Category talk:American novelists/Archive 2

Rewriting the policy?
Many editors seem to be shooting off, intent on reworking the whole way WP categories work. The guidelines say that categories are set up on a tree-based system: " each categorized page should be placed in all of the most specific categories to which it logically belongs". That's the idea behind the system of categorisation. The 'American novelists' cat should fit in with the guideline - why should an exception be made here? Categorisation has its problems but focusing on one small corner outside the context of the cat system is daft and a knee jerk reaction to media coverage. Span (talk) 21:43, 25 April 2013 (UTC)
 * Who decides what is the smallest unit of specificity to which articles should be categorized? Can you point out a policy with regards to that, or is it a matter of common sense and consensus? Transcendence (talk) 21:47, 25 April 2013 (UTC)
 * The point of the policy is that if an article is in a sub-category of some category, it shouldn't also be in the category (because that's just repetitive, and useless). Noel (talk) 01:40, 26 April 2013 (UTC)
 * You're stating the obvious. My point is that the policy states articles should be placed into the most specific category possible, but it's a judgement call on how specific subcategories get. Transcendence (talk) 09:13, 26 April 2013 (UTC)
 * So, the guideline says to place a page "in all of the most specific categories to which it logically belongs". Super, but that doesn't address which categories should exist in the first place. If a "women" subcategory exists for some category of people, but no "men" subcategory, then we get a situation where the parent category is populated by all men. That's not an intended consequence, and a lot of people seem to be agreeing that it's undesirable. Isn't that the situation? -GTBacchus(talk) 22:47, 25 April 2013 (UTC)
 * Perhaps we can also use this moment to discuss "ethnic" novelists being ghettoized into subcategories, while White novelists are listed under the primary category? groupuscule (talk) 22:49, 25 April 2013 (UTC)
 * Agree with your points, GTBacchus (and Groupuscle). Good summary of current situation. I think this issue could do with feeding back up the chain to higher admin levels of WP, since as seems to be the coalescing view, the framework of how lists are assembled seems to be funneling us to outcomes that aren't really satisfactory. I lack enough of a knowledge of WP structure to do this myself, though. DanHobley (talk) 23:41, 25 April 2013 (UTC)
 * I would submit that Distinguished subcategory, or else a new and more specific template, should be applied to all such subcategories. There are plenty of non-diffusing subcats on Wikipedia, and this really shouldn't be all that controversial. -- Visviva (talk) 01:21, 26 April 2013 (UTC)
 * There are few ways to tell that a category is non-diffusing. Also, I think people are ignoring that at some level if we do not difuse we end up with way to large categories. We also end up with lots of categories on some articles.  Neither of these situation are helpful. John Pack Lambert (talk) 03:07, 26 April 2013 (UTC)
 * Why? Why are large categories a problem? What if some people find them useful? How about lots of categories on a single article? Why exactly is that a problem? -GTBacchus(talk) 16:48, 26 April 2013 (UTC)
 * This problem has come up before. And this guideline addresses it: Categorization/Gender,_race_and_sexuality. Among other things, it says, "A gender-specific category could be implemented where gender has a specific relation to the topic." and "As another example, a female heads of government category is valid as a topic of special encyclopedic interest, though it does not need to be balanced directly against a "Male heads of government" category, as historically the vast majority of political leaders have been male by default. Both male and female heads of government should continue to be filed in the appropriate gender-neutral role category."
 * Now, a few opinions: 1) "Overpopulated" categories aren't a bad thing. And it would be much easier to provide some navigational tools to work through the long categories than to deal with the divide-and-hunt strategy that users need to follow once a category is split. (Am I looking for writers or novelists? A woman or a man? What is C.S. Lewis? A.S. Byatt? George Eliot?) 2) The identity category groups (Women writers) probably don't need the same genre-related specificities as the generic categories. So while we may have American novelists/American short-story writers/American mystery writers/etc, we needn't replicate all of those divisions into the "women" and "men" categories. If there's a secondary literature on the category, then fine. And even then gender-plus-genre-plus-country is excessive: Women science fiction and fantasy writers doesn't need to be broken up by country.
 * And one process request: If your arguments are about our filing system, and not about how users access human knowledge, please get over it. Wikipedia is mostly visited by readers, not editors. It's our job to make their lives and visits easier, not to make ourselves happier about the simplicity of our filing system.--Carwil (talk) 03:29, 26 April 2013 (UTC)
 * I was going to make some of these points, but you've expressed things well. Especially the point on "filing"; the purpose of categorization is to make it easier for users to use and understand the articles. The "policy" of finding smaller and smaller subcategories is not the only or primary consideration in assigning categories. Logically a binary split is always going to be weighted on one side, so will not address the issue of a category being too large, so people should not be overstating the importance of this one factor. The guideline seems clear that gender should not be the defining attribute of a category unless it happens to be specifically relevant to the category, so it would not be appropriate to just remove everyone from the central category and label everyone "men" or "women" where it isn't important (i.e. where it wouldn't exclude one from doing a particular job). There may be an academic or cultural distinction regarding some women authors; I would want to defer to an expert on that. The consensus (or at least apparent majority view) expressed on this page essentially concurs with the guideline. There is no need to "rewrite the policy", just to use the actual guidelines, not just an incomplete understanding of them. Avt tor (talk) 04:56, 26 April 2013 (UTC)

*On thinking about this more, I think we should not subdivide novelists by gender. Instead I think we should subdivide fiction writers by gender. There is a large overlap in short story writers and novelists. I think a gender Category:American female fiction writers works, but divisions for novelists and short story writers is a step too far.John Pack Lambert (talk) 03:36, 26 April 2013 (UTC)
 * I am wondering if novelist is the best term. Especially when we have sub-cats such as Category:American romantic fiction writers.  As it is I think Abiola Abrams demonstats why the fight against diffusion has gone too far.  Abrams is in 3 sub-cats of Category:American novelists as well as in that category.  1-African-American novelists, 2-American women novelsts, 3-Category:American romantic fiction writers.John Pack Lambert (talk) 03:41, 26 April 2013 (UTC)
 * I'm curious, why do you think it's a problem that Abrams is in four categories? If I'm looking for American novelists, why shouldn't I find her there? If I'm looking for African-American novelists, why shouldn't I find her there? If I'm looking for American women novelists, why shouldn't I find her there? If I'm looking for American romantic fiction writers, why shouldn't I find her there? Is she not all of these things? Why do you need people to be in a single category? Being any one of those things does not preclude her from being any other of those things. The fact is most things don't belong in a single category. As a reader, if I'm interested in any one of those categories, I'd like to make sure that anyone who belongs is listed there. Otherwise, for example, if I want to find all romance writers, but Abrams was arbitrarily decided to be placed in only African-American writers, I wouldn't find her. Unless, that is, I checked every writer in every category to find the items which actually belong in more than one category but were arbitrarily limited to one. That sounds awful. Also, I'm browsing the guidelines for categorization, and I can't find any guidelines which say an item must belong to one category. The most relevant thing on this topic that I can find is actually warns against overcategorization, saying: "If an article is in 'category A' and 'category B', it does not follow that a 'category A and B' has to be created for this article. Such intersections tend to be very narrow, and clutter up the page's category list. Even worse, an article in categories A, B and C might be put in four such categories 'A and B', 'B and C', 'A and C' as well as 'A, B and C', which clearly isn't helpful." In other words, things don't NEED to belong to just one category. Salspaugh
 * One of the points of sub-cats like Category:American romance fiction writers is that the parent category becomes too large if we do not diffuse it. The whole reason to have such categories is to split up the parent category more finely.  Anyway, too many categories are not helpful.  James Baldwin's 32 is probably above what is reasonable, and he is not as dual level categorized as he could be.John Pack Lambert (talk) 23:20, 26 April 2013 (UTC)

Ethnicity as genre
Going over the various novelsits articles and trying to figure out how to sub-categorize them by genre, I have come to realize that the best way to categorize some, and I hasten to stress some, novelists is that they wrote novels in an ethnic genre. Not all African-American novelists wrote novels that would be classed as "African-American" as a theme, but some certainly did. I am trying to figure out how to best reflect this fact in categories.John Pack Lambert (talk) 08:54, 26 April 2013 (UTC)


 * It's not all that complicated. Genre is determined by the characteristics of the work, not the characteristics of the author.  Neither the gender nor the race of authors, as such, is a genre.  That is why in addition to genre we have categories such as African-American_literature and Women's_writing_in_English and lists of African-American, Asian, female, etc. authors.  Any given author can be of a particular ethnicity or gender and still write in multiple genres, thus falling into multiple categories or lists.  Reality is not tidy.  HistorianK (talk) 11:04, 26 April 2013 (UTC)
 * However, some writers clearly wrote literature that would be described as being "African-American". In some cases that is clearly the genre of what is involved.John Pack Lambert (talk) 23:23, 26 April 2013 (UTC)

Adding link to list
I just added a link to the list, which should have been present before. It would have made for a lot less nashing of teeth and yelling "sexism" if the list had been linked to to start with.John Pack Lambert (talk) 17:47, 27 April 2013 (UTC)

Using a date range as the basis for a set of sub-categories rather than gender or genre or ethnicity
I confess that I am a new editor. I created an account after reading this discussion. I have a background in English literature, and it seems that using gender/genre/ethnicity as a way to break up a large category is problematic not only due to the reasons listed above (inadvertent sexism or racism, ghettoing, implied choices about the American canon, the inability of gender binary to cover LGBTIQ authors), but also because gender/genre/ethnicity doesn't really break up the list all that much. At most, the gender subcategory would break the main list up into three categories, categories that would be, and rightly so, continually argued because the categories are not neutral to large portions of society. As time goes on, these categories would also grow to larger sizes. I would propose an entirely different set of sub-categories for "American novelist" that would be gender, ethnicity and genre neutral and would allow for the sub-categories to grow with time while keeping the size of each sub-category more or less the same. Many literature classes are taught based on a time period, so these categories would help teachers source authors based on time period without being bogged down by visibility issues caused by organizing by other canonizing categories (gender, genre, ethnicity). I would propose using American novelists by a date range, say in fifty year ranges, so if authors overlap the date range of the sub-category, they would still be in only two of these categories. So American novelists, 1800-1849; American novelists, 1850-1899; American novelists, 1900-1949; American novelists, 1950-1999. — Preceding unsigned comment added by MoonofHecate (talk • contribs) 30 April 2013
 * This is already done. We already have and  and . Not sure if we should divide beyond centuries - but we could add  as a child of . In practice, an individual could be classified as LGBT, a woman, and a 19th century novelist - these don't need to be mutually exclusive. --Obi-Wan Kenobi (talk) 21:50, 30 April 2013 (UTC)

Proposal to remove list of novelists from the article
According to a banner at the top of the article, "This category may require frequent maintenance to avoid becoming too large. It should directly contain very few, if any, articles and should mainly contain subcategories." Apparently it hasn't been maintained properly, so the article would be improved by removing the list of articles. Also, it is redundant with this article. Does anyone object? Olorinish (talk) 23:56, 1 May 2013 (UTC)
 * Wait, what? Which article are you proposing to remove? The main problem is not a single list article, it is the thousands of novelists that have yet to be diffused. You could start by putting people in the category for example.--Obi-Wan Kenobi (talk) 00:07, 2 May 2013 (UTC)

I am proposing to remove the list of articles about individual novelists from this article. It is redundant with this page: .Olorinish (talk) 00:18, 2 May 2013 (UTC)
 * This is not an article, it's a category. do you mean the link that takes you to the catscan tool? If you're talking about that, it's not redundant, I'm quite sure they will give you different answers - one is algorithmically determined by category membership, the other is determined by who edits the list... I agree it may cause confusion - the wording - but they aren't the same thing. Perhaps you could propose a better wording? --Obi-Wan Kenobi (talk) 00:47, 2 May 2013 (UTC)

Thanks for the correction, Obiwankenobi. I propose to delete everything in this category page after "American novelist stubs‎ (2 C, 340 P)". I do not know why a catscan tool would be relevant here. Can someone enlighten me? What is the value of the list of articles in this category page when the other page  already exists? Olorinish (talk) 01:08, 2 May 2013 (UTC)
 * Ah, ok, I understand now. Categories and lists are used for different purposes. read this Categories,_lists,_and_navigation_templates. --Obi-Wan Kenobi (talk) 01:23, 2 May 2013 (UTC)

Thanks again for the information, Obiwankenobi. According to the example category given in this article [Wikipedia:Categories,_lists,_and_navigation_templates], it is permissible to have a list at the bottom of a category page. However, this category page currently has so many (3794) articles that casual readers may not realize that there is a list page which could be more helpful. According to the banner in this category page "This category may require frequent maintenance to avoid becoming too large. It should directly contain very few, if any, articles and should mainly contain subcategories." It appears that this page is not being maintained properly. Should someone look into fixing that? In the meantime, perhaps a warning should be put above the list such as "This list of individual authors is intended to assist the categorization process; it is not intended to be complete. For a list intended for use by the general public, see List_of_American_novelists |List_of_American_novelists." I forget how to put the label in text properly, sorry. Olorinish (talk) 01:48, 2 May 2013 (UTC)

Why is the war about novelists
Why are people attacking the novelists situation when we also have Category:American writers and a subcat Category:American women writers?John Pack Lambert (talk) 16:39, 27 April 2013 (UTC)


 * Because it's the category that got noticed. It has been said by at least one higher-level editor than me (which are most of you all) that the consensus outcome on this one topic will probably be applied to others.  HistorianK (talk) 16:44, 27 April 2013 (UTC)
 * This goes against the normal way we do things though. We generally consider categories as sets.  This allows people to make comments related to the whole set, to see what the whole set is, and allows for better notification.  While this might not be undernotified, it has been notified by extremely biased canvassing, which is highly discoraged.John Pack Lambert (talk) 18:38, 27 April 2013 (UTC)


 * I don't know what you mean by "categories as sets," but if that's how these things are normally handled, Mr. Lambert, why aren't you urging people to comment in whatever the appropriate place is, instead of scattering your comments wherever discussions are going on and never mentioning that more appropriate place?   HistorianK (talk) 20:53, 27 April 2013 (UTC)


 * If this is a template to follow for other categorizations, then I'd hope that we could solicit input from the few remaining members of WP:WikiProject Novels and in fact move the discussion there and rather than deciding here at the obscurity of the category talk pages without many watchers. Truthkeeper (talk) 18:46, 27 April 2013 (UTC)


 * Well if someone wanted they could create a CfD about Category:American women writers, but no one has.John Pack Lambert (talk) 02:33, 3 May 2013 (UTC)

Preparing an RFC
A Request for comment has been suggested at Administrators'_noticeboard/Incidents as one way to resolve the various categorization controversies going on within Category:American novelists. I've created this space to draft the "brief, neutral, and complete statement of the issue" required to call an RFC. My first draft will appear below.--Carwil (talk) 04:09, 5 May 2013 (UTC)


 * American novelists is a large category. Should there be a category which attempts to list all notable American novelists, and where should it be located? What are the appropriate subcategories for American novelists? Which subcategories should be diffusing (if someone is a member of the category, they are removed from Category:American novelists), and which should be non-diffusing?


 * There are already guidelines on these issues at Categorization, Categorization of people (WP:COP), and Categorization/Ethnicity, gender, religion and sexuality (WP:EGRS). According to the last of these, subcategories defined by ethnicity, gender, religion and sexuality, should be non-diffusing. A recent community discussion decided on: "merging the categories back together at Category:American novelists, while keeping the women novelists seperate Category:American women novelists because it is a recognized field of study in the literature."


 * Input is requested on the following, with sample !votes for polling purposes
 * Should there be a category which attempts to list all notable American novelists, and where should it be located?
 * No action — The comprehensive list should appear at List of American novelists
 * List linked — The comprehensive list should appear at List of American novelists. A link to it should appear at the top of Category:American novelists
 * List linked, root category emptied — The comprehensive list should appear at List of American novelists. A link to it should appear at the top of Category:American novelists. No novelists should appear at the root level of Category:American novelists since all are fully diffused.
 * Root category is comprehensive list — All American novelists should appear at the root level of Category:American novelists so that all can be seen there.
 * Novelists by name is comprehensive list — No novelists should appear at the root level of Category:American novelists since all are fully diffused. However, all American novelists should appear in Category:American novelists by name for readers who want to browse a complete list.
 * added: Recursive enumeration gives comprehensive list; some authors left in root category — A link to provide all American novelists via a CatScan of subcategories appears (as currently) at the top of the page.
 * added: Recursive enumeration gives comprehensive list; empty root category — No novelists should appear at the root level of Category:American novelists since all are fully diffused. However, a link to provide all American novelists via a CatScan of subcategories appears (as currently) at the top of the page.
 * What are the appropriate subcategories for American novelists? (multiple choice) For each type are they appropriate or not?
 * By genre (historical, science fiction, realism, etc.)
 * By century (18th century, 19th century, etc.)
 * By demographic group per guidelines — per WP:CAT/ERGS this categories are to be created "only where that combination is itself recognized as a distinct and unique cultural topic in its own right'
 * Full categorization by gender — creating all gender categories in addition to Category:American women novelists
 * Which of the above types of categories should be diffusing?

Suggestions for modification or clarifying are welcome.--Carwil (talk) 04:49, 5 May 2013 (UTC)

discussion

 * List linked, root category emptied and By genre, By century, By demographic group per guidelines. I have no real opinion on creating the American novelists by name category, nor on creating gender specific categories. Silver  seren C 05:22, 5 May 2013 (UTC)
 * Nyet. I don't agree with this scope. This has been my biggest issue with the whole media circus - it has been so completely and utterly chauvinistic. We are supposed to be a global resource (yes I realize we have many readers and contributors from the US) - but our article space covers the whole world. So to call a community-wide RFC just to figure out whether the standard rules for diffusion as outlined in WP:Categorization and WP:EGRS will apply to the special snowflakes we call is ridiculous. What the "community" decided as a result of the CFD was that  would, indeed, comply with WP:EGRS! So as far as I'm concerned, the community didn't bring anything new - the end result was just - follow the guidance we've already had in place for some time.


 * Even the media stories around this, that tracked the minute-by-minute dramatic move of women from one cat to another (One author called it "a small step for man which proved to be a giant leap for mankind", seemed to suggest that this was the first time in the history of the wiki that people had been ghettoized so severely, and it was caught it on camera! Bogus! The tree was created in around 2006/2007, and I bet you money that ever since that tree was created, women have been ghettoized within it. This was American writers complaining about American writers - good for them, and thanks for pointing it out - but the problem is way way way deeper, and much more difficult to solve across the whole wiki. Do we want to solve it? I do.


 * If you really want an answer on how to diffuse novelists, first, look at how they diffused the novels! And read the logs, of how they came to that conclusion. Then, if you want more answers, go over to the novelists wikiproject and start a talk discussion there. (FWIW, you should thank the brave souls who are currently diffusing - if it had been properly diffused a month ago, none of this would have made the media in the first place.)


 * But if you want to do an RFC for the whole community, then ask the community some bigger, more important questions, such as
 * (a) What is the best way to rapidly de-ghettoize our whole tree. We may have solved the problem of American novelists, but there are still in my estimate hundreds of thousands of bios that are ghettoized by either gender, ethnicity, sexuality, or religion, but I don't see any large-scale effort to address this, at all, anywhere. People are instead busy debating whether Mark Twain is a 19th-century American novelist or an American novelist, and starting ANI threads about same.
 * In fact, every time I see a bio, no matter which bio - of a woman, an LGBT, or a person of color, I have yet to see one such bio that is not ghettoized in some way or another. Send me a bio, any bio, and I'll show you how they're ghettoized. So ghettoization is endemic. Do we want more NY times articles written about us, or are we content to just declare victory because we moved a few hundred bios up to . Seriously??
 * (b) What are general principles for diffusion/non-diffusion of categories? Should we refine the guidance on WP:EGRS? Does WP:Categorization need a refresh? There is a discussion happening there now, we need help and assistance. But don't just focus on the silly novels!
 * (c) Can we pilot and test out category intersections - for example using the simple templating-approach/interface with catscan tool I've proposed here. There are issues of uptime and server load, but is the community willing to step up to the plate and do some work and pilot something that could actually defuse some of these issues while waiting for wikidata to arrive and save us? I'm sorry for the rant, but I think your scope above is way too small, and doesn't get to the heart of the matter.--Obi-Wan Kenobi (talk) 05:25, 5 May 2013 (UTC)


 * On the above - I agree with Silver Seren. One point though: Please do not create  - if you go to, you will see a link at the top that you can click, and that will give you a full recursive enumeration of all American novelists, sorted alphabetically by name, and you can even generate a CSV file, or XML, or many other output formats. This works, today. Your suggestion is that instead of using this quite useful tool, we will ask editors to change 6700 bios to add a category, and then ensure that every *new* bio also has this duplicative category going forward for the rest of time, all in order to generate the same list you can generate today with a single click. Sorry, it's just a bad idea.--Obi-Wan Kenobi (talk) 05:34, 5 May 2013 (UTC)
 * I've added the recursive enumeration tool to the list of options. Obi-wan seems opposed to posting this more broadly, but perhaps we can organize our discussion here around these options for now. I'll post my opinion shortly, but am waiting for more input on whether this is a clear framing.--Carwil (talk) 11:27, 5 May 2013 (UTC)
 * I would be fine with bringing this whole discussion over to the novelists project talk page. I would be opposed to opening an RFC on the matter - as mentioned if we're going to steal the communities' attention for a long protracted discussion about categorization, let's address the root of the problem and not this special snowflake tree.--Obi-Wan Kenobi (talk) 14:41, 5 May 2013 (UTC)


 * Root category is comprehensive list - the 'click here' link at the top of the Category:American novelists to 'CatScan 2.0β' has the flaw of presenting a sort on the entire author name - the result is a list alphabetized by first name first. Considering the positon of Wikipedia in the datasphere this category is a defacto certification - irregardless of anyone's wishes.   Neonorange (talk) 19:42, 5 May 2013 (UTC)
 * Would you make the same argument for (a) other countries and (b) other jobs? Or is this *just* about american novelists and how that particular category displays? --Obi-Wan Kenobi (talk) 19:52, 5 May 2013 (UTC)


 * Short and quick answer - yes for (a) other countries and (b) other jobs. I have one argument that I am pretty sure I would apply globally: there is no technical reason NOT to have the root category be a comprehensive list containing all possible members AND a comprehensive category IS more useful than a category of categories.  A second argument is that schema in which higher categories are comprensive of all sub-categories and members is easier for a user to process; after all, we are not LISP. In this schema it is not necessary to understand the method of sub-classification AND there is no ambiguity (as there is in the 'category by century' as, effectively, multiple root categories as argued by some). Neonorange (talk) 00:51, 6 May 2013 (UTC)
 * Are you talking in terms of display, or in terms of actual membership? e.g., if I look at, you're saying, all American writers, no matter what sub-cat, should show up. But are you also suggesting the world "American writer" should appear on the bottom of every bio who is a writer? Those are two different things.
 * Also, FWIW, there are technical reasons this is difficult, because cat structures can nest endlessly. For example, I might create a cat for novels set in London, and just for fun, include the category as a member. This would then mean, I have to recursively iterate through several tens of thousands of articles in the  tree and sort them alphabetically in order to create the display you have today. Now, imagine my London 'novels' cat is within, so now every time someone opens the American novels cat, I have to do the same thing - and also dedupe against the recursive enumerations of other subcats which might include other cities, etc, etc. If you want to see how long it takes, just click the link at the top of  - and imagine waiting that long each time someone opens a cat. Now you may say "why don't wikipedia programmers fix that?" - but that is out of our control... --Obi-Wan Kenobi (talk) 01:42, 6 May 2013 (UTC)


 * LIFO - (a) Though Wikipedia has thrived, flexibility and a 'fix it' approach are necessary in the future as in the past (b) The list link takes a long time to load because unlike the category there the Entire list is served up, rather than just the first page with an index bar (c)Is there no requirement for notability of a category? (d) The condition of 'category that is (at some level down) a member of itself' is just not allowed. The index bar could set the level of nexting displayed? (e) Moving to your just prior response I THINK I am saying that display is first page and two indices: a next lower level subcategory index and an index of the terminal nodes of the current level (but perhaps I don't understand where the problems lie - but I do think this indexing structure should present to the user).If you can direct me to further reading on the subject I would be most appreciative; this is my first sustained discussion in Wikipedia so any help is welcome (and thanks for your responses). Neonorange (talk) 20:22, 6 May 2013 (UTC)
 * Hi Neonrange - I think we're veering off-topic a bit into subtle issues of UI design - can we bring this over to my talk page? happy to chat at length there, best, --Obi-Wan Kenobi (talk) 21:28, 6 May 2013 (UTC)


 * There is no reason to treat Category:American novelists as a special cat. We should disperse it all by century, since that is relatively easy and doable.  There is also no reason to list people here who are fully covered by one or more by genre sub-cats.  The big question is actually not even related to this category.  It is clear this category can be reasonably split by century.  It is unclear that splitting many other nationality cats by century would be wroth while.  I am not really sure what the requirements before we try splitting novelits by century are, but that is really what we should be discussing.John Pack Lambert (talk) 21:39, 5 May 2013 (UTC)
 * By genre (but only where defining of writers) By demographic group per guidelines — These are clear. But should be applied with caution. On genres, "romantic fiction novelists" is a not just a genre/writer combination, but a field of practice in which people genuinely specialized and which meets the guidelines at WP:DEFINING (which are narrower than "a RS called them that" and "this a notable category in its own right"). My best guess is that science fiction, children's, YA, fantasy, graphic, western, spy fiction all meet that, but realism, religious, and satire don't. On demography, we should be looking for evidence that there are studies, courses, anthologies, professorships, articles (in the plural) on American novels by XX group.
 * By century is less than ideal WP:COP calls for one (and only one) way of placing people in time, by year of birth. While Category:Novelists by century exists, there are no other nationality subcats of it. Within the people categories, we have Category:People by nationality and occupation and Category:People by nationality and century but no triple categorizations. This is new territory. And ultimately (in "Wikipedia Lists 2.0"), time should be one way of browsing the long list of American novelists, and the list/enumerated data product should be scrollable by date of birth, date of first work, and various other data that are encoded on each page. Precedent suggests we should put them Category:19th-Century novelists without nationality. I'm not viscerally opposed to subdividing American novelists by time, just want to raise caution that this goes beyond precedent.
 * Re: Obiwan, I'm not trying to create something fundamentally different here at American novelists, just apply the guidelines we have to a particular situation. This seems a lot like the heavily populated Category:American physicists. Many, many novelists don't have a WP:DEFINING genre or subdivision. (Again, this is why American bookstores are arranged the way they are, with one enormous section for novels, and a few related sections broken out.) If the world categorizes that way, we should be comfortable doing so on Wikipedia. However, there is one social fact around "novelist" that is different from many other occupations: the generic category is the higher status way of referring to people in the profession. In almost all circumstances, it's more appealing to be designated a "novelist" than an "X-genre novelist." In many cases, since novels are written for an open audience and novelists frequently switch or draw from multiple genres, this is also the most accurate designation. In short, there are complex and shifting POV issues involved in removing someone from "novelist" to "X novelist" (witness this: Robert_Louis_Stevenson), and we should expect people to get bent out of shape about it.
 * Also to Obiwan, I hope the previous paragraph clarifies why American novelists is the upward bound for bubbling up novel writers (novelists who are also journalists etc. being a special case), and therefore Homo sapiens etc. are safe. For the record, WP:COP also specifies the default place to put people: "People are usually categorized by their nationality and occupation."
 * My opinion re: the list/enumeration depends on whether the root category is empty or full, so I'm holding it back for now.--Carwil (talk) 12:27, 6 May 2013 (UTC)
 * However, the guidelines state, and Bearcat who drafted them repeated this in the CFD, is you should never have gendered or ethnic sub-divisions unless the parent can be fully diffused - exactly to avoid this problem that we ended up with. So if we're going to keep and the other ethnic divisions, we have to, according to our current guidance, fully diffuse - either that, or change WP:EGRS. Secondly, I'm not sure why American novelists would be an "upper bound" - the terms American writer and American journalist are bandied about just as much, and I can't see why those wouldn't also be non-diffusing by the same logic you're espousing - in which case, all the novelists must bubble up to the writers cat, and all journalists would be bubbled up to, and then hence to writers, etc. Finally, on the century cats - yes it is unique, but we also have a very large category here, so the century cats are useful for subdividing in an unambiguous way - thus providing the (non-gendered, non-ethnic, and even non-genre) diffusion called for by our guidelines. I agree not all novelists can be stuck in a genre - but that specific problem - that genre is not fully diffusing, is again what is causing so much angst and consternation, as everyone wants their favorite writer to be an unadulterated novelist. The century cats let us do this, and diffuse at the same time. I don't think we should worry about whether a particular genre would be non-diffusing or not however - given the existence of the century cats, everyone should be in one of those, and then in any other genres that fit. Anyway, why not bring this whole convo over to wikiproject novelists, it's a better spot for it, and see if you can get a consensus there - but I would oppose any such agreement that doesn't apply to all novelist categories re: general rules on diffusion - and if someone wanted to diffuse  in the future, a century cat system would be perfectly reasonable as well there. Finally, recall that the  cat was already fully diffused, by consensus, by the novelists project - and they used by-century cats as well.--Obi-Wan Kenobi (talk) 13:36, 6 May 2013 (UTC)
 * Re: bubbling up and upper bounds: First, I'm happy we won't be talking about Humans, Hominina and other superordinate categories any more. Second, Fooian fooers is not a horrible upper bound. In this case we have American writers:American fiction writers:American novelists. We also have American writers:American non-fiction writers:American journalists. In choosing among these levels, we should be guided primarily by WP:DEFINING. Novelists and journalists are first-line items in obituaries. So we don't bubble up above them, unless the subject meets multiple parallel categories (writes novels and short stories; writes for a newspaper and extends her reportage into a book; etc.). Seems clear enough to me.
 * Why do we put extra effort into bubbling up to these categories? Because journalist and novelist are both categories with root-level members (not further specifiable) and are high-status generic categories where unnecessary over-specifying is socially regarded as dismissal. I'm not saying that no one should be subcategorized below and not bubbled up, just that we need to be especially vigilant about seeing whether the subcategory is WP:DEFINING or not.--Carwil (talk) 13:06, 8 May 2013 (UTC)
 * the reason I had brought up was b/c someone really had asked why cant we have a category of all people on earth. As to bubbling up, I think the solution in American novelists is even better - diffuse down to by-century sub-cats. Journalists you can similarly diffuse by state. Basically, if no-one is an American novelist, and everyone is a x-century American novelist, that gets rid of that "social" status issue which is completely OR and subjective - this notion that a scifi writer is not also a novelist is silly, and the idea that some lucky writers get to be unadorned novelists while others must wallow in a genre is also subjective to a terrible degree. The by-century cats solve this problem nicely - everyone is a novelist, and some are also in a genre.Obi-Wan Kenobi (talk) 13:23, 8 May 2013 (UTC)


 * One problem of dividing Category:American novelists into sub-category by century is that the resulting sub-categories are also going to be too large.  These century cats were all created since the "scandal" broke, its not an established method.  I would suggest we could create sub-categories for each decade, e.g., Category:American novelists born in the 1850s, etc.  We already have stubs that do that like Template:US-novelist-1850s-stub.  Then we could divide each decade by the state in which the author was born, and zip code, for more modern authors.  Then under each zip code we could have a category Category:American novelists named Lastname, Firstname for any duplicate names.--Milowent • hasspoken  18:45, 6 May 2013 (UTC)
 * I love this idea! Especially the novelists-by-zipcode - that would be sweet.--Obi-Wan Kenobi (talk) 18:49, 6 May 2013 (UTC)


 * The 5 digit or the 9 digit zipcode? Contact address or legal address? Neonorange (talk) 20:28, 6 May 2013 (UTC)


 * Agree with everything writes above, and the top post on this thread at Ernest Hemingway goes to this question as well. If we need an RfC to establish consensus (and I think we do), then let's decide where to do that, ask an uninvolved editor experienced with RfCs to draft and administer, and then advertise widely. Once consensus is established, we stick to it. Truthkeeper (talk) 15:38, 9 May 2013 (UTC)
 * Be careful what you ask for... last time there was a big community discussion around this (eg. the American women novelists CFD), the result was a half-implemented mess that is still being debated - the only logical conclusion as a result of that community effort was to put every single novelist, no matter what they wrote, into, or to diffuse them all - and neither of those conclusions was understood by most people voting. I'd rather this be sorted first at the Novelists project - if we can't get consensus there, then advertise more broadly.--Obi-Wan Kenobi (talk) 16:49, 9 May 2013 (UTC)
 * It's gone much beyond the scope of Wikiproject novels (probably always was) which is quite small w/ very few active members. That might be the place to hold the RfC though. Truthkeeper (talk) 18:08, 9 May 2013 (UTC)
 * I disagree. This issue has gotten enough attention and caused enough drama that it should be given the broadest airing possible. I would recommend something like Requests for comment/Categorization of American Novelists, and the shortcut WP:NOVELISTRFC or some such, in the form of WP:ARBCOMRFC previously. We then put it up at WP:CENT and at the BLP noticeboard, Village Pump, etc etc... the usual places. We need as much participation as possible to get as clear a consensus as possible, so that we can move forward one way or the other. UltraExactZZ Said~ Did 20:06, 9 May 2013 (UTC)
 * Yeah, I agree with this. Would also like to see someone uninvolved draft. I'm involved, so would prefer not to ask around. Truthkeeper (talk) 21:45, 9 May 2013 (UTC)
 * Do you *really* want to only talk about American novelists? And do you accept or agree with the point I made above, which is that this proposed RFC will do nothing to address the real categorization issue identified by the press (e.g. ghettoization?) Obi-Wan Kenobi (talk) 20:10, 9 May 2013 (UTC)
 * I think that we need to have an RFC to determine once and for all how to deal with this category. If the discussion at such an RFC can draw on examples from other categories in similar circumstances, well and good. I chose a name out of thin air as something to which anyone could point and say "Discuss it there". An overbroad RFC on category ghettoization, as you put it, would be doomed to fail - precisely because the specifics of each category would scuttle any meaningful consensus. "Well that's OK for this category, but what about THAT one?!?" and so forth. If this RFC can come to a reasoned consensus that works, it could be applied to some of the other categories with little drama. Perhaps there are circumstances that differ between this category and others - other RFCs can follow from this one on those specific issues. UltraExactZZ Said~ Did 00:03, 10 May 2013 (UTC)
 * But we already have global guidance on how to deal with category ghettoization, and global guidance on how to deal with diffusing and non-diffusing categories, and exceptions can be given on a case-by-case basis to that guidance - the real issue before us is, how do we actually de-ghettoize the tree? Why are you all shying away from it? Shall we write bots? Have teams of editors that attack different areas? Does the guidance need to be revised? We do have WP:EGRS, but it needs to be rewritten, and there are all sorts of tricky cases that need to be sorted out - the community input here would be worthwhile. Having an RFC, OTOH, to ask the community whether should be a diffusing category or not is, IMO, a waste of the community's time - CFD exists for exactly that purpose, so take it there if you don't like that cat. We've already de-ghettoized the American novelists tree (at least by gender - I'm not sure if anyone has been moving the ethnic categories up), so the rest of this hoopla is people getting way too attached to the  moniker. I'd much rather we focus on (a) When and where can we create a gendered/ethnic subcat? and (b) what are the rules of diffusion/bubbling up from said subcat up to the parent? (c) in which cases do we create a dual-gender split (e.g. have equal male/female cats) (d) what do we do about intersex/3rd gender people and (e) Do we make exceptions to the last-rung-of-the-ladder rule, in certain cases where the parent is not *fully* diffusing, or is that guidance mostly rock solid?
 * I promise you, no media stories will be written about how we heroically bubbled everyone up to American novelists, nor how we heroically diffused everyone down to non-gendered sub-cats - it's a NON-story for the outside world - either result is perfectly fair. But if, in a months time, there are still 10,000 African-Americans who aren't in a non-ethnic cat, well that's another NY times op-ed for you.Obi-Wan Kenobi (talk) 00:25, 10 May 2013 (UTC)
 * I really don't care what the media writes about. If X person is an American Novelist, they should bloody well be in the category that says so. UltraExactZZ Said~ Did 11:37, 10 May 2013 (UTC)


 * What we would really need is a straightforward way of explaining what a given position means. In other words asking questions in the manner of "Should identity-neutral sub-categories be diffusing (i.e. should an article categorized under '20th Century American novelists' be removed from 'American novelists')?" The most likely result, I imagine, will be that we essentially affirm the broad community perspective of the CfD that the issue is not diffusion in general, but unequal diffusion.-- The Devil's Advocate tlk.  cntrb. 21:04, 9 May 2013 (UTC)
 * Keep authors in American novelists and also place them in subcategories (if appropriate). I'm still shaking my head that we're even having this discussion. The essence of categories is to help people find articles and subjects. Just as most subjects on Wikipedia belong to more than one category, most authors can belong in more than one category. Simply keep authors in the overall American novelists category and also include them, if needed, in subcategories. And make no mistake--removing authors from the American novelists category is a political decision which will bring negative attention to Wikipedia. But it's also a decision that does not make sense given what categories here were originally created to do. --SouthernNights (talk) 00:43, 10 May 2013 (UTC)
 * Yes, completely agree with your comment. That's what, imo, the RfC has to about - why is the American novelist category being removed from biographies of American novelists? Why can't novelists stay in the novelist category as well as belonging to appropriate subcategories? Truthkeeper (talk) 01:04, 10 May 2013 (UTC)
 * I've yet to hear a good reason why this is being done.--SouthernNights (talk) 01:15, 10 May 2013 (UTC)
 * Should you want every article to have like five dozen categories at the bottom then it is a great idea. Should you want to have thousands of articles show up in every broad category and have those same articles repeated in categories within that category without any obvious need then it is a great idea.-- The Devil's Advocate tlk.  cntrb. 03:28, 10 May 2013 (UTC)
 * Absolutely. Putting the category at the bottom of the article says "The subject of this article is X". If the subject of an article is, say, an American Novelist, then they should be in the category "American Novelists". THAT's the part that seems to be so difficult to understand. Indeed, that's how this whole debate started - someone wasn't listed as an American Novelist and got angry about it - not unreasonably so. UltraExactZZ Said~ Did 11:37, 10 May 2013 (UTC)
 * So if the media says X is an or an  or an  or an, we should put that at the bottom of their profile? All of these other categories are already mostly diffused. I'm sorry but your example is not very good - especially in this case, where what is being disputed is whether  is diffusing or not. At least half of the characters in "20th-century American novelists" spells "American novelists", so putting someone in that category as opposed to the unadorned version still labels them as an American novelist - I'm not sure what they've lost - they've only gained some specificity. They're not less of an American novelist by being in a by-century sub-cat, any more than an Irish poet is less of an Irish poet b/c they're in .--Obi-Wan Kenobi (talk) 12:17, 10 May 2013 (UTC)

A compromise to consider?
It seems to me that it is seeing the term "American novelist" at the bottom of the bio that is most evocative to people. The other argument that has been mooted, that of "having-one-list-of-all-novelists" is a bit of a red herring, since:
 * we've never had such a list (before this mess started, there were 3000 bios that had never been American novelists at all - they were only in sub-cats, and I've seen no evidence of complaints from readers), and
 * we can today get such a complete list of all novelists by clicking on the link at the top of the category page, no matter how the novelists are categorized - it doesn't matter, as the list is recursive.

Thus, what if we did this:

rename to  or   (as well as the siblings)

That way, the term American novelist is more front-and-center, and will align better with things like and Category:American_diarists which aren't diffused. Then, we can diffuse the parent to eliminate needless duplication amongst parent/child, but those who want to see "American novelist" at the bottom of a category can still do so.

Even better, when a rookie editor comes along and adds a novelist to and nothing else, we can easily spot it, and move it down to the proper genre/sex/ethnicity/century categories. If all novelists are in the parent, it's much harder to find those in need of further specification, without trolling through 10,000 bios.

I'd be happy to support such a rename, and even argue why it should be different than other cats of the same name - due to the evocative power of the term "American novelist".

I welcome your thoughts. Cheers! --Obi-Wan Kenobi (talk) 18:28, 10 May 2013 (UTC)
 * I tend to agree with the "divorced from content" angle, since I doubt that a coherent head article can be written about 19th century American novelists. Which is why I expressed reservations above. Unlike these "ought not to exist" but like signs on bookshelves that read "Literature (A-F)" and "Literature (G-J)," they can help with navigation. I think Obi-wan's parentheses help remind us that this category is basically a navigational aid.
 * I have to add a caveat to my acceptance. We should produce a consensus bit of advice on categorizing within this domain (no, American novelists aren't a snowflake category, but we should try it here and then see if it can expand to similar situations.) Guidelines should include: Everyone should be listed in these parenthetical century cats, and no one should be in the root level. There should be no subdivision of century cats. The CatScan link should remain here.
 * At this point I would regard Obi-wan's new suggestion of Category:American novelists (20th century) and my suggestion of Category:American novelists by name as equally workable ways to resolve this. With no preference for my own over Obi-wan's.
 * Process: I think we need to try to get consensus, but also feel like we haven't done a very good job of bringing in the many people who commented on categorization in the last month. Could we firm up the main alternatives, see who has outstanding objections, and bring in folks from the prior conversations, WP Novels and WikiProject Countering systemic bias/Gender bias task force (which was recently set up to address the ghettoization problem, via notes on each of their pages. (Of course, we should wait on initial comments here as well.)--Carwil (talk) 12:59, 11 May 2013 (UTC)


 * I have no problem with having these by century categories, but I do have a problem with taking novelists out of the main category to put them into by-century categories. As I've discussed elsewhere with Obi Wan I think such categories are useless and would be impossible to define inclusion criteria for properly.  Thus when novelists are placed into them to the exclusion of the main novelists category it will potentially hide them somewhere unexpected to people who are thinking of their centuries differently than the categorizer.  Even though I think they're useless and inconsistent I have absolutely no objection to people making them as long as they don't come chuffing round like automatons taking articles out of this category to put into that category on the basis of their interpretation of WP:CAT or whatever. TL;DR we desperately need an RfC with a lot of eyes on it.&mdash; alf laylah wa laylah (talk) 15:08, 11 May 2013 (UTC)
 * So basically you're saying that the century cats are ok, perhaps you'd like more specification on who goes in what (-- welcome your views and assistance there -- ) -- but you don't think they should diffuse. That's a fine argument to make, but you have to come up with a compelling case, esp since these types of cats always diffuse elsewhere. Usually a non-diffusion argument is because you've specified part of who they are, but you didn't capture the whole essence. In this case, I think we have indeed captured the whole essence - Hemingway is a 20th century American novelist no matter how you slice it, and so are many others. There isn't anything 'left' of Hemingway after diffusing that needs to remain in the parent (whereas with lets say, there are other aspects of them which are not captured in the cat.).--Obi-Wan Kenobi (talk) 01:13, 12 May 2013 (UTC)
 * I think the basic problem here is that you're defending an Aristotelian notion of categorization, in which there are natural categories and definable necessary and sufficient characteristics which determine what falls into those categories. I think that this is a delusion.  I think that Eleanor Rosch, George Lakoff,  Wittgenstein, and their followers have criticized the possibility of such categorization schemes completely convincingly.  Following Rosch I think that the best solution here is to make all categories non-diffusing except for what she calls "basic level categories" (there's a reasonably good explanation here: Prototype_Theory.  These are not definable a priori by necessary and sufficient characteristics but must be determined by analysis of natural ways people tend to categorize things, which can vary wildly from subject domain to subject domain.  I'm not planning to spend any time making compelling cases for novelists or any other group because I think it ought to be left up to editors who work in the area and that categories ought to be considered nondiffusing by default until that conversation is had.  You think that there isn't anything left of Hemingway after you've decided he's a 20th century novelist, but look!  A lot of people disagree with you.  And these are people who write long excellent articles about novelists, too, so they might know something about it after all.  I would say that the reason you can't convince them is because American novelists is a basic level category and every subcategory of it ought to be nondiffusing.  How do I know it's a basic level category?  Because if it weren't you and your category nerd cohort wouldn't be getting such pushback when you try to diffuse it.  You notice that no one at all has argued that American novelists ought to also be put into Novelists, right?  This is how linguists, psychologists, and philosophers of language figure out which are the basic categories.  It's parallel to the famous example about what people say when you ask them what they're sitting on.  They say they're sitting on a chair.  They don't say they're sitting on an early 20th century blonde oak Danish modern chair and they don't say they're sitting on a piece of furniture.  "Chair" is a basic level category.  Why?  Who knows why, but it is an empirical fact that it is.  The fact that Wikipedia categories are supposed to be arranged by default into a conceptually antiquated tree structure is just an artifact of the nonprocess through which they were developed.  It's never going to work because it's contrary to the way humans experience reality.  Maybe a simple way to solve the problem would be to assume that all categories are nondiffusing from today onwards and require that editors who want to argue that a category should be diffused make the argument, and maybe also require that they've actually added some content to any article they want to recategorize just so there's a hope in hell they even know what it's about.  Then the natural process of Wikipedia would let us find the basic level categories through discussion.  Or maybe that wouldn't work, I don't know.  I do know that having algorithmic rules that attract editors who believe unquestioningly that unambiguous categorization is even possible and then allow them to have drive-by categorization wars where their only response is "categories are diffusing by default" is not actually getting us anywhere.  This is already too much so I'm not going to go into Wittgenstein's criticisms of Aristotelian categorization based on family resemblance, but it provides a completely independent attack on the possibility of Aristotelian categorization.&mdash; alf laylah wa laylah (talk) 02:25, 12 May 2013 (UTC)

Yes, we can spend time talking about the theory of categorization, etc., but I'm not defending based on Aristotelian theory, I'm defending current practice and current guidance, which is itself the result of tens of thousands of hours of debate and discussion and work by thousands of editors. I just found this quote on Category_talk:Writers from 2004: "I think it would be better to break them up as is. I.e., writers by country, writers by genre, writers by type. If you want all writers, how about Writers:A ...Writers:Z. Too soon to do it yet, I think. according to Wikipedia:Categorization, if articles are in a subcategory, they shouldn't be in the parent category." - so even back in 2004, diffusion was the accepted philosophy - and even today, the cat is mostly empty.

As I mentioned on your talk page, I agree the system is deeply flawed, and of course I don't think that any categorization is anywhere near perfect, or even that a given diffusive-category actually diffuses properly all characteristics of Mr. X. It's all a vast approximation, given the tools we have in front of us, with a goal of both giving the user useful information, and making it easy for editors to maintain the categorization and correct errors. It's not intended to be the perfect solution, obviously.

But for now, we have the guidance: which states states that all categories should diffuse by default. Remember - if you make a category non-diffusing, that's basically saying, every article in that category - *and all of it's sub-categories* - must be bubbled up to the parent. Non-diffusing is not a casual thing, like "yeah, do it if you like" - if something is non-diffusing, you have to bubble things up - otherwise you're not doing it right.

For example, suppose we divided (which is non-diffusing) into, etc, and thence to  and thence to  and so on. Now, every single article in every single sub-category - no matter how deep those categories go - would logically have to be bubbled up to the parent/grandparent/whatever - and if, in the future, the grandparent cat *itself* became non-diffusing, then you have to go back through the subcats and bubble them up again to yet another parent.

So in general, non-diffusing categories should be leaf nodes, at the bottom of the tree - that's why I think your argument of making them all non-diffusing by default is a non-starter, because non-diffusing is an exhortation to editors that they must bubble up the articles to the parent (or appropriate diffusing sub-cat of the parent), and the chance that editors follow such instruction is slim ... very slim, especially as recursion starts to take hold.

And you saw what happened when editors didn't follow (rather simple) instructions around non-diffusing in one small corner of the tree.

Non-diffusing categories also multiply the possibilities of incomplete or inconsistent categorization; for example, if I expect (as many people seem to) to find all novelists in, are they rudely surprised when they discover 50% of the novelists were never there at all?? If, OTOH, you make all subcats non-diffusing and announce it from the mountain tops, and have a massive RFC, and bubble every single novelist up -- etc - after 3 years, when people have gotten on with their lives, and newbies have come and started adding novelists, etc - once again the parent cat will be a shallow imitation of the whole - but readers showing up there *won't* realize this (because the top of the page will state: "all novelists are here!") - and it's only through using some sophisticated external tools (cat_scan, but used a very specific way) that you can even detect that there are novelists which have not been bubbled to the parent. So now there's this whole additional maintenance issue that gets added with non-diffusing cats - whereby someone or some bot has to ensure that the proper bubbling up is happening. Having considered what it might take to do this algorithmically, I'm not confident a bot could correctly deal with non-diffusing cats and detect when something was out of whack - so it's back to (volunteer manpower), which in the categorization world is extremely rare. I guesstimate there are only about ~20 or so people who do serious, deep restructuring work with categories - and most of them wouldn't care too much for the duplicative work that making everything (or many more things) non-diffusing would imply.

It's just much *harder* to deal with non-diffusing categories, they confuse the heck out of people, and cause extra/repeated work. Again, perhaps a nice idea in theory, and I'm not gonna fight wittgenstein and actually agree with him on many points, but how do you implement a solution like that, or like all-categories-are-non-diffusing, in a volunteer encyclopedia where most editors don't even touch cats in the first place?

I think ultimately, diffusing categories are something that work in practice, even though they don't work in theory.

Your concept of "basic category" is interesting, so I think you should take that over to the category talk page and see if there is something you could do in terms of guidance changes there - try to write it up in a way that is doable without too much philosophical gobbledygook - editors near clear guidance they can follow when setting up cats, and I'd be very worried about a new epic battle around whether is a "basic" category or not. As an example, I took a look at, and there were no birds listed there. --Obi-Wan Kenobi (talk) 05:24, 12 May 2013 (UTC)

What about non-men/non-women novelists?
I understand why we would want to separate novelists by gender, since people are often searching for a specific one, so we should have male novelists and female novelists categories as unique identities, making them equal by having both exist. But what about people that aren't male or female (or are both, ect?). I've read a few people above saying that these people should be relegated to an "Other novelists" category and I have to say that that sounds extremely offensive, because you are relegating people to the naming of being "Other". Isn't there some better naming convention we can come up with that is, at least, minimally offensive? I know we won't be able to satisfy everyone, but we can do better than "Other". Silver seren C 23:34, 27 April 2013 (UTC)


 * There are a lot of potential options: genderqueer or third gender or intersex or bigender ... it's complicated. Transgender is sometimes used as a catchall for the non-binary options, but transgender people also often identify with one or the other of the binary genders, and should be classified according to which they prefer.  Any category that encompasses the non-binary alternatives would have to include a short discussion of the multiple options, and be careful to include only people who definitely reject the binary gender system.  HistorianK (talk) 01:34, 28 April 2013 (UTC)


 * We are talking about a rather small number of people here. I would suggest that transgender novelists be placed in both categories, or in both categories plus a special transgender novelists category. I don't think transgenders should be classified according to what they prefer; sometimes they may be well-known as being of one gender, but prefer the other, or there may be some doubt as to what gender they preferred (at what time in their life? For how much of their life? For what part of their career?) I think the simple thing to do is to put them in both categories. Brianyoumans (talk) 16:36, 28 April 2013 (UTC)


 * No. Transgender novelists who identify as male or female go in the category according to their current identification. This is the standard across Wikipedia. Anything else is disrespectful and misleading.Bonatovada (talk) 01:54, 30 May 2013 (UTC)

"American Women Novelists" must survive
As a largely inactive but longstanding Wikipedian, I just wanted to nip in with one very important point here: whatever solution is chosen for this problem, and I think it's a real problem that does need a solution, there is simply no serious way in which the American Women Novelists category can go. From the perspective of a literature teacher on the college level, the idea is ridiculous. We still routinely teach courses on female authors and people still routinely do research on that specific area. Deleting a category that is an active research topic in the world is farcical. And this goes for race, sexual orientation, and other minority statuses.

I have little particular investment in "American Men Novelists," a category that seems unlikely to be used often as a tool in any practical sense, but if it's the path of least resistance, go for it. But to delete categories that separate sets of things that are active objects of scholarly research and teaching is clearly not a sensible route for Wikipedia to go down. Phil Sandifer (talk) 23:45, 27 April 2013 (UTC)


 * ...so Wikipedia should propagate this particularly pernicious form of sexism so that it will be easier for literature students to participate in their chosen category-of-analysis research? Shall we now begin to cross-reference Wikipedia with syllabi? That will likely be an uncontroversial exercise! Just because something is an area of research does not confer it reality.(See World_disclosure.) ChristineBushMV (talk) 21:09, 23 May 2013 (UTC)


 * What would be the most politically correct feminist option here? Women's studies (previously known as "Gender studies"), African American studies, and ethnic studies are established academic fields in the US. Therefore it is scientifically justified to have subcategories reflecting the research interests of scholars in those fields. This situation exist largely because of the minority push for emancipation. Should we now let feminism have its cake and eat it? (= dual categorization). -- Petri Krohn (talk) 00:09, 28 April 2013 (UTC)


 * This is driving me crazy. It's not "scientifically justified" to have subcategories. Whether or not a subcategory is justified is not a scientific question in any sense of the word. "Scientifically justified" is a nonsense phrase in this context. -GTBacchus(talk) 02:28, 28 April 2013 (UTC)
 * No, what we should do is populate the parent category w/ all the novelists: women, men, martians, whatever, and then disaggregate out from there. Why can't a person be a novelist, a woman/man, white/native/hispanic/asian/etc., write in a specific genre, be born in a certain decade and place? Why can't categories overlap? Isn't that the nature of people? Not allowing categories to overlap is cause of this mess. Truthkeeper (talk) 00:20, 28 April 2013 (UTC)
 * Yes, "not allowing categories to overlap is cause of this mess", but that is how Wikipedia works. We cannot change the rules on this occasion, only because someone in the New York Times wants us to. -- Petri Krohn (talk) 00:36, 28 April 2013 (UTC)
 * We can change the rules any time we want to, because they need changing. It's not because someone in the NYT wants us to (ill-informed - that was an op-ed by a non-journalist who is not "in" the NYT in any sense), it's because we realize it's the right thing to do. -GTBacchus(talk) 02:28, 28 April 2013 (UTC)


 * Nah, I wanted us to 3 years ago much before the press got wind of this. We just don't categorize correctly and that's what the press has picked up. Overreacting to the press is wrong but rethinking how to categorize wouldn't be a bad idea. Truthkeeper (talk) 00:44, 28 April 2013 (UTC)
 * Amen. -GTBacchus(talk) 02:28, 28 April 2013 (UTC)
 * Relevant quote from The Five Pillars of Wikipedia: "Wikipedia has policies and guidelines, but they are not carved in stone; their content and interpretation can evolve over time. Their principles and spirit matter more than their literal wording, and sometimes improving Wikipedia requires making an exception." HistorianK (talk) 01:14, 28 April 2013 (UTC)


 * It would be a mistake to think that I'm making this argument on any point of general principle about minorities or political correctness. I'm making a far more basic point: my students will plausibly need to look up specifically American women authors. This is just a fact of the world. To abandon the category would be to actively decide not to serve information up in a useful fashion. Whatever one thinks of women's studies, the minority push for emancipation, or anything else, we live in the world we live in, and that's the world Wikipedia is trying to serve. Not some theoretical utopia. Phil Sandifer (talk) 02:12, 28 April 2013 (UTC)
 * Absolutely the category should stay. But so should American novelist. My argument is that we need both - by depopulating the parent category we end up with white men only which is what started all this. Categories have to be able to overlap. A woman novelist is still a novelist, no? Truthkeeper (talk) 02:30, 28 April 2013 (UTC)


 * Phil, don't take this the wrong way, but the "my students need this category" argument is the most ridiculous excuse for a category, ever. That is now how wikipedia categorization works - because if we had to create categories for every interesting intersection that would be studied, we'd have way too many categories. I'd suggest you learn how to use WP:CATSCAN and teach your students how to use it, that will serve you much better than depending on wikipedia for accuracy in categorization.
 * Let me give you a very crisp example: African-american gay literature - which is an important topic of study:, about which many books have been written. This is what we might call a quintuple intersection - ethnicity, sexuality, job, nationality, and gender. However, as you might expect, there is no category. Shall we create one, since Dr. X is teaching a class about it this semester and his students *need* to look up all of the African-American gay writers? No no no. However, if Dr. X is clever, he can show his students how to use WP:Catscan, and intersect Category:LGBT_African_Americans and Category:LGBT_writers_from_the_United_States and perhaps Category:American novelists which should give them a pretty decent list to start with.--Obi-Wan Kenobi (talk) 14:21, 28 April 2013 (UTC)
 * Obi-Wan, I have to disagree. "My students will use this category" is an excellent reason to have one. Categories exist for readers to use, so a category that we know readers will use is one we should have. Otherwise, what are they for? Saying that people just need to learn more technology is no good, because we're supposed to serve our readers, not make them do extra work. Implementing CatScan into our search engine - now that would be useful. Implementing category intersection, so that searches naturally use CatScan - that would be excellent. Until we can do those things, implementing categories that we can reasonably expect readers to use is precisely what we should be doing. -GTBacchus(talk) 15:15, 28 April 2013 (UTC)


 * This has been said before, but it's worth repeating over and over again. If we could apply tags to articles instead of categories, none of this would be a problem. Then readers could search by any Boolean expression involving tags that they want, and nobody would be removed from any category just because they also fit in a subcategory. They would just have more tags. -GTBacchus(talk) 02:44, 28 April 2013 (UTC)
 * But we can use categories in the same way - it's really not an issue. The parent cat page shows the levels so it's easy to drop through the levels. The parent cat is also alphabetized so if for instance the person searching isn't sure whether they're searching for an author who is a woman of say Native American ancestry but they know the author's name, they can search alphabetically (or really just search for the article). Having a huge parent cat shouldn't scare people because it can then be disaggregated and sliced and diced down to the tiniest subcategories. Truthkeeper (talk) 02:51, 28 April 2013 (UTC)
 * Tags would still be better. Using categories as they exist now depends on someone creating and maintaining all the categories people might search for. With tags, each search query creates a custom category for that reader, on the spot. Maintenance-wise, it's far easier to keep a short list of tags on an article than a long list of categories for each intersection of properties. As an example, someone could be tagged as "American", "Female", "Hispanic" and as a "Poet". Those four tags do all the work of fifteen different categories: "Americans", "Women", "Poets", "Hispanics", "American Women", "Hispanic Americans", "American poets", "Women poets", "Hispanic poets", "Hispanic Women", "American Women poets", "Hispanic American poets", "Hispanic women poets", "Hispanic American women", and "Hispanic American Women poets". Do you see why I'm claiming that tags are more efficient? -GTBacchus(talk) 02:58, 28 April 2013 (UTC)
 * I absolutely agree that tags are better but taking even baby steps is such a painfully long process here that I think we're stuck with categories. The depopulation of categories is a situation that's frustrated me for a long time. Truthkeeper (talk) 03:06, 28 April 2013 (UTC)
 * Huh. This suggests to me a two-pronged approach. On one hand, work to change the dogmatic idea that any large category must be diffused into subcategories. This is how we address the situation given the status quo. At the same time, begin steps towards getting the developers to implement tags. In the long run, we'll have to do it, because as we've noted, our category system is pretty much broken. This is the twenty-first century; we should get with the program. -GTBacchus(talk) 03:11, 28 April 2013 (UTC)


 * "Assume a theoretically ideal student who knows how to do research well" is a lovely thought, but generally leads to poor results. I'll grant that how to usefully organize all of the various categorizations and tags that might be useful for sorting articles is a challenge. But it's the sort of challenge that Wikipedia, by its nature, signed on for. And removing this category would be a step backwards in that regard.
 * Really, the absence of dual-categorization makes this harder too. But then, the categories system was always a kind of messy kludge. I remember when it came out and we tried, frankly not that well (and I say this having spearheaded it) to sort out the appropriate uses of categories, lists, and infoboxes. All of these were information management tools, and they were barely thought out at the time. And are at this point nearly a decade old. It's time for a larger scale revision to how Wikipedia presents its information.
 * But for now, let's at least generally speaking think about the question of what's useful to readers, no? Phil Sandifer (talk) 17:19, 28 April 2013 (UTC)
 * Yes, agreed. I have to laugh at the suggestion people should just use CATScan, I haven't even heard of that, I can't see the incensed bloggers writing about this topic (many thinking there was some sort of Wikipedia grand council decision behind this) finding something like that. I support implementing tags. BTW I love TARDISEruditorium. :)-- occono (talk) 22:09, 28 April 2013 (UTC)
 * If CATScan can solve a problem you have that is not addressed by the current category system, why is that laughable? You should realize that there are literally millions of valid questions that could be asked of the current system (e.g. give me all articles that are in X and Y, or give me all articles that are in X and all subcats, or give me all articles that are in X, Y, but not in Z) but that cannot be answered by wikipedia today, but that can be answered by CATScan, today. The response to just add more cats to articles is ultimately a losing proposition, as it's a combinatorial problem space: the number of reasonable intersections grows exponentially with every characteristic. If we're talking about scholars (or students) doing serious research, asking them to learn a simple tool is anything but laughable. I completely agree with moving to what you call a tag-based system, but we aren't there yet - this is a workable solution, today.--Obi-Wan Kenobi (talk) 23:12, 28 April 2013 (UTC)
 * Well, try and promote CATScan to non-editors then. I just don't see it taking off, most people never use advanced search in Google for example. But, well, put a link to CATScan somewhere prominent....-- occono (talk) 23:23, 28 April 2013 (UTC)
 * Check this page - see what you think. . --Obi-Wan Kenobi (talk) 23:25, 28 April 2013 (UTC)
 * Well, OK. Do you not think that tool looks a little offputting? It looks really complicated. -- occono (talk) 23:29, 28 April 2013 (UTC)
 * Yeah, but that's just a UI issue. Think about it - if I'm X the American novelist researcher, and I've been dreaming of a list of *all* American novelists, including from all subcats - boom, I now have it. It would be trivial to update the UI to make it a bit simpler, remove some of the fields, and make it look wiki-like - we just need to get some coder time, perhaps work with the tool author to prepare a beginners interface.--Obi-Wan Kenobi (talk) 23:33, 28 April 2013 (UTC)
 * No, it's awful. Worse than what we have and what we have is problematic. Why is there a problem with populated parent categories? Truthkeeper (talk) 23:35, 28 April 2013 (UTC)
 * Because there are too many potential parent categories! This whole notion of 'dual categorization' is from people who don't understand how categories work - there aren't *two* categories are play here - there are dozens - e.g., , , , etc. Once you start bubbling up to the parent in order to satisfy some user who wants a full list, where do you stop? Which parents do you choose? Shall we add 5 new categories to every writer article, just in case? Do we only bubble up for gender (as per WP:EGRS), or do we also now start bubbling up based on thematic cats - so all writers in should also be placed in  - and if that's the case, how do you identify all of the novelists who haven't yet been classified in a more detailed sub-cat? If everyone is in the parent, it becomes quite tricky to identify those in need of further sub-catting. Please ignore the interface for now - consider the FUNCTIONALITY - e.g. a single click, and you have all members including from sub-cats- and it's dynamic, so as category parent/child relationships change, this still works. Your solution of a hard-coded parent-child "dual" categorization is quite brittle, and any change to the cat structure will break it, and possibly create instant sexism! Isn't a better solution worth fighting for? We need to replace our human algorithms (manually adding the "correct" cat to an article) with machine solutions.--Obi-Wan Kenobi (talk) 23:49, 28 April 2013 (UTC)
 * People keep trying to explain this as though I have no brain. I do understand and if you've read Newyorkbrad's comment on Jimbo's page, then you'll understand what I'm trying to say. If there's a problem, we're doing something wrong. I work very much with articles about literature and am very familiar with the categories. That's why I've been commenting: I'm invested in the area. Anyway, I don't think this will go anywhere, but I do think that link to the tool should be removed. Truthkeeper (talk) 00:05, 29 April 2013 (UTC)

Proposal: turn Category:American novelists into a container category. I confess that I was blissfully unaware of this controversy on-wiki and learned of it in the media only recently. I do think subcategories such as "women novelists" or "African-American novelists" are legitimate if they reflect actual categorization in the field of literary studies: are novelists categorized in this way in college courses, anthologies and so on? Even historically, however, it's misleading to create the impression that white male novelists are the norm from which all others depart, which is precisely what the categorization implies if we shunt everybody else to a subcategory. In the main period of the American novel's existence (19th and 20th centuries), female novelists have been as noted, influential, and popular as male novelists, though in lesser numbers canonically. At present, some would even suggest that fiction is primarily a woman's game. If it's important to depopulate the large category "American novelists," then why not make it a container category? Anyone who's familiar with literary criticism of the last thirty years knows that "white male novelist" is an active category of discourse; "white male writers" is far more common, and "male writers" even more so. You can even google "Updike" + "white male novelist". Or see this piece by Katie Roiphe. Search "male novelists" and you get a bounty, including books with "male novelists" in the title. "Masculinity studies" in literature has been a thing since the 1990s. It seems to me that turning "American novelists" into a container category is the only intellectually legitimate solution, if an author can't inhabit both the parent and the subcategory. We can still offer the option of clicking to see the full list, but if gender is a legitimate category in literary studies (and it demonstrably is), then let's sauce the gander too. Cynwolfe (talk) 12:19, 30 April 2013 (UTC)
 * You may want to weigh in here Categories_for_discussion/Log/2013_April_25. However, even if that category survives, it will not mean that men will be placed only in that one - as for now guidance is that gender categories are non-diffusing. I've proposed (elsewhere) that we change this, and always fully categorize/diffuse by gender (if we do one, we do both, and everyone is in a ghetto). However, that's not likely to get consensus support.--Obi-Wan Kenobi (talk) 13:20, 30 April 2013 (UTC)
 * Thank you. As I said, this isn't something I've been following here, though I contribute actively. Cynwolfe (talk) 15:05, 30 April 2013 (UTC)

Categorization quiz
You're welcome to participate in a categorization quiz, as a way of understanding and teasing out the complexity of categorization, esp around gender, sexuality, religion, and ethnicity. The more participants the better! Wikipedia_talk:Categorization/Ethnicity,_gender,_religion_and_sexuality.--Obi-Wan Kenobi (talk) 22:45, 28 April 2013 (UTC)

Classification by century
Category:20th-century novelists has existed since July 2009 and Category:21st-century novelists has existed since February 2010. The American sub-cats may be new, but the parent cats are much older.John Pack Lambert (talk) 21:11, 6 May 2013 (UTC)
 * It should also be kept in mind that Category:19th-century American writers was created last fall, and many other writers by coutry and century categories are older than that.John Pack Lambert (talk) 21:37, 6 May 2013 (UTC)
 * This break out by century makes sense because currently we have a very high rate of overlap between the by-genre sub-cats and the parent category.John Pack Lambert (talk) 21:57, 6 May 2013 (UTC)


 * 'Classification by century' for 'American novelists' has disadvantages when used as a way to depopulate 'Category:American novelists'. Ambiguous, Procustean, and tedious come to mind. More useful to specialists than to the general Wikipedia user.  In the long run not even very useful in reducing the number of terminal nodes in categories by century considering the acceleration in number of novels published by year (or the birth years of novelists, or first publication date, or copyright date or ...).  Not something I care about: UNLESS 'Category:American novelists' is depopulated of 'Author Names' (terminal nodes).  And UNLESS categorizers act in ways that seem divorced from content.  Overlap does not seem to be a bad thing.  Cultural works and cultural workers are always going to have overlap and ambiguity. Neonorange (talk) 02:52, 10 May 2013 (UTC)
 * ooh that's a good new word I just learned - Procrustean. Nice one. :) a question : what do you mean by the first unless clause - eg terminal nodes? I think JPLs proposal above is to empty American novelists and stick everyone into the by-century cats, as well as any applicable genres genders, or ethnicities. And what do you mean by 'divorced from content'? Sorry just need help parsing...Obi-Wan Kenobi (talk) 03:10, 10 May 2013 (UTC)


 * I used 'node' to mean any member of a tree-like data structure. I used terminal node to mean nodes with no children; in the case of 'Category:American novelists' this any 'Novelist Name'.  'Category:American novelist' has nodes that DO have children - 'Category:American graphic novelists' has, as I type this, 71 children, all of which are terminal nodes.  'Category:American Chicano novelists' is at the moment is childless but has potential to have nodes further down the tree. (Just looking at the proposed criteria for membership in that category brings up the question 'Who decides?)  Which leads to 'divorced from content': Wikepedia categorization in some areas require more than just familarity with the cultural and political context - there ought to be oversight by a wide community.  Specifically, removing the 'Author Name' items from the root 'Category:American novelist' to the by century child categories does not really eliminate any 'perceived' overpopulation problem - each century is going to have more members than ALL the previous centuries. The 20th Century list has more members than 17th + 18th + 19th.  The 21st will have more than 17th + 18th + 19th + 20th.  For me authors by century is only of technical interest and an example of fitting to Procrustes Bed B^)  . Cultural works and cultural workers make for fuzzy categories and NOT well formed trees (nodes can have more than one parent is an example of NOT well formed tree). Neonorange (talk) 18:10, 10 May 2013 (UTC)
 * ah ok - thanks. So do you think all novelists should be in, or none? Can you think of another way to fully diffuse the tree (besides centuries?) I agree, the numbers will swell, but I think literature is often studied by centuries, and I think it's useful to have a breakdown by century (even if 21st century will have lots, obviously, esp in 100 years - but hopefully by then we'll have category intersection, and this discussion won't matter anymore!). There isn't any oversight to categorization, except on a per-article level, and if someone wants to put novelist X into genre Y, that has to be disputed based on sources I think. But I agree, we can never fully diffuse on genre alone.--Obi-Wan Kenobi (talk) 18:15, 10 May 2013 (UTC)


 * All American novelists. I think I distrust the category system as it stands (probably for only certain categories; for example, well, your examples of ghettoization). There really should be a way to provide easy oversight to category work. I can think of a few suggestions [projects develop consensus 'signifiers' or 'tags' for fraught classifications?] Otherwise, what is the equivalent of 'cites' for categorization? And what prevents the creeping approach, in the current example of, as you call it, ghettoization?  You've talked about 'intersects' made possible (in the future) by  Wikidata - that's good.  A Wikipedia user can define a retrieval.  But for now, I just feel that adding categorization by century is sort of an end run and avoids the issues. Not as a question of intent, but of effect; otherwise a by-century set is not a bad thing. So, what can I do to help out? Neonorange (talk) 23:34, 10 May 2013 (UTC)
 * There are external tools that let you watch a category, but I've never used them. Otherwise, category membership is determined article by article, so there is no central governance for membership - it's decentralized. The oversight to category work is Categories for discussion, which doesn't have that much participation but is the best place to discuss a particular category - if you want it deleted/renamed/merged/etc. There are no cites for a category, it is expected that if the characteristics as noted in the article are defining the person can be put in the cat. So if someone is called a "rugby player" by sources in the article, and another source calls them "African-American", then they can be put into .--Obi-Wan Kenobi (talk) 14:29, 11 May 2013 (UTC)


 * The jist of the way categorization works is that if someone does not like Category:19th-century American novelists, Category:20th-century American novelists or Category:21st-century American novelists they should make a nomination about these categories at categories for discussion. Until they do they are legitimate, diffusing categories and should be used.John Pack Lambert (talk) 17:46, 11 May 2013 (UTC)
 * I've mentioned that several times, but no-one has taken that up - they seem determined to hold an RFC on this weighty matter.--Obi-Wan Kenobi (talk) 18:14, 11 May 2013 (UTC)


 * Yes, they do seem determined ... possibly because they see the suggestion as an end run in the context of the present discussion. Neonorange (talk) 23:15, 12 May 2013 (UTC)


 * Comment. If Category:American novelists is a container category, I don't see any neutral, easily verifiable way to diffuse it other than "by century". The current note that it won't contain novelists that have been subcategorized is completely useless. So either the category needs to contain all American novelists, regardless of any other subcategorization, or it needs to contain none. If it's diffused, I support neutral diffusion by century. If diffusion is necessitated by the sheer volume of novelists, then if some other nationalist novelist category reaches critical mass, it should be diffused by century too. (As a side note, I personally would prefer that nationalism not be a primary way to categorize writers in the first place; to me it makes more sense to categorize them by the language or languages they write in. Their country of birth is how you categorize them as people, not as writers per se.) Categories by century should not be subcategorized by any other criteria (not "19th-century women novelists"), or we end up with the same problems of ghettoization. (As a second side note, writers should not be categorized by gender on the basis of presumed biological gender; they should be categorized by gender only if the article discusses by means of RS that they are the subject of gender criticism, or appear in anthologies circumscribed by gender, etc. Ditto ethnicity. Such categorization should not displace writers from the main category, whether that's "American novelists" or their century category.) Cynwolfe (talk) 14:01, 13 May 2013 (UTC)
 * On this note: "they should be categorized by gender only if the article discusses by means of RS that they are the subject of gender criticism" - that goes against current practice in my experience. We're in the middle of discussing this, and relevant changes to the guidelines, over at Wikipedia_talk:Categorization, so please come over there to discuss that side point further.
 * FWIW, I also agree that should either have all novelists, or none. The century cats should not be divided further by gender as that would violate the last-rung rule (see last para of WP:EGRS). As to categorizing by language vs. country of birth/nationality, that's a much bigger issue, that would cause lots of other complications if we let it spread to other domains.--Obi-Wan Kenobi (talk) 14:48, 13 May 2013 (UTC)
 * Yes, I know that categorizing writers by language would be a great uprooting; I wouldn't propose it, but can't help mentioning that it would resolve at least some nationalist issues, which are often unpleasant and unproductive. It seems to me that EGRS already should keep writers from being categorized by presumed biological gender, andWP:CAT already says It should be clear from verifiable information in the article why it was placed in each of its categories. Since we don't ordinarily categorize based on biological gender per EGRS, it follows that we should categorize by gender only when information in the article indicates that the writer is the subject of gender studies—and since this represents only one critical/scholarly approach to a writer's work, belonging to a gender category shouldn't exclude the writer from any other category. This discussion is so sprawling that I just occasionally drop by here or there to repeat my two-cents. Cynwolfe (talk) 17:21, 13 May 2013 (UTC)
 * I can see your point - again this discussion is best had Wikipedia_talk:Categorization, as I'm proposing some changes to guidance which would have impact on what you're saying there - in brief, as a reader, if I look at, I'd expect all women poets that we have to be there. Same goes for or  - if we have an article about them, I'd expect them to be there. Doing a 'maximalist' categorization, when the occupation itself is defining, makes more sense to me and is less POV. We should also recall the fact that presence in a gendered or ethnic cat should never remove you from a neutral one, so I don't see any harm in having more rather than fewer in these cats - and other editors will bring them there eventually too....--Obi-Wan Kenobi (talk) 17:39, 13 May 2013 (UTC)


 * Comment Considering the churning movement of novelist names in and out of Category:American novelists AND that categorization by century is obviously not a consensus THEN the Category:American novelists should revert to status quo ante (say, January 1, 2012) UNTIL a consensus is affirmed.  If categories that have large numbers of articles are a problem then present some quantitative evidence.  And keep in mind that computerized knowledge systems have a great capacity to present knowledge in multiple ways TO SUIT THE NEEDS OF USERS. This should be primary; the wishes of editors come second. Neonorange (talk) 16:08, 14 May 2013 (UTC)
 * The person who has done the most editing of putting people back into [{:Category:American novelists]] who are in the century sub-cats says that they totally think the by century categories are OK. There is still not a CfD about the century categories.John Pack Lambert (talk) 21:57, 14 May 2013 (UTC)
 * If you don't like the century categories put this:

On them. The proper way to deal with a category is to discuss it at a CfD. Unless someone does that, there is absolutely no reason to not populate the by century categories.John Pack Lambert (talk) 21:57, 14 May 2013 (UTC)
 * Agree with Neonorange. If we need to have an RfC, then let's just do it. I have neither the time nor the energy, but if anyone else does, that would be nice. The question is simple: do categories have to be diffused, or can biographical articles have multiple overlapping categories, (i.e. American novelist and 20th century American novelist)? Truthkeeper (talk) 00:46, 15 May 2013 (UTC)
 * Cats can overlap. By saying an article has both, what you're actually saying is, is non-diffusing (just like  or ) - or in general, all sub-cats of  are non-diffusing. That's a fine position to take -  is like that - it's just different than most other by-century categories of the same ilk, so you have to defend why this is a special situation.
 * What about my proposal above, to rename as ? --Obi-Wan Kenobi (talk) 00:53, 15 May 2013 (UTC)


 * I sort of like your suggestion to rename as along with your suggestion "The century cats should not be divided further by gender...".  It seems to me that it acknowledges the controversy and avoids 'certifying' in a discriminatory way. But wouldn't it be nice to see quantative data on how Wikipedia users USE categories?  And isn't it amazing what pops up when a section is added to a conversation?  And wouldn't it be nice to get some clarity on "What is to be done?" before discussions of deletions? Neonorange (talk) 02:03, 15 May 2013 (UTC)
 * I don't know where to get rich data, but my 101 research suggests that categories pages in this domain are barely used at all, compared to other pages. Take a look at this - this is page views for the month of March 2013:
 * # Ernest Hemingway: 381,885
 * # Novel 82,194
 * #] The Old Man and the Sea 73,928
 * # List of African-American writers : 7155
 * # List of American novelists: 4982
 * # Category:American novelists: 1736
 * # Category:American journalists: 964
 * # Category:African-American_writers: 510
 * I used March 2013, as I didn't want distortion in the numbers as a result of the media mess. I added the journalists as John Pack Lambert was recently brought to WP:AN for the crime of adding a special bio to a subset of instead of the head cat - as you can see, that page is barely seen more than 30 times a day.
 * This data reveals to me that the much-repeated claim in the media that people are looking at the category is silly - they are going to Novel 80x more frequently, and going to the actual list of American novelists 3x as often. In the case of African American writers, they are going to the list around 14x as often as the category. So if our real concern is to give users a full and complete list of novelists, then let's work on making American novelists and other lists the best they can be, and create templates to link out to these lists to stick at the bottom of various pages, etc. If we're generous, and assume that the only way people got to American novelists was thru the link on the bottom of Hemingway's page, then only 0.45% of people visiting that page actually clicked on that link! In actuality, this number would be much smaller, as they likely came from many other places. So it's probably extremely rare for someone looking at a bio to click on the categories below.
 * I would be all for usability tests but they should be coupled with maintainability testing - e.g. what is the drift in category membership and categorization accuracy if everything is non-diffusing - how easy it is to detect items which should be in the parent cat but which aren't? (the opposite case is easier - it's very easy to tell when someone has been placed in  and not a sub-cat, since it's empty.) If we do think categories are somehow useful, we should also care that they be somehow accurate as can be - and different strategies for categorization will lead to differing levels of accuracy in categorization
 * We can clearly see from many statements made here and elsewhere, that the media and users see 3700 bios in and assumed all are there - even though 3000 are actually missing. So we do our users no good service by such a structure - it was fatally flawed. The only solution is IMHO everyone in the parent, or no-one- and in any case, provide a link out to the recursive enumeration in all cases.
 * My ideal proposed structure would be (1) American novelists empty (2) Every single novelist in a novelist-by-century cat - all 6700 (3) Almost every novelist in a gender cat, if it can be determined from RS (3) Every novelist in one of the ethnic cats, if they fit. (4) Every novelist in one or more genre cats, if they fit. Many will not. This is basically the same methodology they used at .--Obi-Wan Kenobi (talk) 04:10, 15 May 2013 (UTC)


 * Thank you for posting the numbers. Definitely a good contribution. I think I can agree with all you just said.  Especially "We can clearly see ... basically the same methodology they used at ." Neonorange (talk) 20:40, 15 May 2013 (UTC)


 * And March2013 vs April2013 vs May2012 (to date)pageviews for Category:American novelists are 1736 : 23342 : 5333 ! Neonorange (talk) 21:07, 15 May 2013 (UTC)


 * If we decide that cats can overlap, then they can overlap. That's what establishing consensus does. And what needs to be done here. You're free to !vote as is anyone else. If only someone could frame this clearly and get it going, would be nice. Truthkeeper (talk) 01:09, 15 May 2013 (UTC)

RFC or not?
I can't follow the conversation any more. It seems that Carwil's initial, eminently reasonable, post was too far in front of what we're able to discuss cogently yet. Thus I am going to make a simple proposition, which is "We should have an RfC on this category." If there is consensus to have an RfC then maybe we can start working out the wording afterwards. I'm not going to give my reasoning, because it's all explained a zillion times above. I'm not sure how or whether to publicise this proposal, so I'm not going to.&mdash; alf laylah wa laylah (talk) 01:57, 15 May 2013 (UTC)
 * Note: I've informed Wikiproject Novels, Wikiproject Literature, Wikiproject biography, Wikipedia_talk:Categorization and a few editors who have been involved in categories who hadn't yet participated in these discussions: Xezbeth, Nikkimaria and Sadads --Obi-Wan Kenobi (talk) 17:02, 17 May 2013 (UTC)

Proposal
We should have an RfC on Category:American novelists.


 * Support as proposer. &mdash; alf laylah wa laylah (talk) 01:57, 15 May 2013 (UTC)
 * oppose sorry. If we're !voting... I would be up for an RFC on (1) ghettoization, and fixing the issue in the tree in general, and revising WP:EGRS which you may recall caused the media storm in the first place or (2) if it has to be about novelists, make it about the whole writers tree, across all countries - establish general guidelines for structure of the tree for consistency sake - including issues around genre, by-century classification, and which cats should be diffusing/non-diffusing. An RFC over one single cat is too small. If you just want to do novelists, seek consensus with the novels project first and see how far you can get, if that's your only goal.--Obi-Wan Kenobi (talk) 03:31, 15 May 2013 (UTC)
 * Oppose There is no reason to single out American novelsits. If people do not think we should have the by century categories, they should nominate those for deletion.John Pack Lambert (talk) 20:11, 15 May 2013 (UTC)


 * JPL, it's very hard not to regard this comment and your last one demonstrating how to file an CfD as deliberately obtuse. You are effectively saying, "Don't discuss the issue together and try to work out a solution. If you come up with one, I'm not listening until you litigate the process through a CfD." And you're saying it on a page where a dozen of us are attempting to have a collective and constructive conversation. Now, you add to that, don't invite people to an RFC.
 * Look, I understand that the bulldozer approach where one just does things and sees who fights back is often a modus operandi for frequent/fulltime wiki editors, but it is alienating as heck to those who are trying to plan things together. Just inside of this category, we've seen the creation of a subcategory (women novelists), its restriction through a CfD, the creation of multiple (genre) subcats, some of which were removed through CfD's, and the creation of century subcats, which are being discussed here. An appeal was even launched the AN/I to try to get us to talk before we leap into new changes, each of which require moving hundreds upon hundreds of articles. Why in the world can't you commit to a comprehensive conversation?--Carwil (talk) 13:12, 16 May 2013 (UTC)


 * I see where you're coming from, but I think JPL also has a good point - if the desired solution is to get rid of problematic century cats, CFD is the proper forum. If people more or less agree with the by-century cats, then it's really just a question of whether they diffuse or not - in which case this convo should be moved to the novels project and see how far we can get, and do an RFC there if necessary. But CFD is the proper forum for getting community consensus on deletion or merging or renaming of a category if so desired- it's not litigating, it's a place populated by categorization experts and consensus there is usually good enough. I already started another discussion at the novelists project per TKs recommendation on the genre cats, but few have participated to date.Obi-Wan Kenobi (talk) 14:13, 16 May 2013 (UTC)


 * I think the 'novelists category by centuries' can be seen as a way to obfuscate the real problems with categories and avoid accepting responsibility for the weaknesses of the system - including insensitivity to questions of nationality and gender (which are in NO way equivalent to 'genre'). Mechanically bringing up a CFD without dealing with the broader issues shows a reluctance to SEE the underlying problems and is a guarantee that events like the recent media public criticism will repeat (as it should - but please don't let it be for for the same exact problem in this particular category).  If editors showed more understanding of gender and nationality issues it would be a good thing.  It is not possible to force this understanding, but I sure do try to ignore most comments that fail this test.  I do subscribe to Obiwan's broader perspective on categorization. Neonorange (talk) 17:01, 16 May 2013 (UTC)
 * What nationality issues? I haven't seen anyone raise those. And have we not resolved the gender issues in this cat, at least for now? The previous CFD did that - women go in the women novelists cat, and in the parent. Whether they can be henceforth diffused to a non-gendered cat is a different discussion, having nothing to do with gender (since everyone would be treated equally at that point)--Obi-Wan Kenobi (talk) 17:48, 16 May 2013 (UTC)


 * What nationality issues? For one, see section 12  Ethnicity as genre above. Nationality is not the same as ethnicity.  And for the CFD, where's the evidence that second half of 'women go in the women novelists cat, and in the parent' is in practice? Neonorange (talk) 18:22, 16 May 2013 (UTC)
 * The current set up in these cats is that ethnicity is a non-diffusing sub cat of the nationality cat. So an African American woman novelist who wrote her first novel in 2010 would be in, and . If she wrote 3 novels and all were war novels and was called such by RS, she would also be in . Again, this mirrors the approach in  I believe. As for the second part, e.g. "women go in the parent", some (like me) interpret that as "women go into the parent, unless they can be diffused into a cat below" - the same approach applies today to  for example, which is fully diffused on other subcats, so that one is about letter vs. spirit of the CFD. --Obi-Wan Kenobi (talk) 18:31, 16 May 2013 (UTC)
 * Interjection: I've set up an area below for threaded discussion and would like to suggest to move the discussion there so as to keep this area more readable and perhaps make it more friendly to draw in more participants. Feel free, of course, to ignore the suggestion. Thanks. Truthkeeper (talk) 19:15, 16 May 2013 (UTC)


 * Support for the single focused issue of whether American novelists should be kept as a container category so as to avoid segregation by gender and to avoid the issue of remnants that cannot be diffused. Further arguments can be brought forward in threaded discussions if an RfC is held. Truthkeeper (talk) 11:25, 16 May 2013 (UTC)
 * Support — But I'm very open to having this as a RfC on the subcategorization of "X Novelists," as Obi-Wan suggests. I would offer the container category multiple choice options proposed above. And I would suggest some kind of branching rule (like, subdivide by century if more than 400 novelists).--Carwil (talk) 13:12, 16 May 2013 (UTC)


 * Oppose You can't have a local consensus RfC to make this category different from every other category on Wikipedia. Unless you're going to start a global RfC to make it so that categories aren't used at all (because you can't say just use parent categories, as that would just mean we use the one highest category), this is pointless. Silver  seren C 20:54, 16 May 2013 (UTC)
 * Oppose The issue is a lot broader than one category tree and there's no reason as to why American novelists should be an exceptional case, inconsistent with other American people categories and other national novelists categories just because it's the one pounced on. Timrollpickering (talk) 19:10, 17 May 2013 (UTC)

Threaded discussion

 * note re: Carwil's suggestion above to do an RFC on "X novelists" - my proposal was to have an RFC on writers, not just novelists. There are many complexities there as well - eg to what extent are the types-of-writer cats diffusing or not. Quite a mess right now. Plus we have genre trees that don't have national equivalents. The whole thing needs an overhaul, but a broad discussion on same in advance of some good work and thought by a smaller group may not bring much value - better to come to a smaller consensus on a general way forward then propose options to the community at large if there are sticky points.--Obi-Wan Kenobi (talk) 14:13, 16 May 2013 (UTC)
 * Think globally, act locally. Maybe if we can solve this more constrained problem it will serve as guidance for what to do in the more complicated cases.  Maybe we can make this category a shining city on a hill that cannot be hidden.&mdash; alf laylah wa laylah (talk) 14:23, 16 May 2013 (UTC)


 * OK. So, how about a precise, short statement of a the category tree as you would like to see as base camp. Ideally something a bit less baroque than http://biblioklept.files.wordpress.com/2011/03/024_lg.jpg B^) Neonorange (talk) 19:58, 16 May 2013 (UTC)
 * A few years ago, a UN joke was going around along the lines of the British proposing a practical solution to a problem, and the French responding "that works in practice, but what about in theory"? I'm afraid I'm with the French here: the problems are with the system of categorization for literature in general, and looking at only one category won't address the underlying issues—which I see mainly as a failure to distinguish between subject categorization (encyclopedic) and categorizing people (which degenerates into trivial intersections that have nothing to do with any field of study, or tabulation of census-like data). Cynwolfe (talk) 17:48, 16 May 2013 (UTC)


 * R to Neonorange above - I don't have an answer for, it needs much more thought. For , which I have thought about more, my ideal tree would be


 * --- Option 1 ---
 * No novelists in - probably just the list would be there.
 * All novelists in one (or more) by-century cats, based on the century(ies) they were actively writing in. Every single novelist should be in this tree. (e.g., , etc.)
 * Most novelists also in either or  or possibly a 3rd gender cat that could be created (assuming  survives at CFD - if it doesn't, then probably just women in the gendered cat). This is also assuming reliable sources can be found to identify gender - if not identified, they wouldn't be in any of these trees.
 * When applicable, novelists should also be placed in the ethnic sub-tree, per reliable sources (e.g. X is a famous African-American author, so he'd also go into )
 * When applicable, novelists should also be placed in one or more genres, if reliable sources can be found to call out either the thematic grouping of their works or the novelist-genre-categorization of that novelist. So Hemingway would go into, as several sources can be found that call him the quintessential american war novelist, but Melville may not have any specific genre, so he stays out of that tree.
 * This system would be neutral and fair, there would be no gender or ethnic bias (everyone in a gendered or ethnic cats would also be in the century cats at the very least), and no-one would be ghettoized by genre either - genre becomes more of a decoration/extension, rather than a diffusing/defining category for them. If you want a list of *all* novelists, you can click the link at the top to get the external tool which will recursively enumerate everyone, or just browse all novelists (by century).
 * The other advantage would be that someone who doesn't know this system (e.g. the majority of wikipedia editors) could simply put them in, and someone else would be able to later see them hanging out there and move them into the appropriate categories as necessary.
 * --- Option 2 ---
 * The second-choice alternative would be exactly the same as above, except every single novelist (all 6700) would *also* be in.
 * The problem I have with this solution is maintenance - it would be difficult without using category intersect tools in a sophisticated fashion to figure out if some people weren't in the head cat. As a result, users looking at the head cat and seeing 6700 novelists would never know that 1 novelist was missing if someone forgot to put them there. We don't have to wonder if this will happen - it already did - there were 3000 novelists not in the head cat, but media reports and the bulk of users thought it was a list of all novelists. Thus in the interests of the users, we should not pretend to give them a full list unless we really can. Another problem with all novelists in the head cat is that if someone adds a new novelist, and just sticks them there, then they aren't in any genre, gender, or by-century cats - and it becomes quite tricky to identify such new entrants and categorize them per the additional cats that have been created, as it's one new entry out of 6700 (instead of 1 out of 1).
 * Finally, we would have to somehow make sure new entries get added to List of American novelists, as web-log research shows that page is hit 3x more frequently than the head category anyway. The list has an additional advantage, in that it can provide context (dates, famous works), and can also provide redlinks, for novelists who aren't yet in wikipedia but that someone thinks should be - so when people say "give me a list of all novelists", we should point them there, rather than our category system.--Obi-Wan Kenobi (talk) 20:38, 16 May 2013 (UTC)


 * Support the above. We are only ever going to be sure we have added everyone to the appropriate century category if we fully diffuse out by century.  It is also the only easy solution to the complexed questions of how to treat writers who have some very clearly by genre works and some works that are not by genre.  Also, since the general by century cats will not have ethnic and gender sub-cats, those many editors who do not understand non-diffusing gender and ethnicity cats will not incorectly remove categories.  It is a very balanced way to set everything up.John Pack Lambert (talk) 21:03, 16 May 2013 (UTC)
 * I guess I should have been clear. I think that Category:American novelists should end up with only list articles.John Pack Lambert (talk) 00:09, 17 May 2013 (UTC)


 * Support Obi-Wan's suggestion here - the first one, not the "second-choice alternative". -GTBacchus(talk) 23:45, 16 May 2013 (UTC)


 * Comment I like Obi-wan's first choice above (beginning with "No novelists in ..." and "All novelists in one (or more) ..." . Suggest inserting 'also' to make the phrases read "Most novelists also in either ..." and "When applicable, novelists should also be placed in one or more genres, ...". Just in case. Neonorange (talk) 01:24, 17 May 2013 (UTC)


 * Support Option 1 That seems much more simpler to deal with. Focusing on the centuries cats and adding in others from there if applicable seems like a much better way to do it. Silver  seren C 03:41, 17 May 2013 (UTC)
 * Comment I don't understand. Now this seems like an RfC.  Shouldn't we frame it officially as one and advertise it more widely?&mdash; alf laylah wa laylah (talk) 04:13, 17 May 2013 (UTC)
 * Support Option 1, for what it's worth. Were it not for User:Nikkimaria's ownership issues this would already be dealt with. &mdash;Xezbeth (talk) 13:53, 17 May 2013 (UTC)
 * No, Nikkimaria is not the only one who objects to this nonsense. I could just as easily say that if it weren't for your intransigent refusal to apply WP:BRD and discuss category changes on article talk pages this would already be dealt with.  That wouldn't be true, though, because it's mass, institutionalized, complacent intransigence on the part of any number of editors, who can decide for themselves who they are.&mdash; alf laylah wa laylah (talk) 14:01, 17 May 2013 (UTC)


 * Support Option 1 offered by Obi-Wan Kenobi. Let not the perfect be enemy to the good. Neonorange (talk) 23:38, 17 May 2013 (UTC)
 * Support Option 1 with appropriate hatnotes in each of the category talk pages explaining these agreements. And let's invite others to this conversation.--Carwil (talk) 02:46, 18 May 2013 (UTC)
 * Note: I already notified several wikiprojects, shown above.--Obi-Wan Kenobi (talk) 04:43, 18 May 2013 (UTC)


 * Support Option 1 This needs to be 'generalized' to historical biography in general, to be honest, but this would be a good start....a large enough project to work out the issues, and the current 'pointy nail'. Needs to be a more 'general' RfC, eventually... Revent (talk) 04:24, 19 May 2013 (UTC)
 * The "Missing encyclopedia articles" people would be good to get involved in anything decided about this... Revent (talk) 04:27, 19 May 2013 (UTC)


 * Comment - this section wasn't meant to be a vote and some people showed up to !vote after being canvassed. It should be hatted or archived or something. Also adding link to a comment I deleted (out of frustration) that with some sort of reasoning against diffusion. Truthkeeper (talk) 18:32, 22 May 2013 (UTC)


 * Note - this discussion about the RFC is now closed - there was no consensus for an RFC on American novelists. A proposal for a way out is being discussed below. --Obi-Wan Kenobi (talk) 04:37, 23 May 2013 (UTC)

Notification of discussion on CfD notifications
Please participate: Wikipedia_talk:Categories_for_discussion. Thanks! &mdash; alf laylah wa laylah (talk) 00:13, 23 May 2013 (UTC)