Wikipedia talk:Overcategorization/Archive 3

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Category:Jews and Judaism

When we have the categories Category:Judaism and Category:Jews separately a Category:Jews and Judaism and all subcats feel overcategorised. What do you think? --Cat out 05:32, 2 February 2007 (UTC)

I agree. Jews and Judaism sounds completely redundant with the other two. I'd recommend putting J&J up for deletion at cfd. Dugwiki 17:33, 2 February 2007 (UTC)
Probably should be deleted with subcats. I tried CfDing Category:Jews and Judaism by country and Category:Jews and Judaism by city for cfd but those did not go well. A few (what appears to be) panic keep votes prevented deletion. --Cat out 18:26, 2 February 2007 (UTC)
Ah, I didn't realize this was discussed in cfd last year. In reading the discussion, it appears that the keep votes want to use Category:Jews and Judaism as an organizational category to include both subcategories Category:Jews and Category:Judaism (obviously both those categories should remain seperate since one deals with people and the other with religious philosophy and related articles). The analogy here would be that this is like Category:Christianity being the parent category for both Category:Christian people and non-person related articles like Category:Christian history and Category:Christian theology. Looking at it that way, I can understand the keep votes for the parent category. It's debateable whether or not such a parent is necessary, but it's not as appears at first glance a clear cut case of simple redundancy. Dugwiki 19:05, 2 February 2007 (UTC)
I am fine with a Category:Jewish organizations, Category:Jewish history, and etc... they are restrictive to a single topic. I do not understand the point of Category:Jews and Judaism. It sometimes covers history related articles such as the crusades and other times deli restaurants and other stuff. I haven't checked every article but whenever Category:Jews and Judaism or one of its subcat is used, it is always complineted by either Category:Jews or by Category:Judaism or by both. For instance, I really feel Category:Jews and Judaism by city is poor taste and very redundent. See what the parent categories are? --Cat out 07:53, 3 February 2007 (UTC)
I think the idea of Category:Jews and Judaism is to act as a parent category for all those subcategories. My guess is that no article, though, should appear in Category:Jews and Judaism. Rather, the articles should be placed in an appropriate subcategory instead. If you spot articles in Category:Jews and Judaism, you could probably delete that category tag from those articles and substitute a subcategory. Dugwiki 17:56, 5 February 2007 (UTC)
There are oodles of underpopulated subcategories which serve to no purpose. --Cat out 09:52, 9 February 2007 (UTC)

Clarification of "Intersection by ethnicity, religion, or sexual preference" needed

It is worth pointing out in the guidelines that categorization by religion or ethnicity may be appropriate in some situations. For example, ethnicity may play a major role in the careers of actors, and religion clearly plays a role in the careers of priests (ministers, pastors, etc.). The guideline needs to be updated just to avoid something stupid, like the nomination of Category:Roman Catholic bishops for deletion. Dr. Submillimeter 13:14, 9 February 2007 (UTC)

I actually think there are other problems with the guideline, and with how it is phrased on this page. Right now for instance it uses the term LGBT writers to mean writers of LGBT literature which is not inherently ambiguous (and arguably the minority meaning of the phrase). Also, for instance, currently the heading is just "Intersection by ethnicity, religion, or sexual preference"; that doesn't mean just "career" (or more properly occupation) intersection; or if it does, it should say so. And I'm not sure why these categories are only notable if notable in their career; wouldn't it be, notable for the subject of that intersection? And at any rate, that doesn't quite correlate with the Wikipedia:Categorization/Gender, race and sexuality, which speak more to the desirability of creating particular categories, than to the desirability of including people within them. And I think it's difficult to discern a concensus for that practice, although a majority "vote" may produce that result. --lquilter 13:28, 9 February 2007 (UTC)
Just looking at this page again, and the paragraph on this guideline is very confusing and conflates two separate issues.
Likewise, people should only be categorized by ethnicity or religion if this has significant bearing on their career. For instance, in sports, German-Americans are not treated differently from Italian-Americans or French-Americans. Similarly, in criminology, a person's actions are more important than their sexual orientation. While "LGBT literature" is a specific genre and useful categorisation, "LGBT quantum physics" is not.
This is a question of biographical categorizations, but the last sentence is very confusing, because it suddenly switches to subjects (LGBT literature is a subject; "quantum physics", whether LGBT or not, is a subject.). --lquilter 16:57, 17 February 2007 (UTC)

Category:People by medical or psychological condition

I'm moving toward the opinion that all of the subcategories of Category:People by medical or psychological condition should be turned into lists. Perhaps it is time to revisit Wikipedia:Categories, lists, and series boxes (and the essay Wikipedia:Categories vs lists) and make some clear distinctions between the topics that are suited to become categories and those that while not suited for being a category would be acceptable as a list.

For the last year or so, we have been trying to use a criteria for category acceptance that says that each category must be useful, studied and worthy of an article about the topic. The problem, as I see it, is that it in many cases it has been very difficult to draw a line between which categories are acceptable, and which are not. Wherever we draw this line there will be contentious disagreements about keeping some categories and deleting others that are similar. The deleted ones will be frequently recreated and the kept ones will be repeatedly nominated for deletion. I think it is high time that we consider these categories by class and not individually. Using the subcategories of Category:People by medical or psychological condition as an example, some subcategories exist because they meet the criteria for being useful, studied and article worthy while others do not. Also some people may be added to these categories because it is a defining characteristic (Betty Ford and alcoholism?), but it seems in most cases it is not a defining characteristic. I am moving toward the position that we should remove entire classes of categories like this and replace them with lists. We've identified many criteria for overcategorization which should be applied to entire classes of categories. Making the entire class of borderline categories into lists would hopefully lower the level of debate and contentiousness on these CfD pages. Unfortunately, it often seems near impossible to make these sort of decisions about categories. When classes of categories are nominated en masse, many editors are brought to the discussion with the desire to save their favorite category, and not discuss the categorization system as a whole. I was the same way when I first came to CfD. -- Samuel Wantman 07:52, 11 February 2007 (UTC)

  • I concur with this idea. Note that we have been enacting several sweeping changes as of late, e.g. the superheroes-by-team and actors-by-film and such. So give it a try, it may well work. >Radiant< 15:22, 12 February 2007 (UTC)
I generally agree with you that there are some contentious areas in these categories. The only problem is I'm not sure that converting to lists would do much to solve the problem. I'm pretty sure that you'll see the same debates on list article afds that you do on category cfds, such as complaints that inappropriate people are being included in the list/category. I'm also not sure that simply deleting the entire scheme is helpful to the readers, since some or most of these categories are by all accounts legitimate groupings.
Perhaps instead of simply deleting/listifying all these categories, it would be better to try and lay out some general ground rules for what subcategories are reasonable and what types of articles should be included? A good place to discuss this might be the talk page for Category:People by medical or psychological condition, since it is the parent category for everything we're discussing here. Notice that this category has no introduction or description. Maybe by including a description that tightens up the parent category's definition you can cut down on some of the more contentious debates. Dugwiki 17:28, 12 February 2007 (UTC)
At least in a list you can source each person on the list. With a category you never know for sure about that. Which is especially important when it involves living persons. Garion96 (talk) 17:32, 12 February 2007 (UTC)
Actually, categories and article have the same sourcing requirements. Basically you should only be categorizing an article based on verifiable information within the article itself. If the article doesn't mention anything to indicate it belongs to a given category, then it should not be included in that category. Therefore articles which don't include verifiable information about someone having a specific medical condition should not be included in categories involving that medical condition. Dugwiki 17:32, 20 February 2007 (UTC)
I know, but it's much easier to check for sources if it is in a list. For example, a new addition to Category:HIV-positive people is hard to check, a new addition to List of HIV-positive people is very easy to check. Garion96 (talk) 17:35, 20 February 2007 (UTC)

What I'm hoping we can discuss here are the guidelines for making these distinctions. Wikipedia:Categories, lists, and series boxes offers advice about the advantages and disadvantages of each, but is vague about when one should be used and not the other. Can this be improved? Should we consider classes or each category individually? If we consider a class, what would be the criteria for keeping or deleting an entire class? If we don't consider entire classes, how do we decide which categories belong and which do not without this becoming contentious and an energy drain. After 2 years at CfD I'm loosing my patience because policy and implementation often seem capricious. -- Samuel Wantman 23:43, 12 February 2007 (UTC)

Skyscrapers by height

Here's another possible exception to the Arbitrary inclusion criteria, have a look at Skyscrapers by height. Here we have a case where any single cat, taken by itself, is arbitrary. However, taken as a whole, it becomes a useful index category. Thoughts? -- Prove It (talk) 21:13, 11 February 2007 (UTC)

  • This would be a lot better as a (sortable) list. Sort by height, country, build date, etc. >Radiant< 15:19, 12 February 2007 (UTC)
I agree with Radiant. A category would only list the skyscraper's name. But a list sorted by height provides exactly what the reader probably really wants - to compare the heights of specific skyscrapers against each other. Dugwiki 17:30, 12 February 2007 (UTC)

African birds

What does the panel think of categorization decisions such as this? It strikes me that so many categories make the bottom of the page less useful — this bird is apparently found in most African nations. Is there a more economical way to keep track of this information? Should each species be categorized according to every nation in which it's found? I acknowledge that the categories, such as Category:Birds of Nigeria, are useful and interesting. Is there a solution, within the existing software? -GTBacchus(talk) 03:53, 13 February 2007 (UTC)

  • Eek. Bad idea, and a good example of Intersection by location ("Avoid subcategorizing items by geographical boundary if that boundary does not have any relevant bearing on the items' other characteristics"). Since obviously birds aren't intersted in human borders, we should not categorize them by human borders. >Radiant< 11:49, 13 February 2007 (UTC)
    • Actually, I firmly believe that's a case along the lines of "If an article belongs in all or almost all the subcategory of a geographical location, use the parent category," which happens to be related to what Radiant points out. I suggested the same guideline in some flora articles for North America.Circeus 12:43, 13 February 2007 (UTC)
    • I realise I wasn't actually very clear here: It's overcategorization not in the sense that the category itself is useless, but in the sense that it puts a uselessly large number of categories in the article. The scheme is the problem, not the categories themselves.Circeus 12:51, 13 February 2007 (UTC)

Fauna and flora, categories verus lists

Have a look at the following categories and lists found using Special:Prefixindex:

Eek, indeed! What is to be done? With such a large set of categories, surely there was discussion about setting this up, or did it just grow organically? Is it useful? Is there a WikiProject that has expertise to add to this discussion? Carcharoth 15:10, 13 February 2007 (UTC)

  • There seldom is any discussion about setting up a category tree. >Radiant< 15:20, 13 February 2007 (UTC)
    • Based on the non-uniformity of the fauna by US state and fauna by European country categories, I would guess that the categories were developed by multiple people with interests in specific locations. For example, someone spent a lot of time categorizing the fauna of Lithuania. Unfortunately, that categorization is not useful, as the animals are frequently found in many other countries across Eurasia.Dr. Submillimeter 16:48, 13 February 2007 (UTC)

Some interesting examples are Category:Flora of Australia (note the extensive categorisation guidelines) versus Flora of Australia, and Category:Western Palearctic restricted-range endemic bird species and Endemic birds of the Western Palearctic and Template:Endemism in birds. How useful do the people watching this talk page think these pages are? Carcharoth 13:41, 21 February 2007 (UTC)

Category versus Main Article for navigation

A number of cfds recently regarding eponymous categories and cast lists have gotten me thinking about the general topic of how readers presumably use categories versus using links in the main article to find information. For example, let's say I want to find information about a topic. My first likely action is to type that topic in the search field. If an article comes up, then I'll read the article and also click on links within the article to find related information. For example, if I pull up an article about an actor, I'll also get a list of links to all the movies that actor was in.

By contrast, a category would be useful if the category possesses links to related material that the main article doesn't have. For instance, let's say I pull up an article about a science fiction writer. I might be interested in reading about other science fiction writers too, but links to those people won't be in the article I pulled up. Instead, I could go to a category that collects those sci-fi writer links in one place.

Which brings me to the reason I'm mentioning all this here. It seems to me that the primary purpose of categories is to act as a navigational tool for readers. But if a category's associated main article already has (or should have) all the same article links, then why would a reader bother to use the category? They would just go to the main article and find the link they want. And creating a redundant category comes at a cost to the editors in the form of kind of doubling the amount of editorial work needed to implement changes (you have to change the main article AND modify the categories on related articles).

What would you folks think of a section that talks about eponymous categories that recommends eponymous categories only be used when they have a demonstrated practical additional navigational benefit to the reader over the main article? In other words, if a reader can use a main article just like he'd use a category with the same name, then you shouldn't bother creating that category.

This might give some guidance when dealing with eponymous categories about people, and also when dealing with things like "cast list categories" which provide links that are duplicated almost verbatim in their associated main article.

Thoughts? I'm curious to hear feedback, good or bad. If there seems to be general agreement, then I can maybe put up a draft version of a section for eponymous categories. Dugwiki 17:44, 15 February 2007 (UTC)

My rationale in approaching this question it to look for a substantial number of directly relevant articles/sub-categories. 'Substantial' is a subjective judgement, 5-6 articles upwards is my typical level. As to directly relevant, the articles should directly address the person whom the category is named after, be it their life or their work—not simply a collection of trivial connections. Take Category:Jan Smuts as an example—each article here directly deals with the man and his life, as a result I think that it is proper to group them under eponymous categorisation.
Categorisation in this manner, when merited, gives the reader an overview of coverage, something not quite so easily accomplished directly within the article(s) themselves. I think that this kind of direct overview, when used properly, can be considered an 'additional navigational benefit', can't it?
Xdamrtalk 18:03, 15 February 2007 (UTC)
Actually this looks like a good example of the kind of category that can be probably be safely deleted. Category:Jan Smuts consists of exactly six links, including one subcategory link and the link to the main article. All of these links appear in the article Jan Smuts.
So, let's say I'm a typical reader and I'm interested in looking up information related to Jan Smuts. My first thought will almost certainly be to type "Jan Smuts" in the search box. This calls up the main article Jan Smuts. In looking over the article, I have links to the Holism article (which, by the way, should probably not be a subcategory of this, but the other way around) and all four of the "Jan Smuts..." subarticles. By first visiting the main article, I have eliminated the need to visit Category:Jan Smuts , and in fact the article contains more links and more information than the category! Thus not only is the category redundant with the main article, the main article actually appears to be a better way to find articles related to the man.
Thus unless I'm overlooking something Category:Jan Smuts is a case I might recommend category deletion since it seems to serve no practical purpose above and beyond what the main article accomplishes. It also seems to have incorrectly made "Holism" a subcategory of "Jan Smuts", when it should probably be the other way around. The subarticles also should be categorized under other appropriate categories besides simply Cat:Jan Smuts. Dugwiki 20:36, 15 February 2007 (UTC)
And lest you think I simply want to delete EVERYTHING, I'll mention a more positive example. Category:Ronald Reagan looks like a better candidate for a category, because it contains a fairly large number of articles directly related to Ronald Reagan, not all of which appear in his main article, as well as four various subcategories. The thing here is that there are SO many articles directly related to Reagan that it would make it more difficult to navigate it all from his main article or subarticles. Thus I'd be more inclined to keep Category:Ronald Reagan around to help readers go through articles related to him, something the main article might have trouble doing on its own. Dugwiki 20:46, 15 February 2007 (UTC)
Although the life of Jan Smuts is my pet subject on Wikipedia, I didn't create the category. I'm still working on his various articles, so citing his category as an example was probably more an instance of my fond expectations for the future than a wise choice for the present. Having said that, Category:Ronald Reagan seems to be a good use of a category, the sort of use that I was advocating—a good number of directly relevant articles.
The fundamental point is that for such significant figures, with substantial coverage in Wikipedia, an eponymous category acts as an overview, bringing the relevant articles together in one place. So long as the articles are sufficient in number, and of directly connected to the person in question, an eponymous category is justified. Is that something we agree on?
Xdamrtalk 21:12, 15 February 2007 (UTC)
Yeah, I think we're actually in agreement for the most part. An eponymous category should probably meet the following two basic criteria -
1) It should be restricted to articles directly related to the subject. Articles which might be indirectly related, but that don't actually notably mention the subject, should probably not be included in their category. Jesse White (actor), for example, starred in a movie with Ronald Reagan, but he probably shouldn't appear in Category:Ronald Reagan since his article isn't directly related.
2)If the number of article links and subcategory links in the eponymous category comfortably fits within the main article, such that a reader can navigate them simply by reading the article, then the main article is already serving the primary navigational function of the category. In such a case the category can probably be safely deleted. Jan Smuts above would probably be a good example of a category whose links are directly related to the subject, but whose category size is small enough that you really don't need the category to navigate them; you can simply use the main article. Another example would be a category specifically for the cast list of a movie. A movie's article will (or should) always contain the cast list, so there's no need to look at anything other than the main article to visit links for the cast. Thus there's probably no need to create a category specifically to handle the cast of a movie.
Thus a good candidate for an eponymous category would be one with a relatively large number of articles directly related to the main article, and which doesn't simply appear as an easy-to-read list in the main article. Dugwiki 21:41, 15 February 2007 (UTC)
FYI, since I posted my above comments, Xdamr went and dug around and added a number of new articles to Category:Jan Smuts. Given the new content, it looks like this particular category is looking more reasonable. It now includes three times as many links, including an image file and Wiki templates used for his articles. So while the old collection in the category was quite light, this new one looks much more reasonable. Thanks for the update, Xdamr. :) Dugwiki 22:08, 15 February 2007 (UTC)

For the root organisation of this sort of thing, see Category:Categories named after people. I was involved in the renaming discussions for that category, back when it only had 50 or so subcategories. It has a lot more now! Back then, I observed that these eponymous categories act like portals (or as extended 'see also' lists), and that still seems to be the case (compare Category:Culture, Portal:Culture and Culture). My feeling is that, within reason, eponymous categories, links from the main article, "see also" sections, and portals, all serve similar functions. Some readers prefer to browse one way, some prefer another way. As long as things don't get out of control, this sort of thing is not too bad. Carcharoth 13:49, 21 February 2007 (UTC)

More examples. Category:Nature versus Nature, but no Portal:Nature. And (returning to the people example), Category:Tolkien versus J. R. R. Tolkien versus Portal:Middle-earth. Or Portal:Disasters, Category:Disasters and Disaster. Or Portal:Shakespeare, Category:William Shakespeare, and William Shakespeare. Carcharoth 13:55, 21 February 2007 (UTC)

People categories

There seems to be categories for people by their citizenship, by where they live, by the country they were born in, and by the country of their ancestors. If this was not bad enough, the country they live in might also have a state or province category, and a city category. Sometimes the categories are not clearly defined and two or more groupings are combined. On top of that, these categories are often intersected with profession -- especially for actors and musicians. This is adding a good amount of clutter to articles, and has the potential to get much worse. Are all of these necessary? Could we get rid of some or all of the intersections? (The German Wikipedia has done this) Compare the categories for Michael J. Fox in the English and Deutsch versions.

Our version has:

Canadian film actors | Canadian television actors | Canadian voice actors | Canadian child actors | American film actors | American television actors | American voice actors | American child actors | American vegetarians | Back to the Future cast members | American memoirists | Inductees of Canada's Walk of Fame | Canadian immigrants to the United States | Canadian-born entertainers in the United States | Disney voice actors | High school dropouts | Hollywood Walk of Fame | Parkinson's disease sufferers | People from Burnaby | People from Edmonton | People from North Bay, Ontario | British Columbia actors | Alberta actors | English Canadians | 1961 births | Living people

The German version has:

Mann | Kanadier | Schauspieler | Geboren 1961 (People, Canadians, Actors, Born 1961)

Any suggestions? -- Samuel Wantman 01:15, 21 February 2007 (UTC)

I generally support this idea. Some of the "people from" categories are being used to indicate as many places where people have lived as possible. The category list for George H. W. Bush, for example, would be nightmarish if it indicated the 40+ places where Bush lived.
Sorting by nationality also bothers me for a few other reasons. In my profession (astronomy), people are often born in one country and then work in several other places. (I would fall under both Category:American astronomers and Category:British astronomers according to the logic used to categorize people.) This is true for some other professions, especially categories for athletes and performers, but generally not others, such as politician categories.
I would recommend having categories solely for citizenship (citizen of X) and then categories solely for profession. Dr. Submillimeter 09:54, 21 February 2007 (UTC)
To demonstrate my point, I found the following "people from" categories in George H. W. Bush: Category:People from Greenwich, Connecticut, Category:People from Maine, Category:People from Norfolk County, Massachusetts, and Category:People from Midland, Texas. Bush's article also contains Category:Texas Republicans, another location-specific category. Those are five categories (so far) just to indicate where Bush is from. Dr. Submillimeter 12:26, 21 February 2007 (UTC)

I've been trying to sort out how to categorise MJF and other Canadians for a while now. With respect to the overuse of "People from" categories, consider the following from Wikipedia:Categorization of people:

"People are usually categorized by their nationality and occupation. . . . People are sometimes categorized by notable residence, in the form People from Foo (not "Natives of Foo"), regardless of ethnicity, heritage, or nationality. . . . Nationality is reflected by the occupation category (above), not country or county or city of residence. The place of birth is rarely notable."

Although it would take a lot of work to undo the existing intersecting categories, I agree with Dr. Submillimeter's suggestion that we use categories solely for citizenship (citizen of X) and then categories solely for profession (TV actor, Film actor, Astronomer, etc., without appending nationality or any subset thereof). I think this would need to be spelled out as a guideline. --Vbd | (talk) 18:37, 21 February 2007 (UTC)

While I agree that some articles have a lot of these sorts of location-related categories, there are rational reasons for them. For one thing, when a profession related category gets too large it becomes difficult to find people you might be interested in when you aren't looking for an exact name. Thus subdividing the category helps readers more efficiently find related articles. There are two main natural ways to subdivide a professional category: by geographical area, and by specialty. Both can be useful in their own way. Geographical subdivisions are handy, for example, because many readers are interested in finding out about important people within their country or state, and aren't necessarily as interested in reading articles about people outside that area.

Also, local professionals influence their local areas, and the local culture can influence how local professionals conduct their business. Musicians are a good example, as local culture significantly influences the musical style of its local musicians, and local musicians have some of their strongest fan base in their hometown. Thus it would make sense to subdivide musicians by nationality or state or region, depending on how much subdivision is needed, because musicians from the same geographical region are more likely to share certain musical and professional characteristics and are more likely to have overlapping readership and fan interest. Another example is lawyers, where it generally makes sense to subdivide American lawyers, for example, by state of practice because much of American law is state specific.

So while I don't like "clutter" at the end of articles, I think in this case the benefits of subdividing certain professions by geographical area of occupation probably outweigh the problems associated with the clutter in some articles.

That being said, I do think it would be a good idea to lay out some guidelines to make sure geographical subcategories are handled consistently. Personally I think they should generally be by citizenship, either current or former, but that's just my opinion. In cases like law where the profession has very specific geographical boundaries, it's probably better to go by geographical area of practice. Either way, it would be nice to have an article to point editors toward to help the way geographical categories function be a little more consistent. Dugwiki 20:28, 21 February 2007 (UTC)

How about we set up a limit of two for a person's origins? That's how people usually perceive a famous person. So for Bush it'd be Texas, where he's been since age 2, and U.S. MJF would be from Canada, but Americans don't know what province he is from, and he didn't become famous there, so an additional U.S. cat for his citizenship would be sufficient. Xiner (talk, email) 13:53, 23 February 2007 (UTC)
FYI, an article normally shouldn't be in both a category and one of its subcategories when the subcategories completely subdivide the parent. So if someone appears in a state-specific subcategory, they probably shouldn't also appear in the US parent category (eg Texas politicians probably shouldn't also be listed under US politicians). Dugwiki 18:54, 23 February 2007 (UTC)
I agree, but there are instance where they should be. For example, Bill Clinton was an influential politician in his Little Rock days, but he was also important in world politics as a United States president. Thus I think it would be fine to put him in two cats. WP:CAT allows for cases where the parent cat is significant. Xiner (talk, email) 18:58, 23 February 2007 (UTC)
The problem with George H. W. Bush (not his son, George W. Bush) is that he can also be identified with Maine (where he still spends part of the year in the Bush family mansion) and Connecticut (where he grew up) as well as Texas (where he worked in business and which he represented in Congress as an adult). He has also lived long enough in Washington, D. C., (as CIA Director, Vice-President, and President) that he can be identified with that location as well. The above scheme would not work for him. Dr. Submillimeter 19:44, 23 February 2007 (UTC)
Well, it still can, if we agree on the fact that Herbert is not associated with any one U.S. state, and thus belongs in the U.S. cat only. Xiner (talk, email) 19:55, 23 February 2007 (UTC)
I think Dugwiki has confused the issue a little by using an invalid example of categories and subcategories. As far as I can tell, there are no subcategories for politicians by state. And George H.W. Bush is not even categorised in Category:American politicians. But he is in several "People from..." categories, as Dr. Submillimeter noted above.
There are subcategories of politicians by state divided under Category:American politicians by state (eg Category:New York politicians). You're correct, though, that because there is a difference between politics at the state level and politics at the national level that this is an exception to the rule where articles appear in both categories. So I did pick a bad example. Dugwiki 00:20, 24 February 2007 (UTC)
Dugwiki, I apologise. I somehow did not see the subcat Category:American politicians by state under Category:American politicians.
A few random observations: Hilary Rodham Clinton is categorised as an Arkansas politician even though she never held public office there (I've removed her from the category). Bill Clinton is not in Category:Arkansas politicians but is in its subcat Governors of Arkansas. The same is true for George W. Bush, who is not listed as in Category:Texas politicians but is in its subcat Category: Governors of Texas. These seem like appropriate uses of subcategories without redundancies. But they still don't answer the original question.--Vbd | (talk) 01:32, 24 February 2007 (UTC)
What's important to know about where someone is from? I would start with someone's citizenship or nationality. I would then consider where they spent their formative years or influential periods of time in their lives (as in the musician example, above), but I would limit the number of locales. As I noted earlier, the WP:OCAT guideline already states that "[t]he place of birth is rarely notable." Does it matter where they live now? Maybe for certain professions, such as the lawyer example above, but otherwise general not.--Vbd | (talk) 22:04, 23 February 2007 (UTC)
Hmm, I got confused there too. On the original question, I'd be bold and start removing some of the obvious cats from articles I come across (eg Edmonton actors since MJF isn't thought of as such). As for a longer term solution, we need either a new guideline or better policing through some group nominations that will set a precedent. We already have guidelines for categorization though. Xiner (talk, email) 23:05, 23 February 2007 (UTC)
To reply to VBD above, I wouldn't be so sure that locality doesn't make a difference to most professions. Local culture is significantly impacted by local professionals of all types, and people looking for information to compare professionals within their geographical area would also be interested in having things broken down somewhat by location. In particular I think it's reasonable to argue that pretty much any large category of artistic or cultural or competitive sports/hobbies professionals or professionals licensed at the state level would be possible candidates for division by state or national subcategories. That's a pretty wide swathe of careers. Dugwiki 16:22, 26 February 2007 (UTC)
Can you name examples where state lines separate athletes or actors? Xiner (talk, email) 17:20, 28 February 2007 (UTC)