User:Herostratus/Naming conventions (geographic names)

I'm closing the discussion. I'll detail all reasons below. The result is: 1) No change to the conventions used for article text. 2) For article titles incorporating the specific phrase "metropolitan area" the form is "[city name] metropolitan area, [state name]", for example "Columbus metropolitan area, Ohio". Only applies firmly to entities in the United States, only applies where disambiguation is required, applies for terms closely related to "metropolitan area" (see below).
 * 2A) For such titles where the metropolitan area spans two or or more states (or provinces, etc.), the form is "[city name] metropolitan area, [state name 1–state name 2–state name n]", for example "Kansas City metropolitan area, Kansas–Missouri".
 * 2B) For such titles where the metropolitan bears the name of two or more cities, the form is "[city name 1]–[city name 2]–[city name n] metropolitan area, [state name(s) (if included]", for example "Minneapolis–Saint Paul metropolitan area, Minnesota".
 * The use of "Kansas City metropolitan area, Kansas–Missouri" and "Minneapolis–Saint Paul metropolitan area, Minnesota" as examples intended to imply that Kansas City metropolitan area or Minneapolis–Saint Paul metropolitan area require the further amplification, only that if they did the form shown should be used. The use of "and" (as for instance ""Kansas City metropolitan area, Kansas and Missouri") is to be avoided.
 * 2C) State names should only be included when disambiguation is required (per the spring 2013 RfC here which asked the question "Should US metropolitan area article titles include a state name, even when no disambiguation is strictly necessary?" and the conclusion was "No", superseding the previous WP:USPLACE suggestion the state name should be omitted only if the first principal city is one of the 30 U.S. cities cited by the Associated Press Stylebook for stand alone datelines (BTW WP:USPLACE needs to be updated to reflect that).

3) There was little discussion of non-USA entities, so no hard decision is made here regarding non-USA entities, BUT it's strongly recommended that when practicable the same form ("[city name] metropolitan area, [appropriate larger polity name]", for instance "Leeds metropolitan area, Yorkshire" (if disambiguation were required for Leeds)) should be used, UNLESS superseded by some consideration specific to that nation and there's a strong demonstration that that consideration requires the general rule to be overridden. 4) No decision and no change for all other article titles. (Thus, no decision is rendered on whether (for instance) "Columbus, Ohio, train disaster" or "Columbus, Ohio train disaster" or "Columbus train disaster, Ohio" or "Columbus train disaster (Ohio)" or "Train disaster in Columbus, Ohio" or some other construction is preferred.)

In this and all text that follows, I've assumed that "metropolitan area" also includes very closely related terms such as "metropolitan statistical area", "micropolitan area" and "conurbation", and similar terms used outside the USA, such as "census metropolitan area" in Canada, if for some reason these need to be used in titles (when not inappropriate, the use of the uncapitalized non-country-specific generic English term "metropolitan area" is encouraged.)

Scope
The specific question was this:
 * "Recently our guideline Wikipedia:Naming conventions (geographic names) was edited to state that when the city and state is given followed by something else, such as Dublin, Georgia micropolitan area, a comma is used after the state. Should there be a comma there, yes or no?"

And in the Survey section, the question was stated thus:
 * "On just the question of whether one comma or two commas are preferred when a state name is used parenthetically as in "Rochester, New York metropolitan area" versus "Rochester, New York, metropolitan area", in a title or in a sentence, please make a brief numbered signed entry..."

As I expressed in the subsection "Requested closure" at the bottom of the discussion, the questions could be taken to cover various circumstances. However, in reading the discussion and various related discussions, I find that most of the participants, by far, understood the discussion as bearing on article titles only. To expand this discussion to include a ruling on article text would not be proper, there simply wasn't enough discussion for me to gauge community feeling on that. This would be a significant change affecting very many articles and we'd need a new a separate discussion for that, one in which it was specifically made clear that article text was in play.

Furthermore, by far the majority of the examples used, and the discussion, specifically took the question to bear on titles that include the term "metropolitan area" (and related closely similar phrases) -- enough that I did not consider other terms to have been discussed enough that it'd be fair to gauge the community feeling on terms other than "metropolitan area". So I can't render a decision on those.

Besides this discussion, I read through these related discussions, and included them in my deliberations. The titles (and contents) of these discussions further reinforces my feeling that only article titles are in play, and only those containing the term "metropolitan area".
 * Wikipedia talk:Article titles/Archive 42
 * Talk:List of metropolitan areas of the United States
 * Talk:Pensacola metropolitan area and Talk:Pensacola metropolitan area
 * Talk:Brunswick, Georgia metropolitan area and Talk:Brunswick, Georgia metropolitan area

I can't forbid someone taking this discussion and leveraging it as a implied precedent to cover either article titles that don't contain the term "metropolitan area", or article text of any kind, but I wouldn't recommend that, at all. Instead, run a new discussion; pointing to this one as a precedent would be a reasonable argument but no more than that.

Let's get down to brass tacks.

Policy arguments
The controlling authorities are, in descending order of precedence:
 * Article titles (WP:AT), which is a policy.
 * Manual of Style, the Manual of Style style section dealing with article titles. The MOS (Manual of Style) is only a guideline, but a guideline with a great deal of de facto authority.
 * Naming conventions (geographic names), a guideline. According to WP:AT, this is supplementary to WP:AT.
 * WP:COMMA, which is part of the MOS. The extent to which the MOS (outside of the "Article titles" section) applies to titles is debatable.

In summary, my considered judgement is that on this particular issue, the first three of these authorities provides little useful guidance. They don't really prescribe or even suggest to the level of detail sufficient to answer the question we're asking here: should, or should not, there be a comma in a particular case in a a particular circumstance. The fourth authority WP:COMMA does provide guidance to this level of detail, but its applicability to titles is clearly subject to debate.

Let's look at the policy. Article titles opens with
 * "[I]t is sometimes necessary to add distinguishing information, often in the form of a description in parentheses after the name. Generally, article titles are based on what the subject is called in reliable sources". The first section ("Deciding on an article title") reiterates this by opening with "Article titles are based on how reliable English-language sources refer to the article's subject". This in turn is based on the admonition that article titles should be formed in light of "other policies, particularly the core content policies Verifiability and No original research"

So article titles are based on what the subject is called in reliable sources, and we're not supposed to make up our own titles for entities (original research) or use titles that the reader can't independently verify is the actual correct name. This supersedes everything. If and only if what the subject is called in reliable sources is different in different sources (and there's not a clear and overwhelming preponderance in favor of one form over others) is there any need to continue looking into the issue or reading any further into WP:AT. That's what WP:AT says.

However, looking at our actual practice, it's clear that it's not that cut and dried for parenthetical disambiguation. But while preponderance of sources is (according to my reading) not, as a matter of Wikipedia practice, always a case-closed end-of-discussion point for disambiguating terms, it is true that preponderance of sources is something to pay strong attention to. You can't just wave away what the policy says.

OK, so at Wikipedia talk:Naming conventions (geographic names) it looks the preponderance of sources says to use two commas, according to some exhaustive and highly laudable research by User:Dicklyon (thank you Dicklyon!). However, this research covers what style guides and prescriptive grammars say. When it comes to what people actually write in newspapers and official documents and so forth, the answer is a less clear, and it looks like people ignore the style guides and grammars to some extent. My inclination is to take actual usage in text as the reliable sources.

By the way, the argument that this applies only to body text is pretty much knocked for six because the MOS says "The guidance contained elsewhere in the MoS, particularly in the section below on punctuation, applies to all parts of an article, including the title." (emphasis added).

(This leads to some contradictory prescriptions, as when the MOS says "Official names (of companies, organizations, or places) should not be altered" but WP:AT (in its WP:COMMONNAME section) says "Wikipedia does not necessarily use the subject's 'official' name as an article title; it prefers to use the name that is most frequently used to refer to the subject in English-language reliable sources." In cases of conflict, I assume that WP:AT is the operative authority, because it's a policy and because it's specifically tailored for titles. But there's no conflict here because neither the MOS or WP:AT address our specific issue.

(The MOS also says "Places should generally be referred to consistently by the same name as in the title of their article". This doesn't apply to us because adding or omitting a second comma doesn't really change the name of the entity.)

Anyway, User:Huwmanbeing and some others make a cogent and impassioned argument that we have to use common sense and not take this too far. The point was made that our article titles violate many rules that we apply to body text (by not containing verbs or ending with a period, for instance), which is true, and this shows a de facto flexibility on the matter (which is then further avowed to be beneficial). I'm constitutionally amenable to that point -- I don't think our rules were meant to be read as a fundamentalist might read the Bible -- but I'm not willing to just wave away the clear statement "The guidance... applies to all parts of an article, including the title.", so I'm somewhere in the middle on this I guess.

But there is also the argument that our other disambiguation policies carry some weight of precedent -- a great deal of weight, actually. Regarding the latter, User:Frungi makes the case "We don’t consult reliable sources for how to title articles for disambiguation... no reliable source uses the terms “Jumper (film)”, “Jumper (novel)”, “jumper (dress)”, “jumper (computing)”, etc. We have our own conventions for that, and I see no reason that they shouldn’t apply to ambiguous X-politan area names as well."

That's a fair point. Since it's a fair point, let's at least look further down into the policy.

So we then find in WP:AT (in the lede) "When [reliable sources] offer multiple possibilities, editors choose among them by considering several principles: the ideal article title
 * resembles titles for similar articles,
 * precisely identifies the subject, and
 * is short, natural, and recognizable."

and the first section ("Deciding on an article title") expands on this a little, with five bolded bullet points: Recognizability, Naturalness, Precision, Conciseness, and Consistency with other articles of the same type (for which we are then pointed to this page (WP:PLACE) for details on, and I assume only on, this last point (consistency)).

None of these five bolded points are "Grammatical correctness" nor is there any prescription that article titles should be complete sentences or be congruent with normal rules for grammar in text. ("Naturalness" here means "[use what] readers are likely to look for" rather than natural prose stytle and "Precision" is about helping the reader "unambiguously identify the article's subject", not grammatical precision). It doesn't say "ignore rules of grammar" either. It doesn't say anything.

I think a fair assumption is that the prescription of use the name generally used in sources is to serve the important virtue of recognizability, and the other four virtues too, on the macro scale. And so really the prescription "follow reliable sources" is written to handle big differences -- whether to call an article New Orleans or The Big Easy or nouvelle-Orléans -- and not really to dictate small formatting details. The precise formatting used by the preponderance of sources is not something to completely ignore either, it's just the intent of the rule as I read it is not to make this the sole controlling factor.

Drilling down through each of the sections: "Use commonly recognizable names" (WP:COMMONNAME) -- again, I note that all of the many examples give much grosser differences than what we are talking about here -- "Use Guinea pig and not Cavia porcellus" and so on. There're no examples bearing on fine-tuning via comma placement or whatever. The editors who wrote the rule could have included some, but they didn't, and I think that that says something.

And here's something interesting: " Wikipedia prefers the name that is most commonly used (as determined by its prevalence in reliable English-language sources) as such names will be the most recognizable and the most natural." (emphasis added.) To stretch that out and informalize it a bit, the passage could be restated like this: "Remember when we said 'article titles are based on what the subject is called in reliable sources'? Well, we mainly did that because what we really want is the 'most recognizable and most natural' titles, and reliable sources almost by definition provide that. (After all, it's practically impossible to image a case where a clear preponderance of many reliable sources called a place by a certain name, yet significant numbers of readers did not recognize that name.)" This makes sense, and it explains an important reason why we go by preponderance of sources, I think.

So. There are two reasons for using commonly recognizable names:
 * 1) So the reader can find the page, ("natural"), and
 * 2) So the reader, when she has found the page, quickly and accurately perceives what it's about ("recognizable").

For the first purpose, I think that the addition or subtraction of a comma makes essentially no difference given the way our search engine works. It doesn't exactly ignore commas, but the autocomplete on "Columbus, Ohio, metrop" and "Columbus, Ohio metrop" (as well as "Columbus Ohio metrop") are identical, and I think this would be true in 100% of cases or very nearly so. (This is not even considering that, after all, we do have hard redirects.) I think that Google can't easily tell these forms apart either.

For the second purpose, I think that the addition or subtraction of a comma makes very little difference but not absolutely no difference. When a reader accesses the page about the Columbus (in Ohio) metropolitan area, we don't want her to think that maybe she's looking at a page about the Columbus (in Georgia) metropolitan area, or just the city of Columbus proper, or about a city called Columbus in the state of Ohio Metropolitan Area, or a neighborhood in Columbus called the Ohio Metropolitan Area, or any other wrong thing. If these last two seem silly, consider the foreign reader who has an indifferent grasp of English and isn't familiar with the names of our states and so on.

As a practical matter, I'm skeptical about how much the addition or subtraction of a comma is going to matter much. Any comma placement will suffice to differentiate between Columbus in Ohio and Columbus in Georgia, and no comma placement is going to much help the person who thinks that that maybe "Ohio metropolitan area" is the name of a state or a neighborhood.

However, "not much" is not exactly "none", and there is a difference between implying (Columbus)(Ohio metropolitan area) and (Columbus)(Ohio)(metropolitan area), and it mattersw. But then, since we're dealing with a person who's all at sea regarding American geography and English, "Columbus, Ohio, metropolitan area" could be taken to mean something similar to "Columbus, Ohio, great lakes region" (which would be a logical progression in size of entities) -- maybe "metropolitan area" is what Americans call the geographic core of the country and Ohio is a subset of that (see Metropolitan France for instance).

All in all, I'd avoid overemphasizing how much technical grammar structures are going to be helpfully understood by a person in this predicament, and as a practical matter, the person probably going to have to read the first sentence, which is "The Columbus Metropolitan Area is the metropolitan area centered on the American city of Columbus, Ohio".

So to some extent we're grasping at straws here. The fact is that all the solutions, suggested or imaginable, have flaws and are subject to various misinterpretations. If there was a perfect solution we wouldn't be here. We do want to find the solution that is least subject to misinterpretation.

Final note, we're advised via Official names to not pay much heed to legal names, which doesn't really apply here anyway.

"Neutrality in article titles" WP:POVTITLE
 * Doesn't apply.

"Explicit conventions" WP:MOSAT
 * That takes us to this page (Naming conventions (geographic names)) of course. We'll look at that closely later, keeping in mind that it's a guideline not a policy.

"Precision and disambiguation" WP:PRECISION
 * It says that if you can't use natural disambiguation (which we can't), then "add a disambiguating term in parentheses"; but then "With place names, if the disambiguating term is a higher-level administrative division, it is often separated using a comma instead of parentheses" (emphasis added). Note that "metropolitan area" is not a higher-level administrative division. WP:CONCISE says "be concise" and I suppose that an extra comma makes a title a tiny bit less concise.


 * If you take this section straight up literally, my reading is that you might end up with "Columbus metropolitan area (Ohio)", which a couple of editors did suggest -- for instance User:Frungi says (in another discussion) "Rochester metropolitan area (New York); Rochester metropolitan area (Minnesota). This is consistent with parenthetical disambiguation across Wikipedia". He's right, I believe. That didn't gain much traction and isn't on table, though. And by precedent interest groups are permitted to set their own standards -- for instance ship names are usually in the format Soviet patrol boat Tuman rather than Tuman (Soviet patrol boat) per an decision made long ago (dunno why) by WikiProject Ships. It may be the geography-oriented folks just like commas, or it may be just random chance that "Columbus metropolitan area, Ohio" is on the table and "Columbus metropolitan area (Ohio)" isn't. They're not substantively different and even if it made a difference its too late to worry about it now.

The rest of the page, English-language titles WP:UE and on down, doesn't apply to any issues in contention here.

The question of grammatical correctness came up (and also the denial that it much matters for titles), so here I'll lay out my understanding of the matter just so people know where I'm coming from. It's not important and you can skip it.

According to our article Apposition, "Apposition is a figure of speech of the scheme type [a figure of speech that changes the normal arrangement of words in a sentence's structure], and often results when the verbs (particularly verbs of being) in supporting clauses are eliminated to produce shorter descriptive phrases." (It then goes on to double down on the "departure from standard word order" bit.)

I do see "Columbus Ohio" as having eliminated the verb of being, as being shorthand for "Columbus which is in Ohio" or something like that. I'm not seeing there necessarily being any abnormal word order, but then I don't see most of the examples given as showing this either, so maybe that's not crucial or maybe I'm just not seeing it.

If "Columbus Ohio" is an apposition, then "Ohio" is a restrictive appositive, because it specifies exactly which Columbus is being spoken of -- "it limits or clarifies the foregoing one [Columbus] in some crucial way". (A non-restrictive appositive would be something like "Columbus, a beautiful city", I think.)

The article Apposition then reads "In English, non-restrictive appositives are typically preceded or set off by commas, while restrictive appositives are not set off by commas." so thus we have "Bill's friend Alice said... but "Bill's friend, who is pretty smart, said..." requiring the commas (non-restrictive appositive). It gets rather involved.

However, as a matter of actual practice, "Columbus, Ohio" is very common, indicating that it's not an apposition, or it's an exception to the usual punctuation for restrictive appositives, or that as a matter of actual practice people treat "Ohio" as a general parenthetical phrase rather than an appositive. And if it's a general parenthetical phrase it would have to be enclosed by commas, giving "Columbus, Ohio, xxx" for any value of xxx.

And this is just what our Manual of Style (at MOS:COMMA) says specifically for geographical names: "In geographical references... a comma separates each element and follows the last element... the last element is treated as parenthetic."

Further down, User:Dohn joe points to Garner's Modern American Usage to make the case that that "it's not an appositive, really. An appositive 'points out the same person or thing by a different name'...". (If that's true our article (which says that one element serves to "define or modify" the other) is wrong, I guess.) Anyway, following some more points, he concludes that "'City, State'... is itself a single, unified name of a thing - that happens to have a comma included by convention. As such, 'State' does not need a second comma to set it off, because 'State' is an integral part of the name itself".

At any rate, it seems open to interpretation just what "Columbus, Ohio" is, so I'd go with MOS:COMMA, which gives a specific example of exactly the thing we're talking about here, but just keeping in mind that the question of what is "correct grammar" is a little involved here.

Anyway, the other virtues valorized by WP:AT are way more important if they compete with grammatical correctness. For instance, to use an extreme example, I would find the argument "I grant that this title will make it hard for people to find the article or quickly figure out what it's about if they do find it, but its grammatically correct so let's use it" to be quite weak.

That's my personal opinion, which I'll now proceed to basically ignore, since most editors do appear to believe that the example at WP:COMMA represents the correct grammar, and a number of editors do believe that it's important, and since it's a essentially a matter of opinion it's essentially a matter of headcount on this issue.

Survey, and headcount
Italics indicates that that's the editors second choice.

Supporting one comma:
 * User:Dohn joe. Second comma is awkward.
 * User:MelanieN. Reliable sources differ. Argument either way not sufficient to change status quo (which is one comma).
 * User:SarekOfVulcan Titles are not sentences, so appositive rules may not apply.
 * User:Omnedon. Titles are not sentences, clarity is key, and ll in all one comma does this best.
 * User:Huwmanbeing. Per others. Second comma is pedantry. Making a big change for pedantic reasons not called for.
 * User:BDD. Ditto Huwmanbeing.
 * User:Ntsimp. It's not an appositive. Second comma after the state rarely called for.
 * User:ὁ οἶστρος. Because "metropolitan area" is referring to the whole term "Columbus, Ohio", it's (Columbus Ohio)(metro area) and one comma is the best way to show this.

Supporting two commas:
 * User:Dicklyon. Per accepted style guides including our own, state in middle takes commas on both sides. (Open to other constructs which obviate the issue.)
 * User:sroc. Per style guides on proper construction of parenthetical clauses. (Open to other constructs which obviate the issue.)
 * User:HandsomeFella. Per obvious sentence grammar. (Editor tentatively supported (in the discussion sections) "Columbus metropolitan area, Ohio" type structure as a second choice at least better than the one-comma version.)
 * User:Bkonrad. Good grammar.
 * User:Arthur Rubin. State name is parenthetical, which requires that it be set off by two punctuation marks.
 * User:LtPowers. Longstanding grammatical practice and improves clarity.
 * User:Frungi. Per standard English grammar and MOS:COMMA.
 * User:Tony1. Grammar per others, avoids ambiguity.
 * User:Agnosticaphid. It's both wrong and confusing to omit the second comma. ESL users will be confused if its omitted.
 * User:Reify-tech. It's clearer and that's key. (Open to other constructs which obviate the issue.)
 * User:Stfg. More precise and clearer.
 * User:TCN7JM. Proper grammar and clearer.
 * User:Orlady. Grammatically correct. Clearer: this discussion and others shows some are confused by one-comma form.

Supporting "Columbus metropolitan area, Ohio" type structure, as first choice:
 * User:Ohconfucius. Avoid comma overload.
 * User:Frungi. This, or a variant using parentheses, is by far the best option.
 * User:Stfg. The 2-comma form is clunky and hard to parse, while the 1-comma form is misleading.
 * User:sroc. Avoid awkward construction.
 * User:Waggers. Makes much more sense, especially to English speakers outside of the USA.
 * User:No such user. Ditto Waggers.
 * User:Agnosticaphid. Least awkward. Some reservations but still best option.
 * User:Ronan McGurrin. Much more unambiguous and clear cut option than slapping on more commas.
 * User:Chris troutman. Best option.
 * User:David Eppstein. It's the city that names the metro area, not the state, so why separate the two parts of the metro area's name?
 * User talk:76.65.128.222. Preferable, it gets rid of arguments over one or two commas by displacing the whole state. "City (State) metropolitan area" or "Metropolitan area of City, State" OK too.
 * User talk:71.21.81.61. Most logical option. [N.B.: redlinked IP, but sufficient edit history to have standing.]
 * User:Ezhiki. Solves both problems at once.

(Specifically opposed to "Columbus metropolitan area, Ohio" type structure)
 * User:Nyttend. Fails common name test, people don't talk or write like that.

Here's a somewhat recent discussion which is germane regarding, and only regarding, article titles, but not limited to those about metropolitan areas: Wikipedia talk:Article titles/Archive 42 There's a huge of amount of discussion here. Furthermore, there are pointers to many other discussions which I looked at but which I'm not really going to bring in here, there's got to be cutoff somewhere.

The example of Riverside, Buffalo, New York (a neighborhood in Buffalo) is pointed out. User:Born2cycle makes the point: "I don't know of any precedence or reason to use 'common natural sentence form' in article titles". Note that he's talking about precedent and reason, not what our rules say. It is true that the three-word phrase ("including the title") that ties the MOS to article titles doesn't give any reason for doing that; and there's no objective reason that I can think of offhand, and really no one offered a strong objective benefit during the long discussions, and I don't think it's possible to do so; and as thinking persons we're permitted to heavily discount sentence fragments that may be the work of a few people long ago and for which no objective demonstration of benefit may be obtained. That's my reading of Born2cycle's point (OK I expanded it a little), and I think it's a fair point.

As far as individual editors (not included ones who commented in this current RfC) I get:
 * User:Peter coxhead. Two commas, the article title should be styled the same way as it would be in text.
 * User:Born2cycle. Neither, use parens e.g. "Columbus metropolitan area (Ohio)". [Editor did not say if this would extend to "Columbus train disaster (Ohio)".]
 * User:Blueboar. Ditto Born2cycle. Also suggested "Columbus, Ohio (metropolitan area)".
 * User:BD2412. How about "Columbus (Ohio metropolitan area)".
 * User:DanielPenfield. Points to here with approval: "When the state or country's name becomes part of a compound structure, the second comma is dropped". This editor had many other things to say, and I think he was clear enough for me to count him as a one-comma guy. Hope that's right.

Here's a recent discussion which is germane regarding, and only regarding, the use of the phrase "metropolitan area" (and close cognates I assume): and I'll include it too: Wikipedia talk:Naming conventions (geographic names)/Archives/2013/August. It's specifically and only about metropolitan areas.
 * User:Apteva expressed support for the one comma rubrick.

36 respondents. For now, I'm going to put aside User:BD2412, as he had a (useful) suggestion different from the ones in play (and different enough that I can't guess if he'd support one of the others.) For now I'm going to put aside User:Nyttend as he just had an oppose to one option, don't know what he actually supports. That leaves 34. User:Born2cycle and User:Blueboar had early suggestion of "Columbus metropolitan area (Ohio)" which in my judgement is close enough to (the later-suggested) ("Columbus metropolitan area, Ohio" that I'll count them in that camp. This gives, counting first choice only:
 * 9 for "Columbus, Ohio metropolitan area"
 * 10 for "Columbus, Ohio, metropolitan area"
 * 15 for "Columbus metropolitan area, Ohio" (Also 1 specific oppose.)

34 respondents divided 9/10/15 is 26%/29%/44%. Obviously the more choices the more diffuse your spread, and it's hard enough to get supermajority with two viable options on the table and near impossible with three. On the other hand, refusing to stick to a binary vote is good, and I'm not inclined to throw up my hands and say "no consensus possible!" just because people were flexible enough to not be binary. So let's see if we can tease something out of this. I feel I do have to try to do this because a lot of this comes down to opinion. I've done closes where most of the work is figuring out which of many applicable policies apply most strongly, and this ain't one. I need to try tease out a feeling of where people are at for this one.

There's a number of ways to look at this. One is "Columbus metropolitan area, Ohio" has about a 60% to %40 plurality over any single other option, but I don't think that means much. The two less popular options have near identical support, so something along the lines of "remove the least popular single option and refactor the survey" would be too close to "remove an option at random". Instead, I'm going to very loosely group the "Columbus, Ohio, metropolitan area" and "Columbus metropolitan area, Ohio" together for bit. Here and in subsequent exposition I'm going to call these these the "correct grammar faction".

My reason for doing that is that several of the "Columbus metropolitan area, Ohio" folks had "Columbus, Ohio, metropolitan area" as their second choice, but none had "Columbus, Ohio metropolitan area" as their second choice. (Put another way, none of the one-comma folks migrated to "Columbus metropolitan area, Ohio" but several of the two-comma folks did). This, and just the general tone of argument and what seems important to people, leads me to feel that that's how to do it. I just get the vibe that a lot of people who aren't on the one-comma boat are appalled by it, that to the extent that there are two camps, they would be "grammar doesn't apply to titles" vs. "grammar must be correct, however it's done".

OK, so this leaves 9/25 or 26%/74% for the two positions "grammar not so important/correct grammar is important" (or however you want to call it), so we knock out the one-comma folks, and when the remaining unstable compound splits apart we have 10/15 or 40%/60% for "Columbus metropolitan area, Ohio". With these numbers, and on this topic, that's just barely enough to consider a supermajority in play. Certainly not even remotely enough of one to say "case closed" but enough so that I consider headcount a factor.

Anyone who doesn't like my reasoning, we can talk. I'm doing my best to tease some forward movement out of this inchoate and widely-spread series of discussions. Giving up is not high on my list of options.

There was a an attempt at further polling in the "poll" section, which used a first-second-third-fourth-choice paradigm, which is probably a good idea when there are more than two options on the table, but not enough participation to count for much. "Metropolitan area of City, State" was brought up an alternative here, and (who knows?) might be the best form, but hasn't gained much traction.

Strength of argument
OK, moving forward to the discussion.
 * Point: Single comma not permitted by by rules of grammer.
 * Counterpoint: Grammar rules don't apply to titles.
 * How I weighted this factor: A fair amount, in favor of either the two-comma version or the "Columbus metropolitan area, Ohio" version, either of which is viewed as preferable to the one-comma version from the standpoint of pure grammar, I think. The MOS does say in black and white "The guidance contained elsewhere in the MoS, particularly in the section below on punctuation, applies to all parts of an article, including the title.", with the "guidance contained elsewhere" including "In geographical references that include multiple levels of subordinate divisions... a comma separates each element and follows the last element (except at the end of a sentence)" and the example of correct use is "Cooper hijacked a Boeing 727 aircraft that had taken off from Portland, Oregon, and was...".


 * There was some discussion of cases such as "Columbus OH" but we don't use those kinds of abbreviations in article title; and many other points of grammatical detail.


 * Point: Two commas is clearer.
 * Counterpoint: No it's not.
 * Point: One comma is clearer.
 * Counterpoint: No it's not.
 * Point: Putting "Columbus metropolitan area" all together, unbroken by punctuation, is the clearest construction.
 * Counterpoint: No it's not.
 * How I weighted this factor: Somewhat, and in favor of "Columbus metropolitan area, Ohio". Couple things are clear to me: 1) there's no perfect solution here -- each option offers chances for misinterpretation -- and 2) absent an extensive survey of a broad reader population it's not possible to know which kinds of misinterpretations readers are going to have, particularly readers with low knowledge of American geography and maybe of English language conventions. (I discussed both of these points more extensively above.)


 * Because of that, I didn't give a huge amount of weight to this factor. If there was a clear "winner" in terms of clarity, that'd be a huge factor. After reading through the discussion, I think that the point that the entity is (Columbus)(Ohio)(metropolitan area), not (Columbus)(Ohio metropolitan area), is a reasonable point of grammatical logic for what that's worth and militates a bit for the two-comma version, but the point that "Columbus metropolitan area, Ohio" is maybe just a little bit clearer because it puts the actual topic of the article (Columbus metropolitan area) at the front, unbroken by any punctuation, probably has something to recommend it.


 * Point: as WP:TITLECHANGES says, "Changing one controversial title to another is strongly discouraged. If an article title has been stable for a long time, and there is no good reason to change it, it should not be changed."
 * How I weighted this factor: Zero. There is a good reason to change these titles, mainly that the situation is contentious and inherently unstable. There were many many arguments for the various versions, it's not like this is somebody's not-well-thought-out whim. And absent a controlling decision the situation will remain unstable. Not an option.


 * Point re actual usage (as opposed to prescribed usage in grammar guides and style guides): User:MelanieN writes "Reliable sources (not counting the US government, which uses state abbreviations) are split: in a search for half a dozen MSAs where disambiguation is needed, I found that the comma is usually added in formal or legal situations ('Portland, Oregon, Metropolitan Statistical Area'), but omitted in newspapers and general usage ('Portland, Oregon Metropolitan Statistical Area')". Later she writes "What I found was the second comma is commonly used in legal or official use, such as the name of a legislative action or a term in a lawsuit. On the other hand here's the EPA: 'Portland, Oregon metropolitan area'. But I also found that professional journalists find ways to write around it - not to use the actual phrase 'Portland, Oregon statistical area' but rather things like 'the statistical area of Portland, Oregon' - but when they do use the phrase 'Portland, Oregon statistical area' they generally omit the second comma. Example 'Portland–South Portland–Biddeford, Maine Metropolitan Statistical Area'. '[G]reater Portland, Oregon metropolitan area'. On the other hand here's CNN: 'the Portland, Oregon, metropolitan area'... The more I look, the more I find no pattern at all; some Reliable Sources use the second comma, some don't." There was a great deal of back-and-forth on this, some of it dealing with the difference of usage in formal and legal documents vs. news stories, and much else. At the end of the day I feel reasonably comfortable stating that neither version (one-comma or two-comma) can be shown to be clearly preponderant over the other.
 * How I weighed this factor: a lot, to the effect of "reliable sources is not really a factor here" since there's no clear preponderance. And thank you MelanieN! (And Frungi, sroc, and Dicklyon, and everyone else!) As I pointed out above, this actual-usage info carries a lot more weight to my mind than style guides. As I expounded on above, I'm already inclined to take the "use reliable sources" prescription as weak in this case because 1) it's not intended to deal with punctuation, and 2) it's somewhat negated by our de facto actual methods of disambiguating titles. And since there's no clear preponderance anyway, it's basically off the table.


 * User:Nyttend points out a pretty knotty problem, taking Columbus in Georgia as an example. The metro area spans two states. "Columbus metropolitan area, Georgia-Alabama" (or "Columbus metropolitan area, Georgia and Alabama") is not satisfactory because it doesn't define what state Columbus is in, and the implication is that that's necessary (or at least useful) info. There wasn't really any counterpoint given to this.
 * How I weighed this: Quite strongly, in favor of either the one-comma or two-comma versions, because (since we have redirects to help searches) the single most important virtue for a title is to make it instantly clear to the reader exactly what she's looking at. Nyttend argues that using the "[city] 'metropolitan area' [state(s)]" construct works against this when multiple states are involved. He might be wrong, but he might be right, so its a good point. It wasn't refuted.


 * I'll just note that "Columbus metropolitan area, Georgia-Alabama" contains different information than does "Columbus, Georgia, metropolitan area" -- not saying one is better, it's just different info. The first doesn't tell us what state the city of Columbus is in, but it does tell us all the states the metro area is in. The second does tell what state the city of Columbus is in but it doesn't tell us what states the metro area is in. Which is preferable might depend on the individual reader, and which of these mindsets the term "Columbus metropolitan area" puts her in (I'll exaggerate a bit for effect):
 * "Here's an article about Columbus in Georgia, and also (by the way) the contiguous large densely populated area which is technically outside the city limits."
 * "Here's an article about a large densely populated area sprawling over several counties, which (by the way) is named for the large city at its center, Columbus in Georgia."
 * I don't have the answer to that and I guess no one does.


 * The point is made that two commas is awkward, and this is bad. The general prescription is to try to avoid the situation by rephrasing. User:Dicklyon again did yeoman work in looking at various style guides, essentially proving that it is awkward or commonly seen as awkward, at least in style guides. For this quality (awkwardness) style guides carry a fair amount of weight, I think.


 * There's a great deal of discussion in the "Awkward" section, regarding different standards for titles as opposed to body text, and much else. Gets into some closely-argued points of grammar. Interesting and well argued from all points, but not necessarily always bearing directly on the subject of awkwardness. I saw Dicklyon's style guide cites as carrying a lot of weight here.


 * How I weighed this: Somewhat (not a whole lot), in favor of the "[city] 'metropolitan area' [state(s)]" which avoids the one-comma/two-comma problem, and secondarily in favor of the one-comma version which breaks up the flow of the title phrase a little less. It's not nearly as important as clarity, but it's not nothing.


 * In the "Articles and categories affected" section, an editor (User:Apteva, I think) provides a list of usage of in titles. Very good work, Apteva, and useful, and thank you! One thing this shows is precedent. According to Apteva's list, which is certainly substantial but may not be exhaustive, the preponderance of one-comma to two-comma version is 100 to 3, not counting categories and not counting articles which don't use the term "metropolitan area" or something close. That's a large preponderance. (BTW of these 103, only three include the name of more than one city (Canton–Massillon, Ohio; Harrisburg–Carlisle, Pennsylvania; and North Port–Sarasota–Bradenton, Florida).)


 * (User:Apteva then provides a subsection, "Other examples of incorrect usage", which shows some use of the one-comma form. These don't include the term "metropolitan area" and (by intention) aren't balanced by examples of the two-comma form, so not very germane here.)


 * There's more discussion, around the concepts of "correct" vs. "incorrect". User:MelanieN makes the point "English is a dynamic language. Grammar rules change over time through usage" and "What Dicklyon called 'just an informality' is actually an English grammar rule in the process of changing", which I think is well put. The descriptive-vs-proscriptive approach to grammar is never settled, and I think that 100 years ago the prescriptive camp may have held the prevailed, but in this world in these times the descriptive camp has the whip hand, as a matter of practical fact, I think.


 * All this means is to to me personally prescriptive grammar points are not usually convincing so I don't give them too much "weight of argument" merit. For headcount, its totally different and what I think has no bearing: if a lot of Wikipedians are convinced by prescriptive-grammar arguments, then it matters -- a lot, in this case.


 * Getting back to User:Apteva's list, I think it shows two things:
 * To this point, the Wikipedia's editors have "voted with their feet" to go with the one-comma version, and thereby established a precedent, and
 * Changing away from the one-comma version will require a fair amount of work.


 * How I weighed this, for the first point (established precedent): A little -- those 100 titles were created by thinking humans who knew what they were doing -- but only a little. The whole point of this exercise is to establish a new precedent going forward (or sustain and codify the existing one), for the purpose of having the best encyclopedia in the future. It may be that the editors who created 100 one-comma version titles were doing so based on a long and closely-reasoned discussion, or they may have been blindly following a precedent established by whim or at random, or it may have been a precedent originally established from a single source (such as when a bot (User:Rambot) used U.S. Census Bureau records to create many articles back in 2002), or whatever. Not having any evidence at hand of why this situation exists, and since this is a long and closely reasoned discussion, to my mind it takes precedence over existing precedent. We're not slaves to the past.
 * How I weighed this, for the second point (workload): Essentially nil. We're talking about a few tens of man-hours or whatever. I'd certainly not be inclined to let an unsatisfactory situation continue to exist for another ten years or longer to avoid a bit of work now.


 * After a great deal more discussion -- again mostly centered around concepts of grammatical correctness -- The suggestion of the form "Columbus metropolitan area, Ohio" is proposed (actually User:Bkonrad and User:Frungi expressed the opinion that "Columbus metropolitan area (Ohio)" is even better; but User:BDD opined "This form seems to be an invention on our part, which should be avoided at all costs".

A Recently closed RfC
Concurrently with this discussion, there was this RfC, opened on 25 October 2013 and closed (by the proposer) on 8 November 2013, and specifically mentioning WP:PLACE as a page to affecte: The proposal was The proposal didn't pass.
 * Wikipedia talk:Manual of Style
 * 1) Instead of being mandated or not being mandated, MOS should be changed to make it optional as to whether a comma should be included after a date or a place: "Rochester, New York metropolitan" area or "Rochester, New York, metropolitan area", for instance, both being acceptable.
 * 2) If passed, then all pages within a certain Project be the same. This means that the comma would be optional on the "national" standpoint, but one version would be mandated by a discussion from a "local" Project standpoint (all pages within any Project either with the comma or without the comma). This allows members of a project to decide themselves on what to do.
 * 3) This apply to titles and not sentences within an article (a clarification which was made only after the discussion had been live for awhile).

(If one thing's clear from all this, it's that the proposition "Whatever forms are prescribed for article body text automatically applies to titles also" is debatable and not at all universally accepted. So therefore changes to article titles have to be implemented at WP:AT, the "Article titles" section of WP:MOS, and (for places) WP:PLACE. Changes made anywhere else don't apply to titles unless this is specifically stated. It's either at WP:AT or WP:PLACE (or WP:MOS, but only if a change to the "Article titles" section is specifically proposed) that such proposals should be made. It's unfortunate that this proposal was made at WP:MOS and wasn't clearly specified to apply only to titles from the get-go.)

There was some discussion at that RfC, including some of the same editors from this discussion and going over some of the same ground, specifically points of grammar. Since this was bound up with the question "individual projects should have the controlling authority", and also the proposal and its examples included dates, nor did it seem likely that "metropolitan area" was considered a special case, so the applicability of this to our discussion here is limited. FWIW, and not counting editors whom I've already counted here, I get:
 * Support -- "My personal preference is not to punctuate unless it makes it easier for the reader" (but individual projects should decide) -- User:Tony1
 * Support -- "There is no reason to require the second comma" -- User:Neljack
 * Support -- User:SlimVirgin had just "Support per Dohn joe". Again, it's hard to tell if she was mainly supporting the idea that individual projects should decide, or what. But since User:Dohn joe had summarized his lengthy (and cogent) remarks as "We should not mandate an awkward second comma where it is not syntactically required, especially when omitting it is already WP practice, and when omitting it is recommended by a major style guide such as Garner" I gather she's supporting at least the option to go with the one-comma form.
 * Oppose -- "These commas are needed for clarity, and that's why they are required by standard rules of punctuation" -- User:Orlady
 * Oppose -- "If the correctly-punctuated result looks awkward, then rephrase." -- User:Odysseus1479
 * Oppose -- "Although the construction with more commas can be awkward at times, it is often needed for clarity, which is more important than the secondary goal of smooth readability. Perhaps the MoS should recommend recasting a sentence if it is awkward" User:Reify-tech

There was some discussion of, and support for, avoiding the situation by recasting the title (I assume in some manner of the "Columbus metropolitan area, Ohio" form for instance, although this particular form wasn't mentioned). User:Nyttend suggested "PLACENAME, PLACENAME be in a title is [only] when that's the entire name of the article, e.g. Monowi, Nebraska, or when it's at the end of the name, e.g. History of Charleston, South Carolina. We entirely avoid the issue in this manner" which is a useful thought. User:Nyttend has a lot of cogent and useful thoughts, but they don't seem to gain much traction for whatever reason.

Anyway, proposal failed, I can see three new faces in the one-comma camp, one in the two-comma camp, and two in the rephrase-it camp. This discussion was on a different enough premise that I didn't formally add these into the headcount, though.

An open RfC
Here, the proposal is to keep the same rules for commas, but add the advice to recast the sentence to avoid the one-comma/two-comma problem.
 * Wikipedia talk:Manual of Style

It's specifically about article text, but of course ideally WP:COMMA should match how our titles. (That'd be ideal but not absolutely necessary, as long as any differences are spelt out at at the places where titles are described -- WP:AT, the "Article titles" section of WP:MOS, and (for places) WP:PLACE -- using footnotes as necessary, as some rules pages do.)

Its stil open and right now its running 12 in support, 7 opposing (61% support), so its up in the air I guess. Fortunately, based on how I've close this discussion, it won't affect this close. Since this close is basically "recast", then if that proposal passes is just reinforces this one. If it fails it doesn't matter since we've carved out an exception for metropolitan areas and we'll support that with appropriate footnotes in the appropriate places.

Summary
OK then. Here's what we've got. Points that devolve to a policy or guideline are shown in red lettering. Points that refer to the discussion are shown in normal black font.


 * For the two-comma version ("Columbus, Ohio, metropolitan area"):
 * Grammatical correctness, which is somewhat important. (I know that some editors don't agree that the one-comma version is incorrect, but that's a minority view.)
 * The rule WP:COMMA, which I judged as only a little important. (My reasons for giving relatively little weight to the policy are described above.)


 * For the one-comma version ("Columbus, Ohio metropolitan area")"
 * Precedent (most titles use this format now), which I judged as only a little important.


 * For the "[city[-ies] metropolitan area, [state[s]]" (e.g. "Columbus metropolitan area, Ohio") form:
 * Headcount, which is somewhat important. (It'd be very important if there was a clear headcount winner, but there wasn't.)
 * Grammatical correctness, which is somewhat important. (I know that some editors don't agree that the one-comma version is incorrect, but that's a minority view.)
 * Clarity, which is somewhat important. (It'd be extremely important if was clearly clearest form, but there wasn't. I judged this form to be a little bit clearer.)
 * Recognizability, per WP:AT, which I judged as only a little important. This is really just restating the above point, clarity, but by reference to a rule rather than the discussion, so this is kind of double-counting. (It'd be extremely important if was clearly most-recognizable form, but there wasn't. I judged this form to be a little more recognizable, at least in cases where the metro area doesn't span two or more states.)
 * The rules spelled out in WP:PRECISION, which I judged as only a little important. (Actually WP:PRECISION would give us "Columbus metropolitan area (Ohio)" if I'm reading them correctly,but "Columbus metropolitan area, Ohio" is the closest match. I didn't give this a lot of weight because it'd be kind of rules-lawyering to take this too literally.)
 * Lack of awkwardness, which I judged as only a little important.
 * As a negative, paucity of sources using this exact form, which I judged as only a little important. (I discussed at length above my relative discounting of sources as a control over how we disambiguate, as opposed to the actual names for things.)
 * As a negative, the point that for state-spanning metro areas the state in which the core city is in is not shown, which, all things considered, I judged as somewhat important. (Note that applies to only a small subset of articles, though.)
 * There's one other thing. The one-comma version makes some people pound their forehead on the table. The two-comma version makes some people claw at the draperies. This is matter where people's opinions matter a lot. I don't get the feeling that there's as much strong negative feeling around this version (except for User:Nyttend). That counts for something I think.

There's no one "right" answer here. All of the suggested solutions have virtues and weakness, and there's no perfect form. All things considered, I think that balancing consensus and strength of the arguments and weighing the existing rules, and even subtracting for its negatives, that the "Columbus metropolitan area, Ohio" form seems most acceptable to the community.

Re en dash for multiple states
Regarding the admonition to use the en dash rather than the word "and" for metro areas spanning two or more states (that is, "Kansas City metropolitan area, Kansas–Missouri" rather than "Kansas City metropolitan area, Kansas and Missouri", this wasn't discussed and I don't want to shove this down people's throats, and you want to have another discussion on this matter go ahead, but my reasoning was:
 * Had to decide something.
 * And the en dash is parallel with how the cities are named. We'd have ""North Port–Sarasota–Bradenton metropolitan area, Florida" and not "North Port and Sarasota and Bradenton metropolitan area, Florida", I think.
 * And -- this is highly arguable -- but "Kansas City metropolitan area, Kansas and Missouri" (for example) might tend just a little more to imply the existence of the entity "Kansas and Missouri" (analogous to Turks and Caicos Islands for instance) than does the en dash. It's probably about the same really.
 * And the word "and" is a little less succinct than the en dash, and WP:AT suggests succinctness.
 * And if any metro areas requiring disambiguation span more than two states, you'll have stuff like "New York metropolitan area, New York and Connecticut and New Jersey" and that's a lot of "and"s. (And I'm pretty sure we don't want "New York metropolitan area, New York, Connecticut and New Jersey"...)
 * And I could be wrong about this, but I think the en dash is a little more common, for our titles and in the general world, than the word "and" for situations like this.