User talk:Andrewa/Primary Topic RfC

Why this page
This is a place to develop a possible RfC along the lines of User:Andrewa/The third draft regarding avoiding primary topic. Andrewa (talk) 00:18, 9 July 2018 (UTC)

See also (oldest first) which contain some interesting examples, but further discussion on the proposal in general should be here for the moment, and on specific examples and scenarios at User talk:Andrewa/P T examples and scenarios. Andrewa (talk) 05:49, 7 October 2018 (UTC)
 * User talk:Andrewa/Let us abolish the whole concept of primary topic
 * User talk:Andrewa/Why primary topic is to be avoided
 * User talk:Andrewa/The third draft regarding avoiding primary topic

A significant step back
In hindsight my original proposal made it too simple, and was not entirely consistent.

Clarifying the Details section made this obvious. So this is the result. It's a significant change. Andrewa (talk) 16:05, 26 July 2018 (UTC)

Specific changes to guidelines and policies
The proposal does not currently include any specific changes to guidelines or policies, only to principles.

WikiProject Disambiguation/Malplaced disambiguation pages is a good place to start thinking about this as it links to several relevant pages and sections.

But perhaps the RfC should not get bogged down in such details. There's an advantage in first discussing and hopefully getting consensus on the desired effects of the changes. That should be specific enough for start. The devil will be in the detailed wording. Andrewa (talk) 18:50, 7 August 2018 (UTC)


 * This will go at User:Andrewa/Primary Topic RfC. Andrewa (talk) 00:32, 12 August 2018 (UTC)

The two things that need to change are: That is enough. It will have consequences. Andrewa (talk) 01:50, 11 June 2019 (UTC)
 * Whenever a new article is created it should be created at an unambiguous name.
 * An article should never be moved to an ambiguous name.

Cracks are appearing
This discussion on WP:MALPLACED is encouraging. Perhaps others will now look deeper, and perhaps then the whole house of cards will do what those do best.

But of course such rethinks are not easily sold. Lots of people have put lots of work into enforcing WP:MALPLACED, let alone WP:P T. It's a big investment to write off. Needs courage.

''If relativity is proved right the Germans will call me a German, the Swiss will call me a Swiss citizen, and the French will call me a great scientist. If relativity is proved wrong the French will call me a Swiss, the Swiss will call me a German and the Germans will call me a Jew.'' - Albert Einstein (scroll to the bottom of the section). Andrewa (talk) 18:35, 8 August 2018 (UTC)

Wikipedia talk:Disambiguation


 * Archived at Wikipedia talk:Disambiguation/Archive 48. Andrewa (talk)

and

Wikipedia talk:Disambiguation


 * Archived at Wikipedia talk:Disambiguation/Archive 50. Andrewa (talk) 09:00, 1 June 2019 (UTC)

both contain some very interesting observations. I'll comment on their relevance here, in due course. Meantime discussion here is always welcome. Andrewa (talk) 01:40, 22 August 2018 (UTC)

And more
Wikipedia talk:Disambiguation

Wikipedia talk:Disambiguation

Interesting. Andrewa (talk) 08:52, 1 June 2019 (UTC)

Two evils
''That leaves us with a choice between two evils: requiring an extra click from readers seeking NYC, or leading those seeking the state to the wrong article. The former must happen a lot but its consequences are minor. The latter would happen less often but may be more severe: NYC is a big page, and there's a real risk that readers may absorb its facts without realising that they pertain to the city rather than the state. I think we've got it right, but neither answer is ideal and it's a judgement call.''

Very interested in discussing concrete scenarios. Andrewa (talk) 22:23, 11 August 2018 (UTC)


 * One concrete scenario is that having a barely primary topic at the base name also tempts editors to get things wrong. I've just fixed a pile of links to musicians who opened for Prince (William or Harry?), bay horses (I hope none drowned) and Australian events on ABC News (via a telephoto lens?), not to mention several black people. Certes (talk) 23:32, 11 August 2018 (UTC)


 * Good point, thank you... there have been several proposals to raise the bar on P T for that and other reasons.
 * But I'm now more interested in the more basic question... does having a topic at an ambiguous base name benefit readers?
 * It is generally assumed (and was by me until recently) that of course it does. But in the actual scenarios I've considered in previous and now largely obsoleted essays, it doesn't. This bears looking into. Andrewa (talk) 00:26, 12 August 2018 (UTC)
 * If you're still looking for concrete examples, there are several at User:Certes/misdirected links. That page lists some cases where the topic at the base name had inappropriate incoming links which were intended for a different article named "base name (qualifier)".  Some page moves may be in order: the world has far more president (corporate title)s than presidents.  I still wouldn't go as far as the essay suggests, but some movement in that direction would be beneficial. Certes (talk) 11:32, 3 September 2018 (UTC)
 * Yes, still looking for examples. Thank you!
 * And very interested in exactly what  movement in that direction would be beneficial. I have stepped back a long, long long way from my original proposal at User:Andrewa/Let us abolish the whole concept of primary topic, and at each step back the proposal has become less trouble to implement, and at no cost to its objectives... sometimes, they've even been better achieved by the more modest proposal. So this is progress.
 * A bit more in that direction and it would happen IMO. Andrewa (talk) 05:20, 7 October 2018 (UTC)
 * User:Certes/misdirected links came from a semi-automated search for pairs where Basename and Basename (qualifier) both had >1000 incoming links. I've since worked through Basename, Qualifer (mainly U.S. places) with smaller but still significant numbers of fixes.  The main new target was Cambridge → Cambridge, Massachusetts with 224 and the main offender was Punjab with 355 split between Punjab, India; Punjab, Pakistan; and Punjab Province (British India).  There are plenty of miscellaneous cases such as Caterpillar → Caterpillar Inc. (88 articles fixed) but they're harder to find.   I'm not saying that all of these should become dabs: I think it's clear that the insect is the encyclopedic primary topic over the company.  But we should certainly err on the side of putting a dab at the base name when editors disagree for the good reason that the term means different things to different people.  Certes (talk) 10:16, 7 October 2018 (UTC)

Raising the bar
has suggested above we should certainly err on the side of putting a dab at the base name when editors disagree for the good reason that the term means different things to different people and further above I still wouldn't go as far as the essay suggests, but some movement in that direction would be beneficial.

There are several problems with just raising the bar on P T.

One is just simplicity... or lack of it. Just how much disagreement on P T is required to establish no consensus? We risk replacing one contentious waste of time with another that may I fear prove equally problematic.

But the overriding one to me is that it doesn't seem necessary to keep this complication anyway. There simply aren't any scenarios in which having a new or moved article at an ambiguous name is of overall benefit to the readers. But the problem still is, this is counter-intuitive, and so many people have so often argued that having the P T at the base name is of benefit to the readers, either by reducing mouse clicks (it generally doesn't) or by making the P T more prominent in search results (it does, but also makes it less recognisable among those results) or by making it easier to link to the article (it does, a very little, but it also makes it far harder to link to the right article).

There is a valid argument for keeping existing articles at their ambiguous base names, to avoid breaking incoming external links, but it should be noted that it does not apply at all to new articles or those being moved anyway. And there is the principle of least astonishment. But this isn't a problem with existing articles, which should be kept at their existing titles anyway for the preceding reason. It would be a consideration for a new encyclopedia starting from scratch, but we are hardly that (and nor is any future fork).

On the plus side, it may be easier to get consensus to raise the bar on P T rather than just deprecate it. Again I'm unconvinced. We have tried before. Andrewa (talk) 17:51, 7 October 2018 (UTC)

has now also said For these reasons, I think we still need primary topics, though I agree that they should be used less often.

I now agree with this. That's why I'm proposing to deprecate P T rather than abolish it (as I previously proposed... but as well as being unsaleable and an enormous amount of work, that turns out to be unnecessary).

You could even see my current proposal as just raising the bar. But I'm proposing to raise it as high as we possibly can.

This seems to me the most logical solution, and also the simplest, and also the most beneficial to readers. (But it may still be unsaleable.) Andrewa (talk) 01:01, 16 October 2018 (UTC)

Support
I agree completely with this essay. Most WP:RMs I have seen which try to define or to change a WP:PTOPIC are fanclub stuff. I would only add:

The risk of accumulating bad links far outweighs the minor nuisance of requiring readers to make one extra click. (Example: the two discussions at Talk:Jethro Tull, in one of which OP participated.)

Page views and Google hits are bad guides both to current and to long-term significance. They should only be used to support other reasoned arguments. (I recall one case where there were three notable 18th- or 19th-century Scottish botanists with the same name. The argument was made that the one who was PTOPIC was indeed PTOPIC because he got more page views than the other two. Well, duh. The reason for that was obvious, and had nothing to do with relative notability.)

WP:TWODABS is a bad argument unless WP:NOPRIMARY is addressed also.

Something which was chosen as PTOPIC when Wikipedia was just starting out is very likely both to be PTOPIC and to remain so. An article which has only recently been written or moved is very unlikely to be PTOPIC. (There are, of course, commonsense exceptions. As a historical example, when Cassius Clay changed his name to Muhammad Ali, that immediately became PTOPIC: he may have been the most famous person in the world at the time. Similarly, I can imagine cases where something which was PTOPIC loses that status and becomes ambiguous. As a possibly hypothetical example – I haven't checked the history – I think there's a good argument that Big Brother (Nineteen Eighty-Four) was PTOPIC until Big Brother (franchise) gained notoriety.)

One of my tests, which I think OP has seen before: Does just about everyone who knows about a lesser topic also know about the PTOPIC? Example: Tetrahedron and Tetrahedron (journal). If so, there is less risk of bad links being created and more chance of bad links being repaired. Narky Blert (talk) 11:07, 3 September 2018 (UTC)


 * P.S. Have a look at the links to Esplanade and Salt Lake City, and see how many you can spot which relate to Kolkata (often made by the sort of editor who seems to think that what Wikipedia really, really needs is local bus routes and timetables)... total nightmare.
 * Those could be cases for WP:RFC. The problems are so persistent that they may be beyond WP:RMs; which would perhaps be poorly attended, or unrepresentative, and in which I might be the only editor who has fixed several hundred bad links. (If there's one bad Kolkata-related link in an article, there's usually a dozen.)
 * I can envisage screams (SLC, pop. 191,000; Utah, 3.1M; USA, 326M), but IMO Salt Lake City, Utah is not WP:PTOPIC. In Kolkata (pop. 4.5M), and probably also throughout West Bengal (91M) and the rest of India (1.3B), the primary meaning is Salt Lake City, Kolkata. I.e., there is no PTOPIC and Salt Lake City should be a DAB page, just like other well-known examples such as Mercury and New York (yes, I remember the arguments over that one). Narky Blert (talk) 21:12, 4 September 2018 (UTC)


 * I'd forgotten that Jethro Tull RM, excellent catch, discuss at User talk:Andrewa/P T examples and scenarios.
 * And lots of other good points, which I'll get to as time allows. Andrewa (talk) 17:04, 7 October 2018 (UTC)

Another supporter
''I'm sorry, as this is most likely not the spot for my comment, but I got lost with the amount of text and sub-headers so I just placed it at the last relevant-ish header. I'm not a fan of any use of Primary Topics. My main editing subjects are TV, film and media related and it always amazes me how hard people fight to get one item a primary over others. This even goes one step ahead and some editors even argue that an article which itself has a qualifier can be a primary of that qualifier - so " (TV series)" can be a primary, over other TV series with the same name. As a current example of this, see Talk:Wheel of Fortune (1952 game show) where Wheel of Fortune (U.S. game show) is the "primary" for any US game show, even though Wheel of Fortune (1952 game show) is also a US game show. I'd be content even with disambiguation pages always being the primary. Yes, even Paris. If a user is searching the search bar, then Paris, France should be either the first or second results (in any adequate search engine). I've also never been moved by the argument of a user needing to click one more page in order to reach where they want to. So what? Will that kill that user? We click thousand of times each day during a normal web experience, why does Wikipedia have to be very "simple" one-click experience? If this ever gets a RfC, please ping me if you remember.''

Lots of good points, and interesting examples. Andrewa (talk) 05:43, 1 November 2018 (UTC)
 * Surely that should be Paris, France (city) to distinguish it from Paris, France (album), Paris, France (film), etc.? Certes (talk) 01:21, 2 November 2018 (UTC)
 * Good point, there has to be a limit. I usually go for “derivative name use for commercial purposes should be weighted very lowly”.  Whatever the rationale, “no” to Paris, France (city).  NB the French city is PT for “Paris” but that shouldn’t influence the titling decision for best title, “Paris” vs “Paris, France”.  Titles are not just for search purposes. —SmokeyJoe (talk) 02:05, 2 November 2018 (UTC)
 * Paris seems to me to be another case of an exiting article that should remain at its base name indefinitely. I currently give mathematics and London as examples of this.
 * So, are there better examples than Paris?
 * Sorry and  for not picking up on this earlier... as I often say, this is on my back burner. But I ask the same... good point that we can't disambiguate forever. But are there any specific examples of the concept of primary topic helping to clarify how we set a limit on this? Andrewa (talk) 11:45, 31 May 2019 (UTC)
 * Paris, France is almost a meme for mocking the American habit of adding the country name as an unnecessary qualifier. No one outside the U.S. ever says or writes that. Certes (talk) 12:18, 31 May 2019 (UTC)
 * Oh, y'all mean Paris ain't part of the US of A no more? They's still listed as part of North American Territory Overseas (NATO), so ah just figured they's a protectorate or somethin', like the UK. 'Splains why they's speakin' such funny English over there, ah could hardly unnerstan' a word of it. (->
 * I see from the history that the article now at Paris was once at Paris, France, so maybe not such a simple example. But was there ever consensus to move it away from Paris? Andrewa (talk) 19:47, 31 May 2019 (UTC)

Simplicity
''Simplicity we could eliminate all primary topics, and use primary redirects. PT (without a redirect) is the simplest way to get a reader to the article, and similarly for writers. It's not just about navigation, but complexity. There may be reasons to increase complexity, but if search needs improving it's easier to improve search than make the structure more complex than needed (that's my gut reaction)''

I replied ''As an RM regular, I can assure you that P T as it currently stands is anything but simple... In my current view, deprecating P T is a very desirable simplification! But as I said, still early days...''

There was subsequent discussion. Later in that discussion:

''...using an ambiguous title for a most likely topic reduces the complexity to that when there's no ambiguity. It's pragmatic.''

I replied ''No. It simplifies things for readers who are (1) looking for that article and (2) unaware that other meanings of that title may be considered primary by others. For all other readers (including of course any who don't agree with us on what the P T is), it complicates things. Most of the damage can be addressed by redirects, as discussed elsewhere. But this isn't simplicity. The simplest method is, have a DAB (or a redirect to a more general DAB) at every ambiguous article title. But that's not necessarily the best method.''

But here is IMO a better place to explore the issues raised. Watch this space. Andrewa (talk) 09:34, 11 October 2018 (UTC)

three things
Well some things you'd need to address...

The benefits are stated (Improve the chance of readers finding the article they want; Reduce the number of mouse clicks required to find the correct article; Reduce the number of article moves; Reduce mislinkings within Wikipedia) but never proven or even really demonstrated. In fact it says that your proposal is "counter intuitive", and after all we have intuition for a good reason. So a counter-intuitive proposal is going to need especially compelling arguments.

Reduce the number of mouse clicks required to find the correct article, well of course (it is not the only virtue for article titles, but it is a virtue). But you don't provide any useful data or even argument to support that your proposal would, indeed, promote this virtue. Improve the chance of readers finding the article they want is even more important, but... again, crickets.

And such benefit as does seem likely true (to me) seem mostly for editors. I personally don't much care if editors' jobs are made easier, we're here for the reader. That being said, making it easier (and therefore less likely) for editors to make wrong wikilinks does help the reader. However, our internal arguments over article moves don't much harm the reader.

But then one example that you give implies that a link to the term "Prince" unadorned is often written, and this is commonly a mislink (the musician is meant), so I think what you're saying is that the article "Prince" should have been named "Prince (title)" if it was being made now. But assuming that people are still going to link to "Prince" in an article referencing the musician, all this will do is take the reader to a Prince disambiguation page where they then must select their desired article (from a list of dozens) rather than taking the reader to the article about the title Prince, where there's a hatnote. Same number of clicks, and the hatnote is certainly less intimidating than a menu page where you have to scroll around to find the musician

Not only that, but that means that a link to "Prince" alone would never be correct (altho I believe a robot will flag the editor's link to a disambiguation page and tell the, and sometimes they will fix it. Sometimes.) 24.107.115.8 (talk) 05:04, 18 October 2018 (UTC)


 * Good points, thank you. Not sure how best to contact you as an anon, your two edits to this page section are the only ones ever to come from that IP so I guess you use multiple IP addresses. I can't even send thanks to an IP, or I would.


 * Worthy of further discussion, so watch this space. Andrewa (talk) 01:19, 24 October 2018 (UTC)

The third bullet
said The first two dot points are worthy for debate, the third I oppose.

That third point reads All new DABs should be created at their base names unless this would require moving an existing article which still has consensus support for being the Primary Topic.

I'm very interested in exactly why this is a problem. The bolded unless is critical here. This clause is only effective if either there's no existing article at the base name, or if there is one but no consensus exists that it's the Primary Topic.

If the second condition applies, then the article at the base name should be moved and disambiguated. It might perhaps be weakened slightly but significantly so that if no consensus exists either way, there would be no move. But I think that if no consensus exists that a topic is primary, then the default should be it isn't, and it shouldn't be at the base name anyway.

If the first condition applies, and there's no existing article at the base name (which would mostly mean that there's nothing there at all and it's a redlink, but could mean a redirect or a few other things), I don't see the problem at all. Andrewa (talk) 00:39, 24 October 2018 (UTC) Proposal Primary Topic should be deprecated.
 * All new articles should be created at unambiguous names.
 * All article moves should be to unambiguous names.
 * All new DABs should be created at their base names unless this would require moving an existing article which still has consensus support for being the Primary Topic.

Primary Topic should be retained as a valid reason for keeping an existing article at a base name if (but only if) consensus exists that the topic of this article is the Primary Topic of the term.

Commenting ...
 * (#1) All new articles should be created at unambiguous names.
 * Motherhood statement? Ambiguous titling seems stupid.  --SmokeyJoe (talk) 00:47, 24 October 2018 (UTC)
 * (#2) All article moves should be to unambiguous names.
 * Yes, fix the legacy failures of #1. --SmokeyJoe (talk) 00:47, 24 October 2018 (UTC)
 * (#3) All new DABs should be created at their base names unless this would require moving an existing article which still has consensus support for being the Primary Topic.
 * I disagree because for very strong PrimaryTopics (NB. PT has degrees, not bimodal yes/no), if all readers expect the PrimaryTopic to be at that title, it is not a good idea to send them to a DAB page.
 * I further disagree because I think DAB pages themselves need to be titled in compliance with PRECISE. I.e. suffix all DAB pages with "(disambiguation)".  Make *all* disambiguation pages PRECISE, CONSISTENT and RECOGNIZABLE.  CONCISE is good, but good taken to an extreme is no longer good, and in stripping out important information, the information density aspect of "concise" is worsened.  --SmokeyJoe (talk) 00:51, 24 October 2018 (UTC)

Ambiguous titling seems stupid. It does. But here are degrees of ambiguity, as you say, and that's exactly the issue. Primary Topic means, yes it's ambiguous, but not ambiguous enough to matter.

for very strong PrimaryTopics (NB. PT has degrees, not bimodal yes/no), if all readers expect the PrimaryTopic to be at that title, it is not a good idea to send them to a DAB page. Agree. But almost all of those articles already exist, and so are not covered by the third bullet.

''DAB pages themselves need to be titled in compliance with PRECISE. I.e. suffix all DAB pages with "(disambiguation)"'' I have no objection to that in principle. We could then have redirects from the base names to these pages, in fact in terms of User:Andrewa/Primary Topic RfC that's the only option I can see. I'm still not convinced that (disambiguation) is recognizable, but that's a different issue and can be handled quite independently. Andrewa (talk) 01:13, 24 October 2018 (UTC)


 * Not convinced (disambiguation) is recognizable? Of course it is, "recognizable" is a very low threshold word.  There are degrees.  I recall the first time I saw it, well into my time at Wikipedia.  I was surprised (though not astonished) to use such an unfamiliar-looking work, but it recognized correctly as indicating that the page so titled was not a normal page.  I have long seen criticisms of the use of this word, but have never seen even a good alternative suggested.  I maybe like the translation of the Japanese equivalent: "(ambiguity avoidance)".  --SmokeyJoe (talk) 01:53, 26 October 2018 (UTC)
 * As I said, that's a separate issue, and should be discussed as such. I can think of many alternatives that work better for me. But how to tell which works best for the general reader... that's not a trivial task. Andrewa (talk) 20:59, 27 October 2018 (UTC)

Very strong PrimaryTopics
I disagree because for very strong PrimaryTopics... I answered this above but it wasn't really addressed in the discussion.

I think the probability of a new article being on a very strong PrimaryTopic is zilch. Yes, we are creating new Primary Topics all the time. But they're all either controversial, or should be. Andrewa (talk) 23:02, 27 October 2018 (UTC)
 * It's getting rarer but there are exceptions. This year we've had lots of surname articles (Clennell, Gregorini, Steinborn…); some artistic works, mainly songs (Don't Leave Me Alone, The Games We Play, Walk It Talk It…) and several articles on missing topics (Dead heat, Protection, Shuttlecraft…).  Some are even regular and predictable, such as 2018 World Series.  I'm ignoring the BLPs as their claim to PT is usually less clear, but I expect at least one of next year's pop sensations to find their article's title occupied by a long dead and barely notable namesake.  Certes (talk) 01:45, 28 October 2018 (UTC)
 * OK, but these aren't in the same class as London or Mathematics IMO. Perhaps we need to call these very, very strong primary topics or something?
 * If a sufficiently strong P T does come up then it can be covered by . But not convinced that any of these suggestions qualify. 2018 World Series for example is to me controversial in a way that Mathematics and London are not. See 2018 World Series (disambiguation)... Is baseball really the P T from a global perspective? Some parts of the world are pretty keen on Rugby but wouldn't know the difference between baseball and softball. Andrewa (talk) 14:20, 1 June 2019 (UTC)

How ambiguous
My comment above Primary Topic means, yes it's ambiguous, but not ambiguous enough to matter seems worthy of more discussion. (Aren't our own words delicious? But I like your degrees of ambiguity too.)

The issue as I see it is, it most often does matter, to the point that the onus of proof should be on those who wish to assert that Primary Topic should be preferred either in a particular practical instance or in policy. We are at the very least making far too much use of it, and spending far too much time on it, and missing the significance of the damage done, such as when a Primary Topic changes from one topic to another, and the resulting article move sends all incoming external links to the wrong article, and makes most incoming internal links both nonsensical and unfixable. Andrewa (talk) 20:59, 27 October 2018 (UTC)

Commonsense and the Occasional Exception
The naming convention guideline banner reads It is a generally accepted standard that editors should attempt to follow, though it is best treated with common sense, and occasional exceptions may apply.

It might be good to list some current exceptions to current policy and guidelines.

And the proposed deprecation of Primary Topic would possibly have exceptions. It might also be good to consider some hypotheticals, to investigate and perhaps demonstrate how rare these would be. Andrewa (talk) 08:16, 28 October 2018 (UTC)

But some of it is obvious
Using a dab when the PT is in doubt certainly makes life easier for editors. Of course, we should always put readers first, but the easier you make my job, the better I do it and the more time I have left for improving another part of Wikipedia. Certes (talk) 10:47, 1 June 2019 (UTC)


 * Exactly. It makes everyone's job easier in fact... remember the thousands of mislinkings innocently created when the NYS article was at the base name? But if we were to move the NYC article to the base name, regardless of how strong the case was in terms of Primary Topic, there would obviously then be people innocently creating more mislinkings, this time intending NYS and getting NYC.
 * But there will almost always (perhaps even always) be some people for whom the Primary Topic of an ambiguous term will be in doubt. That's half of the reason for the proposal, the other half being, those who do agree with our assessment of Primary Topic are also better off of we deprecate the concept as proposed. Andrewa (talk) 19:46, 5 August 2019 (UTC)

Ambiguous titling seems stupid
Perhaps surprisingly, those are not my own words! But my thoughts exactly.

It's the whole basis of this proposal. OK, the original poster of those words didn't mean it quite that simply (see the context above or by clicking the diff, and the discussion that followed). But I do. Andrewa (talk) 22:37, 1 June 2019 (UTC)

Primary redirects
I know that the proposal is not consistent following this edit and my tidying it up has not fixed this. In progress. Comments of course welcome. Andrewa (talk) 05:12, 11 June 2019 (UTC)

Recognisability and policy consistency
It occurs to me that WP:AT and WP:disambiguation may not be consistent.

An unambiguous name is inherently more recognisable than an ambiguous name.
 * An ambiguous name as an article title is recognisable only to those who:
 * Agree that it's the primary topic.
 * Expect others to also agree on this.
 * An unambiguous name is recognisable to everyone.

Interesting? Andrewa (talk) 23:46, 20 June 2019 (UTC)

At the very least, ambiguous names seem to be contrary to the spirit of WP:AT... despite it explicitly allowing them in the specific case of there being a Primary Topic. Andrewa (talk) 00:49, 23 June 2019 (UTC)

Previous discussions
See User:Andrewa/Primary Topic previous discussions. Andrewa (talk) 23:48, 21 June 2019 (UTC)

Fix the problem you know
It is a proven principle of software maintenance (it's probably got a name and a dozen PhDs describing it, but I worked it out all by my little self):

Fix the problem you know

Often I have confronted several problems with a program. There were big problems I didn't understand, and a little one that I did.

So I fixed the little one that I knew how to fix.

And as if by magic, the big problems that I didn't understand went away too.

This is applicable to this RfC. We can and should fix the problem with new article names. We understand that and can fix it.

So let us do that, and see whether the problem of existing long-standing names is then any problem at all. Andrewa (talk) 00:39, 23 June 2019 (UTC)

That principle could be restated (not as clearly IMO but it's a useful other perspective):

Be as logical as possible

The more I think about it the more deliberately choosing an ambiguous article title seems to be quite simply illogical. The function of the article title is to identify the article topic, Ambiguous titles don't do that very well... that's what ambiguous means!

And it's even what WP:AT currently says too... Article titles should be recognizable, concise, natural, precise, and consistent. An unambiguous article title will be recognised by all. An ambiguous one won't be. Again, that's what ambiguous means. Andrewa (talk) 18:43, 3 November 2019 (UTC)

All new articles should be created at unambiguous names
I have a question regarding this All new articles should be created at unambiguous names. Say our article is a new TV series called Cars cars cars!, do we create it at the base name, or do we create it with a disambiguation - Cars cars cars! (TV series). If the answer is with disambiguation, do we then create it with the base "(TV series)" or with the more specific "(American TV series)" / "(2019 TV series)"? --Gonnym (talk) 12:38, 3 July 2019 (UTC)
 * I would think that Cars cars cars! would be suitably unambiguous. There is no particular likelihood of another article being created that could go by that name. If a notable soundtrack album or other spinoff were to justify a separate article in the future (even by being discussed as a future event in reliable secondary sources), deal with that then.
 * On the other hand if the foreshadowed spinoff already has notability when the first article is created, disambiguate immediately. But only if. And create both articles, and a good stub for one of them (normally I guess the spinoff but not necessarily) is fine.
 * If a spinoff later becomes notable, its article is of course created at an unambiguous name. But the original article doesn't need to move. No reason to do that. But essential to add a hatnote to the original article. If another topic ever becomes primary (that second article or any other topic), then disambiguate the original article too and put a DAB at the base name.
 * Or, if consensus were to be that the original article had simply lost its primary status, then at that stage disambiguate the original article. That would I think be rare, just because it's not all that important, but a wikignome might raise such issues, normally by a bold move in the first instance. You don't need admin rights for such a bold move over a redirect with no other history, and it's likely to be uncontroversial, again just because it doesn't matter very much.
 * A redirect from the unambiguous name to the original article should be created at the first excuse to do so. Redirects are cheap. But if one still doesn't exist when the second article becomes justified, essential to create it then.
 * Does that all make sense? Andrewa (talk) 19:15, 3 November 2019 (UTC)

Propensity for misrecognition
recently commented I have long argued that an important underappreciated underdocumented criteria against PT is propensity for mis-recognition by large particular audiences (eg school children in India).

I couldn't agree more! I have tried to cover this in the proposal. But it's a work in progress. How can I cover this point better? Andrewa (talk) 02:21, 16 July 2019 (UTC)

A challenge
Another editor has suggested that P T is often OK, often so-so, and often unhelpful.

See User talk:Andrewa/P T test cases for more on this and a challenge: Can we find any examples where the current setup is better than my proposal? Andrewa (talk) 18:55, 30 July 2019 (UTC)

Clicks
The essay asserts without basis: Most often [Primary Topic] merely increases the number of mouse clicks required..

Please explain. --В²C ☎ 00:20, 1 August 2019 (UTC)


 * For example, if a reader arrives at the wrong article because for them the meaning of the ambiguous term differs from out decision on Primary Topic, and there's a hatnote pointing to the DAB, they need an extra mouse click to go to the DAB. If they'd been taken straight to the DAB they'd have been one mouse click better off. Andrewa (talk) 19:32, 5 August 2019 (UTC)
 * That supports the notion that PT sometimes increases the number of mouse clicks required, which nobody disputes. It does not support the Most often part of the assertion, which is the problematic part, . --В²C ☎ 20:12, 5 August 2019 (UTC)
 * I'm glad that you don't dispute that. Yes, the most is then the issue.
 * And it's problematic for you, yes, as you hold the opposite view with no more evidence than I have if that. I am of course gathering evidence at User talk:Andrewa/P T examples and scenarios and related pages, and don't intend to actually move the RfC until I have a lot more of it.
 * Nobody disputes that having the article on the TV series The Americans disambiguated caused you extra work, going through the DAB every time you wanted to access the article during your binging, and so needing one extra mouse click on each of these many occasions. But the fascinating thing is that you didn't make any attempt to avoid this extra work, although you found it so very annoying. I think it's reasonable to assume that most readers would be able to avoid the DAB if they wished to, even without any further evidence. Don't you? Andrewa (talk) 21:58, 5 August 2019 (UTC)
 * What do you mean "no more evidence"? Page views. If one use is getting 95% of the page views (to use an extreme hypothetical example just to make this point), and the only other use is getting 5% of the page views, then if we put the first article at the base name  we have 95% getting to the article they seek in one click, and 5% getting to their desired spot in two clicks (assume a hatnote link to their page).  On the other hand if we put a dab page at the base name then we're requiring all 100% to click twice, first to get to the dab page, and second to get their particular article.  So in such a case putting the dab page at the base name is what increases the number of mouse clicks required, and putting the first article at the base name reduces the number for the 95%, and leaves it at two for the 5%. There is similar reasoning to go with any case where there is a primary topic per the usage criteria. --В²C ☎ 23:41, 5 August 2019 (UTC)
 * That would be the case if everyone simply went to the base name, but there is no reason to think they all will. Some at least will see an article title that matches their sought topic exactly, and go to that instead. That's one advantage of having articles at unambiguous names. The question I suppose is, what proportion? We've seen the example of yourself repeatedly clicking on the link to The Americans even after you had discovered that it led to the DAB, and being repeatedly disappointed when you arrived there. I think such strange behaviour is likely to be rare. But it may not be. Andrewa (talk) 11:11, 6 August 2019 (UTC)
 * 100% of those searching with the base name will go to whatever is at the base name, by definition. That's the context that matters: people searching with the base name as the search term. For example, in deciding whether the article or the dab page should be at The Americans, what we consider is people searching with the base name "the americans". 100% of them will go to whatever is at The Americans. So our question is: overall, will there be fewer clicks if we send all searchers of "the americans" to the dab page. or to the article about the TV series?  That's how we end up putting the article at the base name. --В²C ☎ 17:14, 6 August 2019 (UTC)
 * IMO that's exactly the assumption that underlies P T, and it is simply untrue and this is simply demonstrated. See User talk:Andrewa/P T test cases. Andrewa (talk) 19:52, 7 August 2019 (UTC)
 * To be clear, you're saying it's "simply untrue" and "simply demonstrated" (that it's untrue) that we will get fewer clicks with the TV series "The Americans" at the The Americans basename than with a dab page there? If not, then what? If so, you don't demonstrate that at the linked section. You simply allude to a hypothetical seeker of the 1961 series who is confused; that doesn't address total number of clicks depending on article layout.  --В²C ☎ 21:56, 21 August 2019 (UTC)

Rationale
This edit seems to raise questions best discussed here. Will the claimed benefits (see the Rationale section of the proposal) actually be achieved? They are: (And there's now a fifth line to the rationale which I didn't think was necessary, as it's covered elsewhere in the proposal and isn't a motivation for change, but another editor wanted it clarified and it does no harm.) Andrewa (talk) 19:17, 5 August 2019 (UTC)
 * Improve the chance of readers finding the article they want.
 * Reduce the number of mouse clicks required to find the correct article.
 * Reduce the number of article moves.
 * Reduce mislinkings within Wikipedia.

Theory
It occurs to me that so far, my rationale and motivation is almost entirely practical. It just seems to me that the current system fails too often.

But there's another way of approaching it: A theoretical side. I'm touching on this when I claim that ambiguous titles are illogical, and contrary to the spirit of WP:AT.

We don't even have a workable definition of Primary Topic. It's in practice whatever the community decides by consensus should be at an ambiguous base name. And the guidelines have been rather volatile lately. Those both should give us pause.

And the current guideline says exactly that, and has for many years, but that section has been significantly changed over the years as to what is regarded as evidence, see December 2014 and [https://en.wikipedia.org/w/index.php?title=Wikipedia:Disambiguation&oldid=404249200#Is_there_a_primary_topic? December 2010] for example.

More to follow. Andrewa (talk) 07:04, 13 August 2019 (UTC)

How it would work
Some good questions have been asked. Mostly that show a need to clarify the proposal rather than to change it in any basic way,

One contributor criticised the arbitrary date for grandfathering. There is none such.

Suppose a new band called themselves Metric space and became notable. We'd create Metric space (band), but we wouldn't move the existing article just because its title had become ambiguous. If they then recorded a chart-topping self-titled album and there was enough material to justify an article on it too, we'd then create Metric space (album) and a DAB at Metric space (disambiguation).

But we still wouldn't move the DAB to the base name, or the older article away from it, just because the term had become ambiguous. So long as it remains Primary Topic, it stays where it is. But now suppose the band becomes bigger than the Beatles. A petition for the UN to grant them pensions for life gains 100 million signatures in the first 24 hours. A rumour that one of them is secretly married causes tens of thousands of teenage suicide attempts. The Pope proposes all the members for immediate sainthood. Paris is officially renamed Metricity in their honour. They are clearly now the Primary Topic, so we move the math article to Metric space (mathematics), but we still don't move Metric space (band). Metric space is now ambiguous, so only the DAB could be moved there.

And in hindsight, we realise that we should have disambiguated the original article on the mathematical topic much, much earlier! That should have happened as soon as it was no longer the Primary Topic.

But how about completely new names? A professional wrestler called The Zonk becomes notable, and gets an article. A hairstyle and a cruise ship are both named after him, and get their own articles too, and we create a DAB at The Zonk (disambiguation). The wrestler retires and is fading into history, and there's an RM to disambiguate his article, but no consensus as to whether he's still Primary Topic. He might be. So no move.

Get the idea? There's no arbitrary date. And an article at an unambiguous name is trivially the Primary Topic of that name. Lots going on! Andrewa (talk) 12:46, 13 August 2019 (UTC)

Trivially Primary Topic
Here's another little clarification... any article at an unambiguous name is, trivially, the Primary Topic of that name.

This sounds obvious but has some subtle consequences. Andrewa (talk) 22:34, 13 August 2019 (UTC)

Does P T even apply to disambiguated names
Another rather subtle question.

Some current discussion at the RfC at wt:DAB. See here if it has been archived. Andrewa (talk) 22:40, 13 August 2019 (UTC)

Doesn't make sense
Surely you agree that your points at User:Andrewa/Primary_Topic_RfC apply as much to existing articles as to new articles. Yes? Then why does your proposal not apply to existing articles too? If you've already answered this, please point me to the section. Thanks. --В²C ☎ 22:00, 21 August 2019 (UTC)
 * They do apply of course. But we need to look at the net benefit.
 * The disruption of changing names of existing articles is considerable, while the disruption of deprecating Primary Topic only when choosing new names is far less.
 * Almost all of the benefits of completely abolishing Primary Topic are realised by just deprecating Primary Topic when choosing names for new articles and new names for existing articles, as proposed.
 * I've answered this many times, including in reply to you several times. But is that enough now? Does that make sense? Andrewa (talk) 20:27, 22 August 2019 (UTC)
 * So, if there is an existing primary topic but another new article arises with the same ambiguous title then that primary topic is subject to the new rules and is to be disambiguated? --В²C ☎ 18:43, 23 August 2019 (UTC)
 * That's a rather strange question, but it explains why it Doesn't make sense to you, so thank you for asking it. Again, I think all would be clear if you would read the proposal. But I guess others will also want to dismiss it without reading it too, so it's I guess good to discuss it. Both Origin of Species and the Bible suffer the same fate... I've read both, but nearly all of those who claim (perhaps trusting Thomas Huxley or his followers in this) that they are in conflict readily admit to having read only one of them, and even to see no need to actually read the work they wish to criticise and dismiss. It's an interesting POV, and a challenging one to constructively discuss with either side! But I diverge.
 * That article would be disambiguated if and only if we had consensus that it was no longer Primary Topic, exactly as now. All that is changed in the "rules" is what then happens to the base name and the new article name.
 * No article is ever deliberately given an ambiguous name. That and its consequences are all that is changed.
 * And just BTW, how we assess Primary Topic is not covered in any way. If you can sell the idea that we should go by page stats alone, then that's a different issue, and in no way affects the proposal. Andrewa (talk) 21:12, 23 August 2019 (UTC)
 * The reason I asked is because... logic.
 * You say ("They do apply of course") your points at User:Andrewa/Primary_Topic_RfC apply as much to existing articles [including those with primary topics] as to new articles.
 * Almost all of the benefits of completely abolishing Primary Topic are realised by just deprecating Primary Topic when choosing names for new articles and new names for existing articles
 * Now, how do you realize "all of the [#1] benefits" per #2 if you don't (at the very least) disambiguate ambiguous primary topic titles with new (disambiguated) names when a new article arises with that ambiguous name? And, by the way, I say at the very least because there will be new articles with names the same as those of only a small fraction of articles with primary topics currently - so how do you realize "almost all of the benefits of completely abolishing Primary Topic" when the vast majority will remain as primary topics? --В²C ☎ 21:29, 23 August 2019 (UTC)
 * Very poor logic IMO.
 * The first problem is that you are inaccurately paraphrasing what I said. I hope I didn't say they apply as much to existing articles (my emphasis), because that is dangerously ambiguous. This becomes important when you similarly shift from Almost all of the benefits of... to all of the [#1] benefits" per #2 (my emphasis on both). If I'd said what you claim that would be a valid point.
 * But I don't think I said that at all. And as any logician could have told you, false premises lead to completely useless logic. What we have above is rhetoric.
 * So, instead of attacking what hope I didn't say and clearly didn't mean to say, would you like to discuss the actual proposal, as intended? Andrewa (talk) 22:11, 23 August 2019 (UTC)
 * I'm not attacking. Just seeking clarification about the applicability of a particular part of your proposal to existing articles. You seemed to agree ("They do apply of course") with my wording "apply as much to existing articles as to new...", and I didn't think there was a significant difference in meaning with or without the "as much", especially since the referenced section, User:Andrewa/Primary_Topic_RfC, makes no such distinction, implicitly or explicitly. So are you now saying that your Rationale section does not apply as much to existing articles as to new articles? If so, don't you think it would be helpful to clarify that distinction there? --В²C ☎ 22:18, 23 August 2019 (UTC)
 * You were wrong on many grounds to think that these paraphrases didn't affect the validity of the argument. Logically, they destroy it. There is no middle ground when it comes to this sort of reasoning. Andrewa (talk) 00:55, 24 August 2019 (UTC)
 * And a clarification as to why that Rationale does not apply as much to existing articles as to new articles would be helpful too. --В²C ☎ 22:21, 23 August 2019 (UTC)
 * Frankly, you don't seem to be trying to understand. The rationale now includes (in answer to your earlier questions) Preserve existing links both internal and incoming. I thought it unnecessary to add that as we have previously discussed, but it's now there at your request. Now, how could that possibly apply as much to existing articles as to new articles?
 * The impact is significantly different for existing articles, as I thought I'd made clear in my very first reply in this section. So it makes sense to me to treat them differently. But not to you, evidently. Andrewa (talk) 00:55, 24 August 2019 (UTC)
 * What's clear to you and what you think you've made clear to me elsewhere is not necessarily clear in the actual proposal. Thanks for clearing up that section; I think this - with the addition of that clarification to Rationale - was relatively productive. That said, I think it's a rather moot point overall since redirects to previous titles address most issues of preserving the integrity of links both internal and external. But, unless I'm missing something again, that's just as true under the current system as it is with your proposal. So, practically speaking, I'm not seeing much significance difference here. The daily drop in page counts for the previous title to the The Americans, even the Bing and Yahoo still reference it, is telling. I continue to monitor it. --В²C ☎ 01:17, 24 August 2019 (UTC)
 * Yes, you're missing a lot. These redirects may not be there at all, and elsewhere you've said they're unimportant. See the very first bullet here. But simpler just to avoid P T. Andrewa (talk) 02:53, 24 August 2019 (UTC)
 * And the third bullet which I've just clarified may upset you. Andrewa (talk) 03:00, 24 August 2019 (UTC)
 * What I meant of course is that those redirects are not intrinsically important to users — of course they’re important — critical really — after a title change. But that’s indirectly for hidden technical reasons.   —В²C ☎ 04:45, 24 August 2019 (UTC)
 * It seems to me that any redirect that helps a user to find the article they want is intrinsically important to users. Not to you? Andrewa (talk) 22:23, 25 August 2019 (UTC)
 * No, to me too. But, please answer this. Putting aside the effects of artifacts that arise as a result of a title change, how do redirects at parenthetically disambiguated forms of a title for an article at a base name help anyone find the article they want? —В²C ☎ 13:26, 26 August 2019 (UTC)
 * Done. I see no point in forking that discussion here. Andrewa (talk) 01:05, 28 August 2019 (UTC)

Piped wikilinks
This is in response to Talk:The Americans. Interesting impact that I had not previously considered. Andrewa (talk) 20:30, 22 August 2019 (UTC)

Some more mess to clean up
Two more things the proposal should try to address, and probably already does. They are related! Andrewa (talk) 00:57, 28 August 2019 (UTC)

Incomplete disambiguation
I just stumbled upon Partially disambiguated page names which documents and summarises discussion on this apparently perennial issue, most recently discussed at wt:DAB.

The recent discussion meandered a bit, and I'm not even sure whether anyone was aware of that summary of previous discussions... if they referred to it, I may just have missed that in the discussion at wt:DAB. Andrewa (talk) 15:04, 5 September 2019 (UTC)

R from incomplete disambiguation
Template:R from incomplete disambiguation reads in part ''This is a redirect from an incomplete disambiguation, a page name that is too ambiguous to be the title of an article or other project page. Such titles should redirect to an appropriate disambiguation page (or section of it), or to a more complete disambiguation. But hang on... if it's OK to redirect it to a more complete disambiguation'', why can't this (supposedly) incomplete disambiguation just be used for the article title? Which is another take on above really.

We're having a bet both ways. Andrewa (talk) 00:57, 28 August 2019 (UTC)
 * Because an article (or dab page) can have only one title. Every other name it has, which may be incomplete disambiguations for a dab page, has to be a redirect to it. For example: Mercury (novel) is a partial dab to the full dab at Mercury. —В²C ☎ 07:07, 28 August 2019 (UTC)
 * Yes, Mercury (novel) redirects to a section of the DAB, as it should. That's a redirect to an appropriate disambiguation page (or section of it).
 * And there should be no other option. The idea that it could instead redirect to a more complete disambiguation, in this case Mercury (2005 novel) or Mercury (2005 novel), is what I mean by having a bet both ways, and IMO is just plain wrong. Andrewa (talk) 11:48, 28 August 2019 (UTC)

DAB titles
See here. Food for thought? Andrewa (talk) 00:27, 5 September 2019 (UTC)


 * If there is no primary topic, should the disambigation page be assumed to be the primary topic?
 * Victoria might be guessed to be Queen Victoria by readers influenced by the British Empire.
 * Victoria might be guessed to be an Victoria (Australia) by children in Melbourne, Victoria.
 * Victoria is currently the title of a DAB page. Many readers might mistake it for their guess at a Wikipedia primary topic.
 * Victoria (disambiguation) would be a precise title, all readers will recognise that it is a disambiguation page and not one specific possible topic. — Preceding unsigned comment added by SmokeyJoe (talk • contribs) 08:04, 5 September 2019 (UTC)
 * Exactly. No, it shouldn't be. For anyone who thinks that there is a Primary Topic, to land on a DAB at the base name is astonishing. I guess to some extent to be redirected to a DAB is too, but less so. A DAB is similar to a soft redirect, it could even be regarded as type of soft redirect. Andrewa (talk) 14:58, 5 September 2019 (UTC)

Defence of the other criterion
I missed this edit until too late... the section has just been archived.

Exactly. Worthy of further consideration. Andrewa (talk) 20:41, 8 September 2019 (UTC)

Articles that exist and could exist
Very interesting discussion at Wikipedia talk:Disambiguation.

Initiated by myself with the naive comment ''This is probably (I hope) a very quick discussion (for a change). It's claimed here that "We disambiguate article titles based on the articles that exist and could exist." I don't think that's even remotely true. Very interested in other views.''

I was wrong on both counts. A tangled tale unfolded that led to the conclusion that we do disambiguate article titles based on the articles that exist and could exist, and a subsequent RM has confirmed this.

This reduces the impact of my proposal a great deal. New articles can and should be created at unambiguous titles even if no article currently exists for the shorter, ambiguous title. That it would seem is existing policy and practice. Andrewa (talk) 16:52, 10 September 2019 (UTC)

Seeking and searching
Much discussion recently which I'll link to in time! Is using the Search Wikipedia box searching? Some say no rather emphatically.

But there are many ways of seeking an article. One I use regularly without thinking is to Wikilink it in an edit and then test the Wikilink using the Preview button.

I recently did this looking for a magazine called Knowledge that I used to read as a small child. But Knowledge (magazine) was a dead end, not even a useful hatnote. When I finally found it at Knowledge (partwork), the article read Knowledge was a British weekly educational magazine for children which was assembled in blue binders into an encyclopedia. (my emphasis)

So I think that my seeking method was reasonable. Not sure how to improve the navigation in this case... maybe a hatnote.

The point is, there are many ways of searching (seeking?) and we probably can't even imagine them all. But having logical article titles should help! Andrewa (talk) 22:45, 12 September 2019 (UTC)
 * Searching? In encyclopedias, starting with physical Brittanica books. I have commonly done these things:
 * Go to a title. If I have a fair idea the topic will be a listed topic, I would go to it alphabetically. This is like the Wikipedia Go box. Redirects serve it well, Try “doves” for example.
 * Browse. Turn pages and see where my eyes run to. Brittanica was more page turning than goto, but combined with gototitle above, I would jump from topic A to topic B when reading topic A piqued my interest in B. Wikilinks, navboxes, the category system serves this need. Portals have always been disappointing.
 * Search. This used to be the index at the back. You search for a topic if you don’t expected to be a titled topic. I think Brittanica had an entire bound volume for searching. It was usually disappointing, taking you to a mere mention with little information, and no wikilinking to similar information.  Search tools are one of the greatest things of the computerised modern world. —SmokeyJoe (talk) 23:08, 12 September 2019 (UTC)
 * Thanks! And I'd add one that I use (and suspect you do too):
 * Wikilinking. As mentioned above. Is that covered by one of your methods above?
 * I just used it not too successfully. The redirect currently points to a now deleted anchor that I guess existed in February 2014. But I think it's still a useful way of finding articles in general. Perhaps it's covered by Browse?
 * I guess you include external search engines such as Google under Search. Are there other ways of searching in this restricted sense? That's the sort of question I'd like to investigate here. Andrewa (talk) 18:57, 14 September 2019 (UTC)
 * Wikilinking for me works under browsing. I find it doesn’t serve well for seeking or searching, maybe because I am looking for things now covered, I find wikilinking tends to lead me in circles and dead ends as a search tool.  Wikilinking is actually the dominant method for browsing, for me, and I reckon for nearly everyone.  Alternatives are the category system, which work but are not a pleasure to use, and portals, which fail to serve.
 * For searching, I use both google with “Wikipedia” and the internal Wikipedia search engine, both work. Mostly when I take the trouble to use the internal Wikipedia search engine, it’s because I want to search namespaces other than mainspace, for some Wikipedia related purpose.
 * For me, “searching” implies using some method to quickly find something that I want. The thing I want has to be preconceived, otherwise it is browsing.   I find that that google is very good, and when it fails it is usually because I am looking for something that isn’t there.  The Goto box fails as a search because it fails for topics that are not articles, although redirects to article sections help considerably.  Google never fails where the Goto box succeeds.  My beef with retitling arguments based on making the Goto box work better is that it compromises other things for a poor search tool. —SmokeyJoe (talk) 03:53, 15 September 2019 (UTC)

Was the NYRM no P T decision correct
Interesting comment:

IMO the New York Wars ended in the correct consensus – different primary meanings in at least UK and US, and possibly within US, so no PTOPIC.

But... surely if a term is ambiguous, it will almost always have different primary meanings to different groups of people? If we accept that as a criterion for no PTOPIC, there won't be very many primary topics.

Agree that the NYRM2017 decision was correct in terms of improving Wikipedia.

But what got me thinking more deeply is that IMO it's not the correct decision in terms of our current policies, guidelines and practices. New York City is the primary topic. And the more I dug into the logic of P T, the shakier the whole concept became. Andrewa (talk) 21:21, 28 September 2019 (UTC)
 * One editor in the NYRM wrote: "'New York' in sources can be contextual, but when something is from 'New York', it more often refers to the larger locale—the same as Luxembourg and Luxembourg City. The state, by its common name, is the primary topic for 'New York'." The hell it is, in both cases!
 * In this recent newspaper article, chosen because it was the very first hit in a Google search for 'Boris Johnson Luxembourg' (a very recent news event), 'Luxembourg' means the state in the headline and in the second and sixth paragraphs, and the city in the first and fifth paragraphs; but only if you already understand the context.
 * As to the suggestion that 'when something is from "New York", it more often refers to the larger locale'; as a Brit, words fail me. In UK, 'New York' almost invariably refers to the city. I doubt whether as many as one Brit in one hundred could name the state capital, distinctly less famous in song and story.
 * As one historical parallel, during the British Raj 'Bombay' could mean either Bombay or Bombay Presidency depending on context. As another, 'Swan River' during imperial days might mean, depending on context, Swan River itself or Swan River Colony: either modern Perth in a limited sense or modern Western Australia in a broader.
 * In Shakespeare's Henry V, Act III Scene 6, Montjoy says, "England shall repent his folly", and in reply Henry says, "Though France himself [...] stand in our way". Both pronouns are accurate: they refer to Henry V of England and Charles VI of France in person, as embodying their countries, not to those geographical areas. Context is indeed everything, and sometimes needs to be explained. Narky Blert (talk) 00:52, 29 September 2019 (UTC)
 * If context is required, then there is no PrimaryTopic. For New York and Luxembourg, PrimaryTopic doesn’t apply, because they are not independent topics, if you consider extended history, which an encyclopedia should do.  For New York and Luxembourg, it is just a question of the best titles.  There is no astonishment or ambiguity, as anyone who knows one region/city knows that name can also be used the other way.  It is not a case of an homonym, but a variation on detail of usage. —SmokeyJoe (talk) 02:18, 29 September 2019 (UTC)
 * For New York and Luxembourg, PrimaryTopic doesn’t apply, because they are not independent topics... You have made this point of independent topics before, and I still don't get it. Many ambiguous names refer to related topics... wave for example. Currently it's the name of an article which might be disambiguated , which is presumably the Primary Topic... or is it? Several of the other meanings are related, some quite closely. At what point does it become a variation of usage?
 * Several of these other meanings, while not homonyms of wave in the sense of physics, do have sufficient independence that Primary Topic does seem to apply under current policy, guideline and practice. So where do we draw the line?
 * If we're going to say that Primary Topic is only an issue for homonyms, again there won't be many Primary Topics left. Most ambiguous names have an element of polysemy to them, including wave and New York. Andrewa (talk) 14:50, 29 September 2019 (UTC)
 * There is no astonishment or ambiguity, as anyone who knows one region/city knows that name can also be used the other way. I don't think that is true. New York City is far more famous than New York the state, and so some would know of the city but not the state. Andrewa (talk) 21:04, 29 September 2019 (UTC)
 * I don't recall that particular quote from NYRM, but it's an appeal to the rejected HLJC criterion so I can probably guess who said it, and they're still active and as aggressive as ever so I'm not going to ping them here.
 * Context always helps. But does that justify using an ambiguous article name? In running text we just use the pipe trick to remove the disambiguator when it's not required. Andrewa (talk) 15:14, 29 September 2019 (UTC)
 * I've just discovered that the Original Dixieland Jass Band played the St. Louis Blues, but can't find the result. Nor can I discover what happened when Sidney Bechet and Hank Williams each took them on single-handed. Narky Blert (talk) 00:08, 30 September 2019 (UTC)
 * Very droll. Of course I'd like to solve this by not having any article at an ambiguous name.
 * A more difficult (and completely off-topic) one to explain to a layman is My name is Vic Dudman. Vic Dudman likes beer. Therefore my name likes beer. (Original to Vic himself, under whom I was highly privileged to study.) Andrewa (talk) 06:20, 30 September 2019 (UTC)

Homonyms vs polysemes
Very interesting issue raised here above. It's been raised before and I confess I didn't understand the point being made before.

It doesn't actually mention polysemes, just contrasts homonymy to being a variation on detail of usage. NYS vs NYC is not a case of an homonym, but a variation on detail of usage.

Homonyms are words which sound alike or are spelled alike, but have different meanings. It gets a bit more complicated, but as a written encyclopedia we're actually concerned with homographs (words that share the same spelling, regardless of their pronunciation). I think that's what homonym means here, and that distinction simplifies things a great deal. Whether or not they are homophones doesn't concern us when choosing article titles.

Is a variation on detail or usage a polyseme? Polysemy is thus distinct from homonymy—or homophony—which is an accidental similarity between two words (such as bear the animal, and the verb to bear); while homonymy is often a mere linguistic coincidence, polysemy is not. So the usage of New York for both the city and the state is polysemy. The articles aren't all that much help... the one on polysemy is very technical, and both articles are a bit vague as to the precise meanings of the terms, explaining that linguists themselves can't agree.

So linguistics is not a great help, but it's a little. We have one useful term... homograph. Article names are homographs if they are identical and the reason for this is nothing to do with their meanings. Row, bear and stalk are given as examples of words that have such unrelated meanings.

And the claim of seems to me to be that PrimaryTopic doesn’t apply unless the meanings are unrelated... that is, unless the meanings are genuine homographs. Is that an accurate paraphrase? If so I begin to understand I think.

But I still can't see why this makes any difference to us in choosing article titles. All we're concerned about is what our readers and editors will take the term to mean. Why they attach these meanings doesn't matter. Andrewa (talk) 23:32, 29 September 2019 (UTC)


 * polyseme! A new word for me.  I do enjoy these conversations.  --SmokeyJoe (talk) 23:58, 29 September 2019 (UTC)
 * New to me too, as far as I can remember... but it's twenty years since I studied or taught tertiary linguistics and this is an area that has exploded in the meantime. My speciality was semantics which is relevant but that was in the early stages of the ongoing explosion in artificial intelligence... I actually helped implement and document a very successful neural network based business system but it used no semantics or semiotics.
 * I'm surprised and disappointed that homograph seems the only useful term I've yet discovered (according to Wikipedia anyway)... but I think it helps a little, by clarifying the distinction between ambiguous terms that are homographs, and those that are not. Andrewa (talk) 05:13, 30 September 2019 (UTC)


 * Andrea wrote: "{u|SmokeyJoe}} seems to me to be that PrimaryTopic doesn’t apply unless the meanings are unrelated... that is, unless the meanings are genuine homographs. Is that an accurate paraphrase? If so I begin to understand I think."
 * That's correct. SmokeyJoe, like every reasonable Wikipedian, appreciates that Long term significance is a critical aspect of PrimaryTopic.  NB PrimaryTopic is, as is well agreed, a bad term.  Perhaps substitute "Primary Topic" for "Occupant of the base name title".  It presumes to be the main topic for that title, all others as minor alternatives, for all reasonable readers.
 * "Long term significance" is not a useful discriminator for two topics are are so intrinsically related that they share the same long term significance. The is exactly the case for "New York".  New York was a (renamed) young colony that became a great city.  New York was a (renamed) young colony that became a great state.  The critical component of PrimaryTopic decision-making factors out.  Both equally qualify, both prevent the other from being chosen.  --SmokeyJoe (talk) 00:08, 30 September 2019 (UTC)


 * Higher-Level Jurisdiction Criterion shortcut WP:HLJC. A reasonable notion.  It could be used as an alternative to WP:PT when WP:TP fails, due to the competing articles being intimatly related.  It is difficult for New York because NYState is the higher-level jurisdiction, but NYC has more points in its favor. HLJC is, I think, a valid criteria for one to consider, but I suspect that it is always prone to not be decisive in difficult cases, and what's the point of a formal criterion if it only works in easy cases?  It is too simplistic.  -- SmokeyJoe (talk) 00:14, 30 September 2019 (UTC)
 * I think that the HLJC was something that it was good to discuss but now we should leave it as it is and in hindsight always was, dead in the water. It has no consensus support despite a vigorous campaign to gather it. The reason it doesn't have my support is that it's useless at best and often counterproductive. You seem to think that it's useful in theory but not in practice. While I disagree that it's useful in theory, as long as we agree that it's not useful in practice I'm happy with that outcome. Andrewa (talk) 05:02, 30 September 2019 (UTC)
 * No, I wouldn't say "useful in theory". I might prefer to say that: Like the Geocentric_model of the universe, it is frequently in agreement with reality".  (I had to use Google to find that section).  Frequently correct is not the same as useful, because it includes "dangerous", as in "a little knowledge is a dangerous thing".  --SmokeyJoe (talk) 06:11, 30 September 2019 (UTC)
 * Disagree that NYC and NYS share the same long term significance. NYC is one of the most important cities in the world... many say the most. NYS is one state of many (unless you live there or close by). Andrewa (talk) 05:17, 30 September 2019 (UTC)
 * What if I rewrite that: NYC and NYS share the same a lot of long term significance?  --SmokeyJoe (talk) 06:05, 30 September 2019 (UTC)
 * NB PrimaryTopic is, as is well agreed, a bad term. Perhaps substitute "Primary Topic" for "Occupant of the base name title". It presumes to be the main topic for that title, all others as minor alternatives, for all reasonable readers. A very dangerous assumption and not generally true.
 * I'd propose base topic. Base topic is just the topic of the article placed at the base name. No other assumptions. Hopefully that decision, like all others here, is made by consensus, with arguments assessed through the lens of policy.
 * But I'm not convinced that it's well agreed that Primary Topic is a bad term. Perhaps unfortunately, but I think the problems go deeper than the term. The term is a symptom, not a cause. Andrewa (talk) 05:38, 30 September 2019 (UTC)

wave
Andrea has mentioned wave a few times. It is worthy of its own consideration. It is a troubled case.


 * Wave (physics) as PrimaryTopic is a bad PrimaryTopic.
 * (1) Wave (physics), the current occupant at Wave, is over-specific. A primarytopic worthy lede could be written, but is not.  The body of the article is extremely focused on physics & mathematics.
 * (2) There is a pre-datingPRIMARYMEANING wave "move back and forth," Old English wafian "to wave, fluctuate" (related to wæfre "wavering, restless, unstable"), from Proto-Germanic *wab- (source also of Old Norse vafra "to hover about," Middle High German waben "to wave, undulate"), possibly from PIE root *(h)uebh- "to move to and fro; to weave"
 * (3) The DAB page contains only related topics/terms (eg Wave (gesture), related to the PRIMARYMEANING, not directly to wave (physics)) and derivative topics mainly dozens of commercial products). There is no real ambiguity.  Every page alludes to the PRIMARYMEANING.


 * My thinking is that wave should be a WP:DABCONCEPT page. --SmokeyJoe (talk) 23:57, 29 September 2019 (UTC)
 * If you're right and there are no homographs of that broad sense of wave, then agree. That's a very sensible solution that covers all the topics. Of course it won't get universal support, but will in particular be opposed by those with mouse click phobia. Andrewa (talk) 05:56, 30 September 2019 (UTC)
 * Let the people obsessed with mouse clicks volunteer instead for Google. Wikipedia is not a search engine.  Search engines index Wikipedia great.  --SmokeyJoe (talk) 06:04, 30 September 2019 (UTC)

P T and disambiguation
Fascinating discussion archived at Wikipedia talk:Disambiguation/Archive 52. It refers to the RfC at Wikipedia talk:Disambiguation/Archive 51 which closed with There is a consensus in favour of Option 2, and that the standard for making disambiguated titles such as Foo (bar) a primary topic among all Foo's that are Bars should be tougher than the standard for titles that don't have any disambiguator. Andrewa (talk) 04:59, 29 October 2019 (UTC)

Naming of DABs
I'm coming around to the view that all DABs should be distinctively identified as such (as many are by the flag (disambiguation)).

This cuts across the idea that a DAB should be at an ambiguous name. And this is not necessary anyway. Perhaps this is my third epiphany. What is needed is simply that the destination of the ambiguous term is the DAB.

But I'm not convinced that (disambiguation) is the way to go. It's jargon. Maybe, the main namespace (AKA the article namespace) should be reserved for articles, and DABs should be in a separate navigation namespace?

We might for example broaden the use of namespace 100 (currently the portal namespace) to include DABs. We'd still need hard redirects from the article namespace. An alias of DAB could be created, similar to the wp alias of namespace 4.

And again, this wouldn't need to happen overnight. All that needs to happen is that new DABs are created in namespace 100 with a redirect from namespace 0, and that whenever a DAB is moved for some other reason, it is moved to namespace 100.

There are a few complications... what happens of there's already a portal by that name? New portals could be named XY Portal just as wikiprojects are named WikiProject XY in the Project Namespace. Maybe keep the pseudo-disambiguator (disambiguation) to resolve conflicts with existing portals.

It would not be good to complicate the RfC unnecessarily. But clarifying it, to say that it's not what is at the ambiguous term (in namespace 0) that matters but just what its destination is, seems a good thing. Andrewa (talk) 15:27, 29 October 2019 (UTC)