User:Dicklyon/Whither Recognizability?

Whither Recognizability? – an essay to collect and explore the history of the recognizability provision in WP:TITLE.

The early years, 2002–2008

Recognizability was born in this diff: [1] by User:Mav (added general statement) 5 May 2002

Generally, article naming should give priority to what the majority of English speakers would most easily recognize with the minimum of ambiguity, while at the same time making linking to those articles easy and second nature. Beyond this general statement, the most important sections to read are the first few sections: Free links, Simplicity, Precision, Capitalization, and Pluralization.

This recognizability provision persisted in the lead with little change or dissent through the middle of 2008, as Wikipedia's top-level naming guidance.

(bold is usually my emphasis, not in the original):

25 Jan 06 Nutshell added by User:Francis Schonken: [2]

and the summary in the lead was expanded to cover "readers over editors" as an alternative way of looking at recognizability:

Another way to summarize the overall principle of Wikipedia's naming conventions:

Names of Wikipedia articles should be optimized for readers over editors; and for a general audience over specialists.

24 Jan. 2007, still in the lead, small wording change from summary to justification for the readers over editors, etc.; by PMAnderson: [3]

This is justified by the following principle:

22 Nov. 2007, the wording was tweaked. [4] Tony1 took out "to the majority of English speakers" and PMA put back "to the greatest number of English speakers".

Generally, article naming should prefer what the greatest number of English speakers would most easily recognize, with a reasonable minimum of ambiguity, while at the same time making linking to those articles easy and second nature.

and corresponding nutshell update a few days later:

Lasting in the lead through the middle of 2008, that was a pretty stable 6-year run as a guiding principle, with little modification but a hint of concern about the question of who it should be recognizable to.

End of May 2008 reference version

2008, moving toward COMMONNAME and sectioning it off from the lead

On 27 June 2008, Pmanderson added this odd provision [5]:

Wikipedia determines recognizability by what verifiable reliable sources in English use for the subject.

which seems to be the start of a campaign to convert recognizability to common name – converting a general goal to a counting-based methodology – a campaign that accelerates in coming years.

Then [6] 10 July 2008 PBS made it into its own section (first one after lead), but the message remains the same, and is still in the nutshell, too:

Use the most easily recognized name

Generally, article naming should prefer what the greatest number of English speakers would most easily recognize, with a reasonable minimum of ambiguity, while at the same time making linking to those articles easy and second nature.

This is justified by the following principle:

The names of Wikipedia articles should be optimized for readers over editors, and for a general audience over specialists.

Wikipedia determines the recognizability of a name by seeing what verifiable reliable sources in English call the subject.

Early 2009: starting to fight

12 Jan 2009 PBS reverts Kotniski with edit summary "Use the most easily recognized name: reverted the changes to this section as there needs to be a general consensus (on the talk page) before changeing the wording of a key policy section" (even though the changes were basically neutral wording tweaks, it appears).

This idea that one shouldn't thrash a policy page is one that Kotniski never really accepted; nor did some of the others.

2 Feb 2009, Born2cycle (in only his second edit of the policy page, after a first edit in 2006) adds crazy complexity to the simple recognizability provision, changing it to read:

Wikipedia determines the recognizability of a name by seeing what verifiable reliable sources in English for the topic call the subject when addressing non-specialists in the given field.

It was immediately reverted by Hesperian with summary "No consensus for that. Not even the slightest hint of any consensus for that through three months of discussion." PBS finished the revert.

The great turmoil of September 2009

This Sept. 1 2009 reference version still looks like the 2008 state, though the nutshell got updated by Dhaluza in Aug. 2008 to this:

By 1 Oct. 2009, however, the article is locked for edit warring, at this nearly-unrecognizable version with this nutshell

and this section with recognizability bullet re-purposed to support a more COMMONNAME-like approach instead of recognizability to the greatest number of readers:

Deciding an article name

Article titles should name or describe the subject of the article and make Wikipedia easy to use. Article titles do this if they are:

• Recognizable – Use names and terms most commonly used, and so most likely to be recognized, for the topic of the article.
• Easy to find – Use terms that readers are most likely to look for in order to find the article (and to which editors will most naturally link from other articles).
• Precise – Be precise, but only as precise as is necessary to identify the topic of the article unambiguously.
• Concise – Keep it brief. (Even when disambiguation is necessary, keep that part brief.) A good article title is to the point. This makes finding and recognising the article easier (and makes life easier for editors linking to it).
• Consistent – Similar articles generally should have similar titles. This may be true of a series of articles sharing a common topic or articles describing different topics but from a common field.

What was this about? How did it happen? Who favors this demotion of recognizability to an also-ran position and re-purposed it to support source counting? Let us examine the nearly 500 changes to the policy in September...

Early September – converting the recognizability section to COMMONNAME

5 Sept. 2009 Kotniski demotes recognizability to a subsection in a General principles section following an Overview section. PBS reverted all that later that day, but Kotniski put it back, and prevailed, for now.

Kotniski's new overview also mentions recognizablility. On Sept. 7 it further muddled together the goal of recognizability to readers with the methodology of measuring recognizability by looking at sources:

Generally, the objective is to give articles unambiguous titles that readers will most easily recognize, where recognizability is determined by what reliable sources in English call the subject. The principles and conventions listed here set out in detail how that objective is achieved.

On 7 Sept., GTBacchus pulled recognizabilitiy and other considerations into the Overview as bullet points, introducing the "Principle of least astonishment":

• Recognizability - articles are titled in such a way that most English speakers will understand from the title what the article is about. This is often cited as the "Principle of least astonishment".

On 9 Sept., PBS took it back to approximately the 30 Aug. version, but after a revert war with Kotniski, the newer version stuck (Hesperian noted that BPS "opposes everything").

Then on 9 Sept. Kotniski converted recognizability to common name and further evolved it to read as a most commonly used provision, specifying a process rather than a goal:

Use common names

Convention: Name articles in accordance with what the greatest number of English speakers would most easily recognize – usually the most commonly used name in verifiable reliable sources in English. The names of Wikipedia articles should be optimized for readers over editors, and for a general audience over specialists.

We still have the recognizability bullet, but not the section, which from now on is a WP:COMMONNAME section.

Born2cycle – unstoppable meets immovable

B2C did 31 edits in Sept 2009, the most active editing month on record. And 13 in October. He replaced recognizability by common name on 10 Sept., with misleading edit summary:

• Common name - Wikipedia article titles identify the name most commonly used to refer to the topic of each article. When naming articles, use the name most commonly used to refer to the topic by most English speakers. This is often cited as the "Principle of least astonishment" and following it achieves consistency in article naming throughout Wikipedia.

Arthur Rubin reverted that "repurposing of guideline". In subsequent diffs, B2C tries again, claiming "Restoring long held primary importance (see 2 week old and 2 year old and 4 year old versions of this page) of most common name is not repurposing!" but Pmanderson reverts him.

B2C tries again. This time Pmanderson goes along. But Arthur Rubin puts it back again. PMA rewrote it more, claiming it is "no change of guidance":

• Recognizability – Good article titles will convey to English speakers what the subject of the article is; this implies that articles will use common names. This is often cited as the "Principle of least astonishment".

and then PMA linked Principle of Least Astonishment.

Still on 10 Sept., Pmanderson removed the Overview heading, putting the bullet points all into the lead, and rewrote the recognizability bullet again to read:

• Recognizable – Good article titles will convey to English speakers what the subject of the article is; this implies that articles will use common names. This is often cited as the "Principle of Least Astonishment".

11 Sept. Hesperian removes recognizable from the nutshell, and mangles the bullet to read:

• Recognizable – Article titles that use familiar names for the subject are highly effective at conveying that subject to the reader. This implies that articles should use common names. The most commonly used name for a subject is also usually what most readers would expect to be the title, unless it is pedantic or vulgar. This is often cited as the "Principle of Least Astonishment".

but PBS undid it.

11 Sept. B2C attempts to rewrite recognizabililty to call for "the name most commonly used":

• Recognizable – Proper article titles specify what the name of the subject of the article is in English; this implies that, when possible, an article title will be the name most commonly used in English to refer to the subject of the article. This is often cited as the "Principle of Least Astonishment".

WP:COMMONNAME still has "Name articles in accordance with what the greatest number of English speakers would most easily recognize" at this point.

B2C gets rid of "recognizable" and "precise" bullets, buries them in paragraph:

In order to communicate with the reader, article titles need to be recognizable and specific – Proper article titles specify what the name of the subject of the article is in English; an ideal article title, by this criterion, will be the name most commonly used in English to refer to the subject of the article. This is analogous to the Principle of Least Astonishment.

PBS put back the "recognizable" bullet, but not "precise", per talk discussion. PMA put back "precise" later.

Having had no luck yet, on 12 Sept. B2C tries again to make COMMONNAME dominate over consistency, but PBS reverted with edit summary "Reverted good faith edits by Born2cycle; (I'm trying to assume good faith): You are the only edit who thinks that Common Name is or should be the primary criterion."

So far, no net impact of B2C. Recognizability remains in good shape, in the nutshell and in the first bullet in the lead.

Mid September – a little repair

13 Sept. PMA adds old guideline info to recognizability bullet (was only in COMMONNAME):

• Recognizable – Good article titles will convey to English speakers what the subject of the article is. This is often cited as the "Principle of Least Astonishment". Our readers are mostly not specialists; article names should be optimized for readers over editors, and for a general audience over specialists.

but then he took out the part about our readers being non-specialists.

15 Sept. PBS removes "for readers over editors", saying "as no-one has yet articlualted a reason for keeping it".

Late September – anonymous contributor

25 Sept, User:Rannpháirtí anaithnid (Irish Gaelic for anonymous contributor, who made no edits here outside an 11 day period of this month) demotes recognizability from the lead, to new section with heading "Deciding an article title", and reduced it to

• Recognizable – Use names and terms most commonly used, and so most likely to be recognized, for the topic of the article.

This is essentially what B2C was trying to do: reduce recognizability to commonname, making it about most common usage in sources rather than about recognizability to a large number of readers.

The policy page was protected 1 Oct. at this version in a continuing edit war between B2C and Arthur Rubin, mostly. This locked in Rannpháirtí anaithnid's re-purposing of the recognizability bullet for commonname, except that the nutshell went back to something like an early version at some point:

Deciding an article name

Article titles should name or describe the subject of the article and make Wikipedia easy to use. Article titles do this if they are:

• Recognizable – Use names and terms most commonly used, and so most likely to be recognized, for the topic of the article.
• Easy to find – Use terms that readers are most likely to look for in order to find the article (and to which editors will most naturally link from other articles).
• Precise – Be precise, but only as precise as is necessary to identify the topic of the article unambiguously.
• Concise – Keep it brief. (Even when disambiguation is necessary, keep that part brief.) A good article title is to the point. This makes finding and recognising the article easier (and makes life easier for editors linking to it).
• Consistent – Similar articles generally should have similar titles. This may be true of a series of articles sharing a common topic or articles describing different topics but from a common field.

In early 2012, this section created by Rannpháirtí anaithnid remains, in approximately that form.

Oct. 2009 – more turmoil after unlock

After the unlock there was a raft of changes by Kotniski, some changes by B2C reverted by PMA, and then B2C putting in another call for the most common name used, but not in the recognizability provision. Blueboar reverted that one.

On 8 Oct., B2C adds junk in the top of the section, further diluting the recognizability, etc. PMA reverted that a few edits later.

All twelve edits by B2C on 7–9 Oct. were reverted, by PMA, Blueboar, and Arther Rubin. Not much impact, but lots of activity. The only net change in October is in the paragraph in front of the recognizability bullet, most of which is in a footnote thanks to Kotniski so it doesn't dilute it visually:

Every Wikipedia article must have a unique title. Ideally, these titles should be:

• Recognizable – Use names and terms most commonly used, and so most likely to be recognized, for the topic of the article.

...

B2C stopped, and the second half of the month was pretty quiet.

Aug. 2010 – planting the seeds of dissent

A relative stability lasted until 17 Aug 2010, broken by this undiscussed Kotniski rewrite of the whole "Deciding an article title" section, which left recognizability looking like this:

• Recognizability – an ideal title will confirm, to readers who are familiar with (though not necessarily expert in) the topic, that the article is indeed about that topic.

The edit summary Deciding an article title: an alternative formulation for your perusal and possible reversion ('easy to find' and 'precise' seem not to have any meaning not covered by the other crietria) doesn't mention recognizability, but does say it's tentative and OK to revert.

Kotniski's comments on talk a few minutes later didn't mention this change, but he did say "As you will have noticed, I've tried to rewrite the initial section on criteria in a way that seems to me slightly more logical (diff, new version of page - feel free to revert). Starting a new section for comments.

PBS reverted 18 Aug. Knepflerle put it back, saying it was just tidying up. PBS took it out again, Kotniski put it back again. Apparently it was not OK to revert.

But nobody commented on the change to recognizability even though it bore little relation to any previous versions of this provision.

The May 2011 correction

On 21 May 2011, Ohms law changed it, taking out Kotniski's "to readers who are familiar with (though not necessarily expert in) the topic", to this simple bullet:

• Recognizability – article titles should be the most recognizable description of the topic.

Note that "most recognizable" links to WP:COMMONNAME, so it's still sort of a commonname approach.

His edit summary "Deciding on an article title: Changed Recognizability point based on discussion on the talk page" refers to this conversation, where the provision was poked at this way:

If you say it is "slightly winged" from "article titles should be the most recognizable description of the topic", then why don't you just say so, plain and simple? And skip all this nitpicking ("readers who are familiar with (though not necessarily expert in"... although not complete ignoramuses; while keeping in mind some of them may be ignorant but they hold the are not.....)) ... Kaligelos

After the edit, PMA chipped in with

Not most recognizable; recognizable is enough, and the superlative is not particularly meaningful. If there are two names about equally common, we can choose between them. Septentrionalis

and changed it to read:

• Recognizability – article titles are expected to be a recognizable name or description of the topic.

The discussed version lasted pretty well until Dec. 2011, even through B2C's 12 Dec. revert of a COMMONNAME dispute, as:

• Recognizability – Is the candidate title a recognizable name or description of the topic?

The Dec. 2011 flareup

In this 20 Dec. edit, B2C changes it to:

• Recognizability – Is the candidate title a recognizable name or description of the topic to someone familiar with (though not necessarily expert in) the topic?

with edit summary "Restore original meaning/wording which was, apparently inadvertently, removed in May 2011. See talk WT:Article titles#Clarification of recognizability lost".

How he thought the May 2011 correction was "inadvertent" is hard to see.

The linked talk section says "I don't see the question of 'recognizable to whom?' being addressed there." But it is addressed, or the wording about it is at least, in the May 2011 discussion as shown above. The actual concept of 'recognizable to whom' had never been discussed elsewhere, as far as I can tell, so this seems like a funny argument.

The real reason for the change is up front there, to get an advantage in an ongoing argument where recognizability was mentioned as a concern: "Above, it is being argued (in essence) that a title like Crime Patrol, since it is not universally known, does not meet the 'recognizability' criteria, and, so, should have more precision (or predisambiguation) in order to be more recognizable to readers and editors unfamiliar with the topic."

Actually, the point where recognizability was cited by Tony1 was not crime patrol, an RM that B2C asked to re-open after it didn't go his way (Talk:Crime_Patrol_(TV_series)#Requested_move). Rather, it was a discussion of Life Safety Code, per this comment by Tony. I didn't agree with Tony's RM on that one, and solved it by instead clarifying the article topic, but this was clearly a case where B2C wanted to rewrite policy to win an argument, and that could not be tolerated.

So far, after a lot of turmoil and locks, B2C has been rolled back and the policy is locked at "Is the candidate title a recognizable name or description of the topic?" while the ArbCom seeks a better approach to sorting this out. See Wikipedia:Arbitration/Requests/Case/Article_titles_and_capitalisation and Wikipedia:Arbitration/Requests/Case/Article_titles_and_capitalisation/Evidence.

Capitalization in title policy – collateral damage

During 2002–2008, the title policy has a strict lowercase provision (at shortcut WP:NAME#Lowercase since June 2008 and WP:LOWERCASE since some time in there, and for a long time as the very first item in the very first section; see [7]):

Lowercase second and subsequent words

Convention: Unless the term you wish to create a page for is a proper noun or is otherwise almost always capitalized, do not capitalize second and subsequent words.

or in the later part of those years:

Convention: Do not capitalize second and subsequent words unless the title is almost always capitalized in English (for example, proper names).

User:Stevage added the example Art Nouveau in 2006 (it's not a very good example, as it's often lowercase in sources).

This "almost always capitalized" criterion is what I remembered and worked with for years. Where did it disappear from modern TITLE policy? Ah, it was in that same tumultuous Sept. 2009 mess, in this edit by Kotniski. He re-purposed the "almost always" from its use bolded above to this:

• Capitalization. The initial letter of a title is almost always capitalized (for exceptions, like eBay, see the technical restrictions page). However subsequent words in a title are not capitalized unless they would be in normal text: an article about video games is called Video game, not "Video Game". (But see the rules on bird naming.)

in which the criterion "unless they would be in normal text" leaves the capitalization completely open to guesswork.

I can find no hint of a discussion about this edit (but the talk page is a tumultuous mess this month, so it's possible I missed it).

Subsequently, people who don't like lowercase argue that it's just an MOS guideline, not policy. Policy is fragile. Some maintenance and repair is needed.

Statistics

The tool at wppagehiststat.pl can be used to get statistics over time and editors. The most active editing month, by a huge margin, was September 2009, TITLE having 481 edits that month, out of a 10 year total of 2820 edits.

This other tool might be easier for showing user stats:

• Editors active on TITLE (with first and last edit dates there):

267 (191/76)	Kotniski	2009-01-12 10:38	2012-01-12 09:51
252 (234/18)	Pmanderson	2006-12-15 19:06	2011-12-03 02:46
168 (157/11)	Philip Baird Shearer	2005-03-10 19:56	2012-01-24 07:44
 98 (95/3)	Born2cycle	2006-08-13 18:06	2012-01-24 06:43
 98 (97/1)	Blueboar	2009-10-07 21:13	2011-12-24 19:20
 94 (70/24)	Francis Schonken	2005-10-04 14:53	2009-11-09 22:07
 48 (31/17)	Mav      	2002-04-28 13:43	2003-03-22 23:07
 42 (38/4)	Hesperian	2009-02-03 03:41	2011-08-14 23:50
 39 (38/1)	Tony1     	2007-11-10 00:30	2011-12-21 10:35
 38 (35/3)	Noetica   	2009-07-14 00:53	2012-01-24 06:08
 29 (25/4)	GTBacchus	2007-01-16 08:19	2011-08-25 22:44
 27 (23/4)	Xandar   	2009-09-06 23:21	2009-12-21 01:11
 24 (24/0)	Rannpháirtí anaithnid	2009-09-17 18:48	2009-09-27 19:54

• Editors active on the talk page

1061 (1039/22)	Born2cycle	2006-11-06 16:50	2012-01-30 19:44
 974 (948/26)	Pmanderson	2006-08-27 19:15	2012-01-27 22:16
 812 (812/0)	Blueboar	2009-09-28 14:07	2012-01-29 14:35
 629 (549/80)	Philip Baird Shearer	2006-01-03 21:31	2012-01-30 02:12
 620 (595/25)	Kotniski	2008-06-07 08:04	2012-01-23 23:42
 530 (519/11)	Hesperian	2008-12-23 03:01	2011-12-26 01:03
 309 (305/4)	Gavin.collins	2010-02-10 16:22	2010-08-21 12:12
 237 (237/0)	Bus stop	2010-05-23 04:28	2010-08-23 14:36
 236 (226/10)	GTBacchus	2007-01-24 08:59	2011-10-14 20:43
 235 (192/43)	Greg L   	2008-12-31 22:00	2012-01-30 20:03
 216 (215/1)	Dicklyon	2008-02-15 01:30	2012-01-30 22:06
 212 (209/3)	WhatamIdoing	2009-06-10 22:40	2012-01-27 23:35
 194 (194/0)	Tony1     	2007-11-10 00:31	2012-01-26 13:27
 179 (141/38)	Francis Schonken	2005-11-09 15:26	2009-09-24 02:57
 171 (170/1)	Xandar   	2009-08-16 23:16	2010-08-21 12:21
 124 (111/13)	Mike Cline	2010-06-29 18:25	2012-01-30 21:54
 123 (106/17)	Bkonrad  	2006-09-26 20:33	2012-01-30 21:10
 115 (97/18)	Noetica   	2007-11-26 00:25	2011-12-30 02:46
 101 (101/0)	Rannpháirtí anaithnid	2009-09-16 20:56	2011-09-05 20:31

In both cases, I've cut off the list at User:Rannpháirtí anaithnid, because his small contributions were so critical to the history.

Edits of a specific user on a specific page can be viewed with this tool (here configured to look at me, Dicklyon, on TITLE). Notice also that I show up on the talk page editors, about a factor of 5 below the most active.