaning? Would you be willing to withdraw your proposal to narrow us down to three? Remember, we can always revisit later. EEng 06:55, 9 October 2016 (UTC)

I'm not willing to withdraw my proposal, because it's the only one that says "link only to targets that correspond to the meaning intended by the quote's author" and I think that it is important to say that, because it is fundamental to not misrepresenting the author. I said "only when the author's meaning is explicitly known" because if the author's meaning is not explicitly known then assuming it would original research. Typically we would know the author's meaning from the context from which the quote was taken. However I concede that the word "explicit" might be unnecessary, ie "... only when the author's meaning is known (rather than assumed)" would probably suffice. (Whether the author's meaning is "known" can be argued per-case if necessary.) Mitch Ames (talk) 07:30, 9 October 2016 (UTC)

Actually "verifiable" is probably better than "known (rather than assumed)". Mitch Ames (talk) 08:09, 9 October 2016 (UTC)

Note that #Random break to avoid derailing other discussions below makes a valid point. I propose that we should explicitly state whether any link was added, or in the originaleg as we do for italics, per MOS: § Italics within quotations. Such a statement would be in addition to "Version MA" above.

To allow for existing links in a quote, I'd change "with Wikilinks in" to "when adding links to", and make the whole thing:

(Version MA2): Take special care when adding links to quotations: link only to targets that correspond to the meaning intended by the quote's author, and only when the author's meaning is verifiable. Ideally, link from text outside of the quote instead.
If a quote includes any link, add an editorial note, [link in original] or [link added], as appropriate. Mitch Ames (talk) 08:09, 9 October 2016 (UTC)

But I think "verifiable" is too strong. Think of Opulent Architecture said, "The room is of Alabama marble, with fluted columns and Ionic capitals." -- how would these links be verified? Aren't we justified in adding the obvious links? So would you accept dropping and only when the author's meaning is explicitly known (rather than assumed) and add clearly, so that we have, "link only to targets that clearly correspond to the meaning intended by the quotation's author"? Also, would you accept If possible link the same term from the main text of the article, either before the quotation or soon after? If so, can I strike MA and MA2, and add this to the table as MA3? EEng 08:47, 9 October 2016 (UTC)

But I think "verifiable" is too strong.
Verifiability is a core policy. By adding a link to a quote we are, in effect, stating that "when the author says [linked text] s/he means [link target]", ie we are making a statement about the author's intent. As with any other statement, it should be verifiable. Given that we are effectively putting words into people's mouths, I think verifiability should be stated explicitly. MOS: § Original wording already says "Quotations must be verifiably attributed" - so by extension, any expression of intent (not just the quoted words) ought also be verifiable, and most likely can be, from the same source.

Think of "Opulent Architecture .." -- how would these links be verified? Aren't we justified in adding the obvious links?
If the links can't be verified, they ought not be there. They might be "obvious" to someone familiar with the subject, but there is a risk of WP:SYNTHESIS if the quote source does not define the terms. If the quote source does define the terms, there's the verifiability.

So would you accept dropping and only when the author's meaning is explicitly known (rather than assumed) and add clearly, so that we have, "link only to targets that clearly correspond to the meaning intended by the quotation's author"?

No, I think "explicitly known" (version MA), or better still "verifiable" (MA2) is important.

Also, would you accept If possible link the same term from the main text of the article, either before the quotation or soon after?

I think my version is more succinct, and therefore better.

If so, can I strike MA and MA2, and add this to the table as MA3?
It looks like your simply trying to reduce the number of entries in the last column of the table. While I can see the merit in being able to easily see what's the same and what's different between the versions, that's not a reason to change any particular version.

Why not just write your own preferred version and add it to the table for consideration? Mitch Ames (talk) 09:47, 9 October 2016 (UTC)

I've just struck out MA and added MA2 to the table. I'm not sure that converting it to a table was a good idea. Mitch Ames (talk) 10:01, 9 October 2016 (UTC)

As I raised earlier, I believe distinguishing between added links and original links when quoting written sources avoids the issue of imperceptibly changing a quote. Thus I would favour having a label to identify the provenance of any links. isaacl (talk) 14:47, 9 October 2016 (UTC)

Mitch Ames, I think you're reading too much into the verifiability requirement. If, instead of a direct quote, an article said, According to Opulent Architecture, the room was of Alabama marble, the columns being fluted with Ionic capitals, no one would give a second thought to those links. And this would be even more true if, as is quite likely, the attribution was omitted entirely i.e. simply The room was of Alabama marble, the columns being fluted with Ionic capitals. I just don't see how the selection of links is any more fraught inside a quotation than it is in a paraphrase, or in an unattributed assertion of fact.

I went looking for something about the question of link "fidelity" (for lack of a better term), and to my surprise I can find nothing on point in either WP:V or MOS:LINK. And I have to say that in eight years of editing, although I can recall the occasional disagreement about over- or under-linking, I honestly can't remember a single time there was dispute over what would be the right link. Has your experience been different? If not, are we possibly worrying about a potential problem which isn't an actual problem? EEng 17:11, 9 October 2016 (UTC)

Without quotation marks, there is no expectation of the quotation being a literal copy of the original source. With the quotation marks, there is an expectation that the quote accurately reflects the original. As an example of a traditional exception, generally speaking a publication may make spelling changes to conform with its style guide, though I think current practice on Wikipedia does not do this. This type of change is a purely mechanical one that does not alter the meaning in any way. However, adding a hypertext link does involve editorial judgment: both the destination and the appropriateness of including the link are content questions. Accordingly, I am uneasy with adding links where none were present: it gives the impression that the original source made the decision to select Wikipedia pages as its links. isaacl (talk) 04:38, 10 October 2016 (UTC)

The discussion here is selecting the right links to be added to quoted material, not whether it's appropriate to add such links at all. But since you bring it up, I think only the most naive reader would be under the misapprehension you describe -- the same kind of reader who might think, for example, that a red link means the linked subject is somehow related to communism. If you want we can put a disclaimer somewhere. EEng 05:45, 10 October 2016 (UTC)

The context (#Proposed revision: links within quotes) is about whether it's appropriate to add such links. The MOS policy is still that links ought not be added at all. So we need to get consensus that (1) links can be added, and (2) suitable wording of a new policy that allows such additions, and any constraints that apply. Mitch Ames (talk) 13:19, 10 October 2016 (UTC)

In your first post, you quoted the current guidance, including "violate the principle of leaving quotations unchanged" as being one of the concerns with adding links to quotes. Accordingly, I believe it is reasonable to discuss this aspect. Mitch Ames has proposed adding a disclaimer; is this what you are thinking of and so are willing to go along with? isaacl (talk) 21:23, 10 October 2016 (UTC)

My point was that this section of the thread assumes we're going to remove the current prohibition, and focuses on replacement guideline text. Obviously the whole thing will need to be agreed to in a final round of consensus. EEng 03:58, 11 October 2016 (UTC)

Newspapers use single quotation marks to mean paraphrase. So it is doubtful that there is such an expectation. The use of quotation marks is about our style. It is also common practice to add more explanatory information in square brackets. Our wikilinks do no more than that. Hawkeye7 (talk) 23:38, 10 October 2016 (UTC)

If what you're saying, Hawkeye7, is that links are like bracketed glosses in that readers will readily see that they've been added by us WP editors, I fully agree. EEng 03:58, 11 October 2016 (UTC)

Yes, that is what I am saying. Hawkeye7 (talk) 20:24, 11 October 2016 (UTC)

I think a few years ago, this may have been the case, but nowadays, a lot of online news articles contain a plethora of explanatory links to Wikipedia and pages on other sites (as well as links to internal coverage). I think norms are shifting in reliable sources to take fuller advantage of hypertext, and so English Wikipedia may as well address this problem head on. isaacl (talk) 16:45, 11 October 2016 (UTC)

Assuming you aren't speaking of a single-quote versus double-quote standard (some cultures prefer using single quotes in citations), or scare quotes, I don't know of any examples of what you're saying. Newspapers don't quote when they paraphrase. Scare quotes aren't paraphrases, but a mechanism for signalling emphasis or irony, and is not done in encyclopedic writing. isaacl (talk) 16:45, 11 October 2016 (UTC)

For the conventions regarding quotation marks in newspapers, including their online versions, see here. I take your point about encyclopaedic writing, but I'm not sure everyone understands that either, given your other point about the cultures using quotation marks differently. Hawkeye7 (talk) 20:24, 11 October 2016 (UTC)

The article you cite is describing headline writing, for which concision and fitting the words into the available space take precedence. In actual article text, using quotes to quote someone is specifying the literal words used. isaacl (talk) 04:45, 12 October 2016 (UTC)

"... If, instead of a direct quote, an article said, ... ... I just don't see how the selection of links is any more fraught inside a quotation than it is in a paraphrase, or in an unattributed assertion of fact."

The difference is one of "voice", ie who is saying it, Wikipedia or the author of a quote. There's a fundamental difference between Wikipedia says X, and Wikipedia saying that Joe said "X". As has been pointed out already, unless we state otherwise, including a link inside a quote implies that the quote's author made that link and/or intended that particular meaning (the target of the link). Mitch Ames (talk) 13:07, 10 October 2016 (UTC)

"I honestly can't remember a single time there was dispute over what would be the right link. Has your experience been different?"
Yes - Talk:Red pill and blue pill#Derrida, Foucault. This edit is where I removed the links from the quote, triggering the question and my response on the talk page. I'm saying it's common, but it does happen. Mitch Ames (talk) 13:14, 10 October 2016 (UTC)

I'm sorry, but as Hawkeye7 says (I think) above, readers will readily understand that we have added the links. I cannot endorse a requirement that some sort of [links added] disclaimer be attached everywhere, because that's hopelessly awkward, and I think unnecessary. To add to the mess I've thrown my own Version EE2 into the ring in the table above. EEng 03:58, 11 October 2016 (UTC)

DON'T GIVE UP THE SHIP

Well, here I am again pleading that we not let this discussion peter out with no result. And my resolve is redoubled by this edit [2] I encountered recently, in which someone objected to linking Tic Tacs in Donald Trump's cringeworthy teenage-boy-talk, "I've got to use some Tic Tacs, just in case I start kissing her." So please, can we keep this alive?

I've reproduced all the live proposals below, and am meekly asking whether people can get behind EE2, which is about as simple as I think it can get. Please??? If problems come up we can always add more guidance. EEng 00:55, 20 October 2016 (UTC)

	Caution	Fidelity	Minimality	Attribution
SM	Wikilinks within quotations need special care.	Avoid linking from within quotations in a way that may mislead or confuse the reader,	or when the same link can be provided earlier in the article or immediately after the quotation.
JN	Take special care with Wikilinks within quotations,	to avoid linking to targets that may mislead or confuse the reader.	See if the desired link can instead be placed in the main text of the article, either before or soon after the quotation.
RF	Take special care with Wikilinks in quotations:	avoid linking to targets that may not represent the meaning within the quote.
MA2	Take special care when adding links to quotations:	link only to targets that correspond to the meaning intended by the quote's author, and only when the author's meaning is verifiable.	Ideally, link from text outside of the quote instead.	If a quote includes any link, add an editorial note, [link in original] or [link added], as appropriate.
EE2		Add links to quotations only where the targets clearly correspond to the quotation's meaning.	Where possible, link from text outside of the quotation instead – either before it or soon after.

I suggest that if we want to change the policy, which pick the "safest" one, ie the one closest to the spirit of the existing policy and the spirit of WP:OR and WP:V and see if we can get support for that. Presumably the "safest" one ought to be the easiest to pass, being closest to the spirit of what we currently have. Of course, I believe that my version MA2 is the best for this. Mitch Ames (talk) 02:05, 20 October 2016 (UTC)

Are you seriously saying that in quoting Trump (above), and linking Tic Tacs, we have to awkwardly add a note telling the reader that the link is ours and not Trump's? EEng 02:42, 20 October 2016 (UTC)

we have to awkwardly add a note telling the reader that the link is ours – Yes, just as we would for italics/emphasis, per MOS:NOITALQUOTE. While Trump probably can't include a link in his speech, the web page that is our reference could have included a link, hence we need to say whether the link was added by Wikipedia or Wikipedia's source. Mitch Ames (talk) 03:28, 20 October 2016 (UTC)

"Take special care" is not sufficiently restrictive, in my view. I would write it thus: "Insert a wikilink within a quotation only when it would be awkward to provide the same link earlier in the article or soon after the quotation; and never when there is a risk of misleading or confusing readers as to the quoted author's or speaker's original intention." Tony (talk) 03:09, 20 October 2016 (UTC)

Insert a wikilink within a quotation only when it would be awkward to provide the same link earlier in the article or soon after the quotation – I think that's excessive, but would support it as a second choice to my own version. However I still think that misleading or confusing readers is simply too vague; link only to targets that correspond to the meaning intended by the quote's author is quite specific (and deliberately does not require us to make judgements about whether our readers will be confused or misled). Mitch Ames (talk) 03:39, 20 October 2016 (UTC)

Mitch, I think it would be too vague in isolation; but my suggestion qualified it: "and never when there is a risk of misleading or confusing readers as to the quoted author's or speaker's original intention" (my italics). 03:46, 20 October 2016 (UTC)

I have no particular objection to the EE2 version.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  09:04, 23 October 2016 (UTC)

Anyone else??? EEng 23:25, 29 October 2016 (UTC)

Random break to avoid derailing other discussions

Comment: I'm here explicitly intending to avoid !voting either support or oppose, but… Just as a data point, I am inclined to be opposed to links inside quotations because they cause too high a risk of problems related to neutrality, original research, and the case (mentioned above) regarding extant links in the quoted material. Especially as regards the latter issue, the idea that these can be disregarded because such sources are unlikely to be reliable sources is already incorrect and will only get more so over time.

All the most prestigious university presses now have substantial digital services for their output (that would traditionally have been in paper form, in a printed book or journal, and static). A prime (counter)example here would be Oxford Scholarly Editions Online which contains digital-first (i.e. authored for the medium) "books", including links. OUP has a ton of related online services and a growing number of them are "digital-first"-type services that will take advantage of the ability to hyperlink. Similarly, a lot of prominent scientists in many fields are now blogging, both on personal blogs and on blogs affiliated with a relevant institution; both of which are reliable sources.

One prime example of each, from my field, would be Stanley Wells, perhaps the most prominent Shakespearean scholar of our time, who posts articles like Digging up Shakespeare by Stanley Wells; and the Shakespeare Birthplace Trust, that maintains the blog Blogging Shakespeare with contributions by both the curators at the Trust and from affiliated or visiting scholars.

I don't think allowing links inside quotes, except in exceptional circumstances, can be done safely and should be discouraged. Guidance in the MoS for the need to explain parts of quotes should point at ways to do it without altering the quote (e.g. link outside the quote; in a footnote attached to the quote; inside editorial insertion marked by square brackets; etc.), much like the guidance for elisions and similar aim to make any modification by Wikipedia clear and unambiguous.

As an example, while digging up the links for this comment, I ran across this article by Colin Burrow (a well known author in Shakespeare studies) on the OUP literature blog. In it he links the text "John Florio’s translation of Montaigne’s Essays,"—which we are now proposing it would be ok to decorate with inline, in-quote, wikilinks to John Florio, Michel de Montaigne, and Essays (Montaigne)—while Burrow actually linked it to another article, by William M. Hamlin, at the same blog. Did we mean to imply that Burrow and OUP linked to those articles on Wikipedia (quite an endorsement!)? What if the article Burrow had linked to was one that argued that Montaigne had a ghostwriter for his essays, that is, one that in some substantial way actually disagrees with Wikipedia's article on Montaigne? There are a nearly endless variety of such issues, most of which will amount to nothing worse than slight confusion for some readers, but where some few cases will be flat out disastrously misleading.

Anyways, I don't have the spare cycles to participate in the actual discussion here, so I'm instead leaving this comment here so that those who do may opt to take it into account (or ignore it, obviously, as the case may be). Apologies for the interruption. --Xover (talk) 19:05, 7 October 2016 (UTC)

PS. And then there are those who take the MoS guidance and arrive at this as a good and proper course of action (note particularly the first part of the edit summary). And, no, this is not really an argument for allowing links inside quotes. --Xover (talk) 19:38, 7 October 2016 (UTC)

I think Xover makes a very good point. MOS: § Linking and the proposals to change it assume that the original material being quoted did not itself include a link - and that is not necessarily the case when we quote from a web page. Perhaps the solution is to add an editorial note to any quote that includes a link – [link added] or [link in original] – eg as we do for italics, per MOS: § Italics within quotations. Mitch Ames (talk) 07:44, 9 October 2016 (UTC)

See Version MA2 in #A bold alternative above. Mitch Ames (talk) 08:12, 9 October 2016 (UTC)

For several years, I have been removing wikilinks in quotes because of this MOS provision, and sometimes, I am regretful. In This edit of McCulloch v. Maryland, I regretted that, without the link, readers would not guess that Western Shore is an article, and many readers probably do not know that the term refers to an area of Maryland west of the Chesapeake Bay. I don't think any reader would have thought that the link was in any way original to the 1818 legislation. Unfortunately, the article did not have any place outside of quotes to link to Western Shore, so I followed the rules and removed what was probably a benign and helpful link. If we are going to change the rules, this is an example of a link that I would favor being allowed.

In this edit of Walther Hewel, I regretted losing the links to Wilhelm Canaris and Abwehr. I did not regret losing the links to artichoke, venison, or Gemütlichkeit. The last of the three, while a foreign word in English, is one that I think curious readers would search on their own, so I had no regret there. I would disfavor any changes to the MOS that encourage or allow links to artichoke and venison. I wouldn't mind changes to the MOS that allowed for links to Wilhelm Canaris, Abwehr, and Gemütlichkeit.

The rules should mandate not linking within quotes if the link can be anywhere else in the article, not just within a short distance. After all, per MOS:DUPLINK, "Generally, a link should appear only once in an article, but if helpful for readers, a link may be repeated in infoboxes, tables, image captions, footnotes, hatnotes, and at the first occurrence after the lead." At minimum, wikilinks within quotes should be used only (1) in compliance with a conservative reading of Wikipedia:Manual of Style/Linking and (2) where there is no place else in the article to place the link. Further limitations may be warranted. —Anomalocaris (talk) 00:49, 11 October 2016 (UTC)

Yes, common nouns like "venison" would rarely be linked—perhaps in a food article, once, especially if the technical distinction from other cuts is important. On linking within quotations: if the link can't be placed on the same word outside the quotation (often a good fix), we come down to what are good things to link in quotations, and what are not. Tony (talk) 01:23, 11 October 2016 (UTC)

Measures in support of MOS:CURLY

I was surprised to find in an audit of my recent edits and a sampling of random pages that approximately 15% of articles contain one or more curly quotes or apostrophes. There have been previous discussions (2005, more 2005, 2010, 2016) on the problems with curlies. While promoting curlies has always been rejected I'd like to know how much support there is for MOS:CURLY in discouraging use of curlies or even a partial purge of curlies from mainspace.

Full disclosure: I use dial-up, 20 year old hardware and occasionally use a text-based web-browser. Just to say that editors like myself exist as an active part of the Wikipedia community. Also, I've been doing a lot of apostrophe-related typo fixes.

I would appreciate input on the following approaches:

1. To have a bot convert curly apostrophes and quotes to straight apostrophes and quotes in the mainspace. (It would have to exclude articles belonging to categories like Category:Punctuation where curlies should be preserved, and exclude cases with adjacent apostrophes to avoid italic and bold issues.)
2. To have a bot check recent edits for curlies and, when found, leave a message on the editor's talk page (similar to DPL bot when an editor links to a disambiguation page) alerting them to MOS:CURLY issues and linking to a page with instructions for turning off curlies in popular software packages. (Would have to take measures against spamming.)

Like most, I didn't think that curlies were too big of a problem. However, it could be pervasive, effecting some 700k articles. And it wouldn't hurt to inform the small minority of editors introducing curlies into articles. If it's agreed that there are problems with curlies, why not fix it? I'd appreciate your thoughts. - Reidgreg (talk) 17:22, 26 September 2016 (UTC)

Please, not another bot gnoming about. If this isn't already something AWB raises an alert on, that might be a good idea. EEng 19:03, 26 September 2016 (UTC)

Just to point out the "why" of curliest in the present, it's a combination of lazy copy-and-paste of quotes from websites, and/or editing in a word processor and then copying over to the Wikipedia editing window. In both cases, it's just some careless editing and can easily be done as part of routine gnomish edits. Is it more pervasive that we'd like, sure, but I don't think a bot just for that task is needed. oknazevad (talk) 22:04, 26 September 2016 (UTC)

But who wants somebody constantly looking for curly quotes in articles? Philroc^{My contribs} 13:10, 7 October 2016 (UTC)

The editor says "15% of articles contain one or more curly quotes or apostrophes" which seems pretty high to me. Some (unknown) number of articles must contain no quotation marks, so of articles which contain quotation marks, curlies must appear in over 15%. However, yeah, it's not a big deal; I fix them when I see them (and feel like it) but I don't know if its worth worrying about. And, one small advantage of curlies is they signal a possible copy-paste job which alerts me to look for possible copvio. Herostratus (talk) 22:13, 26 September 2016 (UTC)

Thanks for the replies. The percentage did seem high (15.6% of 500 edits and 14% of 50 random articles)), which led me to believe that AWB editors aren't keeping up. Thus the two-pronged approach to clear the backlog and inform lazy careless unaware editors. Copyvio trumps the curly issues, if they can be used as an indicator, though I imagine (or hope) you get a lot of false-positives that way. - Reidgreg (talk) 16:29, 27 September 2016 (UTC)

I don't think that curly quotes are part of AWB's general cleanup. Mr Stephen (talk) 18:01, 27 September 2016 (UTC)

A somewhat related issue I came across today is AWB, as a part of general clean-up, changing ″ and ′ to ″ and ′ respectively. Since these are hard to distinguish from quotation marks visually (curly or straight, depending on the font) I wonder if AWB should be allowed to do this. Jc3s5h (talk) 17:09, 27 September 2016 (UTC)

I've noticed a correlation between curly quotes and copyvios, so it may not be entirely helpful just to "correct" them blindly. --Mirokado (talk) 18:14, 27 September 2016 (UTC)

Agreed. A large percentage of them are due to copy-pastes from other sources. Only a minority are due to people using external editors that auto-curlyize the quotes.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  10:01, 30 September 2016 (UTC)

::I agree too. What should we make the bot do about a possible copyvio? Philroc^{My contribs} 13:20, 7 October 2016 (UTC) Actually, I agree with @Reidgreg: on the fact that if we use curlies as a copyvio indicator, there will be a lot of false positives. Philroc^{My contribs} 13:25, 7 October 2016 (UTC)

Above: Just to point out the "why" of curliest in the present, it's a combination of lazy copy-and-paste of quotes from websites, and/or editing in a word processor and then copying over to the Wikipedia editing window. / Above: Only a minority are due to people using external editors that auto-curlyize the quotes. Sometimes it's none of these. One article I often edit has had curly quotes since its start. I carefully curlify my quotes when I add them. This change from normal practice amuses me. As far as I know, nobody has objected to either the abundance of curly quotes in that article or to my addition of them. I've never thought either that I ought to decurl the quotes there or that I should curl the quotes elsewhere. Perhaps it would have been better if Mediawiki had provided a preprocessor for the (X)HTML Q tag; it doesn't, and this really doesn't matter, compared with (for example) the endemic PR work that infects articles with "prestigious", "legendary", and similar bullshit. -- Hoary (talk) 00:12, 3 October 2016 (UTC)

Ohconfucius's scripts, which I run, zap the curlies. I think 15% is a conservative estimate. Tony (talk) 00:31, 3 October 2016 (UTC)

If you'd like to zap the curlies within my article, be my guest. But really, why should people bother with this when their intellects are easily up to the removal of "iconic", "is recognized [by whom?] as", and suchlike twaddle? -- Hoary (talk) 01:09, 3 October 2016 (UTC)

@Hoary: Having MediaWiki, or at least en.wp's own site-side CSS and Javascript, do something intelligent with <q>...</q> has been added to my to-do list. I think it would be desirable for numerous reasons to markup up inline quotation with this element and to allow people who insist on curly quotes in their output get them. However, this won't do anything for non-quotation uses of quotation marks, e.g. "scare quotes", song titles, etc.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  07:32, 3 October 2016 (UTC)

@Reidgreg: @EEng: @Herostratus: @Oknazevad: @Mr Stephen: @Jc3s5h: @Mirokado: @SMcCandlish: @Hoary: @Tony1: Hey, maybe we could make the bot put in this template that I made that tells people that there are curlies that need to be replaced, but also says that curlies are an indicator of a possible copyvio, and that the article needs to be checked as well. Philroc^{My contribs} 14:29, 7 October 2016 (UTC)

The template looks good though I'd be concerned about putting it on, potentially, 1 in 6 articles. It seems to me that a particularly clever bot could look for curlified text, google it, and then put up {{copypaste}} or a custom template with its results, though this could generate false-positives from sites that mirror wikipedia. I'll try to investigate more as I'm able. I don't suppose if anyone reading this knows whether there are bots looking for copypaste edits? - Reidgreg (talk) 20:26, 15 October 2016 (UTC)

@Reidgreg: Wow, it took you a long time to respond. Now I think we should make the bot remove the curlies, and then make it put in some sort of template about curlies being an indicator of a copyvio. But what template? Philroc^{My contribs} 13:06, 17 October 2016 (UTC)

No. Use a category or something. We've got enough shrill templates no one heeds. EEng 13:36, 17 October 2016 (UTC)

@Reidgreg: @EEng: How about a template AND a category? Philroc^{My contribs} 16:01, 17 October 2016 (UTC)

@Reidgreg: @EEng: Wait, maybe we could add a parameter to Copypaste for the bot. It would change the template to say that the bot removed curlies, and then say that the curlies were an indicator that what the template says already might've happened. Philroc^{My contribs} 16:34, 17 October 2016 (UTC)

I strongly agree with EEng that a hidden maintenance category should be used instead of a loud template at the top of the article. Given that there are 42,000 articles with prominent templates at the top of unreferenced articles in Category:Articles lacking sources from December 2009, and they have been there for seven years, putting a template at the top of articles with curly quotes is extreme overkill.

Anecdotally, I clean up curly quotes where I see them, and it is rare that I see evidence of copyvio. Do we have evidence for that assertion, made above? I find that they are often scattered through the article without a pattern, or copied into citations from titles of articles on the web. I recommend that we first stop the bleeding by working with the writers of automated tools like reFill, which used to copy curly quotes unmodified into citations (and has since been fixed). – Jonesey95 (talk) 15:51, 19 October 2016 (UTC)

The "evidence" is other editors' good-faith statement that they keep finding copyvios this way. I'm one of them. It is okay that your personal experience differs from that of others. It just means your editing doesn't overlap much.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  09:07, 23 October 2016 (UTC)

As an aside, don't forget that (if you are going to start this) there are several other glyphs that should usually be replaced. We have, in no particular order but starting with the commonest

• ’ right single quotation mark
• ‘ left single quotation mark
• “left double quotation mark
• „ right double quotation mark
• ` grave accent
• ´ acute accent
• ‛ single high-reversed-9 quotation mark
• ′ modifier letter prime
• ‚ single low-9 quotation mark
• ‟ double high-reversed-9 quotation mark
• ″ modifier letter double prime
• „ double low-9 quotation mark

plus at least two Arabic glyphs that look very similar, and the guillemots. Some of these come in from cut-and-paste, but some are added at the keyboard. The primes also have a valid (in the MOS sense) use, of course. Mr Stephen (talk) 22:07, 17 October 2016 (UTC)

@Mr Stephen: I can't really start this until it's clear what the bot is going to do if it finds one of these glyphs. Is there anything that you think the bot should do? Philroc^{My contribs} 22:49, 17 October 2016 (UTC)

Actually, I know what the bot will do if it finds one of those glyphs. It will put in {{Copypaste}} with a parameter which changes the notice to talk about how the bot found curlies and changed them, and it also will say that curlies are the sign that what the template says happened. See its sandbox and its testcases. Also, @Reidgreg:, can you give us a better statistic of how many articles have curlies? When I do a BRFA, I'll need to say how many articles are affected. Philroc^{My contribs} 12:29, 19 October 2016 (UTC)

Sorry, Philroc, for some reason I got a notification from your last ping to me but not the several preceding it. (Drop a note on my talk page if my account is active but I'm not responding.) I was also considering a maintenance category but decided to shelve it until I could investigate more and develop a better proposal. There's no point if this doesn't ultimately make it easier for human editors. I found curlies in 15.6% of 500 edits and 14% of 50 random articles. When I have time I'm hoping to get a bot to investigate a larger random sample and generate some data that I could then follow-up with manual implementation of a fix, and see how it works over time. There has been a bot request on this subject which I added to while opening this discussion - and I'm glad I did, as this is much more complex than I'd first thought. I want to be responsible to the community's concerns, and some significant issues need to be addressed. Discussion of technical implementation might be a conversation for the bot request page or elsewhere, though. I don't want to take up any more time from the busy people here who have already patiently explained their concerns. - Reidgreg (talk) 13:34, 19 October 2016 (UTC)

You don't need a bigger sample. The data you have so far tells us that about 14% of articles have curlies, give or take about 5% or so. I think that tells you what you need to know. EEng 14:24, 19 October 2016 (UTC)

@EEng: That's the problem. I have to get a margin of error less than 5% if I want the BRFA for this to get accepted. Philroc^{My contribs} 15:47, 19 October 2016 (UTC)

I didn't realize you were working against an imposed standard. Sounds like you know how to do these calculations, but in case you don't, if you add 50 more to your sample then the give-or-take figure will be at most exactly 5%, and almost certainly far, far smaller. Just for my curiosity, how do you draw the sample? EEng 17:16, 19 October 2016 (UTC)

@EEng: It's not a standard. In fact, there was no standard in the first place. I just don't want a wildly inaccurate value.

By the way, I'm not the one who made the value, Reidgreg did. Philroc^{My contribs} 17:32, 19 October 2016 (UTC)

You know what, I'll just take the value as it is. Philroc^{My contribs} 17:37, 19 October 2016 (UTC)

• I'd still be interested to know how the sample is drawn -- where did you get the 50 random articles from? EEng 17:44, 19 October 2016 (UTC)

@EEng: As I told you earlier, I wasn't the one who got that value. Reidgreg did that. Anyway, from that percent, I have been able to estimate that about 740,000 articles on all of Wikipedia are affected by curlies. Philroc^{My contribs} 19:40, 19 October 2016 (UTC)

For what it's worth, an insource search shows 14,000 articles with one of the four basic curly quote marks in article space. Insource searches do not always work well, though (and this seems way too low based on my experience), so someone should run a search on a database dump if we want data we can trust. If the true count really is 14,000, a supervised bot run should be able to handle it nicely. – Jonesey95 (talk) 19:48, 19 October 2016 (UTC)

Well, there's a teensy discrepancy between your 14,000 figure and the earlier estimate of 14%*5,000,000=700,000. EEng 20:12, 19 October 2016 (UTC)

@EEng: Yeah. Downloading the latest database dump (it's from October 3rd) to Dropbox right now. It should take maybe 6 hours to get my number. Philroc^{My contribs} 21:00, 19 October 2016 (UTC)

I make it about 15% (316 in 2000 random pages). Mr Stephen (talk) 22:33, 19 October 2016 (UTC)

Well, now I'm estimating 740,000 - 830,000 articles affected by curlies even though the insource search says about 14,000. I decided to not use the database dump since it would take too long. So which source should we trust? Philroc^{My contribs} 00:24, 20 October 2016 (UTC)

Well, if someone would please tell me how the 2000 articles were selected, we could use that figure, but I'm not hearing any answer. EEng 00:32, 20 October 2016 (UTC)

• Again just for my interest, how are you drawing these random 2000 pages? EEng 23:46, 19 October 2016 (UTC)

AWB's 'random pages' option. Regards, Mr Stephen (talk) 07:54, 20 October 2016 (UTC)

@EEng: Try pinging people like how I'm pinging you instead of just asking a question. Philroc^{My contribs} 01:56, 20 October 2016 (UTC)

@EEng: Now you have your information. So which source should we trust, my estimate from AWB or an insource search? Philroc^{My contribs} 12:26, 20 October 2016 (UTC)

Well, I've posted a query at mw:Help_talk:Random_page to see if "random" really means random, but on the assumption that's really true, then based on the 316/2000 sample mentioned, the true proportion is about 16%, give or take about 1% or so i.e. 16%*5,300,000 = 850,000 articles +/- 50,000. (I'm not sure if that was based on just double-quote curlies, or the larger list of curlies Mr. Stephen gives above.) EEng 18:27, 20 October 2016 (UTC)

@EEng: But what about the insource search? Philroc^{My contribs} 18:48, 20 October 2016 (UTC)

@EEng: I think we should just go with the insource search because it actually looks at articles, but what do you think?

I'm note sure how I became the arbiter of this, but anyway... At this point, we don't know the technical details of the operation of insource search, and we don't know the technical details of the operation of the random page function, but if I had to choose I'd believe the random sample. Search machinery often has hidden mysteries (doesn't recognize the search string in certain contexts, limits the size of the result set, ...) but even if there's something hinky to the random sample (skewed toward newer articles, skewed toward popular articles, ...) it's very hard to see how any of that would be hugely correlated with presence of curies.

On the other hand, I have to say that 16% seems awful high. Mr Stephen, did you personally examine the 316 you thought had curlies?

Oh, wait... Mr Stephen, were you looking for your full list of curlies? That might explain the difference in results, since the insource search was only looking for the four basic curlies. I wonder if Mr Stephen's list includes something which is legitimately used in a lot of articles. You'd better go back to your 316 and see exactly what character was being used, and how. EEng 19:49, 20 October 2016 (UTC)

@EEng: Speaking of your query about the random article feature really being random, I found your answer. Here! Philroc^{My contribs} 21:21, 20 October 2016 (UTC)

OK, then, despite the (very surprising) oddities mentioned there, that tells us that the random feature is fine for our purposes, which leads me to want to believe the 316/2000 result which I analyze above. However, I still won't be fully comfortable until I hear from Mr Stephen about whether he was searching his whole long list of curlies, and what can be gleaned from looking at the actual hits to see what characters are being found and how they're used. To repeat, I suspect that one of those characters has some technical use we're not thinking of. That would have implications both for the estimate of # of articles affected, and for the design of the bot. EEng 22:58, 20 October 2016 (UTC)

Let me ask him for you. @Mr Stephen:, EEng wants to know if you used your whole list of curlies to make the article count. Philroc^{My contribs} 23:03, 20 October 2016 (UTC)

@EEng: BUT WAIT. It's 12:09 am where Stephen lives, so he isn't up all night to answers your questions. Philroc^{My contribs} 23:10, 20 October 2016 (UTC)

BUT WAIT... Between us we've pinged him three times, so let's just have a nice break until he gets back to us. EEng 23:19, 20 October 2016 (UTC)

OK. Philroc^{My contribs} 23:36, 20 October 2016 (UTC)

OK, got that. I will not be at my PC until later today. I will have a look. Yes, I used the longer list but in my experience the great majority of curly quote s are the ordinary ones. Mr Stephen (talk) 10:00, 21 October 2016 (UTC)

I've decided to download the database dump, except on a 1TB hard drive. I'll get back to you later after I search for curlies. Philroc^{My contribs} 21:49, 21 October 2016 (UTC)

Here we go then, 25 in 200 pages, so 12.5%. Of those 25, Olia Lialina, Amer Fort and Warren Sonbert used unusual quotes, the rest looked like ordinary curlys to me. Mr Stephen (talk) 22:24, 21 October 2016 (UTC)

• diff: Olia Lialina: clean up, straight quotes,see WT:MOS using AWB
• diff: Warren Sonbert: clean up, straight quotes,see WT:MOS using AWB
• diff: Amer Fort: /* Early history */clean up, straight quotes,see WT:MOS using AWB

• diff: Stanwick, Northamptonshire: clean up, straight quotes,see WT:MOS using AWB
• diff: Edward J. Bles: clean up, straight quotes,see WT:MOS using AWB
• diff: Karl Ludwig d'Elsa: clean up, straight quotes,see WT:MOS using AWB
• diff: The Sims 3: Showtime: clean up, straight quotes,see WT:MOS, typo(s) fixed: carrers → careers using AWB
• diff: Broquiès: /* The seigneurs of Broquiès */clean up, straight quotes,see WT:MOS using AWB
• diff: LSU Communication across the Curriculum: clean up, straight quotes,see WT:MOS, added orphan, underlinked tags using AWB
• diff: Adelaide Hasse: clean up, straight quotes,see WT:MOS using AWB
• diff: Mike McCready: clean up, straight quotes,see WT:MOS, typo(s) fixed: tv → TV using AWB
• diff: John Shepherd (scientist): clean up, straight quotes,see WT:MOS using AWB
• diff: 1962–63 Duke Blue Devils men's basketball team: clean up, straight quotes,see WT:MOS using AWB
• diff: Foster's School: clean up, straight quotes,see WT:MOS using AWB
• diff: Max du Preez: /* Awards */clean up, straight quotes,see WT:MOS using AWB
• diff: Information search process: clean up, straight quotes,see WT:MOS using AWB
• diff: Cair Paravel-Latin School: clean up, straight quotes,see WT:MOS using AWB
• diff: Averruncus: /* top */clean up, straight quotes,see WT:MOS using AWB
• diff: Hitchin: clean up, straight quotes,see WT:MOS using AWB
• diff: Ethnic cleansing in the Bosnian War: clean up, straight quotes,see WT:MOS using AWB

• diff: 2012 Munster Senior Football Championship: clean up, straight quotes,see WT:MOS using AWB
• diff: Phil Brown (footballer, born 1959): /* Preston North End */clean up, straight quotes,see WT:MOS using AWB
• diff: 2013 Bengali blog blackout: /* top */clean up, straight quotes,see WT:MOS using AWB
• diff: Contentment: clean up, straight quotes,see WT:MOS using AWB
• diff: George Lefroy: clean up, straight quotes,see WT:MOS using AWB

Why is this percentage so wildly inconsistent anyway? Philroc^{My contribs} 23:54, 21 October 2016 (UTC)

If what you're saying is that the 316/2000 = 16% and the 25/200 = 12.5% are inconsistent, they're not, because the small sample of 200 has a give-or-take figure of 2.3% "or so", meaning it wouldn't be surprising if it were twice that i.e. 4.6%. And 12.5% + 2*2.3% takes you to 16% and beyond. I conclude that the sample of 2000's estimate of 16% +/- 1% is reliable, and I suspect that there's something hinky about the count returned by the insource search. (All of this, of course, relies on others' actual examination of individual articles for curlies, which I have to take on faith.) EEng 05:14, 22 October 2016 (UTC)

• On the basis of my extensive gnoming, 15% is very believable. The script I use cleans it up. But that's separate to the issue of whether MOS should prescribe curly or straight glyphs. I recall very strong reasons for the rule to use straight. Does anyone have a link to that discussion, which was years ago? Tony (talk) 06:20, 22 October 2016 (UTC)

  Well, you were involved in the discussion back in 2007, but debate goes back to 2005 and 2006. Hawkeye7 (talk) 06:52, 22 October 2016 (UTC)

EEng, if you go back about 9 paragraphs from the top, I started with an audit of 500 edits from my contributions. (I was working on apostrophe-related typos at the time, so I personally checked those). I understand that sample was not totally random (all articles had an obscure typo), so I checked 50 random articles from "Special:Random". I don't know the algorithm, either, but I got consistent results so I felt 15% (or even more roughly, 1/6th) was a good enough figure for discussion. What I wanted from a larger sampling was not to confirm this number, but to generate a list with other fields, personally check for copypaste, and look for any other correlations that might be used as indicators (and possibly also generate a list of editors making the copypaste violations). But before going there I wanted to check with some WikiProjects for issues I may not have anticipated. Tony1, if you look at the first paragraph of this discussion, I linked to four sections of this talk page's archives discussing straight and curly quotes. - Reidgreg (talk) 19:08, 24 October 2016 (UTC)

I reviewed the 25 edits listed above. (I'm not very good at this, still learning how to use the copyvio tools.) I found 3 cases that look like copyvio: Cair Paravel-Latin School appears to be copied from the school's website, and has had copyvio problems in the past; The Sims 3: Showtime has text matching a gaming site, difficult to tell if it's a copypaste or a backwards copy; and Contentment attributes a source but does not indicate a direct quotation. As for typography, in 7 cases the curlies should have been wikified instead of straightened. (For example, in Stanwick, Northamptonshire I believe some of the curly single quotes should have been changed to double quotes and others to italic markup.) I was hoping to treat the disease rather than the symptom, as it were, but curlies are indicators of a lot of different problems, many of these articles having multiple issues. Curlies seem to be an indicator of general unfamiliarity with Wikipedia practices. - Reidgreg (talk) 14:47, 27 October 2016 (UTC)

English (International) keyboard layout found.

46.130.34.150 (talk) 19:47, 6 October 2016 (UTC)

There is a keyboard layout called “English (International)”, a enhanced version of US Layout. It can type most languages in Latin script including Vietnamese and Chinese, a limited set for IPA letters (enough for English IPA), lots of dead keys, smart quotes, various (mostly African and IPA) phonetic letters, dashes, Greek letters (no diacritics) and other symbols. There also superscript/subscript numeral characters too, but unlike plain US-International, there is no pre-composed fractions. English (International) requires fractions to use used by slash and super/sub-scripted numbers.

The English (International) layout can be used in a US-Keyboard. English (International) characters are not labeled. The AltGr can be on right alt key, but some US keyboards label “AltGr” without the “Gr”.

I think Wikipedians should use this layout, since it has useful Unicode characters. It should be downloaded and installed. This would allow to type smart quotes and dashes easily in Wikipedia. Instead of "a text", we would use “a text” instead.

http://kbd-intl.narod.ru/english/en — Preceding unsigned comment added by 46.130.34.150 (talk) 19:47, 6 October 2016 (UTC)

The MOS says we shouldn't use curly quotes, and ease of typing is not the only argument against them. nyuszika7h (talk) 20:36, 6 October 2016 (UTC)

Not all Wikipedians use a US keyboard of any description. My top row goes !"£$%^&*()_+ compare that to your !@#$%^&*()_+ and you'll notice the pound sign has been replaced by a hash and the quotes by an at sign. The physical layout is also different. On the right hand side you have 3 keys ([]\{}|), 2 keys (;':") and 3 keys (,./<>?), UK keyboards have 2 ([]{}), 3 (;'#:@~) and then the same 3. Even if you made touch typing a prerequisite for editing I'd still be hitting Enter for \| and trying to complete lines with #~ Martin of Sheffield (talk) 13:42, 7 October 2016 (UTC)

This isn't an MoS matter.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  09:03, 23 October 2016 (UTC)

Policy on names of organizations

The policy on names of organization (such as expanding abbreviations and omitting the "Ltd") found at Wikipedia:Manual of Style/Cue sports § Respect for official organization names should be introduced into the main style manual and not remain ghettoed in the cue sports policy section. See also Wikipedia:Manual of Style#Names which suggests that the cue sports recommendations should be taken as general. Best wishes. RobbieIanMorrison (talk) 09:56, 9 October 2016 (UTC)

• Agree - possibly some minor edits might be required, but generally Wikipedia:Manual of Style/Cue sports § Respect for official organization names ought to apply everywhere. Mitch Ames (talk) 10:08, 9 October 2016 (UTC)
• Essentially the same advice is given regarding the publisher as used in citation templates, e.g. Template:Cite web #Publisher. That should make the guidance already sufficiently ubiquitous to be be worth applying throughout the encyclopedia. --RexxS (talk) 20:47, 15 October 2016 (UTC)
• Agree with the above editors. Makes sense to adopt the general principle. LK (talk) 02:33, 21 October 2016 (UTC)
• Agree, and let's try to normalize any inconsistencies between the MOS:CUE and WP:CS1 details. I wrote almost all of MOS:CUE. The majority of it is careful interpretation of existing rules (and a few bits of it may need revision, since most of it dates to the 2000s), but some of it was intended to fill MOS gaps that come up frequently in pool/billiards/snooker writing. It also turns out that much of it (other than pool-specific stuff like the difference between "the 9 ball" and "playing nine-ball") comes up in all sports writing, and I've long intended to split that material out into a broader MOS:SPORTS. I'm glad that at least one little bit of it also happens to fill a gap in the main MOS itself.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  08:41, 23 October 2016 (UTC)

And/or

Not sure I want to start a huge thread on this minor point, but I'm not sure of the value of recent edits re and/or (whole spread here). For example, the current formulation is misleading: "and/or" does not always mean the same thing grammatically as simply "or". Personally, I'd happily allow it, as the less bulky and proscriptive the MOS is when it comes to perfectly good and useful phrases, the better. But if the bar is going to be maintained, surely the first para of the old wording on its own is clear and succinct enough. The second para there, and the latest version as a whole, both seem a little excessive in terms of detail and justification. N-HH talk/edits 18:09, 9 October 2016 (UTC)

N-HH, I agree with you that "'and/or' does not always mean the same thing...as...'or'". But wouldn't you also agree that writers often use "or" to mean "and" or "or", as in the example of "trauma or smoke"? (If someone dies of smoke inhalation, wouldn't there necessarily be some trauma?) If it seems important to be precise (as in a medical article), then the examples EEng gave would be useful. I like EEng's wording in the first paragraph. I think "and/or" is vague, unclear, and not encyclopedic. Regarding the second paragraph, I think EEng addresses an important point. However, the last part,

• for the other case write: either wild dogs or dingoes inhabit ....

does not completely clear things up. Readers may figure out that "one case" means when two words refer to the same animal and "the other case" is when the words refer to different animals. However, in this example, it sounds like wild dogs may inhabit this place, or dingoes may inhabit this place, but never at the same time. How often does that really happen? I think the addition of "either" does not help (except in that particular situation I have just described, in which case I think the situation could be described even more precisely), unless further explanation is given. Maybe it's just this particular example.  – Corinne (talk) 21:54, 18 October 2016 (UTC)

In strictly logical readings, "or" is already inclusive. So I don't think "and/or" is needed to express inclusive or, and I don't think that's how it's ordinarily used either.

Instead, I would say that "and/or" is a useful sort of equivocation, when equivocation is called for. It means something like "maybe I mean 'and', and then again maybe I mean 'or'; I'd have to think about it more carefully, and I don't think it's important enough to interrupt the flow of thought right now".

But while that's a sometimes-justifiable expedient in speech, we ought to be able to avoid it in encyclopedic writing. We have all the time we need to think about what exactly we want to say. --Trovatore (talk) 22:19, 18 October 2016 (UTC)

My personal preference is that and/or be a marked form, used only where it is very important to convey that there are two options with a definiteness beyond what and provides. In English, by the way, and is the default mortar between the listed bricks; or is exclusive and rather narrower in meaning. Non-native speakers I often see doing what most other languages do: the opposite, where or is the default, and is not exclusive ... and is exclusive in those languages, conveying the sense that the items must all pertain, even simultaneously. Tony (talk) 02:49, 20 October 2016 (UTC)

• Side note to Tony1: Having no idea what position you'd take, I'm hoping you'll find the time to contribute to #DON'T_GIVE_UP_THE_SHIP. EEng 02:55, 20 October 2016 (UTC)

Hmmm. That restaurant serves breakfast and lunch. (But that doesn't mean that you have to eat both when you visit.) Tomorrow I will eat breakfast and lunch. (Strongly suggests both.) Tomorrow I will eat breakfast or lunch. (Suggests only one, otherwise there is no need to make the statement.) Tomorrow I will eat breakfast, or lunch, or both. Well, maybe not so bad. Are there any examples where or both doesn't sound right, or is otherwise inconvenient? Gah4 (talk) 03:22, 20 October 2016 (UTC)

Good examples, Gah, that illustrate my point about English versus other languages. I think "or both" and or all might be clunky or ambiguity-prone in a longish list. But then again, A, B and/or C is ambiguous (all three, or A plus either B or C; and maybe other combinations if the author/reader is being vague). Even my favoured serial comma doesn't make it watertight. Tony (talk) 03:52, 20 October 2016 (UTC)

I don't really agree with Gah's examples. "Tomorrow I will eat breakfast and lunch" doesn't merely suggest both, it requires both. If I say that, and then I miss either meal, then I have made an untrue statement.

However, if I say "tomorrow I will eat breakfast or lunch", and then as it turns out I eat both, then my statement was true.

That's what I meant by my earlier comment that, in strictly logical readings, "or" is already inclusive. I stand by that.

It's true though that there are more subtleties than are covered by this purely truth-functional analysis. If I say "over the hill there is a herd of sheep or goats", and it turns out that there is a herd of sheep, and also a herd of goats, then my statement is true. However, if there happens to be a mixed herd of sheep and goats, if such a thing exists don't look at me; I'm a city boy, then the truth value of my statement is not so clear, because arguably such a mixed herd is neither a herd of sheep, nor a herd of goats.

So in this sort of (somewhat unusual) context, one might be tempted to use "a herd of sheep and/or goats" to indicate a single herd, that may consist exclusively of sheep or exclusively of goats, but may also be a mixed herd.

But I think that's too much meaning to try to jam into "and/or". In speech I have no problem with it; in encyclopedic writing, I think we should go to the effort to say what we mean. --Trovatore (talk) 04:23, 20 October 2016 (UTC)

Well, the content has moved on again since I originally started this. It's been trimmed, which is good, and is less definitive about what "or" means, which is also good, but it does currently still say "or" also implies "both", ie it is inclusive. But it simply isn't in usual speech and writing, at least not always, and I'm quite surprised to find people insisting it is here too. The dictionary definition of "or" is to join alternatives, eg "I'll finish that today or tomorrow". There's reams of grammatical debate around the fact that it can be either inclusive or exclusive, depending on context – and hence why sometimes more precision, such as that afforded by "and/or", or by adding "or both", is helpful. Anyway, this is all a bit academic on the assumption that we want to maintain the bar on "and/or" as a solution to that (which as I say, I'm not sure is necessary, but there you go). N-HH talk/edits 07:39, 20 October 2016 (UTC)

There are too many different contexts and possible meanings involved here; it could be we're just misunderstanding one another. Let's take a very specific case. Do you think "I will have breakfast or lunch tomorrow" could ever turn out to have been a false statement, simply because I ate both meals? I don't. --Trovatore (talk) 07:44, 20 October 2016 (UTC)

By the way, as a separate point, the "finish today or tomorrow" case is not really a counterexample. If you somehow finished the task both today and tomorrow, that would not falsify your statement that you would finish it "today or tomorrow". You may say that it's impossible to finish it both today and tomorrow, by definition of the word "finish", but even if that's so, my assertion is still true, because "if A then B" is always true if A is impossible (see ex falso quodlibet). --Trovatore (talk) 08:03, 20 October 2016 (UTC)

"Tomorrow I will eat breakfast and lunch" doesn't merely suggest both, it requires both. Not that I thought this when I wrote the example, but say I eat brunch. Does that count as both? Or say I eat breakfast food at lunch time? In programming, the usual rule is to choose the one that is most readable, if there is a choice. (It is easier to write unreadable programs than unreadable English.) The actual reason for the example is that it doesn't really matter much. As long as one eats enough food during the day, it doesn't matter much what you call it. But I am sure that there are many examples where the distinction is very important. I am going to fill my gasoline tank or smoke a cigarette, but definitely not both! Now, for the sheep or goat, there might be a geep. Gah4 (talk) 07:48, 20 October 2016 (UTC)

If you eat brunch, and then claim that that makes your prediction that you would eat "breakfast and lunch" true, I wouldn't call that a lie, but I would call it a quibble, a play on words. The usual meaning of "I will eat breakfast and lunch" is "I will eat breakfast, and I will eat lunch", not "I will eat a combination breakfast/lunch". You're toying with the ambiguity of the English language to sneak in the second reading.

However, if you take the position that brunch is breakfast, and that it is also lunch, then I would grant that, if I accept your position, then I must also accept that eating brunch makes the prediction true. --Trovatore (talk) 07:55, 20 October 2016 (UTC)

We're entering the realms of theoretical abstruseness now, none of which helps with the issue at hand. My position is as set out in my comment above. There's no need to debate the minutiae of meaning of specific examples; all that matters is that other examples prove the point that there can be ambiguity in some cases, or clear but potentially different meanings in other cases, around the use of "or" on its own. As I pointed out, this is not a novel, radical or [sic] controversial statement. N-HH talk/edits 08:38, 20 October 2016 (UTC)

OK then, let's take the example in the current text: write simply trauma or smoke inhalation (which would normally be interpreted to imply or both). We don't need to settle the general question; do you agree that this is correct as written? I think it is. --Trovatore (talk) 08:44, 20 October 2016 (UTC)

No, as I said in my last-but-one comment, which I already referred back to, I think it's too definitive about what can be inferred from that example. And even if the example used was more obviously unambiguously inclusive without embellishment, the ambiguous ones would still exist, so it would not be representative of the point. You need an ambiguous example, ie one where "and/or" would otherwise help with clarity, if you want to use one to offer alternative phrasing to people to get the same meaning across (eg "X or Y or both"). N-HH talk/edits 08:52, 20 October 2016 (UTC)

So if 20 people suffered trauma only, 10 suffered smoke inhalation only, and 5 suffered both, you would consider it defensible to say that exactly 30 suffered trauma or smoke inhalation? --Trovatore (talk) 08:59, 20 October 2016 (UTC)

(To be clear, my position is that it's indefensible to say the number is exactly 30. The number is unambiguously 35.) --Trovatore (talk) 09:00, 20 October 2016 (UTC)

I said I'm not going to discuss the minutiae of examples, nor is there any point for this purpose, although I would say that the example in question is also problematic because it has multiple subjects, so there is the added confusion between the collective and the individual experience. Examples can help clarify but they can sometimes just muddy the waters and create more tangential debate. The basic point here, which is widely acknowledged in grammar writing, is that "or" can be inclusive, exclusive or ambiguous as to which. If the MOS is going to ban "and/or" as a means of clarifying ambiguity, it needs to simply state the existence of the ambiguity in some contexts (rather than give an example which is, purportedly, not ambiguous and thereby imply the ambiguity cannot exist) and suggest that: to be clear about exclusiveness, "either X or Y" should be used; and to be clear about inclusiveness, instead of "X and/or Y", "X or Y or both" should be used. It kind of does half of that now, so the problem I identifed above has been ameliorated, but as we have discovered, the example currently cited is making this all more complicated than it needs to be. As I also stated initially, I think this is all proscription/prescription overkill, but that's another debate, and this one has taken long enough. N-HH talk/edits 09:24, 20 October 2016 (UTC)

If you're not going to discuss examples, then I don't know how I can know what you mean. "Exclusive" and "inclusive" are too vague without examples, and so is the notion that there is some problem that and/or can solve. Maybe there really is; I don't know. But I'd like to see an example. (By the way, I don't agree that either...or is exclusive; in my example, there are still 35 people that suffered either trauma or smoke inhalation, and this is unambiguous.)

I tend to agree with you about the MoS being too prescriptive, sometimes. I'm personally OK with the idea that someone might make a judgment call to use and/or somewhere; I'm not going to try to go tracking them down. But if I see it, I'm going to try to figure out what the justification for it is. --Trovatore (talk) 09:36, 20 October 2016 (UTC)

Generally agreed with where Trovatore and Tony1 are coalescing on this. While most uses of and/or are just sloppy writing, there is a marked usage, which most often comes up in formal logic, mathematics, computer science, law and policy, standards documents, and other formal writing, where the expression precisely and concisely conveys "this, that, or both, but not neither", and this is sometimes necessary in encyclopedic writing. It doesn't do readers or editors any service to force them to use a long-winded alternative to and/or. MoS has plenty of "usually don't do x", or "almost always prefer y", or "generally avoid z" instructions, and one more won't cost us anything, while it would reintroduce a small bit of flexibility. I have to note that much of MoS's current emphatic never/always phrasing was engineered by a single editor, who as indeffed back in Feb. It would not hurt at all for us to undo some of these almost entirely undiscussed changes, especially given that the number one complaint about MoS has been alleged rigidity both of MoS's wording in places and of its application. While there are some things we do want consistency on (like most punctuation matters other than optional commas, avoiding loaded terminology, avoiding informal slang, etc.), when it comes to everyday writing choices that reflect a wide range of styles that are still within the encyclopedic register, don't confuse anyone, and don't lead to editorial conflict, then elimination of instruction creep is probably a good idea.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  08:59, 23 October 2016 (UTC)

Quotations

This is why MOS:BLOCKQUOTE is not working for us. Hawkeye7 (talk) 01:46, 23 October 2016 (UTC)

You speak in riddles. How is what not working for whom? EEng 02:39, 23 October 2016 (UTC)

(edit conflict)The linked discussion does not indicate any problem of any kind with MOS:BLOCKQUOTE. Someone is excessively quoting copyrighted material, and is being asked to paraphrase and compress. This is entirely normal (in any publication, not just WP).  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  09:02, 23 October 2016 (UTC)

The editor does not understand what a block quote is. Part of the problem is the lack of prominent quotes. The quoting is not excessive and I am not going to substitute my own opinion for the quoted text. Hawkeye7 (talk) 06:30, 24 October 2016 (UTC)

Pronunciations for Latin taxon names

As most biologists will tell you, there is no "correct" pronunciation for Latin taxon names since there is no standardized pronunciation for New Latin. As our article on New Latin states: "New Latin had no single pronunciation, but a host of local variants or dialects, all distinct both from each other and from the historical pronunciation of Latin at the time of the Roman Republic and Roman Empire." Nevertheless, some people still add IPA pronunciations for taxons or {{pronunciation needed}} tags.[3] This is misleading, however, as it implies that there is a "correct" pronunciation. To quote a couple sources:

• "Latin is now a seldom-spoken language, and we do not know precisely how it was spoken in the Roman world. Many scientific names are words that were not a part of ancient Latin and would sound as foreign to the Romans as Latin does to us. Many English-speaking botanists pronounce Latin names as if the words were written in English. This is known as the Traditional English system. There are many variations and these are often passed on from teacher to student… On the other hand, most classicists and many European botanists prefer Reformed Academic Latin in which strict rules govern the pronunciation of particular letters or combinations of letters… Although there is no consensus among botanists of the world regarding the pronunciation of vowel sounds, there are some general guidelines…" — Vascular Plant Taxonomy
• "These rules cannot satisfactorily be applied to all generic names and specific epithets commemorating persons. About 80 per cent of generic names and 30 per cent of specific epithets come from languages other than Latin and Greek. A simple and consistent method of pronouncing them does not exist… Botanical Latin is essentially a written language, but the scientific names of plants often occur in speech. How they are pronounced really matters little provided they sound pleasant and are understood by all concerned." — Botanical Latin: History, grammar, syntax, terminology and vocabulary

With this in mind, I would like to propose adding the following sentence to the Animals, plants, and other organisms section:

Do not include pronunciations for Latin names, as there is no standardized pronunciation for New Latin.

Any thoughts on this? Kaldari (talk) 05:03, 24 October 2016 (UTC)

Oppose Not sure if that's right. Where there's a long name in wide popular use that people find hard to pronounce – as in dinosaurs and some garden flowers – it seems sensible and encyclopedic to inform readers of the most common pronunciations. After all, there's no standardized pronunciation for anything (try "either", "controversy", ...). Chiswick Chap (talk) 08:13, 24 October 2016 (UTC)

Oppose While it might not be common knowledge among biologists, few people are experts in multiple fields, and palaeolinguists have done substantial reconstruction work to discover how Latin was spoken in Roman times. Rhialto (talk) 08:36, 24 October 2016 (UTC)

That's irrelevant, though: taxon names are not in classical Latin; they often feature letters and consonant clusters wholly or virtually unused in classical Latin. AlexTiefling (talk) 08:59, 24 October 2016 (UTC)

The counterpoint to that is that if we are saying these are not classical Latin (or even 'New Latin') but simply modern pronunciations, then pronunciation guides can still be provided on wikipedia using a descriptive approach of examining what is the actual common pronunciation in use. Rhialto (talk) 09:46, 24 October 2016 (UTC)

I don't disagree - I'm simply saying that while I strongly endorse the use of reconstructed classical pronunciation for classical Latin, it can't be relied upon for New Latin, which has its own contexts and standards, not all of which are fully codified or necessarily compatible with each other. Taxonomy is a good example of this - I wouldn't expect taxonomic Latin to be pronounced particularly similarly to, say, the Latin text of papal letters. So there almost certainly are norms of pronunciation, but classical models are no particular guide as to what they might be. AlexTiefling (talk) 10:09, 24 October 2016 (UTC)

Alan's point is entirely correct. There's essentially no connection at all between reconstructed pronunciation of historical Latin, and scientific circles' often conflicting pronunciations of names in ISV ("New Latin" is a misnomer, as ISV terms are often Greek- not Latin-based, or a mishmash). The liturgical Latin of the Roman Catholic Church has of course has some influence on ISV pronunciations, but they're still diverging rapidly.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  03:28, 27 October 2016 (UTC)

"New Latin" is not exactly a misnomer, since although the sources of scientific names for organisms may be Greek, the names have to conform to some degree of Latin morphology – more so in the case of botanical Latin than zoological Latin, it's true. Peter coxhead (talk) 09:30, 27 October 2016 (UTC)

• Comment I agree with Chiswick Chap that in some cases it is both sensible and encyclopedic to inform readers of the most common pronunciations – with the stress on the plural. The problem is that there are many pronunciations of the Latin names of taxa – a discussion of how the ending of plant family names, -aceae, is pronounced came up with four or five commonly used variants, which differ both between and within English-speaking countries. So a properly neutral presentation might have to list what seems to me to be too many alternatives. So although I don't favour a "ban", I don't think we should encourage the addition of pronunciations of scientific names. Peter coxhead (talk) 12:36, 24 October 2016 (UTC)
• Comment - I have to agree with comments above, that although a ban seems excessive, opening cans of worms is also not an attractive prospect. There is some discussion of pronunciation at Botanical Latin, with scholarly source. New Latin is not the best reference for how taxon names are pronounced. Botanical Latin, as has been pointed out above, has some peculiarities, particularly in the naming of taxa. If any pronunciations are to be added, I would like to see sources, but I doubt that there are many good sources, and know that there are some that I would not wish to emulate, since they reflect the peculiarities of exactly one author. Sminthopsis84 (talk) 23:34, 24 October 2016 (UTC)
• Oppose Bottom line is people will naturally want to know how to pronounce (or one way of pronouncing) a name. If editors of a given article can figure out a way of doing that selecting a usable pronunciation from available sources, let them. EEng 00:00, 25 October 2016 (UTC)

If editors of a given article can figure out a way of doing that, let them – no, there needs to be a source for each and every pronunciation; it's not for editors to generate pronunciations. Peter coxhead (talk) 07:51, 25 October 2016 (UTC)

I've modified my comment to avoid the potential misreading that I meant editors would devise the pronunciations themselves. EEng 08:33, 25 October 2016 (UTC)

• Oppose. While there may be no one true way to pronounce such terms, we can still provide the most common one (or ones, where several variants are frequent), as most dictionaries and other such works already do.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  03:28, 27 October 2016 (UTC)

Provided that an appropriate range of dictionaries are consulted, fine, but what I often see is one dictionary based on one ENGVAR used to add one pronunciation that is not in widespread use, which is not helpful. Peter coxhead (talk) 09:33, 27 October 2016 (UTC)

List of Loanwords

I have made/ am making a list of all english loanwords at WP:LWN, should I file it under proposed guidelines and make an RFC for it, even if it just a list? Iazyges Consermonor Opus meum 23:27, 28 October 2016 (UTC)

See Category:Lists of English words of foreign origin. —Wavelength (talk) 23:57, 28 October 2016 (UTC)

@Wavelength: I am using those as a resource, I am creating one central bank for all loanwords. Iazyges Consermonor Opus meum 00:02, 29 October 2016 (UTC)

Why should MOS include such a list? EEng 00:09, 29 October 2016 (UTC)

@EEng: Building off of WP:MOS#Common usage in English, which says that loanwords are not to be italicized, the list gives a list of said words not to be italicized, basically that if there is a dispute about whether a word should be italicized based on its foreign-ness, they can look at the list and either go "Yep, not a loanword", or "oh it's a loanword, it should not be italicized." While there are lists separated by the language(s) they come from, this would act as a central list of all of them, so a quick page search would ideally immediately tell you if the word was a loanword or not. Iazyges Consermonor Opus meum 00:42, 29 October 2016 (UTC)

You could search an online dictionary and find out the same thing. Is there any evidence that editors are having trouble working this out for themselves? Absent that, I don't see any need for such a list, and if MOS doesn't need something added to it, then it needs to not have that thing added to it, because it's far, far too big as it is. This strikes me as one more way editors can argue over nothing, and we don't need that. EEng 00:52, 29 October 2016 (UTC)

@EEng: Arguments over it happen, I admit infrequently, often on stub or small pages. The list wouldn't get in the way at all, it would just add {{See-also|WP:LWN}} to the MOS page, which would bring you to a full list of loanwords for ease of use. I don't really see how it can cause an argument. Iazyges Consermonor Opus meum 00:58, 29 October 2016 (UTC)

Barely on topic, why is the shortcut WP:LWN and not WP:LNW?

—Trappist the monk (talk) 01:07, 29 October 2016 (UTC)

@Trappist the monk: Thats a good point, I made it to sound out to Loan, but you're right it should be LNW, Ill change it now. Iazyges Consermonor Opus meum 01:10, 29 October 2016 (UTC)

If you don't see how it can lead to arguments then you haven't been hanging around MOS very long. I'm sorry, but based on what you're saying there definitely should not be such a list appended to MOS. (And that's what you're doing when you create Wikipedia:Manual of Style/List of Loanwords.) According to the page itself you're planning to include words like chemistry, and that's ridiculous. English has, I would guess, anywhere from 40,000 to 400,000 words that might qualify as loanwords, making the whole project infeasible anyway. Please rethink this. EEng 01:12, 29 October 2016 (UTC)

@EEng: Alright, how would you resolve it? Perhaps make the list merely about words commonly inappropriately italicized? Iazyges Consermonor Opus meum 03:56, 29 October 2016 (UTC)

One thing MOS should not be doing is dispensing information that applies generally to all formal English writing, and that's what this is, especially since there's no evidence that this is a significant recurring problem that editors aren't resolving easily on article talk pages. Do you have diffs showing otherwise? EEng 04:10, 29 October 2016 (UTC)

To expand on EEng's point; whether a word is a loanword or not is not a matter of style. Whether loanwords should be italicized or not is. I appreciate that the OP wants to inform more editors of our style of not italicizing loanwords, but a style guide is not the place to define what a loanword is. Primergrey (talk) 04:46, 29 October 2016 (UTC)

MOS:LQ is self-contradictory

Source text:

Bierstadt Lake is surrounded by a thick pine forest, and is ringed by sedges that give it a very serene appearance.

Wikipedia prose:

A dense pine forest encircles Bierstadt Lake, and the lake "is ringed by sedges that give it a very serene appearance."

This MOS:LQ statement says to put the period inside: "Include terminal punctuation within the quotation marks only if it was present in the original material, and otherwise place it after the closing quotation mark."

This MOS:LQ statement says to put the period outside: "If the quotation is a full sentence and it coincides with the end of the sentence containing it, place terminal punctuation inside the closing quotation mark. If the quotation is a single word or fragment, place the terminal punctuation outside."

Give-and-take is a good thing at Wikipedia, but useful compromise does not mean creating self-contradictory guidelines. How would you go about resolving this problem? RfC? ―Mandruss ☎ 19:40, 29 October 2016 (UTC)

This might be just me, but I would prefer to avoid this sort of quote altogether. Either quote or don't. Preferably don't in this case; just paraphrase the whole thing. --Trovatore (talk) 19:49, 29 October 2016 (UTC)

Yeah, my concern is more for the guideline than for the article. I'm not here for article help or writing advice. The point is that we have a self-contradictory guideline, and that does more harm than good. ―Mandruss ☎ 19:55, 29 October 2016 (UTC)

Well, but it might be self-contradictory only in a case that shouldn't happen at all, which isn't so much of a problem. It's true though that I wouldn't want to actually ban that sort of quote, so sure, I'll buy that there might be at least a theoretical problem. --Trovatore (talk) 21:51, 29 October 2016 (UTC)

@Mandruss: the two extracts are confusing when read together, I agree. I don't think they are strictly logically contradictory if you pay careful attention to the only in the first of your extracts from MOS:LQ. It does not say "Include terminal punctuation within the quotation marks if it was present in the original material". It is attempting to impose a condition, namely that only if the terminal punctuation was present in the original material should it be included within the quotation marks. It would be better to word the first extract in the negative: "Do not include terminal punctuation within the quotation marks if it was not present in the original material; if it was not, place it after the closing quotation mark."

There's also an issue as to what is meant by "fragment" in the second extract; does it mean that removing any number of words from a sentence always leaves a fragment, or (as I interpret it) does "single word or fragment" imply that the fragment is a small part of the sentence, e.g. just a few words. I'd place the period/full stop inside in your example, but outside in Bierstadt Lake has been described as having "a very serene appearance". But whether this is what was intended or not is another matter! Peter coxhead (talk) 21:48, 29 October 2016 (UTC)

Thanks. Perhaps we can all agree that a guideline is not effective if its application requires close study, parsing, and analysis on its talk page. That would be a good starting point. ―Mandruss ☎ 21:55, 29 October 2016 (UTC)

• I have every confidence you guys will work this out. I just want to say two things:

• This LQ stuff is certainly one of the most effective uses of editor time anywhere on Wikipedia. Millions of articles have been immeasurably improved by the guidance it gives. Really. I really mean that. No kidding. Seriously.
• Trovatore's idea that "this sort of quote" should be avoided is nonsense. And that I really do mean. I don't want to sound harsh, but I'm constantly amazed by the narrow ideas people have about what constitutes good formal writing.

EEng 23:23, 29 October 2016 (UTC)

Re bullet 1: For Pete's sake, how is it helpful to be sarcastic and then state unequivocally that you are not being sarcastic? Either the guideline is important enough to be made more helpful than harmful, or not important enough to exist. I don't really care which, but I do feel strongly that one or the other needs to be chosen. ―Mandruss ☎ 00:51, 30 October 2016 (UTC)

I was being sarcastic when I implied I wasn't being sarcastic. LQ was devised by people who mistake English punctuation for a programming language. But since we seem stuck with it I hope you guys iron this wrinkle out. EEng 03:24, 30 October 2016 (UTC)

LQ was devised by people who value accuracy - what's inside the quotation marks, including punctuation, should match the original text. If it were devised by programmers we'd have punctuation inside and outside the quotation marks:

Did Darla say, "There I am?"?

No, she said, "Where am I?". — Preceding unsigned comment added by Mitch Ames (talk • contribs) 04:08, 30 October 2016 (UTC)

No, LQ was devised by obsessives whose preoccupation with an imaginary problem has put them on a crusade against forms every good publication uses, and with which everyone is familiar e.g.

"I like vanilla," she wrote.

-- instead insisting on idiocy like

"I like vanilla", she wrote.

or maybe (for all I know)

"I like vanilla.", she wrote.

Every reader understands that punctuation at the boundary of a quotation might be modified in certain standard ways as part of the transition to the non-quoted material. LQ doesn't promote "accuracy" but rather turns the familiar and attractive into the unfamiliar and ugly, for no reason. EEng 07:23, 30 October 2016 (UTC)

• I agree with Peter Coxhead. There's no contradiction. It's just a case where things could be made clearer. Jimp 03:04, 30 October 2016 (UTC)

• I agree with Mandruss, and I remember raising this issue here twice, starting in mid 2014. The point about LQ is not clear, so it's hardly a surprise to see that MOS:LQ is rarely adhered to. JG66 (talk) 03:35, 30 October 2016 (UTC)

I'd change MOS:LQ to: Include terminal punctuation within the quotation marks only if it was present in the original material and is grammatically required by the quoted text; otherwise place it after the closing quotation mark. So I'd write:

A dense pine forest encircles Bierstadt Lake, and the lake "is ringed by sedges that give it a very serene appearance".

The full stop is outside the quotes because all though it is present in the original text it is not grammatically part of or required by the quote - the quoted text is not a full sentence, so does not require a quoted full stop. This is consistent with Keep them inside the quotation marks if they apply only to the quoted material and outside if they apply to the whole sentence. In this case the full stop terminates the entire Wikipedia prose sentence, not "only the quoted material". Mitch Ames (talk) 04:24, 30 October 2016 (UTC)

I would tend to go the other way and include the period in the quote. Maybe it's not "required", but putting it outside suggests to me that the quoted sentence did not end there; and it did. In any case, I think what we have here is an option to do it either way, neither of which is incompatible with the principle of LQ. Maybe we can express it in a way that removes the contradiction, without necessarily being more prescritive than necessary to keep it "logical". Dicklyon (talk) 05:16, 30 October 2016 (UTC)

The correct way to show an option either way is: "The period may be placed inside or outside the quotation mark, at the editor's discretion." Or just say nothing at all. ―Mandruss ☎ 05:50, 30 October 2016 (UTC)

The point of the guideline is not to "show an option either way", but to specify a single consistent way of doing it. If you wanted to change the guideline to "at the editor's discretion" that would be a separate proposed change to the guideline. Mitch Ames (talk) 05:58, 30 October 2016 (UTC)

I agree, and I was not and am not proposing that. I was replying to Dicklyon's comment. As I said, my only concern is eliminating the apparent contradiction to avoid AT conflict and circular editing, both of which are wastes of editor time. Alternatively, scrap the guideline, but that would be a harder sell I think. ―Mandruss ☎ 06:03, 30 October 2016 (UTC)

putting [the period] outside suggests to me that the quoted sentence did not end there" — On the contrary, the point of LQ is that the absence of the period inside the quote does not imply anything about the original text other than what is actually quoted. The source text could have been "... and is ringed by sedges that give it a very serene appearance, reminiscent of ..." and it would not matter; the Wikipedia prose would be just as correct and have exactly the same meaning. Mitch Ames (talk) 05:55, 30 October 2016 (UTC)

• Comment - In my opinion, people who oppose the existence of a guideline should initiate an RfC for its removal. Otherwise they should refrain from disrupting discussions about its improvement. ―Mandruss ☎ 06:13, 30 October 2016 (UTC)

Applicaion of MOS:TENSE to "first" status that is no longer "only" status

Our article A Midsummer Night's Dream (1909 film) currently reads It is the first film adaptation of the eponymous play. This to me reads like a contemporary statement that implies there were no subsequent film versions. Would It was the first film adaptation of the eponymous play not read better? Is this an IAR scenario, or would formatting the R to cover cases like this be better? Hijiri 88 (聖やや) 05:09, 30 October 2016 (UTC)