The following discussion is an archived debate of the proposed deletion of the article below. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review). No further edits should be made to this page.

The result was keep. —fetch·comms 00:56, 18 September 2010 (UTC)[reply]

List of largest empires[edit]

List of largest empires (edit | talk | history | protect | delete | links | watch | logs | views) – (View log • AfD statistics)

(Find sources: Google (books · news · scholar · free images · WP refs) · FENS · JSTOR · TWL)

This article is fundamentally Original Research. Furthermore, this cannot be fixed via article clean-up, because the fundamental concept always will be original research. Qwyrxian (talk) 06:16, 10 September 2010 (UTC)[reply]

Following is the detailed explanation I posted on the article's talk page about 1 week ago:

So, normally I'd just take the article to take an article I thought violated policy to AfD, but this is a particularly tricky article. My concern is that the article is now, and will always by definition be, original research. Two things point me to this conclusion.

1. The very long intro section clarifying how difficult a ranking of this type is to make. This whole section is pure OR--it's not a discussion of how other sources have found it difficult to define the size of empires, it's actually a discussion about how we, here, on Wikipedia, find it difficult to define the size of empires (or, even, what exactly constitutes an empire). That section reads to me as exactly the sort of thing I would expect to read in a scholarly article covering this same topic, not in an our encyclopedic reporting of those scholarly articles.

2. The fact that we have an ordered list, but the specific order is based on multiple disparate sources. To me this is a clear violation of WP:SYN. We're treating all of the different measuring systems found in the various sources as working from similar premises, similar methodologies, and similar definitions. But we have every reason to suspect that that is simply not true, as that simply isn't how academic research works, especially in analytical (as opposed to experimental) research.

Thus, I believe that this article is a definite violation of the prohibition on original research. I further think that there is nothing we can do to make it not OR, unless we could find all of the information in studies that used identically methodology and measuring devices. Qwyrxian (talk) 06:19, 10 September 2010 (UTC)[reply]

Strong delete. Excellent analysis about this list being inevitably OR and SYN. I tried to fix the list in the recent past, but I soon found my best and in fact only weapon was to delete whole chunks of material altogether.

You hit the nail on the head: Any such hierachical list lives from applying an order to the referenced numbers, but given the vast array of different sources, authors and methods, any such sorting is bound to remain purely subjective and thus merely reflecting the views of the latest editor who bothered to edit it. Ironically the best-researched figure in the list is also the one which reflects most the epic failure of this list: there are 27 different estimates on the population of the Persian Empire, only to have 26 of them being ignored in favour of an alleged most "preferred" number.

Soon, inclusionists who are prepared to ignore the impossibility of such a listing for the 5th time will flock in and vote through the article on the grounds of the topic being notable alone, but mark my words: as interesting as the topic may be to our quantifying age, this article is destined to be indefinitely tagged as pseudo-scientific compilation of random numbers — because, the way it is set up, it cannot exist in another way. Regards Gun Powder Ma (talk) 16:57, 11 September 2010 (UTC)[reply]

• Note: This debate has been included in the list of Lists-related deletion discussions. -- • Gene93k (talk) 17:45, 11 September 2010 (UTC)[reply]

• Comment. I'm not sure I understand the problem. Surely the purpose of this list is not to obtain a ranking (...and the winner is...). Do the objections of the nominator also apply to, for instance, List of largest cities throughout history, List of countries by homicide rate, and United States military casualties of war? These articles are based on fewer sources, but nevertheless the information presented has been obtained from various sources that do not necessarily use the same definitions and methods. But, surely, that is also true for other non-listlike articles, such as Philosophy, or Cuisine of the United States.  --Lambiam 20:52, 11 September 2010 (UTC)[reply]

  No, because List of largest cities throughout history actually consists of three different estimates which are put side by side without preferring one over the other on unclear grounds. Not sure what you mean with that "the purpose of this list is not not to obtain a ranking", because it exactly does that. What is it then? Gun Powder Ma (talk) 22:00, 11 September 2010 (UTC)[reply]

  The purpose of the article, as far as I see, is to present a list of the largest empires in world history. Obviously, you can't do that without some measure of largeness, and as there are several reasonable measures, such as area and population, it presents both. One could also imagine having separate articles, like we have List of countries by GDP (PPP) per capita next to List of countries by GDP (nominal) per capita, but somehow I doubt that would satisfy you. It should also be possible to present both in a multi-column list, sortable by different fields.

  List of largest cities throughout history indeed consists of data from different sources put side by side in a multi-column format, but this is possible only because of the extremely limited number of sources used: just three. If the article had not confined itself to this unmotivated selection of sources, we might have seen very different estimates; for example, for Nineveh the sources consulted give 100.000 and 120.000, but the book Nebuchadnezzar by Geneviève Tabouis estimates 300.000, and a number as implausibly large as 600.000 is given by many sources, while some think 100.000 is too large. In my opinion the limited number of sources is a weakness rather than a strength of the article. I mentioned several other articles which are based on many more sources, in which the results are not presented "side by side", and so your reply does not apply to those.  --Lambiam 14:14, 12 September 2010 (UTC)[reply]

• Keep The complaints above are insufficient reason to delete as it is our clear policy to retain articles for improvement rather than delete them because of some supposed imperfection. The claim that the article cannot be worked upon without OR is incorrect. Firstly, there are plenty of sources which are cited such as Taagepera's notable analysis. Secondly, the use of simple arithmetic to rank estimates is not OR, as is well-established by precedent here. Colonel Warden (talk) 23:45, 11 September 2010 (UTC)[reply]

You are correct that we should preserve and improve. I'm arguing that it is fundamentally impossible to do so in this case without violating WP:OR. Specifically, I'm arguing that comparing numbers of this type from two different books is not actually just simple arithmetic, but a form of original research which is, in fact faulty. Maybe two examples might help. If Source A considers an empire's largest point to be the one at which it had complete and definite control over all points of its empire, while Source B considers an empire's largest point to be the one at which it had staked the most claims, even if those claims were not well held, then comparing the numbers between these two sources and then ranking them produces a flawed result. Alternatively, if Source C measures an empires size by making estimates based strictly on census data, while Source D measures an empires size based upon interpolations from theoretical population densities and farming methods, then there is no way for us to compare the numbers. This type of fundamental difference in quality and methodology in the different sources is inevitable, because the list, by definition, will contain empires which are relatively recent and for which fairly accurate data can be determined, and empires which are lost in antiquity and for which all data is derived. So while we can use simple arithmetic to compare a few numbers and then rank them, when we do so we're actually doing a whole bunch of implied research about the scientific comparability of these numbers. This is why the article List of empires, which order the empires based on an arbitrary standard (alphabetically in English) is not original research, but a hierarchical ranking must always, and improperly, be OR. Qwyrxian (talk) 01:36, 12 September 2010 (UTC)[reply]

• There is no fundamental problem because we can present the rankings and statistics of each source separately - in distinct, sortable columns, say. There seem to be enough comprehensive analyses such as those of Taagepera and Maddison to make this approach quite feasible and so your fundamental argument fails. The rest is a matter of ordinary editing per our policy which is explicitly tolerant of imperfection. Rome Wasn't Built in a Day. Colonel Warden (talk) 10:01, 12 September 2010 (UTC)[reply]

• Delete per norm! --Vinie007 11:25, 12 September 2010 (UTC)[reply]
• Keep. I find the arguments for deletion quite unconvincing, as they would seem to apply equally to many fine articles whose content is based on multiple disparate sources.  --Lambiam 14:14, 12 September 2010 (UTC)[reply]
• Keep. The article may need improved and deletion would remove a valuable resource. G.R. Allison (talk) 14:48, 12 September 2010 (UTC)[reply]

• Keep - It seems a shame to lose an article that has obviously had a lot of work put into it and that I personally found quite interesting. Would it make people happier if the word 'largest' was removed, as this seems to be the biggest issue? It could just be a list of empires with size by area and population (those figures shouldn't be too difficult to source).--Ykraps (talk) 15:54, 12 September 2010 (UTC)[reply]
  • We already have a list entitled List of empires. If the info could be formatted properly (i.e., so that it fits on the screen), then conceivably the info could be added there. Qwyrxian (talk) 21:52, 12 September 2010 (UTC)[reply]
    • I have no objection to combining the two articles providing we don't lose any information. As long as the facts and figures are there, people can draw their own conclusions as to which was the largest; if that's what's important to them. Lists of this nature are always going to attract debate. The list of largest buildings for example. What is a building? How do we measure its size: height, volume , floor space? Do we think these articles should be deleted?Ykraps (talk) 07:27, 13 September 2010 (UTC)[reply]
• Keep. Interesting and useful resource. The counter-argument that it takes insufficient account of conflicting views is an argument for article improvement and expansion, not deletion. We could and should have extra columns for contrary opinions, mainstream-vs-fringe views, mean-vs-outer limits arguments, etc. Lots of excellent list articles like this in Wikipedia. Jamesinderbyshire (talk) 20:48, 12 September 2010 (UTC)[reply]
  • Apologies, but that is not at all what the concern is. The concern is that it's OR/SYN to combine the data in such a way that a hierarchical list is created. Qwyrxian (talk) 21:52, 12 September 2010 (UTC)[reply]
    • How can it be synthesis to simply order sizes in a descending list? I didn't bother going on about that point as I see no merit in it. There's no reason why such a list can't be generated and no reason why other lists based on other criteria can't be generated, in either the same article or other similar list pages. Jamesinderbyshire (talk) 07:34, 13 September 2010 (UTC)[reply]
      • Comment: "How can it be synthesis to simply order sizes in a descending list?" Quite simple, because there is no established view on the order whatsoever, therefore any order which forces some entries above/below others is pure synthesis. Basically, the editor(s) can, will and must choose the order to their liking by selecting the figures to their liking (and ignoring others). What, for example, keeps me from downsizing the Persian Empire to 10 million people or upgrading it to 80 million? This inherent problem cannot be solved given the huge spread of the data, the complete lack of consistent criteria by which the most disparate sources are used. Colonel Warden above obviously has not understood the basic flaw, most probably because he did never work on such a list. Gun Powder Ma (talk) 08:16, 13 September 2010 (UTC)[reply]
        • Yes, lists like this always involve multiple intracacies of source evaluation, criteria, scope, comparison criteria, etc. This applies just as much to "established" lists such as population. What I personally would like to see would be a sophisticated timelined-moving chart at different points in history, like the ones in books or what used to be (?) Microsoft Encarta (don't know if that's still around) but I suggest that would stretch Wikipiedia. In the meantime, why not build on this? Delete is effectively throwing hands up in surrender and as this is an important topic worth covering, I reject that approach, as apparently do most others here. Jamesinderbyshire (talk) 09:44, 13 September 2010 (UTC)[reply]
• Keep as per Colonel Warden. Edward321 (talk) 01:46, 13 September 2010 (UTC)[reply]
• Keep There might well be a lot of work to be done, but the argument against is based on the assumption that the synthesising is of the 'bad' variety prohibited under WP:NOR. I think this has always been a misnamed policy because all articles rely upon 'good' synthesis of one sort or another under WP:V. Wiki-Ed (talk) 09:40, 13 September 2010 (UTC)[reply]
• Keep This seems similar to List of rivers by length, which also arguably involves synthesis and original resource and has a "very long intro section clarifying how difficult a ranking of this type is to make", yet is an interesting and useful page. Personally, when I see pages like these on Wikipedia I take it for granted that the rankings may or may not be absolutely correct--that it is probably impossible to make an absolutely correct ranking. Nevertheless I find it useful to have a list of "largest empires" or "longest rivers" that is "close enough" for getting a sense of things. Do these pages involve synthesis and original research? Maybe. Is it impossible for them not to involve synthesis and original research? Maybe, maybe not, I'm not sure. Is the gathering together of the world's largest empires or river useful? Definitely. Is it important that the ranked listing be absolutely correct? Not really. Should these pages be deleted? Definitely not. Pfly (talk) 10:57, 13 September 2010 (UTC)[reply]
• Comment: I am aware that consensus will probably be against this deletion, but I really want to try one more time to express my initial point, which almost no one is addressing. Let's take Gun Powder Ma's example of the Persian Empire, which independently reliable sources report with a population of between 10 million and 80 million. Looking at the current list (List of largest empires#Largest empires by population), and assuming no other changes were to be made, that means it should be somewhere between tied for 11th, and significantly below last, as the current "last" number is 48.4 million. This means that if 10 million is correct, one could argue it shouldn't even be on this list, as it's not one of the "largest." Now, the fact that we have contradictory sources is not normally a problem--in a standard article, we would include both sources, and report both numbers. But in a ranked list, how would that be possible? We literally cannot choose where to put it on the list, or even to decide if it belongs on the list, without choosing one reliable source over another. Any choice we make will be OR, because it will be us saying "Source X is correct, source Y is incorrect." This is not about it being difficult, it's about it being impossible without violating OR. I'm not throwing my hands up--I'm saying this list cannot be done according to WP's policies. To me, there are only 3 solutions possible given WP policies. 1) Delete. 2) Delete, but move all of the relevant data into a non-ranked list (such as List of empires). 3) Keep, but make a separate list for each source. #2 seems like both the most work and the most "correct"--I'd recommend moving the article to someone's name space who actually wants to take on the process of transferring the data and reformatting the table at List of empires. #1 seems like the solution if no one wants to volunteer to actually make an article that meets policy. #3 seems ludicrous to me, because we shouldn't have an article that is essentially a copy of a variety of original sources (i.e., Subheading 1 is "Source A's ranking of empires by population", #2 is "Source B's ranking of empires by population," ad nauseum), especially since such an article wouldn't even really be that "helpful." Obviously, the result of this debate will be decided by consensus, but even if the article is to be kept, I would strongly appreciate someone providing, either here or on the article's talk page, a clear explanation as to how we will make non-OR based decisions about how to maintain a hierarchical list in the face of sources that disagree. Qwyrxian (talk) 02:05, 14 September 2010 (UTC)[reply]
  • Comment:I still don't think these arguments are sufficient to warrant deleting the article. Gunpowder Ma makes the point that he/she could upgrade the Persian Empire to 80 million. What's to stop me changing the list of tallest buildings by selecting a different definition of a building. For example the OED defines a building as a structure with walls and a roof, this includes the CN Tower which at 1814 feet is now 2nd. But why would I want to? As long as the article is clear about how the ranking is obtained and transparent about any doubts over figures (which it is), what's the harm? To talk about deleting the article is setting a precedent that could endanger all ranked lists in Wikipedia.--Ykraps (talk) 16:04, 14 September 2010 (UTC)[reply]
• Keep - perhaps if it was not for all these deletion attempts some editors would have more time to improve the article. It is certainly notable and useful. BritishWatcher (talk) 21:07, 14 September 2010 (UTC)[reply]
• Keep - Encyclopaedic and useful. It won't necessarily be original research forever: there could be a source which also lists empires this way. --Joshua Issac (talk) 01:14, 15 September 2010 (UTC)[reply]
• Keep I don't see the compelling reason to delete. Moncrief (talk) 21:59, 15 September 2010 (UTC)[reply]
• Keep all articles and lists require some degree of judgement, and this is not unreasonably OR. I mention this is the 5th nomination after four successive keeps, some almost unanimous.The last one was only 12 months ago. I would regard a sixth nomination as abuse of process. This one is not, because a much more extended argument is made than previously & it's not unreasonable to ask for reconsideration under such circumstances. DGG ( talk ) 06:00, 17 September 2010 (UTC)[reply]
• Delete - OR indeed, and a playground for nationalists, too. Konstock (talk) 11:02, 17 September 2010 (UTC)[reply]

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review). No further edits should be made to this page.