Wikipedia talk:No original research/Archive 57

This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 50

←

Archive 55

Deaths using primary sources

We have a number of anomalies at Wikipedia:Database reports/Living people on EN wiki who are dead on other wikis where we have someone as still living and other language versions of Wikipedia have sourced their death from US social security data. I'm sure most local papers would happily publish a death notice for a 1950s sports star who dies locally. But they may not do so online. So sometimes we can source obits from local papers etc, but not always. Is this one of those rare occasions where it would be OK to use a primary source? Obviously as these are for existing articles the GNG doesn't apply to this and we would just be adding facts. Feedback at Wikipedia talk:Database reports/Living people on EN wiki who are dead on other wikis would be appreciated. Ϣere SpielChequers 11:49, 21 October 2011 (UTC)

Let me cut and paste here the example I gave on WSC's talk page. "I edited Danny Wagner and declared him dead based on the U.S. Social Security Death Index. That's an awful reference since it just claims that some guy named Daniel E Wagner died on that day but doesn't tie the info to the basketball player. However, the birth dates match (though the birth date info is not exactly based on high-quality sources) so let's say I'm 90% sure this is right. In the end, because I can't find a better source of information, I'm adding info (he's dead) which I know full well is potentially incorrect, yet is less likely to be wrong than the current info (he's alive). It seems like a net positive but it feels like such a sin..." Pichpich (talk) 12:25, 21 October 2011 (UTC)

Based on some primary sources, sure. Based on this particular primary source—well, it's kind of dubious. If I were desperate (and had carefully checked, a process that includes figuring out his last known residence, finding out the name of its local newspaper, and using all reasonable efforts to determine whether they published an obituary), I might use it with suitably weasel-worded phrases and a clear explanation of the problem on the talk page so that it's easy to fix the occasional mistake. WhatamIdoing (talk) 19:12, 28 October 2011 (UTC)

RfC on "verifiability, not truth"

There is an RfC here on whether to remove from the lead of Wikipedia:Verifiability that "the threshold for inclusion in Wikipedia is verifiability, not truth." The RfC is likely to close in a few days, so if you want to comment please do so soonish. Many thanks, SlimVirgin ^{TALK|CONTRIBS} 23:51, 28 October 2011 (UTC)

I previously announced this on October 6th (now in archive 56). The proposal is a lot more complex than just removing the line in question... but you can read the proposal and the rational for it at the RfC... in any case, I echo SV's call for you to come and comment. We are coming up on 30 days, so it will close soon. Blueboar (talk) 02:34, 29 October 2011 (UTC)

Slim's summary is not accurate. Of the portion which Slim claims that the proposal is removing from the lead, the actual proposal RETAINS all but two words of it in the lead, and moves the two words ("not truth") into the following section.

It's even more complex than that... the proposal does change the wording of the sentence SV is concerned about, but it seeks to retain the concept behind that sentence, and tries to explain that concept more clearly by expanding it in an entirely new section of the policy. Please just go to the RfC, read the proposal and the rational that accompanies it, and make your own decision. Blueboar (talk) 14:50, 29 October 2011 (UTC)

Note: the link above in the first sentence should read that the RfC can be found at Wikipedia_talk:Verifiability#RFC - Compromise proposal re first sentence Unscintillating (talk) 19:29, 30 October 2011 (UTC)

Have to ask

Why? Why no original research? Obviously we can't have people making assertions based on personal knowledge which they merely claim to have and could have made up, but if you can confirm the validity of your research then why not use it? 193.146.58.181 (talk) 23:06, 3 November 2011 (UTC)

Who is the 'you' in "if you can confirm"? And why should we include things people just though up at breakfast and went outside and checked for themselves? We need proof that facts are of some interest and that they have been checked. The verifiability requirement is a low bar if anything for that. People's own original research fails verifiability, this policy actually allows a small extra over what verifiability allows so the bar is really quite low. Dmcq (talk) 15:34, 7 November 2011 (UTC)

No original research

This policy can't have good consequences: it just stops any research, and more the useful than the useless. If something appears, why to wait for ages for it to be publicated in primary sources, secondary sources, verifiable sources, infinite sources? You can actually wait for ages, and information in Wikipedia will become very old, never containing recent research. — Preceding unsigned comment added by 46.146.21.196 (talk) 20:33, 8 November 2011 (UTC)

Please use "new section" to add a new topic to a talk page. Research is good, but it has to be conducted elsewhere because editors have no method to assess what research is useful and what is crackpot nonsense. Instead, editors rely on reliable sources. Johnuniq (talk) 01:28, 9 November 2011 (UTC)

Indeed research is good, publication of research is essential. In fact, there are 1,000s of scientific journals for that. These journals have a quality control system designed to evaluate the quality of original research (imperfect as that system may be). Wikipedia has a quallity control system which is not designed (at all) to evaluate the quality of original research and for that reason does not publish such research. Arnoutf (talk) 18:38, 10 November 2011 (UTC)

Possible enhancement of WP:CALC

Hello,

I'm writing with regard to this discussion (a RfC is also open, please express your opinion). In short, the issue is that we have a table with numbers from several sources and a line containing median of these sources as a summary. An editor argues, that the median line should be removed since it violates WP:OR via WP:SYNTH and that WP:CALC does not apply. Strictly speaking, he's correct: several sources are combined via an operation that isn't very simple. However, I feel that this strict interpretation falls outside the spirit of Wikipedia, since no position is advanced and no new information is provided - the median line is only for illustrative purposes, so that a reader doesn't need to read entire table to build his own summary in his head. I think this should be allowed somehow. Since I don't see ways to solve this via loose interpretation of the current policies, I propose an addition to WP:CALC similar to the following: Calculations, summarising several sources for illustrative purposes, are allowed if no position is advanced using that information, no new information is provided and the change is supported by consensus. Disclaimer: I am involved in the abovementioned discussion, I'm writing here because I think it's a more general issue that could be solved wikipedia-wide. 1exec1 (talk) 14:24, 7 November 2011 (UTC)

Computing the median of a set of numbers is a very simple operation. WP:CALC is meant to prevent people from doing complex calculus and physics computations, not to stop them from taking the median of a list of numbers. — Carl (CBM · talk) 14:32, 7 November 2011 (UTC)

I agree with that point. However, I don't think that if we strictly interpret the policy, it allows this, especially since the current wording is "<...> adding numbers, converting units, or calculating a person's age <...>". Also, in this case WP:SYNTH is somewhat violated. Because of that, I feel that we probably are extending the guidelines too much. Hence in my opinion we need to build consensus and maybe to clarify the guideline. 1exec1 (talk) 14:40, 7 November 2011 (UTC)

Why are we putting out stuff like this if our sources don't? I don't support extending WP:CALC to cover this. Dmcq (talk) 15:11, 7 November 2011 (UTC)

Because summarizing information helps our readers understand the material: "Most US Presidents were just under six feet tall" or "US Presidents have ranged in height from 163 cm to 193 cm" is far easier to understand than "US Presidents have had the heights of 193 cm, 193 cm, 189 cm, 188 cm, 188 cm, 188 cm, 187 cm, 185 cm, 185 cm, 185 cm, 183 cm, 183 cm, 183 cm, 183 cm, 183 cm, 183 cm, 183 cm, 183 cm, 182 cm, 182 cm, 182 cm, 182 cm, 180 cm, 180 cm, 179 cm, 178 cm, 178 cm, 178 cm, 178 cm, 177 cm, 175 cm, 175 cm, 174 cm, 173 cm, 173 cm, 173 cm, 173 cm, 171 cm, 170 cm, 170 cm, 168 cm, 168 cm, and 163 cm."

Summarizing mathematical information is no different from summarizing any other kind of information. The job of an encyclopedia editor is to summarize information, not to regurgitate exactly what his or her much longer, much more verbose sources say.

1exec1, I think there is a widespread consensus to permit simple descriptions like this; in fact, I've never heard anyone except Dmcq claim that simple, obvious descriptions of mathematical information are inappropriate. You will find a brief description of what's normally permissible at WP:NOTOR#Simple_calculations. WhatamIdoing (talk) 15:37, 7 November 2011 (UTC)

Good points. Now consider that you did not have a complete set of "presidential heights" but say, only a sample of 10 presidents. Suddenly it becomes important how those 10 presidents were selected by the source. If they happen to be the 10 most recent presidents you can certainly calculate the median of their heights, but you must be very careful not infer (or convey) anything about the entire distribution of presidents because of the influence of modern living standards. Now consider that you had samples not just from "U.S Presidents" but just "presidents" - some U.S. presidents, 2 russian and a president of the National Restaurant Association. You do not have the luxury of a complete set anymore. Now your median (and any other statistical function) will depend on the selection criteria - if any. My point is that your example is oversimplified and trivial. This debate is about the WP:OR policy and the set if trivially simple cases is far outnumbered by the set where editors will err when applying statistical function inappropriately.Useerup (talk) 16:02, 7 November 2011 (UTC)

WP:NOTOR#Simple_calculations is just an essay and even its talk page isn't exactly in consensus. Dmcq (talk) 10:52, 11 November 2011 (UTC)

Computing the median is a moderately simple operation. However, deciding when the median is applicable is not a simple and straightforward decision when compared to calculating a persons age from his date of birth. The latter is uncontroversial and unlikely to be debated. The former assumes a great deal about the numbers over which the median is calculated: Are the numbers proper statistically unbiased observations? Is the median or the mean the proper way summarize the observations? Have the numbers been collected in a way which is consistent with the median being applied? What is the objective for applying the median (remove statistical outliers? communicate a distributuon?). The median is a statistical function like the mean/average function, and like the mean it is frequently misused and its applicability is far from as obvious (or "routine") as e.g. unit conversions. Case in point: The Usage_share_of_operating_systems calculates the median for each operating system and the others column. When adding those numbers - including the others column, one would expect to arrive at 100%. But when used that way the sum of the medians may even exceed 100%. What does that mean for comparability of those numbers? In my opinion, using statistical functions on numbers derived from sources implicitly interprets those numbers is a novel way, even more so when performed across multiple sources. Disclaimer: I am the editor who raised this issue in talk:Usage_share_of_operating_systems (and several others) at least in this latest round. Useerup (talk) 15:39, 7 November 2011 (UTC)

So, according to you, if we used mean instead of median, there would be less problems? Am I correctly understanding your point? 1exec1 (talk) 16:34, 7 November 2011 (UTC)

I must have expressed myself poorly. No, I don't think any statistical function can be considered "routine". As soon as you enter statistics you have to be careful with a number of things such as error sources, potential biases, and general theoretical applicability. There no risk of bias, error and inapplicability when considering the examples of routine calculations of the current policy (calculating age, converting units). There is a big risk of not just WP:OR but also downright misrepresentation when editors who are not experts in statistics start making assumptions on data and apply statistical functions. Some editors (evident in the archived discussion of the article in question) believe that median may be a good way to neutralize recognized bias of the sources. Useerup (talk) 16:58, 7 November 2011 (UTC)

I don't think this is statistics. This is an application of some simple function to several numbers to get one number in order to increase readability. 1exec1 (talk) 17:39, 7 November 2011 (UTC)

If it is not the statistics median, which one of Median (disambiguation) are we discussing? To me, median is squarely a statistics function and it cannot be applied indiscriminately. Especially so if it is applied across multiple sources not sampled by a reliable source but rather selected by Wikipedia editors.Useerup (talk) 18:35, 7 November 2011 (UTC)

Saying that the range of heights among US presidents is 163 to 193 cm is also "statistics". That's why our article on the subject is titled "Range (statistics)". Identifying the range is no less "statistics" than identifying the median or the mean—and all three of these concepts are taught to children around age 10, so I really do believe that they fall within the definition of "simple". WhatamIdoing (talk) 15:17, 8 November 2011 (UTC)

It is a logic fallacy to generalize from an arbitrary simplified example. This debate is about changing a policy (a generalization). You push an oversimplified example of US presidential heights which is not representative for the real scenarios where the policy will be applied. Consider a situation where you did not have the full set of heights, but only "some presidents" referenced by various sources. Are you now comfortable computing height median, averages heights, minimum and maximum? You shouldn't be, because once you are using sampling you should make absolutely sure that the samples are collected in a way which does not introduce bias or errors. What about the situations (such as operating system usage share) which are inherently sample based? Maybe you learned about mean and median at 10, but I don't know any at 10 who understands the wider field of sampling, bias, errors, etc. In fact, Usage share of operating systems to see how wrong even grown-ups can get it. Useerup (talk) 15:57, 8 November 2011 (UTC)

I've come round to the idea that this can be considered as an illustration rather than as verifiable content. However it should be placed just beside the data in the article that it summarizes and I'd prefer to see min an max markers in too as the median gives a false idea of accuracy. Some figures vary by a factor of more than three. Also placing with the data will make certain all the caveats about the measurements are beside it as well. Illustrations should be beside what they illustrate. Sticking it at the top gives undue prominence. Dmcq (talk) 18:21, 7 November 2011 (UTC)

Why the median, min and max then? Why not use the mean and standard deviation? Mean is also a way to summarize (illustrate?) and condense multiple values, and standard deviation conveys the spread. Thinking about it, why not the lower and upper quartiles? Which would be the proper way to illustrate the data? Can you recognize that a debate about this is likely? A debate on how to apply statistics and interpret results. The point I'm trying to make is that we should only ever summarize numbers when it is verifiable safe to do so. Using any type of statistics function conveys the position that the sources have been selected in an unbiased and statistically sound way and that the statistics function is meaningful in the context. If this is clear (verifiable) from the source then, yes, it could be considered merely an illustration. But when the sampling in the source is not verifiable unbiased or (worse) when the sampling based on multiple sources selected by WP editors, IMO that crosses the line. We should not allow this as a matter of policy. That policy would need to be written by an expert in statistics in order to prevent (accidental or deliberate) misuse. My feeling is that the policy would quickly spiral into the intricacies of statistics - a rather large topic for a policy. Useerup (talk) 19:08, 7 November 2011 (UTC)

Yes having the median in the main text is definitely wrong. There's just too many problems with mixing up a bunch of figures from different sources and different standards. I think probably best would be to remove the median in the table and only show min and max points in the graphic. Dmcq (talk) 19:22, 7 November 2011 (UTC)

Useerup, if you aren't using a complete data set, then you can still provide a summary of the information that you do have. You just need to clearly identify what you're doing, exactly like you would clearly identify that your summary of a position held by some reliable source as being the position held by that specific source, if there is any reasonable doubt about whether that source represents the whole field.

So you don't say "The One True™ Median"; you say "Median^[11]" followed by what would be the 11th note under that table, which would say something like "^[11]As calculated from the data in this table". WhatamIdoing (talk) 23:50, 9 November 2011 (UTC)

I would like to rephrase the point in order to make the problem more evident.

In occurrence of statistical information from different well-established reliable sources there always is a problem of summarising the data. This problem exists with such data regardless of attempts of Wikipedia editors to address it in statistics-related articles and regardless of taken approach. According to the current regulation such data should be either left as is (as per WP:OR) or excluded (as per WP:NOTSTATSBOOK). The idea of proposal is to develop a common instrument that would help the editors to summarise such heterogeneous statistical data in order to provide reader with summary, as the reader may happen to be less aware of statistical problems then the editors of Wikipedia, who would use such instrument.

The proposed way of addressing the problem would involve establishing a common practice of giving a summary to the statistical data in form of synthetic numerical values and providing user with proper information about the synthetic nature and possible accurateness issues in relation with such summary. The method of summarising statistical information should also be explicitly stated in a relevant guide.

Dmitrij D. Czarkoff (talk) 02:56, 11 November 2011 (UTC)

And doing so would still be original research. Why don't you publish your own meta analysis of this in a reliable source elsewhere. Fifelfoo (talk) 03:11, 11 November 2011 (UTC)

You didn't get my point. I don't care the statistical issues per se (apart from one I came from here, but it's insignificant for this discussion), I care the ability of Wikipedia to fulfil its primary goal — to be a reliable source of encyclopaedic information.

My claim is that this kind of OR (if at all) is needed to be excused (just like WP:CALC) from the WP:OR policy as it is important for preserving the quality of Wikipedia as the source of information (as opposed to list of references).

Surely, such type of OR (if any) should be used with a proper caution.

Dmitrij D. Czarkoff (talk) 03:25, 11 November 2011 (UTC)

No, surely it shouldn't. Wikipedia is neither a statistician nor a bureau of statistics; and, surprisingly enough, the mission of an encyclopaedia is not to publish original work as a result of professional grade statistician work. Go publish yourself, be damned or praised; and we'll quote you. Fifelfoo (talk) 03:31, 11 November 2011 (UTC)

Collecting summary thoughts

This issue has not arrived at a consensus and I have put in a bit about supports and opposes to generating a median and sticking that into the article at this. Please have a read of the issue and make a decision. If support then I believe a change to WP:CALC will be necessary to fall into line with what is done in the article. Personally I'm against that. Dmcq (talk) 00:04, 10 November 2011 (UTC)

And first click here to actually see what everyone's talking about. The question is whether a table of 88 separate data points about ten different operating systems from eight different sources is improved by adding a single line at the bottom that identifies the median estimate from each of the eight sources, or if the readers should be left just the raw data. WhatamIdoing (talk) 02:17, 10 November 2011 (UTC)

I understand enough about statistics to know that anyone who thinks that statistical calculations are simple doesn't understand enough about statistics. I'm with Dmcq on this one as it's very easy to draw incorrect and / or misleading conclusions from processing statistical data properly. And to make sure that a proper job is done smacks to me of OR. Now there will be some grey areas, obviously, between when it's simple (like the presidential heights) and not simple (as when the set is incomplete and sampling has been done). In these cases WP:CALC can be used to cover them. So: No change to WP:CALC is my vote. --Matt Westwood 22:58, 10 November 2011 (UTC)

I also agree with Dmcq and Matt Westwood. And in particular, the median line in the table at usage share of operating systems needs to be removed. It is not obvious that taking the median is statistically valid, because the sources are not sampling from the same population. As the footnotes to the table say, some sources are sampling primarily from Europe, one mostly from Germany, others leave out mobile operating systems, etc. For this to be a valid statistical operation, you would need additional data to show that browser usage in Europe is the same as in Germany and the same as in the US, and you would either need all of the data sources to sample both desktop and mobile operating systems, or you'd need to correct for that.

You know, I think I'll just take that line out. I don't see any way to rescue it without OR. Ozob (talk) 23:19, 10 November 2011 (UTC)

Oh, and statements like "Clicky Web Analytics does not publish desktop/mobile split so mean of Net Market Share and StatCounter figures (6.36% mobile) used in lieu" sound really bad to me. It causes you to overweight the data published by Net Market Share and StatCounter. Ozob (talk) 23:24, 10 November 2011 (UTC)

Yes I'm dead set against having the median in that table. However I'm just about okay with them using the median to set up a graphic as an illustration provided it is made clear it is a Wikipedia illustration. Those are allowed to just look like what they're supposed to represent without having an authenticity stamp on them. Examples also can be made up by Wikipedia editors. I think there is a general problem shown here in that they want a let out for making a summary and I have some sympathy for that. If we had a good way of delineating a section saying this is our summary it might help a number of articles. We'd need to have a guideline about it and bit in the OR policy referencing it. It would need a bit of thought setting it up to stop people sticking silly things in Wikipedia and making certain people didn't mistake summaries for verifiable data. In such a section we could have a general let out like an illustration that editors have to agree it is a reasonable representation of text elsewhere in the article as a summary. I would in fact disallow any cites with a summary, just references to other sections of the same article. Without a summary policy this business about the median could only be allowed under WP:IAR I think - and I'd still want much better markers saying it is just our own made up summary of the data. Dmcq (talk) 23:33, 10 November 2011 (UTC)

The articles are actually very good examples on why WP:CALC is sensible and may even need to be strengthened rather than loosened. The statistics are being used wrongly, an abstract number (the median) is calculated which is claimed to be nothing more than "a median" - but it nevertheless is used to compare to other medians - ignoring the fact that the medians no longer represent a share of anything (sum of the medians exceeds 100% for browser shares). A naïve argument often pushed during the debates is that median is a simple computation even a 10 year old can perform; a position which totally ignores all of the caveats of sampling statistics. To avoid this, I propose that the WP:CALC policy/guideline stresses that computation over multiple sources is always WP:SYN and not acceptable unless a reliable source directly supporting the concrete calculation exists. By "directly" I would like to explicitly rule out some source on statistics which under some interpretation (by WP editors) can support it. Directly as in those specific sources. Useerup (talk) 23:53, 10 November 2011 (UTC)

Oh hell!!! There is Usage share of operating systems, Usage share of web browsers process going on, and in the meanwhile:

Some believe they can fork discussion without noticing other parties and
~~others believe they can act though there is an ongoing dispute resolution process?~~ It seems that the actual edit was done two minutes before the dispute resolution was started. Though the talk page has a fairly clear indication of it.

What's wrong with You, people?

Dmitrij D. Czarkoff (talk) 01:06, 11 November 2011 (UTC)

A participant from the debate realized that there indeed may be a problem with WP:OR and he/she started a process to attempt to modify or add to WP:CALC to allow for statistics functions like median. Please understand that this process can very well run in parallel. Any content or policy dispute must be handled with respect to current policies. Opening an RFC for changing the policy has no bearing on the original dispute, which must be settled within current guidelines and policies. Anyway, it looks like the amendment will have a hard time getting passed. Useerup (talk) 01:16, 11 November 2011 (UTC)

This discussion clearly originates from that one and has a direct impact. Wikipedia is supposed to be a collaborative effort, so why do You and Dmcq try so hard to kill collaboration? Now You two have pushed the edit to a page out of due process, which is really a nasty game. And this all is happening while You both know that there is an ongoing process of dispute resolution! — Dmitrij D. Czarkoff (talk) 01:50, 11 November 2011 (UTC)

If you are referring to this page, you may want to investigate who started this discussion and why. Thanks. Useerup (talk) 07:02, 11 November 2011 (UTC)

I did. The particular provoking thread (that resulted in inappropriate edition) was started by Dmcq. And it was Your obligation to inform the parties of discussion You've started that there is an ongoing parallel discussion You participate. — Dmitrij D. Czarkoff (talk) 10:20, 11 November 2011 (UTC)

In a nutshell: No. Fifelfoo (talk) 01:59, 11 November 2011 (UTC) They can try, but this is really obviously dodgy statistics and Original Research, and I doubt that you're going to be able to muster consent, when as an outsider I take one look at calculating a median from disparate data series, when the underlying data and methodologies aren't exposed, but only the totals. Firstly: to conduct such a rectification of data would be original statistics research. Secondly: to conduct such a rectification of data without the exposed datasets and methodologies is original statistics research that is fundamentally wrong. Fifelfoo (talk) 01:34, 11 November 2011 (UTC)

Ask yourself what is wikipedia's goal? To provide reliable information. Does presenting a central tendency of disparate data helpful to the readers? The alternative is to present a bunch of numbers that few want to look at or to fight with editors which one represents the "best" numbers. The central tendency from the table also known as a "summary" is used in quite a few other pages. It is a compromise between options that are not ideal. The compromise helps people understand the "usage share of web browsers". Daniel.Cardenas (talk) 05:08, 11 November 2011 (UTC)

If the reliable sources suck; you have to write a sucky article. The alternative, original research, is unencyclopaedic. Fifelfoo (talk) 05:22, 11 November 2011 (UTC)

The reliable sources do not suck. With out a summary the alternative is a bunch of numbers. Showing a central tendency is helpful, informative, and very encyclopedic. Daniel.Cardenas (talk) 05:29, 11 November 2011 (UTC)

Inventing a central tendency from sources which are incommensurable is a reaction to the fact that there is no reliable commensuration of those sources. The sources suck for what you want to write, and so you're suggesting original research (and fundamentally flawed original research at that). Fifelfoo (talk) 05:37, 11 November 2011 (UTC)

A simple calculation from commensurable sources. These are very professional sources that people pay money for to see the details of the data. Suggesting a summary of the information for those that want a summary and details are available for those that want that too. Daniel.Cardenas (talk) 05:50, 11 November 2011 (UTC)

I agree with Fifelfoo's comments. The value this statistic will bring to readers is exaggerated, and relaxing the policy is unnecessary and unwise. Given the excessive oversell or pushback on this, one might ask why it is that this "bunch of numbers" (which are so supposedly so baffling that they won't make sense to readers who aren't handheld via the median) are being included in the first place. Professor marginalia (talk) 06:16, 11 November 2011 (UTC)

No one has suggested the numbers are baffling and won't make sense. What people have said is that there are a lot of numbers. A summary is common in wikipedia. For example see wp:lead. Since this article is about numbers it makes sense that a summary is mathematically based. Daniel.Cardenas (talk) 06:22, 11 November 2011 (UTC)

A median isn't a "summary". It's a statistic. It is easy to derive a median, however it isn't in each and every case an appropriate statistic. Professor marginalia (talk) 06:33, 11 November 2011 (UTC)

From Median#Medians_in_descriptive_statistics:

"The median is used primarily for skewed distributions, which it summarizes differently than the arithmetic mean."

Daniel.Cardenas (talk) 06:40, 11 November 2011 (UTC)

Proving my point. It's a very bad idea to relax the policy here so as to change the "routine calculation" clause and allow editors to go out data gathering on their own, collect numbers for a data set and perform any kind of statistical analysis on it for inclusion here. "Routine calculations" are already allowed. Discussion is the proper venue for hashing out whether or not a given calculation qualifies as "routine"-existing policy is sufficient. Relaxing policy so as to end run potentially frustrating discussions over whether not a calculation is "routine" is a bad idea. It just opens the door to more original research and editors disputing over more complex calculations. Professor marginalia (talk) 07:00, 11 November 2011 (UTC)

It doesn't have to be done in a pandora's box fashion. Daniel.Cardenas (talk) 07:06, 11 November 2011 (UTC)

The issue before us here is changing WP:CALC because -let's be honest- there is currently a single dispute waging in two articles over whether a wp calculated median should be included and the WP:CALC policy is being cited to exclude it. There have already been arguments over how disparate the data may be and whether the median is valid or meaningful. Pandora's box popped open pretty quick - even under the existing policy. And relaxing the policy won't make those sorts of disputes disappear. It will only make them more complicated. Professor marginalia (talk) 07:27, 11 November 2011 (UTC)

Seems your point is that you are worried about pandora's box opening up. My perspective is how can we provide descriptive summary of a bunch of numbers. Applying what I perceive your arguments are to wikipedia, we would never have it, because it would create a big complicated mess and editors would do whatever they want when they want. Daniel.Cardenas (talk) 07:51, 11 November 2011 (UTC)

No, I'm saying that the existing policy, limiting calcs to simple, routine calculations provided there's consensus is more suitable for dealing with a disputed median, and that relaxing the policy to allow users to add one where "several sources are combined via an operation that isn't very simple" is a very bad idea. Professor marginalia (talk) 14:11, 11 November 2011 (UTC)

(ec) Oppose We summarise text by picking out the most significant facts, already presented in the original. We might collate information, but we don't combine information. If it were useful to combine, someone would have already done it. Uniplex (talk) 06:38, 11 November 2011 (UTC)

I think you can agree that procedure won't work with a table of data. Otherwise tell us how you would summarize the summary table. Daniel.Cardenas (talk) 07:06, 11 November 2011 (UTC)

I think you misunderstand: our principle is not simply to summarise; however, summarising text is in line with our principles (because it is done by selection). Summarising figures by selection would also be fine; combining disparate figures is synthesis. Uniplex (talk) 07:49, 11 November 2011 (UTC)

How do you define disparate? All the numbers are measuring usage share of web browsers. Daniel.Cardenas (talk) 07:53, 11 November 2011 (UTC)

They are collected and compiled using different methodologies, sample sizes, sampling techniques, significance corrections and the like. As such they're not commensurable; but, we can't commensurate them anyway as we are not able to synthesise works like this. You have had this repeatedly explained to you by editors lying outside of your content dispute. You have entered into IDHT territory. Fifelfoo (talk) 07:56, 11 November 2011 (UTC)

Seems a professor of statistics agrees with me. Daniel.Cardenas (talk) 03:59, 16 November 2011 (UTC)

Yes, indeed. As long as it is not presented as the median of browser usage share (the statistical population), but just the median of the selected sources. Useerup (talk) 04:27, 16 November 2011 (UTC)

We don't actually allow summarizing in text even, more what we allow is a précis because our summaries have to have the individual bits verifiable. If all the members of a family had red hair and we know that because we have citations noting each individual member had we can't summarize saying they all had red hair. This is what I was saying about we should be investigating having some sort of policy or guideline about straightforward summaries which would deal with this problem too. Dmcq (talk) 11:16, 11 November 2011 (UTC)

I'm not sure it's any different from "normal" editing: when we summarise, as always, we maintain verifiability and a neutral point of view, and we "do not combine material from multiple sources to reach or [even] imply a conclusion not explicitly stated by any of the sources". You're right, précis might be a better term to use. Uniplex (talk) 12:24, 11 November 2011 (UTC)

As the graphic of Dmcq shows, all sources in this case quite agree on/reflect the same general tendencies.

As long as they do so, the median can hardly "reach or imply a conclusion not explicitly stated by any of the sources", however inappropriate the method may be in general. As has been already pointed out in the discussion to the article, the numbers become increasingly meaningless and misleading when the single values diverge much from each other. And it's _then_, that selection of sources etc. matter and a "simple calculation" becomes a complicated one. But with largely varying base numbers (and in result a meaningless mean), probably nobody would have tried to provide such a mean value.

I would think current policies already cover this well enough. Even age calculations can become non-trivial, if every year has a different length, the exact date of birth is not known, the person lied about his/her age most of their life and others did likewise...

All WP:CALC wants to say: it's not synthesis if you calculate something like the age 20 years after a guy was born, it's downright trivial (and it says that rather well.) All you have to do is decide if it's trivial in this case. No policy changes needed. Iridos (talk) 23:15, 28 November 2011 (UTC)

Discussion and poll with implication for WP:CALC at talk:Usage share of web browsers

On the talk page of Usage_share_of_web_browsers an editor is conducting a poll to see whether there is consensus for or against calculating the median of usage shares of wen browser across multiple sources collected by Wikipedia editors.

This topic has been discussed before on this page (above), but you may want to chime in with your opinion and for/against. Some of the editors on this page are actively involved. Some have been involved before but may be unaware of the poll as the editor who initiated the poll did not see fit to notify previous participants. I consider this notification relevant to the discussion as I assume that editors on this page have formed opinions on the matter and may contribute.--Useerup (talk) 01:21, 18 December 2011 (UTC)

Original images...diagrams and graphs?

I just noticed that the page on original research on German Wikipedia ignores images altogether. So I was delighted to find Wikipedia:No_original_research#Original_images here. But this also mainly seems to cover photographic pictures, while I had diagrams and schemata in mind that indeed illustrate/show/depict theories or models. That can be chemical formulae or a schematic showing the contents of a cell, really anything that depicts non-trivial information in a schematic way.

I think these images need a citation in the image description just like text describing the same information would. The citation need not necessarily link to another scheme looking just like this (it could, though), but also to a sufficiently detailed description that enables readers to verify the information contained. Do we agree on that? Iridos (talk) 11:16, 15 October 2011 (UTC)

The last paragraph at Wikipedia:Images#Pertinence_and_encyclopedic_nature describes best practice for this issue. The short answer is that it's not absolutely "required" (some diagrams are incredibly simple, and providing a source is like citing a science textbook to say something like "the sky is blue") and the source is not always (or even normally) listed in the article, but we do very much appreciate being able to identify the source for complex diagrams. WhatamIdoing (talk) 16:16, 18 October 2011 (UTC)

We're allowed to not cite illustrations if they illustrate the text and don't add something new. For instance in dolmen the illustrations show things that are obviously dolmens according to the description in the text so we don't need a source saying the pictures are of dolmens. In Internal combustion engine the animations show things which correspond with the descriptions of two or four stroke engines. In fact few images need citation because we practically always have text which describes things and if there isn't text it isn't won't noting. If a picture or diagram is used to back up something the it definitely needs a citation. Dmcq (talk) 22:52, 18 October 2011 (UTC)

An interesting case that came up recently was something like an African womens' hockey team. It showed them carrying babies i a village and didn't look like an african womens' hockey team. So it needed a citation. Dmcq (talk) 22:58, 18 October 2011 (UTC)

Sorry for the late reply, been away a bit (and took this back out of archive, I think what I did is obvious from the edit comments, hope that was the correct way)

I don't see that Pertinence_and_encyclopedic_nature helps clarify the matter.

Obviously, common sense has to be applied - and for photographic images (if they are not manipulated) all proof that one can possibly require is, that the object shown is indeed the object it is claimed to be - so things like pictures of dolmen were specifically what I had not in mind. While most photographic images need no source, I feel that most schematic representations do.

As for the Internal combustion engine - in those relatively complex cases, the picture usually has a "higher precision" than the text. It shows more, because the text only describes some details. So how do you verify the correctness of File:Arbeitsweise_Zweitakt.gif, if not by checking a source? What about the size of the exhaust pipe compared to the motor, is that realistic? Does the text cover that? Also, the air/fuel mixture (green) is not only injected into the upper combustion chamber, but also below the piston. Is that correct? Is it covered somewhere in the lengthy article? How would I verify it? (Mind, I'm not challenging anything, this is just an example you picked)

Or, how about the structure of Clevudine or LSD - sure, that can theoretically be deduced from the substance name, in practice it's not so easy for more complex molecules.

Even worse: what if the picture is _not_ in an article, but you want it included. So you find a very convincing graph on commons, depicting a complex process, but it is unsourced. As I just demonstrated, the article text usually does not cover every single detail of any moderately complex graph (but it should, according to you). So a) how do I verify its accuracy and how would I include a "complete" description in an article?

I think in principle what I think is covered by the sentence: "Original images created by a Wikipedian are not considered original research, so long as they do not illustrate or introduce unpublished ideas or arguments, the core reason behind the NOR policy."

Schematic drawings usually do "illustrate or introduce ideas or arguments" and only citing a source for them can show that they are not unpublished ideas or arguments. Iridos (talk) 15:04, 17 November 2011 (UTC)

Diagrams by editors are quite often wrong or misleading and they can be removed just saying why. If there's something wrong they aren't a good illustration! The original research policy is about citing images that back up the text or imply something that the text doesn't say. Basically illustrations are welcome. If they imply too much there's a problem but a bit extra clarity doesn't hurt anyone. Just because something is in commons doesn't mean they're suitable for use everywhere, if they are not cited they are only suitable for use where they fairly faithfully illustrate the topic or section they are associated with. Dmcq (talk) 15:45, 17 November 2011 (UTC)

Hm, where do you see the difference to text, which usually should be have its own citation (if not trivial). "If the text is wrong or misleading, it can be removed/changed just saying why"; "If there's something wrong, it's not a good text!".

I couldn't say if the combustion engine image above is inaccurate, nor off-hand if a structual formula of Clevudine or LSD is correct. We both could probably verify it against a source.

I just found Wikipedia:Attribution#Original_images, which states in the last paragraph Images that constitute original research in any other way are not allowed, [...]" (It also says that pictures have enjoyed an exception, but I would think that is true mainly for photographs, which I agree do not need not much verification)

WP:V does not seem to mention images in any way, it does not only cover text, though, but states that "material challenged or likely to be challenged must be attributed to a reliable published source". Iridos (talk) 20:09, 17 November 2011 (UTC)

Yes, images have been challenged before, and have failed verification and been removed as a result. Uniplex (talk) 21:49, 17 November 2011 (UTC)

So is there anything about how they have to obey the general rules for citations if they contain information that is likely to be challenged (or just plain in need of verification to ensure correctness)? 79.234.91.137 (talk) 14:14, 18 November 2011 (UTC)

Wikipedia:Images#Pertinence_and_encyclopedic_nature and WP:OI describe everything needed. This can be extended to examples in text as well. They do not need citations if they simply apply what has been describe in the cited text and don't say anything new. Dmcq (talk) 14:40, 18 November 2011 (UTC)

OTOH, if the contents of the diagram cannot be deduced from the text, by someone not knowledgeable in the subject, then the diagram needs to be referenced to a reliable source. In the case of Clevudine, it would seem that citation is needed. Uniplex (talk) 14:21, 19 November 2011 (UTC)

So you can't figure it out from 1-[(2S,3R,4S,5S)-3-fluoro-4-hydroxy-5-(hydroxymethyl)oxolan-2-yl]-5-methylpyrimidine- 2,4-dione? ;-) I don't remember seeing anything like the requirement you talk about and even if there were the requirement would be for verifiability not references. I'd like a reference to be put in but I don't think it is eligible for removal just because it doesn't have a reference. Dmcq (talk) 14:49, 19 November 2011 (UTC)

You're right; though this may not quite be "is the sky blue?", it didn't take much to source a similar-looking diagram. Case by case basis really; if you suspect an image may be wrong and are unable to verify its accuracy, then tagging with cn and/or talk-page discussion are options. Uniplex (talk) 16:45, 19 November 2011 (UTC)

If you really do think it is wrong then just remove it and say why you think so. You don't have to prove anything any more than the person who put it in had to prove anything but it certainly helps to have a reference if there is a dispute! Dmcq (talk) 19:32, 19 November 2011 (UTC)

Actually WP:OR and WP:RS say that the person putting it in has to prove it, not the person taking it out. And I think that should be leading in these cases as well, the more so since creating an image is a lot of work, and the creator may not be willing to let his/her work go easily. Arnoutf (talk) 17:10, 20 November 2011 (UTC)

WP:OI is part of the OR policy. WP:RS is a guideline, WP:V is the nearest policy. Wikipedia:Images#Pertinence_and_encyclopedic_nature which is part of WP:IMAGE is also a guideline and is specifically aimed at images. Dmcq (talk) 17:33, 20 November 2011 (UTC)

The thrust of Wikipedia:Images#Pertinence_and_encyclopedic_nature is entirely in a different direction, it's not surprising that it doesn't talk about verifiability at all.

Another type of graph that needs a source are those displaying data, e.g. File:Eisenach_inhabitants.JPG or File:World_pulp_bleaching_1990-2005.svg. The graph itself (i.e. the exact representation) can be "original", but the data must be sourced, else the graph is close to worthless (specially so, as for economic data different sources will not yield the exact same numbers).

I kind of see it as a logical consequence of WP:V and/or WP:NOR that cretain types of images should have a source. As those policies are about "content of articles", this would only be required when an image is included in an article (which can be a long time after the upload) and only fore images "likely to be challenged". What I do miss, is WP:V addressing the issue at all and WP:OR addressing it very clearly. Iridos (talk) 20:56, 28 November 2011 (UTC)

I'm nit certain exactly what your point is. Those types of graphics should only normally be used where the text beside them cites the data, it is not so useful to have a cite on the graphic when the data is not used in the accompanying text. If the data is not referenced in the text then there's usually no justification for the graphic as it does not illustrate what the text is talking about. Dmcq (talk) 22:33, 28 November 2011 (UTC)

The text would not contain the same wealth of information as the graphics, so the logical place for the citation would be the graphic. Also use of the graph in sister-projects (governed by the same principle of WP:V) would be prohibited, if they cannot find the data for the graph. Maybe I will try to rephrase later; I had thought I made my point rather clear in my last comment. Iridos (talk) 23:23, 28 November 2011 (UTC)

I do notice that File:World_pulp_bleaching_1990-2005.svg is being used wrong on the English Wikipedia. I haven't looked at the others. It is used in a section of Bleaching of wood pulp, however that section of the article doesn't talk about anything related to that at all. If it is to be there then it should be beside something talking about what it is illustrating and the figures should eb referenced there with the illustration saying also what it is referring to. The citation though can be left to the text if the illustration is obviously referring to the same stuff. Dmcq (talk) 23:49, 28 November 2011 (UTC)

The article does talk about it, just not in this paragraph, but after the lower-level headings following below. Let's not get hung up on the topic of the examples, though. This one was made for de.wikipedia and is sourced. If it was not, I think inclusion in en.wikipedia would have been problematic. 79.234.91.181 (talk) 17:53, 29 November 2011 (UTC)

MOS:IMAGE has been rearranged recently. You need to read Wikipedia:Images#Image_description_pages, paying particular attention to the paragraph that reads, "Reliable sources, if any, may be listed on the image's description page. Generally, Wikipedia assumes in good faith that image creators are correctly identifying the contents of photographs they have taken. If such sources are available, it is helpful to provide them. This is particularly important for technical drawings, as someone may want to verify that the image is accurate." WhatamIdoing (talk) 00:32, 29 November 2011 (UTC)

Ok, thank you. The bit about technical drawings is what I have in mind. I think everyone who did think about it for a bit more would say that certain images have to carry a source (other than "selfmade") on the description page. This is of special importance if one would want to re-use the graph/scheme/image in question on one of the different-langugage sister projects.

To put it in one sentence: All pictures that portray/reflect a theory or model of any kind or (make the reader?) "reach or imply a conclusion" need to be sourced.

I see that as a consequence of current policy at least when the image is included in an article, but it is not spelled out very clearly.

Oh, when you write "Those types of graphics should only normally be used where the text beside them cites the data, it is not so useful to have a cite on the graphic when the data is not used in the accompanying text." — do you mean that all the data should be in the article as an actual table? That would seem a bit excessive to me. The graphic depicts the trend well enough, the exact numbers are not of such great interest. Iridos (talk) 13:01, 30 November 2011 (UTC)

Of course people should list their sources when putting graphics into commons. It very much decreases their usefulness otherwise. However just having the description page of a graphic contain that information is not enough for a graphic inserted in an article, it still needs to look like something the article is talking about or have a note under it indicating why it is there and where it came from. So basically for a bunch of figures they have to come from somewhere and a citation given somewhere in the article to that. Some figures may be actually in the text but they still need a citation saying where they came from. You can do some things with graphics like for instance group some small figures as 'others' which you might find trouble justifying in text but make for a better illustration and of course one can round or otherwise style things like using small people in a barchart of a poll and none of that requires any citation. Dmcq (talk) 15:48, 30 November 2011 (UTC)

It very much decreases their usefulness otherwise. is just my point exactly. It does so up to the point where it is useless.

And of course I agree that the article should also talk about whatever pictures it contains... actually, that's still the wrong way around, the correct way to say this is: the pictures should illustrate the things the text talks about. And yet, the citation for the graphic need not necessarily be the one mentioned in (or even useful for) the text. You can cite a source for the text that "use of chlorine was reduced by 90% from 1990 to 2005, because of environmental concerns and stricter laws[citation]", while the source that has the data used for the graph does say nothing about the reasons behind the trend (or even what the trend is), it has _just_ the raw data. Iridos (talk) 01:28, 2 December 2011 (UTC)

More relation to OI

Recently there have been a discussion over an original image of a diagram such as the one seen in here. I honestly believe the image being used wrong. the sources being primarily provided through 1st party and 2nd party sources and barely covering ground by using one particular source. Not only that but the mere idea of it relies to some level of subjectivity. Original Images such as charts and diagrams can exist a lot more and potentially unnecesarily due to how WP:OI stands. I highly suggest we make it more precise or maybe more requirements to allow Original images such as charts and diagrams. Like for example, published ideas that have been noted significantly. If I am completely wrong on this please enlighten me, because it seems a lil too easy for anyone to make an OI despite potential lack of necessity and removing unecesarry subjectively useful free content tends to be a hassle as the main reason not being removed is the fact that they're free content and not Non free content. Again, if I'm seeing this the wrong way, please enlighten me.Lucia Black (talk) 12:31, 17 December 2011 (UTC)

It does look pretty definitely wrong. It is saying things that are not supported by cited text and it isn't cited itself. It is just plain original research. By the way I think we might have a copyright problem with this as well but I'm not sure about that side of things. Dmcq (talk)

The diagram has been defended by the sources used above it. But the issue I still see is that diagrams like these can be easily made if people took WP:OI a lot more. I still think we should make clear distinction or add more requirements for OIs.Lucia Black (talk) 21:10, 17 December 2011 (UTC)

Well I've asked there for an instance from the diagram, where is there a WP:Reliable source that points out that the back bow may be the main secondary or tertiary colour? Hopefully that will show how good the basis for the diagram is. Dmcq (talk) 22:34, 17 December 2011 (UTC)

I still think WP:OI should be updated to something less vague.Lucia Black (talk) 23:53, 19 December 2011 (UTC)

As a side point: something violates NOR if no reliable sources in the entire world has ever published the material before, not if the current version of the article does not happen to contain a citation to a reliable source. A zero-citation article can fully comply with this policy. WhatamIdoing (talk) 02:04, 21 December 2011 (UTC)

Be that as it may, there should not be a zero-citation article. Perhaps I am ultimately referring more to WP:Verifiability than to this particular policy discussed on this Talk Page. A zero-citation article such that you mentioned can not satisfy WP:Verifiability, can it? The Mysterious El Willstro (talk) 04:25, 21 December 2011 (UTC)

I'm saying the level of the OI should be used more significantly. Such as multiple sources that verify the "published ideas". It takes one source to source one original graph and diagrams of fictional things. The necesity of an original image should still be taken count of. If the published ideas are subjective then maybe find more that agreeing.Lucia Black (talk) 04:32, 21 December 2011 (UTC)

Original image Template - and Wikimedia Commons guidelines?

Original images illustrating unsoucred statistics is a very common problem. To start with, {{citation needed}} templates should be placed in the image description in each Wikipedia article where such an illustration is appearing, . But this is mainly a Wikimedia commons problem. The person who originally uploaded the illustration to commons should provide the factual source - rather than the person who embedded the illustration in an article. We need to develop some kind of template to be used at wikimedia commons for requesting the factual source.A policy or guideline should be developed at Wikimedia commons. And a routine should be developed for where and how to place footnotes at Wikimedia commons, how to deal with sources in different language versions of an illustration, for importing citation templates from various language versions of Wikipedia to commons, etc.Are there any good examples of footnotes at Commons today? Mange01 (talk) 14:12, 20 December 2011 (UTC)

I certainly think people uploading to Commons should be prompted for and supported in providing citations backing up details in any pictures they upload and tags available to mark that a graphic would be much better if such sourcing was available or for marking as citation wanted where the person wasn't sure it was wrong or generally would like to see a source. It wouldn't reduce the need for citation in Wikipedia but overall it would help to reduce problems and make graphics easier to reuse. Dmcq (talk) 16:37, 20 December 2011 (UTC)

Wikipedia is not in a position to dictate to Commons, just like they aren't in a position to dictate to us. But sources are good in my books, and we're already encouraging our editors to include them when they upload images. WhatamIdoing (talk) 02:06, 21 December 2011 (UTC)

I copied this discussion to commons village pump.

In a BLP

In a BLP about a controversial videographer, Mr. X, there are several (9) sourced statements in the article body to the effect that "Journalist So-and-so said Mr. X's work is deceptive" or "Journalist So-and-so said Mr. X's work is misleading". In the article lead, this is summarized as "Mr. X's work has become widely seen as deceptive". My question is whether this kind of summarizing is a violation of NOR because it's SYNTH, or if it violates any policy for any reason. My own feeling is that this kind of generalization is not permitted, because I don't see how you can claim a view is "widely held" by naming 9 (or 100) people who hold such an opinion. I'd appreciate other comments. Thanks! --Kenatipo ^speak! 07:42, 16 December 2011 (UTC)

That kind of summary is disallowed by avoid weasel words. The opinion should be attributed. Pick two or three of the most prominent journalists and quote them and attribute the quotes to the people that said them. -- Schapel (talk) 16:17, 17 December 2011 (UTC)

Since this is a BLP, we do need to be overly cautious. If we are going to state that a videographer's work is "widely" seen as being deceptive, we need a source that explicitly discusses the number of other people who view the videographer's work as deceptive. For the lead, I would suggest a more neutral summarization... along the lines of: "Mr. X's work has been criticized for being deceptive or misleading." This is accurate (apparently), but avoids the weaselly (and potentially POV pushing) quantifier "widely". Blueboar (talk) 16:57, 17 December 2011 (UTC)

(edit conflict w Blueboar). Thanks, Schapel. I told them that something along the lines of "Some journalists find O'Keefe's work deceptive" would be acceptable in a BLP lead, but their mindset is "saying 'O'Keefe's work is deceptive' is like saying 'water is wet'" so they think their sourcing in the body is adequate. --Kenatipo ^speak! 17:06, 17 December 2011 (UTC)

Thanks, Blueboar. It's the James O'Keefe article in case you didn't know. My review of talk page history indicates the article has serious OWNership problems based on the number of editors that have raised NPOV concerns but have been driven away. The OWNers aren't budging on the "widely seen as deceptive" issue either. --Kenatipo ^speak! 17:10, 17 December 2011 (UTC)

Ah... then the best I can suggest would be an RFC to get a lot more people involved. However, I suspect that it will be a "ten-foot pole" situation for the average editor (ie, they would not touch it with one)... which will leave the arguing to POV pushers on both sides. And it will probably take about five years of dispute resolution to resolve into a neutral article. Blueboar (talk) 17:39, 17 December 2011 (UTC)

I'm shocked to hear you speak so cynically, Blueboar; I hope this place isn't getting you down. "Causa sui" also suggested an RFC. I'll think about it. Thanks. --Kenatipo ^speak! 21:10, 17 December 2011 (UTC)

These kinds of questions are usually best posted to WP:NORN rather than here.

In general, if the view honestly seems to be widely held, then you can summarize it that way. Although that may not be the best solution, it is probably (barely) acceptable. In looking for better alternatives (which I encourage you to do), one thing you'll want to keep in mind is that "Some people say" is normally interpreted as "Less than half of people say". If half or more of journalists (who have expressed an opinion on this point) find his work misleading or deceptive, then you'd violate NPOV and NOR both by trying to pass off a majority opinion as one that "only some" hold. WhatamIdoing (talk) 02:12, 21 December 2011 (UTC)

The reverse (passing off a minority opinion as a majority one) is equally against both WP:NPOV and WP:NOR. I just figured that we should complete that example. The Mysterious El Willstro (talk) 04:43, 21 December 2011 (UTC)

Could use some input

If anyone is watching this, can you please chime in on a discussion at Talk:San Diego State Aztecs men's basketball#Notable players to explain why adding an arbitrarily chosen list of players from a school is considered original? Thank you. Jrcla2 (talk) 18:18, 21 December 2011 (UTC)

Very neutrally put.--JOJ ^Hutton 18:58, 21 December 2011 (UTC)

The concept of primary and secondary sources is not logically consistent.

The concept of primary and secondary sources is not logically consistent. Please discuss at Wikipedia:Primary and secondary source paradoxes in law related articles. PPdd (talk) 20:38, 1 January 2012 (UTC)

The way I understand Wikipedia's policies about primary and secondary sources, and reliable sources, is that Wikipedia, or any encyclopedia for that matter, is a very highly condensed form of information. It does not attempt to compile all possible information about a topic, but gives an introduction to the material. Futhermore, Wikipedia provides links to sources where others can verify that the information is correct and can get more detailed information. The idea that seconadary sources should be used in Wikipedia rather than primary sources is because we want Wikipedia to focus on the very most important information. If a judge or attorney said something that no newspaper or magazine reported on, it probably isn't worth adding to an encyclopedia article. That's why Wikipedia says secondary sources should be used instead of primary sources. Don't try to make the rules theoretically "logically consistent" -- make them pragmatic in that they will lead in practice to better encyclopedia articles. -- Schapel (talk) 23:37, 1 January 2012 (UTC)

WP:CALC phrasing

The phrase "the arithmetic and its application correctly reflect the sources" seems to be taken sometimes as saying something passes WP:CALC if there is a consensus that some arithmetic has been applied correctly. I think there should be more emphasis on the obviousness and meaningfulness of the calculation. One melon and one grape make two fruit, but saying two fruit is only meaningful if one is concerned about different types of fruit, however it is just silly if one is talking about amount of food. How about:

This policy allows routine mathematical calculations, such as adding numbers, converting units, or calculating a person's age, provided there is consensus among editors that the calculation is an obvious, correct and meaningful reflection of the sources. See also Category:Conversion templates

Any thoughts? Dmcq (talk) 09:11, 5 January 2012 (UTC)

Good point. I think that the policy should also point out that a calculation involving numbers or observations from multiple sources is improper synthesis and thus not allowed under WP:OR unless a RS directly support such calculation. Thus, I would be careful with using the plural form "sources" in your suggested phrasing. --Useerup (talk) 12:03, 5 January 2012 (UTC)

I think straightforward statistics are okay in many cases though one has to be careful with them, I wouldn't try banning them outright. What I object to is people taking statistics on figures arrived at by using different methodologies and measuring different things which are only broadly similar. The mathematics is okay when just applied to those figures but the meaning is extremely unclear. It fails in the 'application' part of the original wording but I think that needs to be made much more explicit and clear. Dmcq (talk) 12:41, 5 January 2012 (UTC)

Actually, I did not think of statistics in particular. I am certainly not against using statistics on numbers from a single source. I can imagine a number of scenarios where it would be ok to calculate the mean, min/max and - yes - even the median or quartiles. But similarly I can come up with only few scenarios where even adding numbers or summing across multiple sources would be ok. It seems to me that going from a single source to multiple sources we should raise the bar somewhat. The single-source case would certainly benefit from your suggestion. I just believe that the multiple source case should be restricted as it will almost invariably represent synthesis.--Useerup (talk) 16:30, 5 January 2012 (UTC)

Yes multiple sources is the major problem, and even with a single source like OECD you get loads of different countries getting figures their own ways. My usual response to people wanting to calculate their own thing and stick it into Wikipedia is why are you doing this if nobody else has been bothered to note it? How about 'If data items are joined together their compatibility must be WP:verifiable.' Dmcq (talk) 11:57, 6 January 2012 (UTC)

Not sure where to raise this

There is a question about the lack of third party sourcing at Talk:Calvary Chapel#Lack of third party sourcing. The argument by two editors is that the primary sources are being used to describe the subject and as such are adequate, while on the other side is an insistence that more secondary sources be provided to support claims. Could we please have a few people look at the article and the discussion and weigh-in? --Walter Görlitz (talk) 08:47, 7 January 2012 (UTC)

Wikipedia:SOPA_initiative

The Wikimedia Foundation has already hired a K Street lobbyist to lobby for "civil rights/civil liberties" but some political activists now want to use wikipedia.org as their advocacy vehicle. Besides the obvious WP:NPOV concern I see this initiative as likely to eventually undermine the philosophy behind WP:NOR if it isn't reined in. Why? Because the general philosophy behind no original research is that Wikipedia is passive, we follow instead of trying to blaze a trail. If Wikipedia turns crusader with respect to some legislation being debated in Washington, it's going to lend support to crusaders with respect to Wikipedia content who will use "ends justify the means" thinking to say original research is OK if it serves some libertarian agenda. I see "no original research" as a particularly "stay on the reservation" sort of instruction that would be particularly threatened if Wikipedia as whole has set a precedent of going off on its own. This Talk page is not the place to discuss SOPA but persons familiar with WP:NOR arguments may wish to express themselves on the appropriate SOPA pages.--Brian Dell (talk) 23:25, 15 January 2012 (UTC)

NPOV regards content of articles, and conflating that with SOPA is incorrect. People do not to think about issues outside their immediate area unless the issues become big news—it would be irresponsible for Wikipedia to fail to inform readers of the likely consequences of lobbyist-driven legislation that directly affects Wikipedia. Johnuniq (talk) 00:34, 16 January 2012 (UTC)

Is aggregation of primary sources synthesis?

Hello, I was wondering if aggregation of primary sources was considered synthesis. Consider the following two examples where the citations are primary sources:

A. Studies showed that result A was attained[23][24][25][26].

B. Studies showed that result A was attained[23][24][25][26], though some studies claimed result B was attained [27][28][29].

My personal preference is that these are exactly what belongs in encyclopedic knowledge. Though, some people are claiming these are "border line" synthesis. What is the consensus opinion about topics like these? Gsonnenf (talk) 17:09, 20 January 2012 (UTC)

If [27][28][29] are reliable sources they can be considered significant viewpoints and an article need to mention them. See WP:DUE.--Useerup (talk) 17:20, 20 January 2012 (UTC)

It all depends if they are primary sources or not (see WP:PRIMARY). If it's medical studies, and they're primary, then it's often synthesis to draw conclusions from different studies. Per WP:MEDRS, you should rely instead on review articles, which are secondary. Jayjg ^(talk) 19:23, 20 January 2012 (UTC)

Jayjg, you have a conflict of interest on this subject as you are a person I am in dispute with. I recommend you let people who are not involved in the dispute answer. I already presented a strong case that using evidence and quotes from WP:PRIMARY, WP:MEDRS, WP:SYNTH and WP:DUE that aggregation is not synthesis. I came here to get a neutral POV from people not involved in the discussion. I believe your answering of the question was in bad faith because you did not state your conflict of interest. Gsonnenf (talk) 23:15, 20 January 2012 (UTC)

It can be difficult to give a correct answer with only an abstracted situation. What is the article in question, and has it already been discussed on a talk page? Will Beback talk 23:35, 20 January 2012 (UTC)

The article is circumcision. There were many instances removal of primary sources added to a secondary source, or were presented when a secondary source wasn't available on a specific topic. The talk page contains elements of the most recent debate under #Primary Research and #Recent additions.Gsonnenf (talk) 23:44, 20 January 2012 (UTC)

I don't see a problem with citing medical studies with a form like "According to a study by X, Y is true", provided the opinion cited to X is literally the opinion X gives (i.e., without extra interpretation). Interpretation of the result beyond the words used by X requires a reliable secondary source. Zero^talk 00:30, 21 January 2012 (UTC)

This is illegitimate SYNTHESIS, and a MEDRS violation. Fifelfoo (talk) 02:09, 21 January 2012 (UTC)

IMO, primary sources should only be used rarely and with great care. The best use for them is to provide an illustration to something already covered in a secondary source. The idea that it's OK to use them if a secondary source can't be found is all wrong. One big problem with using primary sources is that they often require interpretation, which we're forbidden from doing, and that it can be hard to judge all of the externalities, such as their reliability and relevance. On a well-developed topic like circumcision there's very little reason to use them. Will Beback talk 02:26, 21 January 2012 (UTC)

It is synthesis. TFD (talk) 07:12, 21 January 2012 (UTC)

Looking at the guidlines via WP:SYNTH, i'm not really sure how it could be synthesis. Synthesis appears to be defined as follows:

"Do not combine material from multiple sources to reach or imply a conclusion not explicitly stated by any of the sources. If one reliable source says A, and another reliable source says B, do not join A and B together to imply a conclusion C that is not mentioned by either of the sources. This would be a synthesis of published material to advance a new position, which is original research.[9] "A and B, therefore C" is acceptable only if a reliable source has published the same argument in relation to the topic of the article."

And if we look at the what synthesis is not page WP:SYNTHNOT we can review the following:

SYNTH is not an advocacy tool: If someone doesn't like what was said, and they therefore cry SYNTH, others almost certainly will be right to cry foul. Virtually anything can be shoehorned into a broad reading of SYNTH, but the vast majority of it shouldn't be.

"SYNTH is not a rigid rule :Wikipedia doesn't have them, supposedly. But if a policy gets enforced zealously, it can be hard to tell the difference. The solution is not to enforce policies zealously. Never use a policy in such a way that the net effect will be to stop people from improving an article.

"SYNTH is not presumed: If you want to revert something on the grounds that it's SYNTH, you should be able to explain what new thesis is being introduced and why it's not verified by the sources. You don't have to put the whole explanation in the edit summary, but if someone asks on the talk page, you should have something better ready than "Of course it's SYNTH. You prove it isn't." The burden of proof is light: just explaining what new assertion is made will do, and then it's up to the other editor to show that your reading is unreasonable. But in any disagreement, the initial burden of proof is on the person making the claim, and the claim that something is SYNTH is no exception."

with that said, it would be proper to ask what new assertion is being made by aggregating that is not supported by the sources. Gsonnenf (talk) 16:54, 22 January 2012 (UTC)

WP:SYNTHNOT is an essay. It has no implications for policy, it's just someone's personal views. Since you appear to have a great deal of difficulty understanding WP:SYNTH, instead please review WP:MEDRS, which is very clear, particularly in relation to using primary sources versus review articles. Jayjg ^(talk) 17:24, 22 January 2012 (UTC)

Jayjg, I'm sorry you feel the need to wp:hound all my posts. This was a question about whether aggregating sources is synth. This was not a question on primary sources in WP:MEDRS. That is a different topic entirely. If you would be so kind as to stop following me around to every topic I post on, it would be appreciated.

My question to those who disagree is: What new information is being synthesized by aggregating sources? Gsonnenf (talk) 18:15, 22 January 2012 (UTC)

You are making a medical evaluation from synthesis of primary sources. You have been repeatedly told that you cannot do this. Your I don't hear that behaviour has become disruptive; I strongly urge you to stop it and accept that the community has established WP:MEDRS as the standard for making medical claims . Fifelfoo (talk) 22:48, 22 January 2012 (UTC)

In addition to what Fifelfoo has said, you have accused me of "wp:hounding all [your] posts". I have, in fact, responded to your claims on this specific issue, and this specific issue only, in several (though not all) of the fora in which you raised it. You, on the other hand, have now followed me to a completely unrelated discussion in order to oppose me there. Unlike what I have done, you are now actually WP:HOUNDing me, and that better be the last time you do it. Jayjg ^(talk) 22:59, 22 January 2012 (UTC)

Jaysg, you followed me here, and you followed me to an individuals talk page to harass me. Their were several arguments on the previous page. I wasn't discussing the Primary source here, I was discussing aggregation, not primary source in wp:medrs. Your inability to understand that I am asking a question far different than I am asking is unproductive and feels like harassment and bad wiki lawyering. If you have something to constructive to say about the actual question, please by all means say it, then let the discussion continue without disrupting it repeatedly. Thank you.Gsonnenf (talk) 05:47, 23 January 2012 (UTC)

" Is aggregation of primary sources synthesis? " yes it is illegitimate synthesis, and it is original research. Fifelfoo (talk) 06:04, 23 January 2012 (UTC)

Yes Fifelfoo, I understand from your earlier comment that you believe it is synth. Other people have commented that it isn't synth. What I'm trying to find out is "what is being synthesized?", excuse me for being obtuse but I can't seem to figure out what would be reasonably synthesized in this case. I'm sure the people who think it is Synth have a rational explanation to this they would like to share with those people who do not believe it is synth. I'm also sure this article will stop when people lose interest., though 2 days is a very short amount of time. Gsonnenf (talk) 06:36, 23 January 2012 (UTC)

It is synthesis and not permitted. This is a well-documented subject and so there are ample secondary sources available which must conform to WP:MEDRS. Mathsci (talk) 09:47, 23 January 2012 (UTC)

Here's my take. Assuming arguendo the sources are secondary, or are appropriate primary sources (assuming such a thing exists), and that they are not being misrepresented, it is not only appropriate but required to outline both results or points of view. The caution has to be in representing sources accurately without synthesis or new interpretation. That said, however, caution must be applied in how you list the different points. Any elaboration on the differences would be OR without a source for that elaboration. If you use a conjunction to join the two statements, which conjunction is used should be a careful choice, so as not to imply either side is more important. As a final reminder, though, it's the fact these are primary sources that requires the greatest caution. SamBC(talk) 14:43, 23 January 2012 (UTC)

open research

open research must be mentioned. the confusion between wikipedia and an open research tool is often made by newbies — Preceding unsigned comment added by Mehmedmed (talk • contribs) 22:39, 4 February 2012 (UTC)

Well the no original research policy is mentioned all over the place and I'd have thought a clue might be in the 'No'. I don't think extra words help to make things clearer, saying open research might give people the idea that other research is okay for instance. I think a Just Say No approach is better. Dmcq (talk) 22:53, 4 February 2012 (UTC)

people will get less confused by this policy if their energy are redirected to Original Research wikis. 'go there' is more constructive than 'just no'.

People violating WP:OR are simply doing open research in the wrong place. I had no clue about open research before finding it by chance by myself. It saves the energy of everyone, as I no longer even try to contribute to WP.Mehmedmed (talk) 00:00, 5 February 2012 (UTC)

Please do not keep inserting the stuff when two other people disagree. It may be useful but this is a policy about putting stuff into Wikipedia. It is not a howto about doing research. People violating WP:OR do it for all sorts of reasons and I'm pretty certain very few of them are doing anything worthwhile or willing to put their efforts elsewhere. I'm glad you actually have found a place to expend your efforts on something useful but I would think that your case is very rare. If a person wants to do research they can search for research and find out about it. Wikipedia says it is not the place for research. The policy is not the place for this any more than the passport office is the place to advise you about a good hotel. Dmcq (talk) 00:37, 5 February 2012 (UTC)

I am afraid you hold too negative feelings on the intentions of contributors facing 'WP:OR': it is merely people who think they have some knowledge to share, and go to the most famous website where such contribution is possible. They just leave WP when they find that the true story is more complicated (weird rules, etc...), and go spending their time elsewhere. I thought I could divert some of those people to websites they were really looking for.

But I have to admit I do not hold a better view on editors bullying contributors with 'WP:OR': they implicitly convey the message:'do not introduce your own ideas in articles: if you really were smart enough to have interesting ideas, you would not be here on WP: you would spend your time publishing in prestigious journals'.

You are just confirming my fear, and this is why worthwhile contributors no longer come to WP, they prefer writing for prestigious journals, or Scholarpedia, etc...(when they are given the opportunity to hear about the latters). If you are too negative on this issue, I just have to give up too.Mehmedmed (talk) 20:06, 5 February 2012 (UTC)

Well thanks for expending the effort trying to be helpful even if others have different ideas about accepting your advice and good luck with wherever it is that you have gone to with your original research. If some other site eventually becomes more successful with a different way of doing things that's life but variety is strength, it is better if different sites do different things. Dmcq (talk) 23:35, 5 February 2012 (UTC)

Wifione general approach is wrong

Wifione wants to

Require multiple secondary independent sources for an analytic claim. So we can't quote an expert, writing a secondary source, who has a different view than the rest of the experts, even if we also include the mainstream view.
Analysis of a primary source must be from multiple secondary independent sources. Not all primary sources gather enough attention for multible, findable, secondary independent sources to exist.
The editor's attitude toward primary sources appears to be that primary sources really ought to be prohibited, but a primary source can be allowed as an interim measure until a secondary source is found.

I advocate removal of most of Wifione's edits to this policy. Jc3s5h (talk) 16:02, 8 February 2012 (UTC)

Well from that little bit it seems you do not believe in any sort of citation at all. Could you please substantiate what you say please by pointing out what has prompted this and what you want changed in the policy about it, after all this page is about improving the policy page. Or just say why you wrote it and we might be able to work something out from that. Dmcq (talk) 16:10, 8 February 2012 (UTC)

I see something that has been inserted in the primary sources aboutr requiring secondary sources rather than a secondary source. I'll revert that. Dmcq (talk) 16:18, 8 February 2012 (UTC)

Quotes taken from this version.

"Any interpretation of primary source material requires reliable, independent secondary sources for that interpretation." [Emphasis added.] This statement changed to plural form by Wifione.

"Material based purely on primary sources should be avoided." I couldn't find who added this statement, but the discussion above indicates Wikione supports it. It might be a reasonable requirement for an entire article, or a substantial section, but "material" can mean a short phrase.

New requirement: "Articles may make an analytic or evaluative claim only if that has been published by multiple independent, reliable secondary sources." Again, not appropriate for short claims that are not a central thesis of the article. Jc3s5h (talk) 16:31, 8 February 2012 (UTC)

I've reverted the bits about requiring multiple secondaries for an evaluation of a primary source. The bit about that material based purely on primary sources should be avoided is I believe in line with the rest of that paragraph. Material from primaries should not be brought in unless it is justified by a secondary source referring to it. Bringing in bits from a primary source referred to by a secondary source but unrelated to what is talked about in the secondary source is very problematic, they should only be used to fill in factual details. It is often okay to use them but they should not automatically be considered okay and open season for trawling for facts. Separate manuals say what they consider okay from primary sources and I think a general avoid unless allowed or justified is correct. It stops WIkipedia being filled with trivia that have no weight. Dmcq (talk) 17:11, 8 February 2012 (UTC)

Oh, just noticed this thread. Jc3s5h, you have valid points and I believe Dmcq has undertaken one change you pointed out. In general, you must have seen I have no problems with any version of the policy that is accepted by consensus. The plural was incorporated by me keeping in mind that multiple would mean not one, but at least two sources. I don't think that's too tough a requirement to ask for in the case of an expert who in his primary source article may have a different view from the mainstream. If only one reliable source has reported this expert's divergent view, there is still pretty much a question, on a case by case basis of course, on whether such a divergent view is worth being included in the article. But like I mentioned, I've no issues with any changes/additions as long as they adhere to consensus; and with this particular change, no issues at all. Wifione ^Message 06:13, 9 February 2012 (UTC)

I think that requiring multiple secondary sources is a very bad idea, and I suspect that the fundamental problem is that Wifione doesn't really grasp what the difference is between a primary and a secondary source.

We should not be requiring multiple meta-analyses or systematic reviews to be able to say that pre-term babies who are given corticosteroids before birth are less likely to die than those who aren't. Just one such secondary source should be enough. I could understand (but would oppose) a desire for multiple primary sources (a peer-reviewed report in an academic journal about a randomized controlled trial is a high-quality primary source) to make such a claim, but not multiple secondaries. WhatamIdoing (talk) 18:33, 9 February 2012 (UTC)

WhatamIdoing, as I mentioned above , I sense that we both seem to be arguing for a similar issue from different perspectives. Multiple sources to support a contentious, never reported before or divergent viewpoint is surely a preferred idea (Why would it be, as you mention a bad idea?). Taking your example, if just one secondary source has reported that "pre-term babies who are given corticosteroids before birth are less likely to die than those who aren't," and multiple secondary sources have reported that "pre-term babies who are given corticosteroids before birth are more likely to die than those who aren't," you will have to take a call on how much weight you provide to each assertion - and that's a call to be taken on the talk page of the article, not here. Again, do note that the change to a source has already been made by Dmcq; and I have no issues with the same. Kind regards. Wifione ^Message 05:01, 10 February 2012 (UTC)

Policy on primary sources is too extreme

At present WP:Primary states:

Wikipedia articles should be based on reliable, published secondary sources and, to a lesser extent, on tertiary sources. Secondary or tertiary sources are needed to establish the topic's notability and to avoid novel interpretations of primary sources, though primary sources are permitted if used carefully. Material based purely on primary sources should be avoided. All interpretive claims, analyses, or synthetic claims about primary sources must be referenced to a secondary source, rather than to the original analysis of the primary-source material by Wikipedia editors.

The second sentence:

Secondary or tertiary sources are needed to establish the topic's notability and to avoid novel interpretations of primary sources, though primary sources are permitted if used carefully.

states a useful purpose of secondary and tertiary sources: "to establish the topic's notability and to avoid novel interpretations of primary sources" and suggests the reasonable position that "primary sources are permitted if used carefully". These useful guidelines are followed up upon in the last sentence:

All interpretive claims, analyses, or synthetic claims about primary sources must be referenced to a secondary source, rather than to the original analysis of the primary-source material by Wikipedia editors.

which very sensibly is intended to caution against introduction of "interpretive claims, analyses, or synthetic claims" based upon primary sources.

It then goes much farther stating:

Material based purely on primary sources should be avoided.

This statement requires some revision because it goes beyond the sensible purposes described in the rest of this paragraph to completely ban the use of primary sources unless secondary or tertiary sources are used. This statement is not required by the others, and is unnecessary to achieve the goals already pointed out. I believe it originates in the worry that contributors that compose an article using only primary sources will fall into the traps already identified in the other sentences of this paragraph, and rather than deal with that possibility, an iron law is being applied that goes too far.

There are circumstances when an article that uses only primary sources makes sense. As an example, suppose that an article were written to detail the various state requirements on becoming a district attorney. It could refer only to the laws adopted by each state in this regard, and therefore use only primary sources. It is purely descriptive and contains no interpretation, synthesis, or whatever. There would be nothing wrong with that article, and it could be a useful article on WP. However, this sentence: " Material based purely on primary sources should be avoided." suggests such an article should not be accepted.

I'd suggest that this sentence be deleted from the policy. The other sentences in this paragraph suffice to avoid abuse. Brews ohare (talk) 18:02, 4 February 2012 (UTC)

No I oppose this. Are you really saying there are no secondary sources about the various requirements in becoming a district attorney? Or can you find a less hypothetical example to try an bolster your case please. You're allowed to use a primary source to expand a bit about the facts in a secondary source. Dmcq (talk) 18:31, 4 February 2012 (UTC)

Dmcq: This is a hypothetical case of course, so let's hypothesize there are no secondary sources on this matter. Then, under these circumstances, should the article be refused?

Regardless of the particular example, the dangers in using primary sources already are spelled out, and this blanket assertion is not necessary to avoid these hazards: the policy can be cited using them directly. Brews ohare (talk) 18:37, 4 February 2012 (UTC)

Your hypothesis is that there is a subject everyone will agree is notable and yet there is no secondary source about it. That's basically assuming that there is a problem and then because you hae assumed there is a problem complaining about the problem. We really do need a proper example. Dmcq (talk) 18:56, 4 February 2012 (UTC)

There is no need for a particular example here, other than to aid the imagination. The point is that misuse of primary sources is already covered in the policy without this statement that "Material based purely on primary sources should be avoided." Maybe you can think of circumstances that require this generalization? If so, please provide your example. Brews ohare (talk) 19:03, 4 February 2012 (UTC)

There is actually a need for a concrete example.

As worded the redundancy is an apparent attempt to wikilawyer-proof the section. That is something that is unfortunately needed. To get it removed, something more than a "What if..." or "Use your imagination..." is needed to show why.

- J Greb (talk) 19:55, 4 February 2012 (UTC)

J Greb: Let me see if I have understood your position. I'd try to paraphrase it as follows:

There are situations where primary sources can be misused. Let's call these "the conditions".

Unscrupulous contributors will argue that their contributions are not ruled out by "the conditions", leading to long debates.

Therefore, the practical thing to do, to avoid debate, is to say primary sources can never be used by themselves, but only to supplement what is said in secondary sources.

Maybe some articles that could be useful are prohibited in this fashion, but that is tolerable collateral damage.

Have I got your message? Brews ohare (talk) 20:14, 4 February 2012 (UTC)

Not quite.

The text is, or should be, clear that articles should be built relying on secondary and tertiary sources.
Primary sources can be used, but not as the major, or only, support for the article.
NOR is a policy which means there are going to be close to no cases where an exception would exist. When those rare exceptions come along, they are discussed either on an article talk page or an RFC. But it does require a specific case or situation, not a "Well, what if..."
The PRIMARY section allows a little more latitude but it underlines the caution and care needed when plunking in material from a primary source. At least one example of where this latitude is applied is with articles that include a plot summary of a work of fiction. Even with that, it still boils down to there needed to be a clear example to discus if that latitude should be applied.
In general, editing from primary sources has been contentious. The wording needs to be clear if redundant so as to avoid weaseling to try to avoid the policy.

Now, if you have a specific example of an article or article section that you believe is an exception to NOR in general or PRIMARY specifically, a discussion of that topic would be a good idea to see if there is consensus for it being an acceptable exception. But that would not be grounds for loosening the policy or lessening that cautions in it.

- J Greb (talk) 21:28, 4 February 2012 (UTC)

An example where "Material based purely on primary sources should be avoided" is too extreme would be any article about laws in US states. Many secondary sources could, and would, be used to summarize the most common positions taken by the various states. For example, see Gun laws in the United States (by state). But the details for each state usually refer to sources provided by the government of each state, which of course, are primary sites. If the present policy were strictly enforced, a secondary source that describes each state would have to be found, and only then could a primary source be used to supplement the secondary source(s). Jc3s5h (talk) 20:58, 4 February 2012 (UTC)

If the opinion of a state is not mentioned in reliable, apposite secondary sources, why are we mentioning it at Gun laws in the United States (by state) at all. Your example is an invitation to synthesis. Articles, "based purely on primary sources" are inherently fabulism and original research. Fifelfoo (talk) 21:23, 4 February 2012 (UTC)

Gun law isn't something I'm interested in but what you are describing sounds like what is allowed for primary sources, that they can be used to give precise details of things described in secondary sources. If there is stuff there that secondary sources have shown no interest in that would be a worry. Dmcq (talk) 21:27, 4 February 2012 (UTC)

J Greb, Dmcq: To summarize your position: To avoid "weaseling" or "wikilawyering" WP:Primary is deliberately made overly restrictive with the purpose of forcing any contributor to WP to argue that their specific and particular use of primary sources is OK, and so an exception should be granted.

The problem with this tack is the following: anyone opposed to the contribution will simply quote WP:Primary:

" Material based purely on primary sources should be avoided."

This position will end the discussion as any counterargument will be taken as contrary to this provision.

So far as I can see, such collateral damage is perfectly all right with you two. Am I right? Brews ohare (talk) 22:17, 4 February 2012 (UTC)

You have requested an example of a good article that violates this provision. However, you also already have said that if an example be found it is not an argument to change WP:Primary, but simply an isolated instance for which an exception might be made.

According to my experience, with this language in WP:Primary there is no likelihood whatsoever of an exception being made for any example, however worthy.

That is why I think the wording should be changed to state what is necessary, namely, retain the other sentences in this paragraph that have specific objections, and delete the vaguely general statement quoted above. Brews ohare (talk) 22:17, 4 February 2012 (UTC)

I think that Brews is right about the sentence ("Material based purely on primary sources should be avoided") going too far. Half a sentence counts as "material" on Wikipedia. Any single fact counts as "material". I think what we mean here is something more like "major sections of articles" or "whole articles", not "a couple of words". WhatamIdoing (talk) 22:23, 4 February 2012 (UTC)

I would certainly agree with WhatamIdoing's clarification... indeed there are (limited) situations when citing a primary source should not be avoided, when citing a primary source should be actively encouraged (a quote for example, should always be cited to the original... as secondary sources may misquote or take the quote out of its original context).

So the instruction needs to be modified... however, let's not go too far... while there certainly is a place for primary sources, the core of any article should be based on secondary sources. For Notability reasons if not for OR reasons. Blueboar (talk) 22:54, 4 February 2012 (UTC)

J Greb, Dmcq: Maybe you could rationalize requiring a WP article stating a fact to refer to a secondary source instead of directly citing the primary source in WP? If, as explained in the WP:Primary policy, this fact was built upon to draw an inference or whatever, of course a secondary source is needed. But if all that is done is to cite a Wikimedia article that says "Jimmy Wales is on the board", why do we need a secondary source that says that "Wikimedia says that: 'Jimmy Wales is on the board' (See Wikimedia)"? Brews ohare (talk) 22:31, 4 February 2012 (UTC)

And to pre-empt the response that this is fine for an isolated fact, what if the entire article were to consist of exactly such things, for example, WP:Formal organization? Brews ohare (talk) 22:34, 4 February 2012 (UTC)

Wikimedia invites sociologists in to discover who does what on Wikipedia. Your facts are not straightforward. Fifelfoo (talk) 22:37, 4 February 2012 (UTC)

Your summary of what I said is nothing like what I said. Please come down from the clouds if you wish to have a reasonable debate, I am not given to arguing about the number of angels on the head of a pin. Dmcq (talk) 22:47, 4 February 2012 (UTC)

Dmcq: There is nothing I've said that is not concrete. I've provided a simple question, a simple example, a simple statement of what should be changed. Just who's argument is "in the clouds"? BTW "debate" is not the goal, but "discussion". Brews ohare (talk) 23:12, 4 February 2012 (UTC)

Well your standards for concrete don't satisfy my criteria for that. I'll keep an eye on this and perhaps come back if I see something I consider concrete and think I can con contribute to. Dmcq (talk) 23:16, 4 February 2012 (UTC)

The discussion is a waste of time because it begins with a false example. Suppose for example that no newspapers or other secondary sources had covered the moon landing, the Second World War, the last U.S. presidential election - then neither would we. But their coverage provides evidence of their importance. TFD (talk) 03:31, 6 February 2012 (UTC)

Most newspaper articles are primary sources. See WP:PRIMARYNEWS for more information. Newspapers show "evidence of their importance" by being independent, not by being secondary. WP:Secondary does not mean independent. WhatamIdoing (talk) 04:23, 8 February 2012 (UTC)

Deciding whether primary, secondary or tertiary sources are appropriate on any given occasion is a matter of good editorial judgment and common sense, and should be discussed on article talk pages I am confident that the talk pages of articles would resolve the issue of whether a primary source is appropriate for a particular situation or not. When the matter being reported is not exceptional, promotional or controversial, discussions on the talk page of articles should be able to amicably resolve the usage of primary sources appropriately. I'll strongly recommend not changing the primary source policy here as the basis of BLPPRIMARY and the reason why we excluded mentioning considerable portions of PRIMARY in BLPPRIMARY was because the same was already mentioned here. When I authored BLPCRIME, the same primary issue was considered quite strongly. Would you accuse an individual of being guilty of a crime because a court trial record confirms the same? Or would you wait till a reliable source mentions the same? I mention these questions basically not to get answers, but to reinforce that instead of changing the statement here, one should use talk page discussions. Yet, there is a mid-path that I could suggest. We could change the statements alluding to "material based purely on primary sources should be avoided" to "challenged material based purely on primary sources should be avoided." Does this sound doable? Wifione ^Message 16:12, 6 February 2012 (UTC)

Well that what it is really but the phraseology sounds to me like open season for people to trawl around old court records and suchlike and stick in stuff. I really would not prefer that sort of original research. Dmcq (talk) 20:03, 6 February 2012 (UTC)

There is nothing gained in refusing to allow a primary source for a simple statement of fact (like, "the court record states ...", or "WP says it is structured in an hierarchy consisting of ...") and we all agree about that. However, the present policy says no article can be based entirely on primary sources, and makes no exception for any type of article. That is the issue being raised here. So simply as an example, an article with the subject "State requirements for the office of district attorney" with subsections: "Arizona requirements", "California requirements", and so on for all the states, is disallowed if it uses only state primary documents as a basis of the requirements. Why do that?

In fact, the articles District attorney and United States Attorney, violate WP:Primary, refer only to primary sources, and should be removed.

It is simply ducking the issue to say that no such article has any value, is notable, or that secondary sources can be found that say "according to A, A requires x" so that instead of quoting A we can quote the secondary source that quotes A. Brews ohare (talk) 19:10, 7 February 2012 (UTC)

We do not all agree that trawling round court records for things that haven't been mentioned in secondary sources is a reasonable thing to do in articles. Only if a court case has been mentioned in secondary sources are we entitled to inspect the court records and get details from them. See WP:BLP#Misuse of primary sources. And p.s. WP:WEIGHT applies if it just gets a one-liner so it might have to be omitted for that reason. Dmcq (talk) 23:24, 7 February 2012 (UTC)

I will also note that District attorney does not violate WP:Primary, because it does not cite any sources (either primary or secondary). That is certainly a problem that needs to be fixed, but it is a different problem from the one being raised here. United States Attorney does cite secondary sources... it also cites a lot of primary sources. The question is whether the article overly relies on the primary sources. It may. I suspect that a lot of the material currently supported by primary sources could (and should) be supported by secondary ones instead. If so, we should swap sources. Blueboar (talk) 23:53, 7 February 2012 (UTC) Blueboar (talk) 23:29, 7 February 2012 (UTC)

And this is the most appropriate view (even though I've mentioned that we could add "challenged" to the primary statement). The fact is that the word "avoid" is not equal to the word "prohibited". In other words, certain self published sources (primary) can be used as sources for themselves as per policy here, as long as they are not promotional, self-aggrandizing, controversial and similar exceptional claims. Remember, any challenged claim (even if it is about someone's name or designation etc) becomes an exceptional claim if it is purely supported by primary sources, and this is policy. Therefore, I really don't think there is any issue here. Avoid material purely based on primary sources, taking into account our SPS policy. Wifione ^Message 03:47, 8 February 2012 (UTC)

I don't think that's a good idea. Let me explain why with an example: Did Bill Clinton say, "I did not have sexual relations with that woman", or did he not?

We've got a video of him actually pronouncing those words. That video is a primary source. The video is uncontested in the real world. Nobody has ever claimed that the video is a hoax. The video is, in fact, the most authoritative source in the universe for the question at hand: did he say these words, or not?
So let's pretend that somebody decides to challenge that fact. WP:CHALLENGEs do not have to be based on any rational concern: if it happens, it happens. You could challenge anything you want, for any reason you want. Maybe the person has political motives here. Maybe the person is a crackpot. Maybe the person hopes to erase any mention of sex from Wikipedia. It doesn't matter: challenged is challenged.
Your rule would prohibit us from using the single most authoritative source in the universe to support this fact.

I don't think that's what you want, but that's what you'd get if you said that challenged material could not be supported by a primary source. WhatamIdoing (talk) 04:32, 8 February 2012 (UTC)

If the video was quoted by a reliable source, you could use it. But if the video was uploaded by Mr. Clinton, then it remains a pure and proper primary source. WELLKNOWN; if you don't have multiple secondary reliable sources supporting this claim, then the video source is unreliable. I know of many well known personalities (I'm not including Mr. Clinton) attempting such stuff for publicity. This example may be actually extreme; yet, that's the point I guess. I'm ok with or without including the term challenge, as the current Wikipedia:Verifiability#Self-published or questionable sources as sources on themselves policy takes care of this issue. Wifione ^Message 05:08, 8 February 2012 (UTC)

No, that's not true. The video clip is a primary source no matter who taped it and no matter who published it. "Primary" is not a synonym for "self-published".

We aren't, by the way, talking about a YouTube clip of dubious origins. We're talking about a public speech that was broadcast by practically every television station in the world that has any pretensions to a news program, and that can be found on the websites of news organizations as well as in the National Archives and Records Administration. WhatamIdoing (talk) 18:25, 9 February 2012 (UTC)

I understand what you're saying. I just suspect that we're both arguing for the same issue from different perspectives. The video clip does remain a primary source. But as per BLPPRIMARY, you can use it only to augment a reliable source. Does that sound better worded? Best. Wifione ^Message 04:46, 10 February 2012 (UTC)

as an addon to this discussion, what does "reliably published" mean in this context. Reprints of letters/videos/etc? by reputable sources? This was coming up in a discussion regarding a band, where an editor was justifying use of the bands website under primary sources. Gaijin42 (talk) 21:17, 9 February 2012 (UTC)

The current wording of NOR is fine: avoided does not mean never. One reason why primary sources should be avoided is that Wikipedia is a sitting duck for cherry picking POV pushers who could quote extracts from various complex legal proceedings to suggest a conclusion that is contrary to subsequent developments (maybe politician X was found guilty of molesting children, but a secondary source might mention that X was completely exonerated in the appeal). A video of an interview might show a biologist saying that "evolution is junk" when in fact the biologist immediately corrected the statement to "creationism is junk" when they jumbled talk was pointed out to them after the interview. Primary sources are good for things like illustrating what a secondary source says, and some other limited stuff. Johnuniq (talk) 22:33, 10 February 2012 (UTC)

Text too short to convey meaning

....apparently [1]. Please discuss. B^e——C_ritical 17:12, 6 February 2012 (UTC)

WP:POLICY#Content says 'be as concise as possible—but no more concise. Verbosity is not a reliable defense against misinterpretation. Omit needless words. Direct, concise writing may be more clear than rambling examples. Footnotes and links to other pages may be used for further clarification.' However I haven't any particular objection against either. Dmcq (talk) 20:09, 6 February 2012 (UTC)

Prefer the longer version... trimming to omit needless words is all fine and good ... but I think some of the words that were omitted to create the shorter version are needed. Blueboar (talk) 22:17, 6 February 2012 (UTC)

I also prefer the longer version. WhatamIdoing (talk) 04:34, 8 February 2012 (UTC)

Longer version seems more clear. Perhaps a bit of trim could happen.Gsonnenf (talk) 12:36, 11 February 2012 (UTC)

Educating our editors

Since we seem to be having this problem again, perhaps it's time to insert these sentences into each of the three "definitions" at PSTS:

A primary source can be either independent or non-independent. A primary source can be either self-published or non-self-published.

A secondary source can be either independent or non-independent. A secondary source can be either self-published or non-self-published.
A tertiary source can be either independent or non-independent. A tertiary source can be either self-published or non-self-published.

What do you think? Would this help editors figure out that WP:Secondary does not mean independent or that primary does not mean self-published? WhatamIdoing (talk) 19:10, 9 February 2012 (UTC)

Independent doesn't have much to do with OR policy, it is more about reliable sources or verifiability. Dmcq (talk) 19:17, 9 February 2012 (UTC)

Yup, but you'd never know that from all the people who keep using the words secondary and independent as if they were merely alternate spellings for the same concept. The whole point of these sentences is to help people discover that independence (and self-publication) doesn't have much to do with OR policy. WhatamIdoing (talk) 19:48, 9 February 2012 (UTC)

In my opinion, independent vs non-independent of self-published or non-self-published is much less important than peer-reviewed vs non-peer-reviewed. In addition, with regard to independence, the very term is dubious: independence from whom? Are state-sponsored sources deemed dependent? Are privately owned mass media independent? That is a matter of judgement, and strongly depends on a context. For example, is state sponsored BBC a dependent or independent source?--Paul Siebert (talk) 20:26, 9 February 2012 (UTC)

In the context of this page, I don't actually care how independence is defined. It's not relevant to this policy. All I really want here is for people to stop typing p-r-i-m-a-r-y when they ought to be typing words like "not independent" or "self-pubished". WhatamIdoing (talk) 21:28, 9 February 2012 (UTC)

Clarifying terminology in a policy/guideline where some of the wrongly-interchanged terms are germane to the policy is, I think, appropriate and good to cover in the policy. North8000 (talk) 21:44, 9 February 2012 (UTC)

Probably should be made clear that primary/secondary/tertiary sit along a different axis as first-party and third-party (which is different from self-pub/non-self-pub too.) --MASEM (t) 23:08, 9 February 2012 (UTC)

WhatamIdoing, the essay you've mentioned on Independent sources repeats a few points. I just wish to compare two statements here; this might lead to a better understanding:

Our NOR policy mentions: Secondary sources are second-hand accounts, generally at least one step removed from an event.
The essay mentions An independent source is a source that has no significant connection to the subject and therefore describes it from a disinterested perspective.

I just feel that even the essay you mention (with this line as a leading example) replicates to some extent what is already mentioned. What do you exactly mean by independent which is clearly different from what is mentioned in secondary? Wifione ^Message 05:20, 10 February 2012 (UTC)

Comment: Ah! This is a fascinating question. What you are actually revealing here is a facet of OR. In the consideration of sources, unless a reference's complete origin can be ascertained/traced, your language is too confusing for the average editor to take its meaning. Verifiability seems to be the guideline that has worked very well thus far. Many editors are simply happy being able to verify that a reference actually exists. What I see being proposed here is some new level of scrutiny of references that is inappropriate to Wikipedia.--Djathinkimacowboy 07:59, 10 February 2012 (UTC)

RfC: Policy on primary sources is too extreme

WP allows primary sources to support isolated facts. Comment is requested here upon extending WP:V and WP:Primary to explicitly allow entire articles based upon primary sources, provided these articles contain no synthesis, analysis, or interpretation of the original sources. A summary of a recent discussion follows this RfC at Wikipedia talk:No original research. Brews ohare (talk) 18:46, 8 February 2012 (UTC)

Summary of discussion on this Talk page

I'll try to summarize the views expressed.

Topic The issue was raised that statements from WP:Primary,

Material based purely on primary sources should be avoided.

According to Wikipedia:Notability

if no reliable third-party sources can be found on a topic, then it should not have a separate article

and also, according to WP:V:

Base articles largely on reliable secondary sources. While primary sources are appropriate in some cases, relying on them can be problematic.

These statements are too strong because they do not allow any articles that are made up entirely of primary sources. In WP:V there are these conditions upon using ~~primary~~ self-published sources:

the material is not unduly self-serving;
it does not involve claims about third parties;
it does not involve claims about events not directly related to the source;
there is no reasonable doubt as to its authenticity;
the article is not based primarily on such sources.

The opinion expressed was that the first four of these requirements for self-published sources sufficed to prevent misuse of articles containing only primary sources, and the last restriction was not needed, and indeed would reject some useful articles.

[Educational comment: "Primary" is not another way to spell "self-published". This entire proposal is based on ignorance of what a primary source is. See, e.g., the extensive quotation of WP:SPS, followed by assertions that this is what WP:V says about primary sources. I have struck the erroneous assertion that WP:V imposes these conditions on primary sources and added the accurate bold-faced words to the previous statement. WhatamIdoing (talk) 18:41, 9 February 2012 (UTC)]

Responses Response to these observations were negative, but not directly relevant to the issue. They fall into these categories:

No useful articles like this exist, even in principle, so it is a non-issue. There are such articles, and an example is given later.

If an article like this comes up, it can be dealt with as a possible exception to policy, and no policy change is needed. This is a utopian position; exceptions to policy are very rare. And there is no need to force this situation: just change the policy.

If an article like this comes up, secondary sources always can be found to replace or supplement the primary sources, so the question comes down to whether the article is "primarily" based upon the primary sources. It is not true that secondary sources always can be found; books are written with different objectives than WP articles.

I find none of these objections addresses the point raised, namely there is no need for the blanket refusal to accept articles based entirely upon primary sources, and the four first items in the list of reservations above suffice without the fifth reservation. To assist in this consideration the following example was provided:

Hypothetical example

an article with the subject "State requirements for the office of district attorney" with subsections: "Arizona requirements", "California requirements", and so on for all the states.

If the present policy were strictly enforced, a secondary source that describes each state would have to be found, and only then could a primary source be used to supplement the secondary source(s). The policy has the nonsensical effect of requiring a secondary source and citing: "The book Law in America says 'Arizona law says x (Ref: Arizona law)' " instead of saying directly 'Arizona law says x (Ref: Arizona law)'. WP would not require that for this specific fact, so why should it be required for an article consisting only of such specific facts, with no interpretation, analysis, or synthesis?

In fact, there is an article already on WP United States Attorney, that is somewhat like this, but does have some secondary sources, and present policy raises the non-issue of whether it is primarily based upon primary sources. Such a consideration should arise only under the circumstances described in the first four restrictions cited above.

Are there further comments on the proposal to drop the last-named restriction upon articles consisting only of primary sources?

Comments

If you are not satisfied with the comments above in this same talk page you can raise an WP:RfC about it. Dmcq (talk) 17:54, 8 February 2012 (UTC)

Hi Dmcq: That sounds like a good idea. I have put in an RfC template

Comment: This change in policy will avoid tendentious articles by excluding synthesis, interpretation, and so forth, to allow only strict reporting of the content in primary sources. This policy change will allow articles in WP that are simply descriptions of what sources say about themselves. Examples are articles like WP:Formal organization or United States Attorney. A policy change avoids arguments about whether such articles are "primarily" based upon primary sources. It also avoids the need to find a secondary source S to say things like "Secondary source S says 'Jimmy Wales is on the board of the Wikimedia Foundation' (See primary source Wikimedia)" and instead to say directly 'Jimmy Wales is on the board of the Wikimedia Foundation' (See primary source Wikimedia). Brews ohare (talk) 18:57, 8 February 2012 (UTC)

Reply: Your first statement does not appear to be correct, since policy does not permit tendentious articles that include synthesis, interpretation, and so forth by editors from the content of primary sources. That's explicitly forbidden by WP:NOR. If you mean to imply that because secondary sources themselves are works that include synthesis, interpretation, and so forth from the content of primary sources, that this makes them tendentious, I think you're not going to have much fun working on Wikipedia, which relies almost entirely on secondary sources. For the rest, you don't tell us why we should avoid determining whether an article is primarily based on primary sources. The idea that we should avoid a requirement to use mostly secondary sources, and avoid it just so we can avoid a requirement ot use mostly secondary sources is circular reasoning. I agree that the type of case needs an exception, but the way to handle that is to make an exception, not by throwing the baby out with the bathwater and getting rid of the general rule. — SMcCandlish Talk⇒〈°⌊°〉 Contribs. 23:43, 26 February 2012 (UTC)

Comment: The aspect of notability is missing from all this, and without it, it opens the door to huge number of articles of non-synthetic, straight-up summary of primary sources but without any context. The impression I get from this discussion is we are mostly talking about list-style articles that summarize in a non-synthetic manner several disparate primary sources. It is not that it is bad to do this, but the problem is that there needs to be some reason why we have this list. In the specific example, one certainly can find enough secondary sources to explain what a DA is, and that the route to becoming a DA are diverse by state; once it is established that the routes and means to becoming a DA is a notable topic, primary-sourcing the individual state policies makes perfect sense. But if the list as a whole could not show notability through secondary sources, then the list shouldn't exist in a stand-alone form.
There's a lot of other complex interactions here when you move away from list-style articles, but still, we do expect every article to show notability, and thus show some amount of secondary sources. The ratio of primary to secondary may be vastly skewed depending on the topic (again, for the DA example, with 50 states, that's 50 primary sources, while I'd only expect to have to see 3 or 4 to establish the notability of the pathway to becoming a DA - that is not a problem in terms of this discussion). --MASEM (t) 19:10, 8 February 2012 (UTC)

Masem: I'd agree that some such articles may not be notable, but if they should not be on WP because of that, doesn't Wikipedia:Notability get rid of them? Shouldn't we reject articles with other faults for their actual failings, not because they refer to primary sources? Brews ohare (talk) 19:44, 8 February 2012 (UTC)

Notability requires coverage beyond primary sources. Mind you, I recognize the complex issue where we do know there's secondary coverage but prefer the primary as it is less interpretive (such as the process to be a DA by state), so it is unlikely that notability would affect this type of article. But when only primary sources are available - no secondary ones at all - then it is failing notability so deletion would be appropriate. --MASEM (t) 19:57, 8 February 2012 (UTC)

Notability as written does require deletion, but it also should be rewritten. The important issue here is WP:Indiscriminate, which covers WP is not just a collection of information. Again, shouldn't we reject articles with other faults for their actual failings, not because they refer to primary sources? Brews ohare (talk) 20:06, 8 February 2012 (UTC)

The actual failing of an article that only can be sourced to primary source is that there's no evidence of why it is important to include in a work that is not an indiscriminate collection of information. For example, from the video game area, I could easily make an article about the weapons in a video game, all sourced from the game and strategy guide. But unless sources go into any secondary, analytic depth about the weapons, this information has no importance to the larger picture of being an encyclopedia. --MASEM (t) 20:29, 8 February 2012 (UTC)

Your example can be eliminated from WP using WP:Indiscriminate. An article like WP:Formal organization uses only primary sources but it is useful as a guide to the many pages describing WP organization. An article like the "straw man" above could be similarly useful. Of course, this information could be used for other purposes, in articles drawing comparisons, evaluations and so forth, and those would require secondary sources. Why be restricted to articles making judgements and assessments?

Why assume absolutely no useful articles exist that are based upon primary sources alone? Why not let the applicable policies like WP:Indiscriminate be applied to eliminate inappropriate articles? Why assume we are smart enough to anticipate all possibilities, and rule this class of articles out altogether? Brews ohare (talk) 20:48, 8 February 2012 (UTC)

We need to establish why the topic is relevant to the reader. Without this, it is just datum, and not appropriate to the encyclopedia, as well as a conditioned that can be gamed. To show a topic to be relevant , we need secondary sources to show that's the case, otherwise, it is just original research.

I would argue that most articles that you point to that are presently only sourced by primary sources can be augmented with a few choice secondary sources in the lead to establish that point, ergo establishing notability and of course moving the article from being only sourced to primary sources, avoiding the issues of deletion/merging. --MASEM (t) 21:19, 8 February 2012 (UTC)

Please see my remarks to Dmcq below. Brews ohare (talk) 22:45, 8 February 2012 (UTC)

An article like WP:Formal organization uses only primary sources...

No. There is no such article. There is a project page, but no article. See the "WP:" at the start? That means "not an article". WhatamIdoing (talk) 18:44, 9 February 2012 (UTC)

Comment: Well knowing which books are on which floor in a university would help Wikipedia's educational aims, but should I really be putting in an article about every single library saying what the layout of its bookshelves is? And there are lots of recipe books around, surely being able to find a recipe easily is good and these are in books so shouldn't I stick their contents lists into Wikipedia? And how about the current train timetables? That'll get students to class in time. We really do need secondary sources to stop Wikipedia being filled with stuff that just isn't notable no matter how useful it is. A secondary source is visible evidence somebody finds the stuff notable rather than editors doing original research and deciding things like that for themselves. There is no weight without a secondary source. Dmcq (talk) 20:00, 8 February 2012 (UTC)

Dmcq: You have identified some possible articles that use only primary sources, and should not be on WP. However, WP:Indiscriminate is the relevant policy here: WP is not just a collection of information. I'd suggest that there are some articles of this type that are in fact useful, and rather than be clairvoyant and dismiss all such articles blindfolded, we should dismiss them for their actual failings, using pertinent policies, not because they are based only upon primary sources. Brews ohare (talk) 20:12, 8 February 2012 (UTC)

That is exactly my point. How do you show something is indiscriminate? I gave explanations of why each of my examples was useful and furthered the educational purpose of Wikipedia. Who are you to say that I'm wrong? We could have AfD's all day and night with editors fighting over tons of things like this without some decent bright line. And that minimal requirement is notability as shown by secondary sources otherwise it is a fairly quick exit. Dmcq (talk) 21:15, 8 February 2012 (UTC)

Dmcq & Masem: The purpose for such an article could be (for example) to act as a map locating the topics or sources outlined. Additionally, the article should contain enough summary detail to enable a reader to find what they are looking for. That is what WP:Formal organization does.

The statement of purpose doesn't have to be from a secondary source. In fact, a secondary source for (say) a guide to some classification of web sites, might be hard to find, even if the guide were useful.

If it is difficult to decide whether the statement of purpose is sufficient for inclusion, then maybe that debate should occur. Of course, there may be no unanimity over the value of "Regulations regarding dogs and rabies in American cities". The debate would decide if there was really any interest in the topic. Acceptance shouldn't be decided simply because all the sources were to government web sites.

Possibly, the proposed revision of policies should include a requirement for a "statement of purpose" and provide some guidelines for an acceptable statement? Over time a sensible set of guidelines would evolve, don't you think? Brews ohare (talk) 22:45, 8 February 2012 (UTC)

If we are talking specifically mainspace articles that would fall as "outlines" or the like, we generally do not apply notability guidelines to them, in the same manner we don't apply them to disambiguation pages - any page that helps in general navigation is ok. As part of the larger picture on that, however, if you are creating a navigation page for a very broad topic, it is absolutely certain that the topic is notable.

A statement of purpose does need to be secondary, otherwise you enter into the realm of COI to maintain an article on WP.

But let be me clear: there is a difference between an article that currently has only primary sources but it is known that secondary sources exist that repeat the primary sources; and an article where only primary sources exists and no secondary sources have been found outside of the article itself. The former we'd generally keep, because as long as there's some secondary interest, that's evidence of notability.

Your example of dogs and rabies raises a different problem. The subject of dog and rabies regulations is notable, and you probably can source the local regulations for them and fill the article with primary sources. But would we ever allow such an article? Likely not, not due to sourcing but due to it being indiscriminate - discussing the specific regulations per town (10,000s of them) is excessive for a encyclopedia that is meant to summarize. An article discussing the types of regulations in broad strokes, borrowing examples from specific regulations using primary sources, would be more appropriate here. But this is not an issue because of the primary sources, just the shear number of entries. --MASEM (t) 23:05, 8 February 2012 (UTC)

Comment: Masem: You suggest that mainspace articles that are outlines are not usually subjected to WP:Notability, which is good news. However, this policy does state:

"If no reliable third-party sources can be found on a topic, then it should not have a separate article"

So the fact WP:Notability is seldom used this way probably is because WP:V and WP:Primary come to editors' minds more readily. I'd take it that you would be agreeable to explicitly exempting 'outlines or the like' from these strictures against articles relying upon primary sources alone? Is that so? Brews ohare (talk) 14:05, 9 February 2012 (UTC)

My read of the various guidelines that talk about outlines and other formal organization mainspace "articles" used mostly for navigation is that they are generally not subject to the same content and inclusion policies as they work on the assumption that the outline is not introducing anything new to WP; however, there is disagreement if WP should support these types of "articles" that goes beyond WP:OR and other content policies, as well as WP:N and inclusion guidelines. For purposes of this discussion I would exclude discussion of outline and navigation-style articles since how policy and guidelines apply to them are different, and instead focus on articles with new content and how primary sources apply to them. --MASEM (t) 14:27, 9 February 2012 (UTC)

Masem: I'd appreciate some guidance to find these guidelines exempting outlines and such from WP:Primary, WP:Notability, and WP:V. It may be that these exemptions are really all I am looking for here, and some language in WP:Primary, WP:Notability, and WP:V referring to these guidelines would be adequate to my concept of what is needed here. Brews ohare (talk) 14:38, 9 February 2012 (UTC)

Well, the problem I see with that is that the example being used, the article on the various state requirements for becoming a DA, would not be an outline or similar type of article and that seems to be the type of article we're focusing on. So I don't think this advice is going to help with the concept you're trying to set. --MASEM (t) 14:57, 9 February 2012 (UTC)

Maybe the language you are aware of could be helpful? It seems that you think the definition of "outline or the like" needs extension somehow. I'd guess we could come up with some kind of trial balloon? Brews ohare (talk) 15:10, 9 February 2012 (UTC)

I have begun such a trial balloon approach in a comment further down the page. Please take a look. Brews ohare (talk) 15:50, 9 February 2012 (UTC)

Comment:
[Primary sources] are very close to an event, often accounts written by people who are directly involved, offering an insider's view of an event, a period of history, a work of art, a political decision, and so on... a scientific paper documenting a new experiment is a primary source on the outcome of that experiment. Historical documents such as diaries are primary sources.... Any interpretation of primary source material requires a reliable secondary source for that interpretation.
It seems to me this rule says it all. If, for example, Wikipedia wanted an article about Pepys' diary and used only the diary from its initiation, in simple terms that would merely be one of two things: a pure summarisation using only the diary, or, direct quotes from the diary in a piecemeal way. Either usage is OR without secondary sources, and that is why I think article creation rules should remain as they are now. May I offer a real counterexample with Birthstones: several lists are in the article, including the Gem trade and Jewelers (both of America). However, without secondary and tertiary sources behind them, the article would only look like a commercial advert for the two associations, nothing more. As it is, the article was created as a well-rounded and encyclopedic article about birthstones. A good specific example I can cite is mineralogist Georg F. Kunz's work, which is perfect as secondary reference on the subject and is cited in the article. It cannot stand without Kunz or its other secondary citations- but if one were going to make an article about birthstones using only one source, that seems to me equivalent to the change being proposed here. For all those reasons I oppose this suggested change to the rules.--Djathinkimacowboy 05:23, 9 February 2012 (UTC)

Djathink: You raise an important point, a possible misuse of primary sources to provide a parochial view of a subject by reporting only part of a subject through primary sources. I'd respond to this objection in several ways. First, all articles evolve, and if a first presentation of a topic is based too narrowly on primary sources, those recognizing this fact can add secondary sources (if there are some) or flesh out the presentation with other primary sources, or more balanced quotations from the primary sources themselves. Second, if objections to the article can be raised using criteria of WP policies that are not related to primary sources, those policies should be applied, not the objection to use of primary sources. For example, WP:NPOV comes to mind. Third, there are in place already important restraints upon the use of primary sources quite apart from the condition that not only primary sources be used. I think these conditions (which roughly speaking disallow interpretation, synthesis, and evaluation) should suffice to control major abuse of primary sources. The aim here is not to allow unbridled use of primary sources, but to remove the restriction that "no article shall rely primarily upon primary sources." It may be that removal of this restriction will require additional stipulations upon use of primary sources: that seems fine with me. I simply feel that the present "no article shall rely primarily upon primary sources" is a blunt instrument where a scalpel should be used instead. Brews ohare (talk) 13:54, 9 February 2012 (UTC)

Comment: Summaries of wide reaching government legislation can support an article on there own. Mundane, non-controversial but important legislation may not spawn secondary sources for a time, but a summary of such legislation is important encyclopedic knowledge. When discussing the "asserted position/policy" of a person or group, a primary source is the most reliable source and should be used over secondary sources. Content without notability, etc, is best handled under its respective policy. wp:primary should not be used as a tool to enforce policies covered elsewhere.Gsonnenf (talk) 12:48, 9 February 2012 (UTC)

Are you really saying governments can pass wide reaching government legislation of note without anybody secondary sources commenting on it? Dmcq (talk) 13:20, 9 February 2012 (UTC)

The key point here is the last statement, which is the thrust of this entire thread: WP:Primary should not be used as a tool to enforce policies covered elsewhere. Or perhaps, WP:Primary should not be used as a substitute for enforcing other policies. Brews ohare (talk) 13:34, 9 February 2012 (UTC)

If you want to talk about removing the definition of primary out of the Original research policy that's something else but currently this policy is the place to go for finding what primary sources are. That part is integral to how primary sources can be used. It would help nobody and it certainly wouldn't support editors if we have to go round the houses trying to build that up from the various policies instead of stating it in a straightforward statement in a description of primary sources. The asserted policies of an organisation are only of interest if its policies are of interest and you need secondary sources for that. I think people may be mixing up verifiability and citation here, citations are only needed if verifiability is disputed. Dmcq (talk) 14:10, 9 February 2012 (UTC)

Dmcq: I don't think this thread is about changing the definition of primary sources. It is about replacement of the present language that says more or less "no article is acceptable that uses only primary sources" with language that allows such articles subject to the other restrictions already in place. These strictures include those against interpretation, evaluation, or synthesis. Maybe if "no article is acceptable that uses only primary sources" is deleted, some different wording will be needed. Any suggestions? Brews ohare (talk) 14:17, 9 February 2012 (UTC)

For example, one suggestion arising from Masem's comments would be to replace "no article" with a list of articles that could be used this way: "outlines and the like are acceptable that use only primary sources" subject, of course, to other WP policies and to the requirements for no interpretation, synthesis, or evaluation of the primary sources. Brews ohare (talk) 14:27, 9 February 2012 (UTC)

Dmcq, if you feel something is not notable you can bring it up on a talk page and delete it under WP:NOTE. Some topics with secondary sources are not notable, some topics based on primary sources are notable. Editors should not be forced to hope for some obscure secondary source, such as a brief confirming article that a mundane non-controversial policy was passed, when there is consensus that the primary source is far more authoritative and notable then the secondary source. Having a mix of poorly focused rules enforcing a distinct goal (notability) will only encourage people to game the system and create bizarre intertwined wikilawyer arguments. Gsonnenf (talk) 02:18, 10 February 2012 (UTC)

Comment: Here are some possible approaches to relaxing the present restriction that "no article is acceptable that is based upon only primary sources":

Approach (1): Subject always to all WP policies regarding synthesis, point-of-view, interpretation, evaluation, notability and so forth, articles using only primary sources are allowed, but of course may be amended at any time to include secondary sources or additional primary sources. It is anticipated that most articles of this nature are preliminary, and will become full-fledged articles in time.
Approach (2): Subject always to all WP policies regarding synthesis, point-of-view, interpretation, evaluation, notability and so forth, articles using only primary sources are allowed that are intended only to facilitate the navigation of sources pertaining to a topic. As such the sources may be quoted or outlined strictly for the purpose of guidance to the content of the sources, to assist a reader to find their way to the sources that are pertinent to their interest.

Approach (1) is very open-ended and simply amounts to dropping the restriction altogether, subject of course to other policies.

Approach (2) is restricted, and replaces the outright ban with a more specific allowance of certain types of article. The specification of allowed articles could be made more restrictive or less restrictive: maybe some suggestions could be entertained? Brews ohare (talk) 15:44, 9 February 2012 (UTC)

Brews... Your proposals contain an internal contradiction. You say: "Subject to <WP:Notability> articles using only primary sources are allowed". The problem is WP:Notability says: "Sources, for notability purposes, should be secondary sources, as those provide the most objective evidence of notability. In other words WP:Notability requires at least one secondary source, and therefore indicates that articles using only primary sources would NOT be allowed. To restate your proposals, you are saying: "Subject to WP:Notability (which requires at least one secondary source) it is ok to write an article with no secondary sources". Contradiction! Blueboar (talk) 16:14, 9 February 2012 (UTC)

That's a good way of putting it. If an article is of the type subject to notability requirements, there is no way it can get by on primary sources alone within the article. If the article is of the type that notability does not apply (which presently is only includes disamb pages), then primary sources alone are fine, but usually these types of articles don't need to be sourced anyway since they're navigational in purpose. --MASEM (t) 16:24, 9 February 2012 (UTC)

Blueboar and Masem: Excuse any confusion on this point. As outlined at the outset of this RfC, it is proposed to change the wording regarding this matter in WP:V, WP:Primary, and WP:Notability (and any other places where this restriction pops up). As you point out, it makes no sense to change one policy and not the rest.

Masem: I gather that you believe WP:Notability does not apply to articles of the type described in Approach (2). However, WP:Notability's categorical refusal:

"If no reliable third-party sources can be found on a topic, then it should not have a separate article

does not explicitly exempt navigational articles. Perhaps changes in all the governing policies would suffice that specifically exempt navigational articles from a requirement for secondary sources? Brews ohare (talk) 16:32, 9 February 2012 (UTC)

Again, Secondary does not mean third-party. You need to find out what the term primary actually means before you embarrass yourself any further. Eliminating a requirement for secondary sources will not have the smallest effect on the separate and unrelated requirement that WP:Independent sources have taken notice of the subject. WhatamIdoing (talk) 18:49, 9 February 2012 (UTC)

WhatamIdoing: Of course, you are right that these two things are different. To restate matters more clearly:

Perhaps WP policies can be rewritten to allow navigational articles to be based entirely upon primary sources. This rule is an example of what I have called Approach (2) above.

Brews ohare (talk) 19:02, 9 February 2012 (UTC)

It may be best if you list out some examples what you think are "navigational articles" where this advice on primary sources would be intended, as I have a feeling there's a disconnect here. That may help to clarify what changes you want. --MASEM (t) 19:36, 9 February 2012 (UTC)

Masem: The trial balloon for a "definition" is:

A navigational article is one intended only to facilitate the navigation of primary sources pertaining to a topic. In a navigational article, the primary sources may be quoted or outlined strictly for the purpose of guidance to the content of the sources, to assist a reader to find their way to the sources that are pertinent to their interest. As with all articles, any interpretation, evaluation, or synthesis of views in the primary sources must be substantiated by secondary or tertiary sources.

How's that? Brews ohare (talk) 19:48, 9 February 2012 (UTC)

First, I think it would be helpful to distinguish between an "article" and a "page". The purpose of an article is to present information on a topic. The purpose of a navigation page (such as a dab) is not to present information - it's purpose is to aid readers navigate Wikipedia and find articles on the topics they are looking for.

That said - There is no need to rewrite our content policies to allow navigation pages to be based entirely upon primary sources... because navigation pages are not articles, and our content policies only apply to Articles. There is no reason to allow navigation pages to be cited to primary sources, because navigation pages are not required to cite any sources... Navigation pages are not subject to our content policies. . Articles require sources, but navigation pages are not articles.... they are simply aids to locate articles. Blueboar (talk) 19:50, 9 February 2012 (UTC)

Blueboar: We are running into semantics here. The above is to be viewed as a definition of a completely new category of article. Perhaps we need a new designation. How about this:

A familiarization article is one intended only to facilitate the acquaintance with primary sources pertaining to a topic. In a familiarization article, the primary sources may be quoted or outlined strictly for the purpose of guidance to the content of the sources, to assist a reader to become familiar with the sources that are pertinent to their interest. As with all articles, any interpretation, evaluation, or synthesis of views in the primary sources must be substantiated by secondary or tertiary sources.

The follow up on this is the statement in relevant policy pages that familiarization articles can be based entirely upon primary sources, but if any interpretation, evaluation, or synthesis of views in the primary sources is presented, it must be backed up by secondary or tertiary sources. Brews ohare (talk) 20:02, 9 February 2012 (UTC)

Suggestion: Brews, distinguished body of everyone here, what if someone could distill the best of what has been posted herein and use that to alter the language of the rule slightly? That wouldn't exactly constitute a major change in policy, and it would have the happy benefit of doing what you mentioned about articles: deftly applying a scalpel instead of an ICBM. My small opinion is that the discussion is getting a bit off the rails....--Djathinkimacowboy 07:40, 10 February 2012 (UTC)

The only thing that has come up that I can see is that disambiguation pages aren't subject to the normal rules which is something everybody assumes anyway. There is no secondary source talking about the notable topic of disambiguating Waterford in Ireland from Waterford in Australia or Canada or England or the United States or all the other uses. There is no reason to start sticking something in the OR policy about primary and secondary sources because of this. Is there some point about this discussion about primary sources as far as a straightforward ordinary article is concerned? I don't see it. Dmcq (talk) 10:39, 10 February 2012 (UTC)

Comment I can see a case for an article to be based on a primary source, but backed up by secondary sources. Such an article that is on my "to-do" list is the armoured train ambush during the Boer War in which Winston Churchill was captured. The de facto reference works regarding the ambush are Churchill's own writings (he was in South Africa in 1899 as a journalist). The notablility of the ambush only arose once Churchill became Prime Minister (40 years later). Subsequent writings have been an analysis of what Churchill wrote rather than a rewrite of the event. Martinvl (talk) 16:32, 10 February 2012 (UTC)

We wouldn't be interested in that unless a secondary source had commented on them and an analysis by a secondary source can certainly show what are the important points in his writing. Everything is based on primary sources at the end, that does not mean that articles are solely based on primary sources because secondary sources have pointed them out and commented on them. Writing a section in the article about something in his book that nobody has shown any interest in though would be definitely something to avoid according to this policy. Dmcq (talk) 22:26, 10 February 2012 (UTC)

Comment: Allowing to create (scientific) articles completely based on primary sources would destroy any scientific credibility of WP. We should rely on the established opinions as published in accepted textbooks or review articles in reputable journals, because this is the only way to establish WP:Notability and WP:Undue. Primary sources might probably be used in addition to secondary sources - but only the latter should be used as the foundation of an article --D.H (talk) 11:29, 12 February 2012 (UTC)

Comment: Regular articles will almost never be in such a situation, and even if they are, the addition of some secondary sources will resolve the issue, even if an editor or two will roll they eyes and ask "why bother?". The main reason to bother is, as others have pointed out here, notability (and related problems like conflic of interest and promotionalism of local hooey and made-up crap). Another major one is neutrality: The majority of primary sources on certain topics are overwhelmingly for or against particular ideas with regard to those topics. What could make sense is an exception in the case of bare lists/indexes of topics that are demonstrably notable in their own right, e.g. a "List of 2011 U.S. state attorneys general". Each party on that list is notable as one of the top politico-legal figures in the jurisdiction, and so either will have an article with reliable sources per WP:N and WP:BLP and WP:V or will not have an article yet (or a bad article headed for AfD/Prod/CSD, but that's a temporary matter). I.e., a simple list of contain only primary sources without this being a problem for Wikipedia reliability generally if secondary sources are used in more specific articles on the list members. Cross-reference: This idea is presently under discussion at WT:SAL#Citing sources. Such an exception, probably detailed at MOS:SAL, but at least codified in summary at WP:PRIMARY, would obviate any further consideration of throwing the baby out with the bathwater by dropping the general rule that articles must be based primarily on secondary sources, which is not going to happen.
Real, long-standing test case: (I just posted this over at WT:SAL, but its equally relevant here.) Absent a case of a notable list entry being controversial and someone demanding a source (or the case of including one being helpful in reducing editwarring), it would be far less tedious for editors and indeed readers to not include redundant piles of citations in list and index articles. I've basically been doing this quite explicitly, with <small>''(see [[List entry's article name here|main article]] for sources)''</small> for several years at Albinism in popular culture as an experiment, and it's rarely raised any concern at all. The multi-section list that article mostly consists of cites no sources itself, except for: a) entries that do not have their own articles; b) entries where editors have challenged or might be likely to challenge the entry's membership in the list; or c) the entry's own article doesn't touch on the subject of the list entry (e.g. because the particular portrayal by the actor wasn't career-important enough to be covered in the actor's own bio article). This has been stable for something like 4 years, and I suggest therefore that this is a good model.
— SMcCandlish Talk⇒〈°⌊°〉 Contribs. 03:40, 27 February 2012 (UTC)

A substantial proportion of our less-developed stand-alone lists, particularly lists that contain little than links to other pages, are completely unreferenced. Of course, the requirement is really that it be possible to name a secondary source, not that someone has actually typed the name of the source into the page, but I'm not even sure that's feasible for some of them. It's appropriate for us to have pages like "List of schools in ____", but I'm not convinced that you'll find a true secondary source for that subject. WhatamIdoing (talk) 03:52, 27 February 2012 (UTC)

"familiarization article"

Brews, I have a problem with the entire concept of a familiarization article. First, what you seem to be envisioning is a stand-alone annotated index of sources. I don't think an encyclopedia is the right place for that.

But, let us assume that I am in the minority, and others think it is a great idea... you would still need a secondary source to establish notability. Every stand alone article has a specific topic that is unique to that article. In a "familiarization article" as you envision, that topic would be: "Primary sources pertaining to X". OK... So what makes the topic of "Primary sources pertaining to X" notable enough for its own stand-alone article? You would need to establish notability... and to do that, you would need a secondary source that discusses the concept of "primary sources pertaining to topic X", or at best discusses these sources as a distinct group. Blueboar (talk) 14:59, 10 February 2012 (UTC)

Notability requirements could be changed for these articles, or dropped. For example, what is the notability of WP:Formal organization or any information page on WP like WP:Mathematical symbols? Brews ohare (talk) 15:15, 10 February 2012 (UTC)

Content and inclusion guidelines do not apply to any Wikipedia-space articles. These are non-examples. --MASEM (t) 16:05, 10 February 2012 (UTC)

The only reason WP:Formal organization is not part of the Main space article Wikipedia is that many editors invoked WP:Primary to deny its presence there. Brews ohare (talk) 16:18, 10 February 2012 (UTC)

I'm going to say it; that specific page is highly exceptional, and impossible to categorize in the overall theme of primary vs secondary sources, because it involves navel gazing, COI issues, and the like. Basically, any page in mainspace about Wikipedia is going to be tricky to write about, and why we generally fall back to secondary sources to assure that the article can be written (eg Reliability of Wikipedia.) We can't use that article to establish any precedent here. --MASEM (t) 16:30, 10 February 2012 (UTC)

Hi Masem: I am glad to see that such a thing as an exception is out there. It raises the specter that there may be other exceptions, eh? Brews ohare (talk) 16:43, 10 February 2012 (UTC)

It may be that "navel-gazing" pages are not that uncommon. For example, one could use the description given by any other web site, like Citizendium Knowino in the same way. Also, the description of any corporation's or charity's hierarchy. Brews ohare (talk) 16:52, 10 February 2012 (UTC)

Wikipedia having an article about Wikipedia or any other Wikimedia Foundation sites can be taken as navel gazing without proper secondary sources. Wikipedia having an article on any other website like Citizendium or the like is not navel gazing - they are not part of Wikipedia/Wikimedia. We still require secondary sources to talk about that website in an encyclopedic manner. --MASEM (t) 17:30, 10 February 2012 (UTC)

If I understand your view, if one wished to describe the organizational structure of Citizendium and could find no source describing that structure other than CZ's own documentation, then no article could be placed on WP about this structure, and in fact, any outline of this structure as part of another article would have to be removed if challenged because it relied entirely upon a primary source?

That seems extreme. Brews ohare (talk) 17:42, 10 February 2012 (UTC)

The same position would disallow a description of the governmental structure of Mozambique (presently unsourced entirely) or Chad (presently sourced to the BBC, seemingly less authoritative than Chad itself), or the organization of the presidential primary caucus in Nevada. Brews ohare (talk) 17:48, 10 February 2012 (UTC)

If, using your example, there was no source describing CZ's structure outside of CZ itself, then we would not allow an article on it. That doesn't mean that that structure cannot be talked about in context of the article about CZ, as your other examples do. It is not the case we don't allow primary sources, it is that we simply don't allow articles solely based on primary sources. --MASEM (t) 18:00, 10 February 2012 (UTC)

Masem: Surely that is silly? So if I want to write about CZ organization I can draft the article: "Organization of on-line encyclopedias" with a subsection "Citizendium" that is exactly what would be contained in the stand-alone "Organization of Citizendium" and that is fine? And likewise, in an article on Chad I can discuss its governmental structure sourcing the BBC News seven years ago, but I couldn't put the same thing in "Chad (Governmental structure)" sourcing documents from the government of Chad? Nonsense, I'd say. And something that should be changed. — Preceding unsigned comment added by Brews ohare (talk • contribs)

If you wanted to write about the CZ organization - which is only defined by CZ itself, you would go to the Citizendium (which is notable) and write about it in there; that's the more natural location for that. Chad's government, on the other hand, is covered by secondary sources and can be written about in a separate article. --MASEM (t) 18:47, 10 February 2012 (UTC)

Masem: So WP:Formal organization is more naturally located on Wikipedia? My thoughts exactly. We may be getting caught up in the examples instead of the principle. The principle (not the practice on WP) is that material that is suitable as a subsection of a larger article should not be disqualified from being split off as a separate article. If its support is OK for the larger article, it is equally OK for the stand-alone. If "Nevada caucuses" can exist in "Caucus" it can exist in "Nevada caucuses". Brews ohare (talk) 19:03, 10 February 2012 (UTC)

No, because the level of detail and information in WP:Formal organization is not aimed at a general reader, but for an editor of Wikipedia - it is necessary navel gazing within Wikipedia namespace to allow editors to understand how the work is organized. For the average reader, the organization in general is discussed currently within the Wikiepedia article; no further article is needed. And yes, actually, while material can be included in a sub-section of an article, splitting it off is not appropriate per WP:N and WP:SPINOUT. Only notable topics should be spun out to separate articles. --MASEM (t) 19:15, 10 February 2012 (UTC)

Masem: You are framing the situation differently. Splitting off may be inadvisable for various reasons. I agree that the intended reader of an overall article may have limited interest in a lot of detail that might be appropriate in a stand-alone article on a subtopic. WP:SPINOUT is pretty vague about the appropriate circumstances, and makes no reference to a difference in support requirements. Nothing there is germane to allowing material in a longer article to be sourced to a primary source, but refusing that same information to appear in a stand-alone article simply and solely because its support is a primary source. I understand that arrangement is present practice, but it makes no sense.

Yes, the stand-alone may be inadvisable for many reasons, but if the article falls into the classification of a "familiarization article", primarily relying upon primary sources is not one of them. You also raise the prospect that the stand-alone is not sufficiently notable to become a separate article. Here also, if the stand-alone is of "familiarization article" type, the notability issue should be decided based upon various criteria we have touched upon, and not entirely upon whether or not it relies upon a secondary source. Brews ohare (talk) 21:55, 10 February 2012 (UTC)

I imagine "Chad (Governmental structure)" can be referred to a secondary source. Assuming that were possible, the secondary source would say something like "Chad has a constitution. (See Constitution of Chad)", which could be equally authoritatively directly referenced as a primary source. (Actually, in this example, I don't think Constitution of Chad is a primary source, as that is probably in French.) What is aimed at is removal of the silliness of finding source S that simply quotes primary source P, when source P can be cited directly with equal authority. Brews ohare (talk) 21:55, 10 February 2012 (UTC)

First of all, we're still not talking about practical "familiarization" articles - these are not navigational in any shape or form. These are articles in mainspace, and need to meet content and inclusion guidelines. So be aware, you've still not identified what exceptions would be needed here.--MASEM (t) 22:57, 10 February 2012 (UTC)

Yes. Brews ohare (talk) 16:17, 11 February 2012 (UTC)

Now, let's compare the two cases "the organization of CZ" and "the government of Chad". Barring anything else, as long as we can source these via primary sources that meet WP:V, such could easily be covered in the respective parent articles (CZ and Chad - both are clearly notable topics). The question you're asking is why is there is a issue if these were separate articles.

If we talk about the "government of Chad", that is a clearly notable topic of its own. It is assuredly covered in textbooks within the country and by political science/global government textbooks published elsewhere, at a minimum (I don't have time to do a sources run, but I hope this is an accepted point). Within the article, however, the "best" resource is the primary source - the country's own constitution and laws that describe the organization of the government. So we have a topic that is notable where there do exist secondary sources, but readily sourced to the primary. In such a case, the lead for the article simply needs to establish the notability (and arguable, the notability is so obvious that most people won't flinch at the lack of secondary sources since they are assured it would exist), and the rest of the article could be on primary sources. In the context of this discussion, the resulting article may only include primary sources, but we know that secondary sources exist that talk about the topic "The government of Chad". Note - this is not the issue with the proposal, because this is not the case where only primary sources exist.--MASEM (t) 22:57, 10 February 2012 (UTC)

So, is it correct to summarize as follows? In a longer article, a primary source may be preferable to a secondary source for some matters (just which could be made very clear if desired) and a spin-off of this material to a stand-alone article is OK using the same primary sources provided (i) notability can be established somehow, or failing that, (ii) secondary sources are known to exist, even if they actually are not used. Brews ohare (talk) 16:17, 11 February 2012 (UTC)

Now, take the case of "the organization of CZ". We know CZ is a notable website, but is it's organizational structure notable? Certainly there's no way I'd take the word of anyone on that, like I would with the government structure of a major country. Let's presume that there are no secondary sources that talk about "the organization of CZ". (about "CZ" yes, but not the organization). This article would then have to be sourced to the primary work, the pages of CZ itself. This is the article that fails the policy as only primary sources exist for the topic; this article thus should not exist as a stand-alone article, but still can be discussed within the body of the "CZ" article.

If it were the case that there was significant coverage by secondary sources about CZ's organization, that makes it a different story: now we can include that secondary sourcing as the lead for the article (and here, unlike the Chad case, it would be expected to see that), and then have the rest of the article using primary sourcing. Mind you, the appropriateness of splitting this article off may be questioned: if both the CZ and the "organization of CZ" articles are short, then it is better to keep them as a single article; splitting off should only be done when the topic is notable and size issues start coming into place.--MASEM (t) 22:57, 10 February 2012 (UTC)

Here we discuss spin-off of material where notability cannot be established, or is debatable. For some reason, the criterion for notability here is taken to be the existence of secondary sources. The type of case considered is where only primary sources are available, and it is pointed out that such a spin-out is not presently admissible. There is no doubt that this is present policy.

What is not discussed at all is the subject of circumstances under which this policy should be relaxed. In particular, there are articles that can be sourced only through primary sources, where these primary sources are the best support imaginable for the topic at hand, and where a secondary source, even if it existed, would simply parrot the primary source. The only valid issue here is whether notability for the article can be established without citing a secondary source to establish the topic is worth talking about.

That is possible in some instances (you have provided a couple of them already). So I'd guess the subject of this conversation should include a discussion of how to establish notability in the absence of secondary sources, and possibly even the topic of just how a secondary source establishes notability when it is available. I'd guess the latter is a grossly misused mechanism in practice. Brews ohare (talk) 16:17, 11 February 2012 (UTC)

So to stress this point: the issue is between articles where the only sources in the world on the subject are primary, and articles where at the present time only primary sources are included but secondary sources are known to exist. The latter case does not fail OR policy/N guideline (though really those secondary sources should be added to prevent the issue), but the former one does and should be merged/redirected/deleted to a larger topic. Remember: once within the article of a larger topic, a smaller topic can be discussed to appropriate due depth from primary sources without question. --MASEM (t) 22:57, 10 February 2012 (UTC)

My view of this matter is that the imbedded section achieves notability in the larger article because it is part of a larger discussion or presentation and serves to illustrate some aspect of the main topic. As such, the notability of the section does not apply to the section itself when it is divorced from the larger topic. I believe this is a paraphrase of what you think yourself.

However, I'm not interested here in any kind of subject matter. The material I'm interested in is simple factual description, for instance, the structure of an organization, and not material involving evaluation, interpretation, or synthesis. For such material, primary sources are the most reliable support. The notability of such an article is not established by the significance of its views (it doesn't express any) and so does not require a secondary source to support its views. Its notability could be established by other means, for example, one might claim the structure of an organization was notable because the organization itself is notable. Perhaps, what is needed is a secondary source saying, for example, "IBM is a billion dollar company [Source: Barrons], with the following structure", or "Citizendium is an on-line encyclopedia designed differently than Wikipedia[Source: USA Today]. What follows is an outline of its structure", or "Rabies in dogs occurs [Veternary Science] and municipalities have regulation regarding this matter, as outlined below." These topic sentences establish context, perhaps, but not a thesis that is the article content. How would that fly? How would you adjust this approach? Brews ohare (talk) 16:17, 11 February 2012 (UTC)

You are in the right direction that you should establish the notability of the topic at the start of the article with secondary sources (or make sure that you have identified them on the talk page) to meet both WP:OR (that the topic has been covered and thus the rest is not undue) and WP:N (as general inclusion guidance); once established the rest of the article could be all primary sources though I think there may be some caution about going too far overboard on details from primary sources only, but that's not the issue at this point. But we you have to recognize is what topic that article now is about. If I broke out "Government structure of Chad" from "Chad", the topic I have to show notability is "Government structure of Chad". Similarly, breaking out "Organization of Citizendium" from "Citizendium", I have to show that "Organization of Citizendium" is a notable topic, even knowing that "Citizendium" itself is notable already. If you can't show that, you shouldn't break out the article , but discussion from primary sources is still fine in the original parent article. --MASEM (t) 16:36, 11 February 2012 (UTC)

Masem: I've raised the point as yet not addressed in detail of just how to establish notability. Of course, if you are trying to establish a thesis (an evaluation or interpretation or whatnot) a secondary source will provide support that the thesis is of sufficient note to have received some attention. But what are the other ways to establish notability? You suggested that something like "District attorney" was a notable topic because it was part of the court system or system of justice, and that is in itself notable. Would it suffice for an article like "State regulations regarding district attorneys" to simply have a secondary reference noting what the term "district attorney" means, and then proceed with description based upon primary sources? How could one formulate some suitable approach to deal with such things in general, apart from this particular example? Brews ohare (talk) 02:31, 12 February 2012 (UTC)

No, that would not be good enough. You're basically saying that any sub-topic of a larger topic can be given an article and talked about in depth even if the sub-topic is non-notable because the larger topic is notable. That's basically the idea of inherited notability, and that is not appropriate on WP. --MASEM (t) 12:53, 12 February 2012 (UTC)

No, the question I am raising is that of how to establish notability without invoking a secondary source. You noted earlier that you thought articles that could be classified a "familiarization articles" do exist, and are acceptable. Would you say they are exempt from notability requirements, or is their notability determined without the use of secondary sources? Brews ohare (talk) 13:59, 12 February 2012 (UTC)

There is no way to establish notability without significant coverage by secondary sources, otherwise we are putting undue weight and bias on primary sources. Now, yes, navigational articles like disambiguation pages or outline do not need to show notability. What that problem is is that you're using this term "familiarization article" and applying it to things that are definitely not navigational pages (eg the suggestion of a page about "State requirements for DAs"). As I've asked you before, if you provide examples (existing ones at best) that you think are "familiarization articles", we call tell you if they would have to qualify for notability or not. --MASEM (t) 14:08, 12 February 2012 (UTC)

Masem: So to state your position in my own words: There is no way to establish notability other than using a secondary source. However, some acceptable articles do not need to be notable, and the guidelines to identify these exceptions are sufficiently obscure or subtle that no formal policy can be written to identify them. Rather, in case an article is challenged as not exempt, the matter should be referred for adjudication to Masem, who will make the determination as to whether an exception is warranted. Brews ohare (talk) 15:00, 12 February 2012 (UTC)

This is well-established practice, and takes no guesswork as to what is exempted. Navigational articles like disamb pages or outlines are exempt from generally all other content and inclusion guidelines for articles, because they are strictly that, navigation aids. Any articles outside that are assumed to be normal encyclopedic articles and thus must mean all content and inclusion guidelines need to be met. The problem is that you see to want to push normal content articles as a "familiarization" article to make them exempt from content and inclusion guidelines, which is against current policy. --MASEM (t) 15:06, 12 February 2012 (UTC)

No, that is not my intention. I guess the closest that present guidelines come to this matter is Notability: Lists. That guideline section already admits to confusion on this matter. It makes the recommendation: "Editors are still urged to demonstrate list notability via the grouping itself before creating stand-alone lists"; which is completely unintelligible to me.

As you may recollect, I am not focused here on "normal content" but upon content that is purely descriptive, contains no evaluation, interpretation, or opinion, and is based primarily upon primary sources. Of course, a navigation page like Outline of chemistry is in this category, and kind of resembles a knol collection. That is, a guidance page governing material internal to WP, which is useful to a reader searching for what there is on WP about the topic. To a degree it is a very restricted pre-programmed Google search like this. So I raise this question: why wouldn't one assume that such pages are useful also for outside information sources? Brews ohare (talk) 15:31, 12 February 2012 (UTC)

Because basically, even if the information is fully sourceable to primary and thus verified, if no one else has taken note of it, it is undue weight for us to cover it in depth on Wikipedia. We cannot create an importance for a topic if there is no importance given to it in sources. --MASEM (t) 16:51, 12 February 2012 (UTC)

Another possible candidate for a "familiarization" article is United States Attorney. It does have a few secondary sources, but it appears to be debatable whether any of these establish notability, and it raises the issue of whether it is primarily based upon secondary sources: it doesn't look that way. Yet, I'd say the article has a purpose and should be on WP. Brews ohare (talk) 15:59, 10 February 2012 (UTC)

That's not what most people would consider a navigational article. "United States Attorney" is a notable topic since they're a critical part of the national court system, and clearly backed by secondary sources that establish its notability. --MASEM (t) 16:05, 10 February 2012 (UTC)

The criterion "they're a critical part of the national court system" as a basis for notability is an example of a basis different than requiring a secondary source. It is the major reason for considering this article notable, and would suffice by itself. Most of the secondary sources in this article refer to some particular minor point. So this article seems to fit the bill as an article denied existence by WP:Notability WP:V and WP:Primary that should be on WP nonetheless. It is not primarily based upon secondary sources.

Maybe you don't like the name "familiarization article" (I don't either) but "a rose by any other name..." Brews ohare (talk) 16:15, 10 February 2012 (UTC)

It's not a navigation, familiarization or whatever other name you want to give to call these articles "special". Unites States Attorney is a normal, everyday mainspace article that shows its notability through secondary sources. Some sources are primary, which are fine, but the topic is notable to secondary ones. So this is still a bad example to start from. --MASEM (t) 16:30, 10 February 2012 (UTC)

Masem: It is suitable in the sense of being an article whose "notability" is independent of the presence of secondary sources. That shows that notability can be established in some cases without secondary sources, perhaps on a case-by-case basis?

Are you suggesting that this "bad example" indicates that no category like "familiarization article" is needed? Brews ohare (talk) 16:34, 10 February 2012 (UTC)

No, I'm not saying there's no such thing as a "familiarization article". Just that United States Attorney is not the type of article that would be considered as such. It is a notable topic (you have to consider what full range of sources exist, not just those that are presently listed in the article), and not used for navigation. Just so you know what I consider as such, Outline of chemistry is one example of a navigational, or familiarizational article. --MASEM (t) 16:48, 10 February 2012 (UTC)

There appears to be many articles in the area of officialdom that need some attention: District attorney, Prosecutor, Crown attorney... and may end up being familiarization pages. Brews ohare (talk) 16:15, 10 February 2012 (UTC)

I suppose they could end up the way you envision... the question is: should they end up that way? I would say No, they should not. And if they did end up as you envision, I strongly suspect people would object and re-write them back into normal articles (hopefully citing reliable secondary sources).

I think you need to answer a further question here... WHY would we want to have "familiarization articles"... what would be their purpose? Blueboar (talk) 17:53, 10 February 2012 (UTC)

Of course, an article evolves over time, and one might hope that an article based upon primary sources might become a regular article, for example, if a book happens to be written one day that covers the topic. In the meantime, the article can sit in a different category. At the moment, if any challenge is made to an article based primarily (whatever that is taken to mean) upon primary sources, it must be removed. That means topics like the structure of organizations that have no secondary coverage cannot be written. Of course, there are such articles on WP that escaped challenge, but if there is any opposition (such as arose in trying to include WP:Formal organization in the article Wikipedia) the article is denied to Main space. Brews ohare (talk) 18:07, 10 February 2012 (UTC)

It is clear that good reasons for rejecting an article should not be ignored, but rejection of a "familiarization article", defined as defined above, would not happen simply and solely because its sources are primary. Brews ohare (talk) 18:10, 10 February 2012 (UTC)

Brews, the place to discuss the "structure of an organization" is within the article on the organization itself. The structure of the organization is very unlikely to be considered a notable topic on its own (and in the rare cases where it might be, due to some unique aspect of that structure, we would determine its notabiliy by having secondary sources discuss that structure).

This actually gets us back on track as far as the OR policy goes... I would absolutely agree that if we were writing a section on the structure of an organization (in larger article on the organization itself), it would be fine to base that section primarily on an a Primary source... after all, the most reliable (ie best) source possible for the information is highly likely to be a primary internal document from that organization. We would, of course, have to be very careful not to slip into analysis or draw any conclusions... but as long as what we are writing is purely descriptive, a primary source is fine. But... we are talking about a section. The article on the organization, as a whole, would still need to be mostly based on secondary sources. Blueboar (talk) 23:36, 10 February 2012 (UTC)

Blueboar: Do you think the present policy as written is satisfactory regarding OKing primary sources for sections of articles? I am doubtful. Brews ohare (talk) 02:18, 12 February 2012 (UTC)

I think the present policy is OK as far as permitting the use of primary sources (it explicitly says you are allowed to use primary sources, but cautions that there are limitations. It does a good job of explaining when it is not appropriate to cite primary sources (statements of analysis, interpretation, conclusions, etc.), but I have long thought it could do a better job of explaining: 1) when it is appropriate to cite primary sources, and 2) how to do so in those situations. Blueboar (talk) 02:33, 12 February 2012 (UTC)

Blueboar: What is your take on the list of articles on WP provided here? How is notability established for these separate articles? It seems probable that the way to achieve notability is largely established by forming a "project" like WikiProject Alaska that includes the document. Unfortunately U.S. Attorney for the Eastern District of Kentucky hasn't evolved such a group as yet. Brews ohare (talk) 15:51, 12 February 2012 (UTC)

I opened a thread on Talk:Notability where you may prefer to take up this matter. Brews ohare (talk) 16:57, 12 February 2012 (UTC)

Can someone break and possibly summarize this list and what currently supported options are? I'm having trouble drawing a conclusion from this long debate.Gsonnenf (talk) 23:08, 19 February 2012 (UTC)

Based on what ground the statement "XXX is a term" can be made in WP articles?

From time to time, I encounter the WP articles that start with the statements of that type:

"XXX YYY is a term used for XXX by YYY." (The most recent example can be found there: Soviet occupations is a term used for military occupations by the Soviet Union...).

As a rule, a source used for such a statement contains just a combination of XXX (as a subject) and YYY (as a predicate), and contains no definition of such new term. For instance, the sources 1-3 from the above example contained no definition of "Soviet occupation" as some separate term used for for description of some separate phenomenon. They just tell about occupation by the Soviet Union. In my opinion, taking into account that I have been a witness of numerous attempts to create new "terms" using the same approach, it would be useful to describe some of them as an example of WP:OR.--Paul Siebert (talk) 02:54, 9 February 2012 (UTC)

Hmmm... this could be OR. However, I think it is an OR that could probably be fixed by adding attribution. Change the sentence to: Soviet occupations is a term used by historians to describe various military occupations by the Soviet Union..." (followed by citations showing historians using the term to describe those occupations) and it would not be OR. Blueboar (talk) 03:32, 9 February 2012 (UTC)

In the case at hand, and noting ArbCom decisions and dicta thereon, the lede is proper as is. A plenitude of sources are given in the articles, and thus nitpicking over grammar is non-utile. I suggest you read the ArbCom discussions on all of this "stuff." Collect (talk) 14:07, 9 February 2012 (UTC)

I am not sure I understand what concrete ArbCom decision do you mean (maybe you provide us with the link?), however, none of the sources given in the article define the term "Soviet occupation", which is supposed to be a specific type of occupation to deserve a separate term.--Paul Siebert (talk) 15:35, 9 February 2012 (UTC)

I have come across people making up terms and they are obviously OR but others are not. In the case of soviet occupation it just sounds like clumsy language and unnecessary. We are allowed to use descriptive names for article titles if there is a clear topic and no agreed name or term. Dmcq (talk) 14:19, 9 February 2012 (UTC)

We are definitely allowed to use descriptive names, however, we are not allowed to pretend they represent some separate terms. In connection to that, I propose to describe the difference between descriptive names and the terms. Let me explain that using the following (totally imaginary) example. Consider two statements:

"Medieval negativism is a term used by historian to describe the mainstream philosophical discourse in Middle Age Europe.¹"
"During Middle Ages, the European philosophical thought was dominated by negativism.¹"

(Again, the example is totally artificial, so, please, ignore the factual side of the sentences #1-2). Depending on what the source 1 says, the statement 1 may be allowed or not. If the source contains an explicit definition of "Medieval negativism" as a term, the statement 1 is acceptable. However, if it just contains "medieval" (a subject) and "negativism" (a predicate), and does not define the term, the statement 1 fails WP:NOR. Since I have been a witness of several attempts to introduce such new "terms", the above situation is not unique, so we probably need to explain the difference between the terms and the descriptive names either on the policy or guidelines page.--Paul Siebert (talk) 15:35, 9 February 2012 (UTC)

Paul, your basic interpretation of how this policy works is accurate... the determination as to whether a specific statement is OR or not can indeed depend on the exact language used in the cited sources. If historians actually use a specific term to describe something, then it is not OR to say so. If they don't actually use the term, then it would be OR to say that they do.

However, I don't think we need to change the policy to resolve your concern. The policy already requires that sources "directly support" what is said in an article. There is no need to explain this provision in terms of every instance where it might apply. Doing so is what we call "instruction creep". Blueboar (talk) 17:00, 9 February 2012 (UTC)

That said... I have looked through both the edit history of the Soviet occupation article and your personal edit history... I suspect that your real concern here is WP:NPOV rather than WP:OR If so, I suggest you address that issue directly. As it is, there is an appearance (which may or may not be accurate) that you are trying to wikilawyer about OR in order to "win" a NPOV dispute. If I am misreading things, I apologize... but that is how it appears to an outsider. Blueboar (talk) 18:12, 9 February 2012 (UTC)

By writing that my primary concern is NPOV you were right. By writing that I am trying to wikilawyering to win some dispute you were wrong. I simply found some typical example of creation of new "terms", and I think we need to clarify the policy to make such violations of NOR impossible.--Paul Siebert (talk) 20:18, 9 February 2012 (UTC)

With few exceptions, "XXX is a term" is usually just an example of weak writing style, not a policy violation.

Importantly, the lead is supposed to summarize the article, not provide dictionary definitions. So the best approach in such situations (using Blueboar's example) is usually something like

Medieval negativism was the mainstream philosophical discourse in Middle Age Europe, which was characterized by... WhatamIdoing (talk) 18:53, 9 February 2012 (UTC)

Correct. That is definitely better. However, my argument is that someone may argue that the difference between ##1 and 2 (or your test) is just stylistic one, whereas the difference is more profound: the # 1 is a violation of NOR, and that should be explained clearly.--Paul Siebert (talk) 20:18, 9 February 2012 (UTC)

Comment: My two cents!- OR is a rule generally employed very skillfully. The object here seems simple: if a citation can be found that says 'X is known for' or 'X was defined by' (which gives us a 'term' applying to 'X'), I don't see that as OR. The Medieval example is simply too abstract and niggling for the general purpose of the query. I.e., my understanding is that Wikipedia is not a dictionary. When you speak of 'terms', aren't we trespassing on the dictionary's field? So, what I say is if a citation is found that is clear enough and not a matter of definitions or semantics that lay outside Wikipedia's scope, it is acceptable; otherwise it is OR.--Djathinkimacowboy 07:50, 10 February 2012 (UTC)

I do not propose to consider a situation when such a citation is found. I focus on the opposite case: when editors, based on handful books that use some subject and predicate together, create a new "term". In my opinion, that should be explicitly prohibited: no sourced definition - no term.--Paul Siebert (talk) 15:08, 10 February 2012 (UTC)

What this seems to come down to is: While we are allowed to invent our own phrasings and terminology when creating a descriptive title (with the caveat that if there are existing, commonly used phrasings and terminology for the topic, we should use those instead)... it is not always appropriate to describe our invented phrasings and terminology as being "a term" in the text of the article. ie - We should not use the word "term" unless a source uses the exact phrasing/terminology as a term.

I would say this is a very narrow and nit-picking interpretation of WP:OR (bordering on being Wikilawyerish)... but I have to agree that it is OR. We require that sources directly support the text. In the situations Paul is describing they do not (they indirectly support the text). The solution is not to prohibit the phrasing/terminology... the solution is to re-write the opening sentence to avoid using the word "term". Blueboar (talk) 15:40, 10 February 2012 (UTC)

I agree it seems to be more of a case of clumsy language when some editors attempt to apply MOS:BOLDTITLE, and the solution is as you suggest and easily applied[2]. --Nug (talk) 15:49, 10 February 2012 (UTC)

In re-reading the proposition as originally stated by Paul Siebert, I see the problem as a simple one: no editor can or should infer/invent a new term solely based on two terms seen used together in a citation. In other words, as another wise commentator said, if it isn't verifiable in a cited source, it is OR. I recall being whipped nearly to death once when I tried to describe something. I used the term 'ironically' and got the lash for it. I hope that comparison example is clear. Let the reader infer whatever the cited material tells, and no more.--Djathinkimacowboy 01:40, 16 February 2012 (UTC)

My handy rule for history articles:
- Where the scholarly discourse is dominated such that only a scholarly FRINGE disagree, then the term is the thing. "World war II was a conflict between the Allies and the Axis."
- Where the scholarly discourse is internally divided, it is a term, "The Great Patriotic War is the term used in Soviet historiography to describe the front in World war II between Germany and its European allies and the Soviet Union."
- Where a scholarly discourse is a minority position, it is a term, "Post-fordism is one of a number of terms to describe the structure of employment and the economy in the advanced West after 1970."
- Where a scholarly discourse is FRINGEy, it is a term, "Within Novocastrian provincialist gutter murmurings, "Newcastle Hardcore" is that particular style of Hardcore techno developed in Newcastle, NSW between 1990 and the mid 2000s."
- Where it ain't scholarly and it is FRINGE, it shouldn't be used.
Of course, I strongly suspect (but feel insufficiently read to claim) that "Holodomor" ought to discuss Ukrainian remembrance, whereas the Ukrainian Famine article should be about the object. Consider regarding your own views, and with salt. I think my schema is internally consistent, and reflects the WEIGHT policy. Fifelfoo (talk) 03:09, 16 February 2012 (UTC)

Clarification of what constitutes a primary or secondary source

The description of primary and secondary sources is not as clear as it might be on this policy page. At the moment, the policy says under Secondary sources:

"For example, a review article that analyzes research papers in a field is a secondary source for the research."

That is very clear. What is less clear is whether this statement implies that the research papers themselves are primary sources. A footnote is intended to clarify this matter:

"The Ithaca College Library compares research articles to review articles. Be aware that either type of article can be both a primary and secondary source, although research articles tend to be more useful as primary sources and review articles as secondary sources."

I find that this footnote serves only to muddy the waters. It can only lead to debate over whether a particular journal article falls under multiple categories and which category applies in what instance. I'd suggest it be revised to say the following:

"Research articles are considered primary sources and review articles as secondary sources."

That is crystal clear, and does not leave the door ajar.

The reason journal articles presently are allowed as secondary sources is most probably related to journals being peer reviewed. That review process may be adequate for some technical journals like Physical Review, but even there articles may not reflect a broad consensus of the technical community. In fact, they report original research that by definition may not be conventional in its approach. Unless WP wishes to go into the unlikely business of ranking journals, the simplest policy is to regard them all as primary sources.

Another footnote in WP:NOR is:

"Duke University Libraries offers this definition: "A primary source is a first-hand account of an event. Primary sources may include newspaper articles, letters, diaries, interviews, laws, reports of government commissions, and many other types of documents."

Placing this information in a footnote in this manner suggests that this statement is to inform judgment, and is not an integral part of the WP policy on this matter. In the interest of clarity and to ensure objectivity of article content, I'd suggest that wording of this sort be included explicitly in WP:NOR, for instance, with the statement:

"Newspaper articles, letters, diaries, interviews, laws, reports of government commissions, journal articles, and many other types of documents are considered by WP to be primary sources."

Brews ohare (talk) 18:49, 15 February 2012 (UTC)

Comments on what constitutes a primary or secondary source

Comment: The situation really is muddy. If a policy or guideline requires a precise definition of primary and secondary sources in order to be interpreted, the policy or guideline does not reflect the real world and needs to be changed. Jc3s5h (talk) 18:52, 15 February 2012 (UTC)

Jc3s5h: If I understand your comment, if the real world is confusing, a guideline must be confusing too. If it isn't as muddy as the real world, the guideline must be made murkier. Have I got your meaning? Brews ohare (talk) 18:57, 15 February 2012 (UTC)

No. Since the distinction between primary source and secondary source is murky, any policy based on this distinction will be murky too. The alternative is to base the policy on something else; something not so murky. This may not be possible, but then again a way might be found. Jc3s5h (talk) 19:03, 15 February 2012 (UTC)

Comment: It seems to me that this probably cannot be resolved in a really satisfying manner, as use of the terms primary and secondary sources depends on the concrete context and field. The keyword in list of publication forms in the Duke definitions is "may". I see no way to classify news paper articles, journal articles or commission reports principally as primary sources, that would be nonsensical. For example a commission report that simply collects facts/information is certainly a primary source, however a commission report by scholars that assesses and reviews facts,information or primary sources in general is no different from a a book or journal article reviewing primary sources. In the context of history subjects you may consider all those listed as principally (historic) primary sources, but that works only in this context (due to the time difference to current publications on the subject). If you move to other fields that approach fails. Also I'd like to note that we should avoid getting hung up on (possibly unachievable) formal distinction between primary, secondary and tertiary. It is much more important that editors focus on the quality of the content of a source and its reliability and reputation, that is what matters primarily and not the formal distinction between primary, secondary and tertiary.--Kmhkmh (talk) 19:16, 15 February 2012 (UTC)

It is certainly true that there are gray areas. You suggest that some government reports are reviews, and so are some journal articles. I don't see much problem in identifying these as secondary sources. On the other hand, many journal papers begin with a summary of prior work, that is in a sense a review. I believe excluding such introductory sections from consideration as a secondary source is warranted, especially as such introductions are generally intended to cast the subject in a light to reflect favorably upon the original research that is the purpose of the journal article.

Can you elaborate upon your reservations? Brews ohare (talk) 19:26, 15 February 2012 (UTC)

As Kmhkmh notes, when they're older than about 10–50 years, newspaper articles become primary for history. In Medical Science, research articles are primary, whereas research reviews are secondary. A number of commission reports are primary, where the commission is an involved party, consider the report of the Winter Soldier Investigation. The provision of a brief literature review in historical journal articles is considered to be a secondary source and is privileged especially if the whole work is a literature review for weighting and structure in history articles. Context is all here. Fifelfoo (talk) 22:47, 15 February 2012 (UTC)

Comment: Oppose. The proposed edit " Newspaper articles, letters, diaries, interviews, laws, reports of government commissions, journal articles, and many other types of documents are considered by WP to be primary sources." is a blatant oversimplification. Below is my analysis:

"Newspaper articles." All newspaper articles? Whereas some news reports, where the facts have been presented without any interpretation are, probably, the primary sources, the analytic articles are definitely not primary.
"Letters." Whose letters to whom and about what?
"Interviews" What about an interview of some WWII historian about his work on the history of the attack of Pearl Harbor? Is it a primary source?
"Laws" Yes. It is a primary source.
"reports of government commissions" Sometimes, such reports contain a comprehensive analysis of some event (e.g., of some catastrophe), its causes, consequences, etc. That fully meets a definition of a secondary source ("are second-hand accounts, generally at least one step removed from an event").
"journal articles" In exact sciences, journal articles are divided onto research articles and reviews. The latter are definitely secondary sources. The former contain the author's own data, their discussion, and more or less comprehensive overview of what have been done by others in this field. Therefore, depending on the context, research articles are either primary or secondary sources. With regard to history articles, each of them contain a huge amount of references to numerous archival documents, letters, research articles, books and monographs, which definitely makes it a secondary source.
"and many other types of documents" That addition makes the whole sentence useless. Which concrete other documents?

And finally, let me remind you that the PSTS division exists not for the sake of itself, but as a part of our NOR policy. We cannot interpret raw data, historical documents, laws, etc by ourselves, therefore, the sources that contain no such interpretation should be used with great cautions in Wikipedia. However, the sources where such information have been interpreted by professionals (research/hisotry articles, reviews, letters of one scientist to another, philosopher's diary, etc) is safe to use.--Paul Siebert (talk) 19:28, 15 February 2012 (UTC)

Paul Siebert: It appears that you have your particular definition of primary sources, namely "raw data, historical documents, laws, etc." Perhaps you could express these ideas in a form that would clarify WP:NOR? Brews ohare (talk) 19:35, 15 February 2012 (UTC)

No. I follow the definition of a secondary source as "second-hand accounts, generally at least one step removed from an event". The history article dealing with historical documents (primary sources) is definitely at least one step removed from the event it describes. The interview with a present-days physicist about the Manhattan project is also at least one step removed from the event, and so on. I do not propose my own comprehensive definition of primary sources, I just outline those sources that are always primary. Other sources may be primary or secondary, depending on a context.

And, again, my major point is that we always have to keep in mind that the real reason for the PSTS division is to avoid original research, which mean to avoid making amateurish interpretations of raw data. If some source contains an opinion of some professional, it is safe to use it in Wikipedia, and that is a major point of the NOR policy, in my opinion. Everything else should be built on based that.--Paul Siebert (talk) 19:50, 15 February 2012 (UTC)

Paul, I'll take the opportunity to look at your specific points above:

1 "Newspaper articles." All newspaper articles? Whereas some news reports, where the facts have been presented without any interpretation are, probably, the primary sources, the analytic articles are definitely not primary.

I personally would not accept an interpretive newspaper article as a secondary source. These articles are labeled as opinion pieces and that is what they are. Their only value as a source is as a primary source regarding what the analyst thinks.

Actually even that is imho somewhat fuzzy since, we need to distinguish between "interpretation" as a scholarly (well researched) assessment and interpretation as a mere (political) journalistic opinion piece.--Kmhkmh (talk) 22:35, 15 February 2012 (UTC)

Additionally, the assertion that all interpretive articles are "labeled as opinion pieces" is simply false. If a periodical runs a long series analyzing an issue, e.g., comparing current proposals for tax reform, that's not an opinion piece. WhatamIdoing (talk) 05:06, 23 February 2012 (UTC)

2 "Letters." Whose letters to whom and about what?

Any letter is a personal observation, or a hearsay version of other peoples', and is not a secondary source.

3 "Interviews" What about an interview of some WWII historian about his work on the history of the attack of Pearl Harbor? Is it a primary source?

Yes, it is a primary source regarding his views of his work. On the other hand, the historian's work itself is treated on its own merits as to whether it is primary or secondary in nature.

4 "Laws" Yes. It is a primary source.

5 "reports of government commissions" Sometimes, such reports contain a comprehensive analysis of some event (e.g., of some catastrophe), its causes, consequences, etc. That fully meets a definition of a secondary source ("are second-hand accounts, generally at least one step removed from an event").

Yes, and where these reports are secondary sources it is quite easy to establish. For instance, if they assess the efficacy of government agencies, they are secondary sources, although possibly not reliable.

6 "journal articles" In exact sciences, journal articles are divided onto research articles and reviews. The latter are definitely secondary sources. The former contain the author's own data, their discussion, and more or less comprehensive overview of what have been done by others in this field. Therefore, depending on the context, research articles are either primary or secondary sources. With regard to history articles, each of them contain a huge amount of references to numerous archival documents, letters, research articles, books and monographs, which definitely makes it a secondary source.

Overviews in reports of original research may be accurate, or they may cast prior work to showcase the following original work. As such, these overviews should be viewed as primary sources about what the authors think. In eliminating all but review articles from consideration as a secondary source, not much is lost.

7 "and many other types of documents" That addition makes the whole sentence useless. Which concrete other documents?

Good point; this language was inadvisably carried over from the original source.

Brews ohare (talk) 20:02, 15 February 2012 (UTC)

1. May I ask you, what is the reason for division of the sources onto primary and secondary?

2. As a rule, the reviews and research articles are being written by the same authors (I myself have authored both research articles and reviews in my real life). In connection to that, consider the following situation (i): an author A reviewed the works authored by X, Y, and Z, and published that in the review article; (ii) the same author reviewed the same works and published the same material in the introduction to his own research article. What makes the former source secondary for X, Y, and Z, and why the same text in the latter source is primary?

3. Sometimes, the monographs are the collections (with small additions) of the research articles written previously by the same author. What makes the book (monograph) containing several articles (e.g. as chapters) a secondary source, and why the same research articles published in some scholarly journal separately are primary sources?--Paul Siebert (talk) 20:30, 15 February 2012 (UTC)

Hi Paul: I confess that this topic is more subtle than I had first thought. From a WP stance, the role of a primary source is to support some statement of fact, while a secondary source has a more involved role. The secondary source can be used to establish WP:Notability and a primary source cannot. Personally, I find that nuts.

It may be that on WP there is no other point in separating the two types of source, and it is just nonsense. The only point of importance is that contributors to WP should not intrude their own opinions, and that can be established with any type of source.

So our conversation here is leading me to the view that all discussion of primary and secondary types of sources should be eliminated from WP and WP:Notability should be rewritten entirely to establish different means to establish whether an article is worth appearing on WP. Maybe any topic for an article should be allowed, subject to deletion if they have too low a traffic count? Brews ohare (talk) 20:54, 15 February 2012 (UTC)

No. This is silly and needs to stop. Its tendency is towards disruption. Fifelfoo (talk) 21:23, 15 February 2012 (UTC)

Comment:

I note that this discussion in this section is a few steps removed from the previous section, after Brews ohare has tried to broach similar issues at other boards. --MASEM (t) 21:03, 15 February 2012 (UTC)

Comment: Agree with the above posts by both Masem and Fifelfoo.--Djathinkimacowboy 01:31, 16 February 2012 (UTC)

"Maybe any topic for an article should be allowed, subject to deletion if they have too low a traffic count?". No, maybe not. If your real aim here is to eliminate WP:N and replace it with WP:INTERESTING then please open an RfC to this effect and see it snow-closed as Oppose. Fences&Windows 21:33, 22 February 2012 (UTC)

How to say "no" nicely?

I've been thinking about the issue of Timothy Messer-Kruse and his complaint about how WP editors treated him when he tried to add stuff to the article on the Haymarket affair. Wikipedia:Wikipedia_Signpost/2012-02-13/In_the_news

I think we do have a problem, not with the substance of the OR policy, but with the way we present it. We need to be aware that not all researchers are cranks, and no-one is born with a knowledge of Wikipedia's policies. From time to time, we will get real experts wanting to put new stuff into WP articles. The issue I'm raising is not whether we say no – I agree that we have to say no – it is: How do we say no without appearing to be contemptuous of the expert?

Why not include a paragraph like this...

"The reason for the OR policy is not that WP lacks respect for specialist researchers. Rather it is because of our own limitations: we are a community of non-specialist volunteer editors, but validation of new research work is a task for communities of specialists. That's why WP cannot be the first publisher of a new discovery or analysis; but as soon as significant new findings have been published in an RS, WP can and should tell its readers about them." Kalidasa 777 (talk) 08:59, 1 March 2012 (UTC)

Some suggestions:

1) Suggest a better web site for them to post their research.

2) Include a link to their version of the page on their user talk page. If this isn't possible, make a copy for them in their user space. This will allow them to copy it elsewhere. We can't assume they know how to dig through the history to find it.

3) And, of course, just be nice. Don't act as if they wasted your time and corrupted Wikipedia, but like they made a valuable contribution, but put it in the wrong place, accidentally. StuRat (talk) 09:32, 1 March 2012 (UTC)

It is slightly more complex than just publishing "facts" in RS. It is also a matter of the WEIGHTing of the community of specialists. Messer-Kruse's own work is (at this point in time) not considered by the community of specialists in relation to the broader historiography (given the structure of history, we should wait on scholarly book reviews to tell us how to weight Messer-Kruse's works). Similarly in the medical field we don't rely on RS, as in the publication of new case studies, but on review articles that conclude the results of multiple case studies. We care so much about specialists' contributions that we don't evaluate their importance ourselves, that we only trust third specialists about the significance of a specialist's findings. Fifelfoo (talk) 09:55, 1 March 2012 (UTC)

I edit in a couple areas where I am the RW source cited by other Wikipedians. There is a way to not only for these folks to edit, but to also utilize their expertise in ways that is consistent with Wikipedia policies, albeit not always with how they are sometimes mis-quoted. Don't forget that experts are also usually experts at knowing about a wide range of sources, what is in them, and their context in the subject at hand. Starting with the title of this post, the "no" is just to certain things, not to editing. We should make a little essay somewhere (for linking to etc.) which welcomes them, explains the issues and how to navigate it. North8000 (talk) 10:46, 1 March 2012 (UTC)

This is the start of such an essay, WP:NOR/Workshop. NewbyG ( talk) 10:51, 1 March 2012 (UTC)

I'd be happy to help/work on that. It will take me a day to get to it. North8000 (talk) 14:29, 1 March 2012 (UTC)

Experts can be directed to WP:EXPERT and we probably should have a some links to that from the OR policy and a couple of other places so experts coming to Wikipedia can find it easily. Dmcq (talk) 14:07, 1 March 2012 (UTC)

Thanks, Dmcq – you've saved us from doing something that to a large extent is already done! Yes, WP:EXPERT is a very good page. I think I read it a few years back, but had forgotten about it. I agree we should link to it more -- good stuff is of limited use if people can't find it! North and Newby, please read this one too! Does it already cover all the points you wanted in an essay, or can it be made even better? Another idea... maybe there should be a short welcoming template, that can be put on the user page of a newby who seems to have specialist knowledge, and that template could link to the WP:EXPERT page. Or has that already been done too?? Kalidasa 777 (talk) 19:04, 1 March 2012 (UTC)

I had something completely different in mind, but referencing WP:EXPERT would allow it to be significantly shorter. North8000 (talk) 19:44, 1 March 2012 (UTC)

I started/drafted something there. WP:NOR/Workshop I'd be happy to work on it more if there is an interest / it gains some traction.If so, drop me a note there (I'll watch it) or ping me. North8000 (talk) 11:54, 2 March 2012 (UTC)

Date and time calculations

Should we state here that it is allowable to convert a relative date from a reliable source (such as a newspaper) to an absolute date. I.e. "Last Tuesday" becomes March 20, 2012, when it is used in a story printed on March 25, 2012. Hcobb (talk) 16:49, 1 March 2012 (UTC)

This question has just been asked by another person at the end of WP:NORN and the answer is that in normal straightforward cases this is okay by WP:CALC. If somebody gives a good reason though why the date might be wrong you'll have to discuss that with them on the talk page. Just quoting WP:OR is not a reason to stop this though. Dmcq (talk) 16:56, 1 March 2012 (UTC)

Illustrations as original research / synthesis

Can someone familiar with WPOR comment at Wikipedia:Files_for_deletion/2012_February_28#File:Epsilon_Tauri_b_and_Hyades.png on the acceptability of original research in the form of illustrations? 70.24.251.71 (talk) 04:59, 1 March 2012 (UTC)

Can someone familiar with WPSYN comment at WT:ASTRO#Illustration of the Tau Geminorum brown dwarf companion on regards with improving the image and possible synthesis due to improvement? 70.24.251.71 (talk) 06:07, 9 March 2012 (UTC)

original research on images on Commons

Is there a way to tag images on Commons so that they are not used on Wikipedia articles? Since WP:OI doesn't allow such images, it'd be good to indicate such so that editors don't choose those images to embed into wikipedia articles, or perhaps to get a bot to autoremove such images. 70.24.251.71 (talk) 05:16, 4 March 2012 (UTC)

No there isn't. The idea of "original research" is unique to en.wiki (And possibly the other language wikis) but not to commons, and thus you'd be asking to set a wiki-specific tag for that, which is not going to happen there. If you see an image that is on commons but you believe is introducing original research on an article, it's usually a good idea to gain consensus before removal; the interpretation of what is original research can vary from person to person so consensus is highly recommended. --MASEM (t) 06:11, 4 March 2012 (UTC)

So, if a determination of OR is made at say NORN, do we tag such images on en.wiki (en.wiki file page for an image on commons) to indicate they should not be used on en.wiki for article content? (if we don't have such a tag, then create such a tag for this case)

As some images built from datasets don't indicate the data it is built from (no references provided), would this also be discussed under NORN? (I don't see a REFNB... is there one?)

70.24.251.71 (talk) 11:15, 4 March 2012 (UTC)

It would help if you linked us to the image in question. An image is rarely OR just sitting there in commons. What is often OR is how the image is used in an article (images should be used to illustrate something said in the text ... they should not be used to present information that is separate from the text). Also, very often the OR problem isn't actually with the image, but with the caption that accompanies the image. A photograph of a pile of dead bodies isn't OR... but a caption that identifies the image as: "Victims of atrocity X" (without a source to support this identification) would be. Blueboar (talk) 14:45, 4 March 2012 (UTC)

Also note that OR doesn't apply directly if an image and its caption is a reasonable faithful illustration of the text. Only if it says something new not said in the text can it be OR. Dmcq (talk) 15:10, 4 March 2012 (UTC)

Maps on Commons are often OR. Dougweller (talk) 15:18, 4 March 2012 (UTC)

Well, as I stated, if an image is built from a dataset (like a graph or chart) and doesn't indicate the dataset it is built from... (Unreferenced)

or if the graph or chart extends beyond the dataset it was built from (Original Research)

or if a picture is built from data but provides visual features that are not part of the data used to build it.

Per this FFD, an image was created from an unknown dataset that features data that is questionably not present (coloration) and of a visual appearance that is questionable (astronomers have written theoretical papers saying that some planets in this mass range are likely banded, and not blandly monochrome).

70.24.251.71 (talk) 05:42, 6 March 2012 (UTC)

Offhand, I'd guess that this relates to the color of the exoplanet image discussed at Wikipedia:Files for deletion/2012 February 28#File:Epsilon_Tauri_b_and_Hyades.png. WhatamIdoing (talk) 15:45, 4 March 2012 (UTC)

WP:PERTINENCE is the relevant thing to look at for that. The image does not have to be authentic. It has be be pertinent to the article and be accepted as a reasonable illustration of the topic or section it is supposed to illustrate. The image looks like a water planet to me. Dmcq (talk) 16:06, 4 March 2012 (UTC)

RfC archived; just got announcement; proposal

I just got the bot about this RfC, read it and responded and low and behold in the interim it was Archived! Seven days is just too short a time to archive in on a policy page that is not THAT busy. Let's change it to two weeks. And I do have a specific proposal on this. CarolMooreDC 19:10, 5 March 2012 (UTC)

Um... what RfC? I looked in the archives and don't see any recent RfCs. (No objection to extending the archive time anyway ... just curious as to what you are talking about). Blueboar (talk) 13:14, 9 March 2012 (UTC)

Bot message I got here. Wikipedia_talk:No_original_research/Archive_57#RfC:_Policy_on_primary_sources_is_too_extreme. In short, the example given of laws passed that may not YET have been reported in WP:RS but are of obvious importance does seem to me one possible exception to the rule on no article using just primary sources. Would you agree 7 days is just too short a period of time before archiving?? CarolMooreDC 22:02, 9 March 2012 (UTC)

The RFC started on 08 February and was archived on 05 March. That's 26 days, not 7. WhatamIdoing (talk) 00:02, 14 March 2012 (UTC)

can editors base the naming of articles on original research

e.g., unpublished, undisclosed, so-claimed page-view statistics

e.g. from discussion at http://en.wikipedia.org/wiki/Talk:The_Big_Bang_Theory — Preceding unsigned comment added by 65.94.204.129 (talk) 04:08, 14 March 2012 (UTC)

Editors may not make claims based on original research; claims should be verifiable in reliable published sources. However, editors can choose titles that will convey to the reader what the article is about, and distinguish the topic from other Wikipedia articles, based on the collective preferences of the editors. If someone wants to title an article about a certain television show "The Big Bang Theory" that's fine; the show itself serves as a reference for that title. If someone wanted to title the article "Len the Nerd" on the basis that the editor had a conversation in a bar with the show's creators (but there is no written record of the conversation), that would not be allowed.

Ignore the IP, s/he is forum shopping following a failed move request s/he initiated last month. Parsecboy (talk) 18:20, 14 March 2012 (UTC)

It sounds like someone needs to learn WP:How to lose. Even if the unregistered user is 100% right, the user has already lost and should simply give up, rather than wasting more of everyone's time to achieve the same amount of nothing. WhatamIdoing (talk) 22:10, 16 March 2012 (UTC)

Revert of my edit

Can anyone explain RedPen's comment about this revert? I don't get it. Operative67 (talk) 02:07, 23 March 2012 (UTC)

your suggested wording implied that sometimes we may allow interpretation based on primary sources. we do not, and should not imply that we do. -- The Red Pen of Doom 02:10, 23 March 2012 (UTC)

But you can't be serious. All interpretations are always going to be based on consensus, regardless of the source, if only because consensus will determine just how "primary" a source really is. Interpretations from primary sources is common in some scientific articles I've seen. There are some articles based almost exclusively on primary sources, and conclusions are drawn from them. I happen to think that's as it should be, but it also seems to be common-enough practice. Operative67 (talk) 02:20, 23 March 2012 (UTC)

We can be serious. Your argument here amounts to WP:OTHERSTUFFEXISTS. The fact that some article are badly written and poorly sourced is not an excuse for officially endorsing those bad practices. WhatamIdoing (talk) 05:41, 23 March 2012 (UTC)

But that page should read WP:OTHERSTUFFPERSISTS. And the longer it persists the more legitimate it is likely to be. Surely you don't consider the AZT article to be "badly written and poorly sourced". I consider it excellent, and yet it breaks at least two "rules", so-called: 1) It is based almost exclusively on primary sources, and 2) conclusions are drawn from them not necessarily clear to people of average education. Far from it. Without the primary sources the article would just be a stub. And much if not most of it requires an advanced scientific education to actually understand. But that's perfectly acceptable (or should be) because there is a cadre of editors who clearly know what they're talking about and who make it their business to keep the article truthful. Also (although it's not of central importance), a solid listing of primary sources is, after all, what the average reader expects in a scientific article. Otherwise, many of them would have objected by now.

All I'm really saying is that with a few concise edits (like the one I tried to make but was reverted), Wikipedia's policy regarding primary sources could conform to actual practice. OTHERSTUFFPERSISTS! So acknowledge it and move on. Right now, it has the sound of theology about it, more aspirational than actual. It undermines Wikipedia's credibility. Operative67 (talk) 03:39, 24 March 2012 (UTC)

No responses, so are we good to go with restoring my edit? If no one has anything to say in the next couple days, I'll re-insert may like I did earlier. No edit warring, OK? Operative67 (talk) 02:54, 25 March 2012 (UTC)

I have reverted... we do not (and should not) allow editors to interpret primary sources for themselves (actually, we should not allow them to interpret secondary or tertiary sources for themselves either... our job as editors is to report on the interpretations of others.) Blueboar (talk) 15:47, 26 March 2012 (UTC)

Sorry, but we do (and should) allow editors to interpret primary sources for readers. But I'm not going to pursue this point here anymore, as it seems utterly futile. Suffice it to say, however, that the policy as it stands fortunately does not reflect actual common practice, especially in medical-scientific articles -- nor will it, I can assure you of that. I doubt that it ever did reflect reality, but it sure does seem to be an artifact of deep theological importance in the rarified councils of Wikipedia! I humbly bow three times as I step backwards from the sepulchre -- chewing gum nevertheless. Operative67 (talk) 17:46, 26 March 2012 (UTC)

We do permit surveying, we do not permit original research. Fortunately, many editors do understand the difference and follow our policies. Nageh (talk) 17:57, 26 March 2012 (UTC)

What do you mean by 'interpret' there? It doesn't seem to me to be what WP:PRIMARY says in 'Any interpretation of primary source material requires a reliable secondary source for that interpretation'. My impression was we could use them for straightforward factual stuff and filling in the gaps but interpretation was something secondary sources did. And I'm very much against loosing the hordes as it were in Wikipedia to do interpretation of primary sources. I'm already in a dispute with an editor picking out pages from a wiki which seemingly is allowed under WP:SELFSOURCE and interpreting them. Dmcq (talk) 18:04, 26 March 2012 (UTC)

I think the problem here is a verifiability one, not an original research one. Perhaps a different wording can help. If a {{cn}} is put against an interpretation then a secondary source is needed, I think perhaps that's the sense the proposer wanted? I don't know exactly how to phrase that though. Dmcq (talk) 18:14, 26 March 2012 (UTC)

OK, what I mean (and I may not have been clear) is that it is common practice (at least in sci-med articles) to use primary sources in place of secondary sources. The primary source is used as a secondary source. You could argue that that is an act of interpretation. But on the other hand, if the reader has the educational background to read through the source, the logic in citing that source will become clear. But that logic may not be immediately apparent. You have to study your way through the source. You might even have to purchase the source online (because it is behind a pay-firewall, which is permitted as a source in Wikipedia) or in print. But what's wrong with that? We assume good faith of the editor who posted it, and Wikipedia makes no provision that the source must be conveniently available. Say what you will, this is done all the time. I'm just saying it should be recognized as Wikipedia policy. Operative67 (talk) 18:24, 26 March 2012 (UTC)

Perhaps an example of what you're talking about would help? I would have problems with someone saying of a research paper 'and therefore ...', and saying something that the paper didn't in effect say. Dmcq (talk) 18:32, 26 March 2012 (UTC)

Since I mentioned the AZT article (above), I went there to look around. It has I think one single "secondary source". The rest look pretty much primary. Here's one that's pretty representative in a good way of what I'm talking about. I think it's hard to take issue with. But take your pick. You can see that without the primary sources, the article would be a stub. Hardly encyclopedic. Operative67 (talk) 19:49, 26 March 2012 (UTC)

Some of the other sources are secondary, such as the New York Times article and the textbook Pharmacology (3rd ed.). One could argue whether some of the others are secondary, such as some of the government health agency reports. Finally, you have not explained where the article contains interpretation of the primary sources, rather than just reporting the results contained in those sources. Jc3s5h (talk) 20:15, 26 March 2012 (UTC)

The interpretation is an editorial one. The sentence says that AZT reduces mortality, period. But the source has restrictions and limitations to that statement. But they would derail the sentence and its meaning if it were included. So it's just left in the footnote. I think it's perfectly acceptable, but I do look upon it as interpretation. Operative67 (talk) 20:57, 26 March 2012 (UTC)

Clarifying definitions

OK... I think I see where the confusion is coming from... it is one we have discussed before, but had no solution for. The confusion stems from the fact that different academic fields use somewhat different definitions of the terms "Primary" and "Secondary" Source. What this means is that a source might be dubbed "Primary" by someone a sciences background, and "Secondary" by someone from a humanities background. Now... It happens that most of the editors who first worked on this policy came from a humanities background, and not a scientific background... and they used terminology in a way that reflects their background... What I am getting at is that the majority of the sources used in the AZT article are considered "Secondary" sources per Wikipeida's (humanities inspired) definition. Hope this clarifies things. Blueboar (talk) 20:51, 26 March 2012 (UTC)

Link? Nageh (talk) 20:40, 26 March 2012 (UTC)

Not sure what you want me to link. Blueboar (talk) 20:51, 26 March 2012 (UTC)

You said there was previous discussions on this, I was interested in having a look. (No worries if you cannot recall where and when it was.) Nageh (talk) 21:10, 26 March 2012 (UTC)

Search "Humanities" in the archives... that should give you a fairly good sampling. Blueboar (talk) 21:48, 26 March 2012 (UTC)

That's clarifies matters a good deal. But there is no reason why that couldn't be put into the policy itself, perhaps as a footnote. Operative67 (talk) 20:57, 26 March 2012 (UTC)

Currently, a key sentence reads:

For example, a review article that analyzes research papers in a field is a secondary source for the research.

Maybe that could be changed to:

For example, a peer-reviewed article that analyzes research results in a field is a secondary source for the research.

So the scientists' original pre-publication notes would then be defined as primary sources. I think that's what a primary source really is anyway -- the actual, original documentation. 21:50, 26 March 2012 (UTC) — Preceding unsigned comment added by Operative67 (talk • contribs)

Could you clarify what you see as being the difference between a "review article" and a "peer-reviewed article"? Isn't the first a subset of the second? Blueboar (talk) 23:45, 26 March 2012 (UTC)

Actually I think you've hit on it, you're right. But it isn't obvious -- it wasn't to me until you mentioned it. So I think the thing to do is clarify that point. Once we've done that, this entire problem goes away, and articles like AZT are safe. So why not follow up on your point by posting my second quote, above (...a peer-reviewed article that analyzes research results is a secondary source ...). This should work because, as you point out, review articles are subsets of peer-reviewed articles. So logically we should specify the top category, not a subcategory. Right now the subcategory is mentioned, thus the confusion. Operative67 (talk) 00:11, 27 March 2012 (UTC)

Not really. Any interpretation of a reliable source must come from another reliable source, not a Wikipedia editor. It's just that more violations of this principle have been observed with primary published reliable sources, as opposed to secondary published reliable sources. Perhaps that's because the primary sources tend to have more raw data begging for interpretation, while the secondary sources tend to already contain the appropriate interpretations.

The reason that interpretations of a primary published source must be a secondary published source is that sources that analyze other published sources are almost always secondary sources. The exception would be if the author (or the author's organization) of the analysis was personally involved in the events. So a US Supreme Court opinion almost qualifies as a secondary source, because the justices didn't witness the alleged crime they are ruling on, they are analyzing paperwork that has been submitted to them. But they are not just giving their opinion, they are issuing orders which may ultimately be physically enforced by agencies like the US Marshal Service or even the armed forces; that makes them too involved to consider their opinions a secondary source. Jc3s5h (talk) 00:22, 27 March 2012 (UTC)

Well, are you saying then that this is primary or secondary? If it's primary, then either the article must be severely edited (theoretically -- it wouldn't actually happen), or the policy amended to allow what used to be considered primary sources to be considered secondary ones -- a redefinition, in other words. I would opt for the latter as the simplest, following up on BlueBoar's comment.Operative67 (talk) 00:41, 27 March 2012 (UTC)

That's a primary source, because it describes an experiment conducted by the authors. But the Wikipedia article does not appear to interpret the source, it just summarizes the source. That's OK. If the Wikipedia article, for example, applied a new statistical test to the raw results reported in Fischl et al., that would be interpretation. Of course, I don't have the paper, so maybe the summary in the Wikipedia article is an interpretation. Jc3s5h (talk) 01:10, 27 March 2012 (UTC)

Jc3s5h is correct. A lot of people think we are not allowed to use primary sources at all... that is incorrect. We can use them for descriptive statements as to their contents.

In fact, I think that "description" is the key to understanding how to use any source appropriately. Our job as editors is to describe what is said in reliable sources, whether they are primary, secondary or tertiary... If we wish to include an analysis, interpretation or conclusion in an article, we must find a reliable source that explicitly contains the analysis, interpretation or conclusion we wish to include. We can then describe what that source says and cite it. Blueboar (talk) 02:08, 27 March 2012 (UTC)

That's right -- it's an abstract. Its key conclusion is: These data demonstrate that AZT administration can decrease mortality and the frequency of opportunistic infections in a selected group of subjects with AIDS or AIDS–related complex, at least over the 8 to 24 weeks of observation in this study. And the article-statement referencing it says: The study found that AZT could prolong the life of patients with AIDS. To me, that's a valid interpretation and needn't be changed. But if you are right, it is a primary source. So it boils down to whether the article, which is based on sources like this, violates policy. Again I would argue that the policy should be changed to conform with actual usage. Operative67 (talk) 01:51, 27 March 2012 (UTC)

No, that is not an interpretation... it is a descriptive restatement of what is stated in the source. That is OK. There is no interpretation involved. Blueboar (talk) 02:12, 27 March 2012 (UTC)

I'm going to just apply that idea across the board in the AZT article. So that's that for interpretation. But there is one matter we should also address before closing: how to square the AZT article with this policy statement: "Do not base articles and material entirely on primary sources." It seems like we might still have to do a re-definition OR interpret that statement as allowing one single secondary source (and there is a solid one, the New York Times article) to satisfy the requirement. I say for simplicity we just do the latter by default.

So we're done. The AZT article (and others like it) is in compliance. Thanks for the clarifications. Operative67 (talk) 18:50, 27 March 2012 (UTC)

Is this a primary source, if so, how does the WP:NOR policy apply to it?

Is Thomas Jefferson's document Plan for Establishing Uniformity in the Coinage, Weights, and Measures of the United States a primary source? If so, how does the WP:NOP policy apply to it? As can be seen, we have an entire article dedicateed to this document. I have no problem with the document, just how Wikipedia rules define and relate to it. Martinvl (talk) 13:04, 29 March 2012 (UTC)

Yes, it is a primary source. It is appropriate to cite the document directly for a basic unadorned descriptive summary of its contents, or for any quotes from it ... but... any interpretation or analysis of the document or conclusions about it must be cited to secondary sources. The bulk of the article should be focused on what others say about Jefferson's plan, not just an account of its contents... and for that you need secondary sources. A good rule of thumb is... When in doubt, find a secondary source. Blueboar (talk) 13:56, 29 March 2012 (UTC)

Web search results

I would have thought citing search engine results is original research, but at this Talk: page another user suggests it is acceptable [as evidence for the relative frequency of different phrases]. I note {{Google templates}} is not to be used in the article namespace, and WP:LINKSTOAVOID deprecates external links to search results but I found nothing explicitly relating to references. Perhaps this needs to be explicitly mentioned in Wikipedia:No original research, Wikipedia:Verifiability, and/or Wikipedia:Identifying reliable sources. jnestorius^(talk) 14:37, 2 April 2012 (UTC)

Citing the relative frequency of search engine results cannot by definition be primary research, since all it does is count (and, if followed through, locate) items of previously published information. It is a perfectly reasonable piece of secondary research to cite when the issue at hand is, precisely, the relative frequency of different usages for (in the instance under discussion) various names for one place. Brocach (talk) 16:56, 2 April 2012 (UTC)

Probably WP:IRS. It's definitely not something we should use - not only does it vary with the search engine, the way you phrase the search, probably the country and even time of day perhaps (I've found the same search turn up different numbers of items when I run it later), it's going to churn out rubbish - it will find self-published stuff, mirrors or printed copies of our articles, you name it. Dougweller (talk) 15:57, 3 April 2012 (UTC)

I think it's only appropriate on talk pages for deciding among editors what to name the article, what other phrases also used to describe topic of article, various other uses of phrases or descriptions. And then you have to be careful that you actually have narrowed your search requirements. Searching John Francis Smith III without quotes will get you many more returns than searching "John Francis Smith III" with quotes. Though it might be helpful to mention that here very briefly for clarification. On the other hand, enough people always seem to know it when the topic comes up. CarolMooreDC 05:16, 4 April 2012 (UTC)

Dougweller- obviously to have any value as an indicator of relative frequency of usage of two (or more) versions of a name etc., the search terms would have to be used with precision (e.g. in Google, with quotes), and one should (as I always do) use several search engines. When reasonably consistent results are returned, that does not constitute original research, but it can provide evidence worth citing as to which of the alternative forms is more widely used (on the web, at least). CarolMooreDC, I understand your expectation that this sort of 'source' should be confined to talk page discussions, but in the case at issue, the article included discussion of various forms of a place name, so it made sense (in my view) to put it in the body. Brocach (talk) 23:57, 7 April 2012 (UTC)

Please note that the Wikipedia:No original research policy only applies to article content.
— V = IR (Talk • Contribs) 04:15, 8 April 2012 (UTC)

If we want to know which terms are most frequently used we should use a source that tells us. Counting Google hits is original research. A high number of hits for a term may be found because it is popular with bloggers, the hits may be using the words to mean something different, etc. Also, a search for "County Down" will register a hit for "Jones County, Down County". TFD (talk) 01:05, 20 April 2012 (UTC)

Talk:Trollhunter#Move to Trolljegeren?

Is it an original research to conclude one source as more reliable than other sources for article titling in terms of English translation of non-English titles? Discuss more by clicking above. --George Ho (talk) 15:47, 24 April 2012 (UTC)

Dividing line between "Published/Reliable" and "Original Research" ?

I understand that Wikipedia has a policy of not putting on here statements which cannot be directly attributed to a reliable, published source... my question is, wouldn't a now-published, now-reliable source had to have been considered "Original Research" itself, until somebody else decided to publish it and declare it reliable, at some point in time? No, I am not trying to create any kind of personal-belief-driven arguments, I'm genuinely interested in discussing this policy and learning what divides a "Reliable, Published" source from something that may be considered at present to be "Original Research" simply because nobody has published it yet or declared it to be reliable. Thank you in advance for being patient and understanding with my question... Blozier2006 (talk) 22:18, 19 April 2012 (UTC)

OR primarily refers to content in WP, meaning research normally has to be published elsewhere before it can be integrated into WP. So researchers are not supposed to publish their results via WP.

In addition there is the problem how WP treats original research in external sources, which is what you are talking about. This is a somewhat grey area that requires some common sense. If possible WP tries not to rely on such original publications, but (in doubt) wait for additional publications (in particular review articles), which indicate that the research of that original publication has become established knowledge (collecting/describing the established knowledge and providing free access to it, is the primary goal of WP). However depending on exact the context of the original publication, you may still want to use it to include the most recent authoritative information on subject and/or if you can be sufficiently certain that the original publication is correct anyway. For instance in many cases can original publications in (prestigious) peer reviewed journals be considered as reliable enough to be considered "established knowledge" already and hence may be used as a source. However in some cases in particular in heavily disputed fields it might not be appropriate to cite original publications and also if you cite an original publication do in doubt describe it as a claim or opinion of scholar X rather describing it as a fact (=established knowledge). Also in addition various fields in WP may have additional recommendation how to deal with original publications. Another aspect to consider here is notability. Many opinions/claims in original publications without a reception in the academic community may simple not be notable and hence not to be included in WP and consequently not to be used as a source either.--Kmhkmh (talk) 22:57, 19 April 2012 (UTC)

(e/c) If I understand your question, I think you have a confusion about to what material the standards of original research and reliable sources apply. Original research is a standard applied to material we wish to add to the encyclopedia and not to the external source. Whether a source is reliable looks to the external source and not to the material we wish to add (except to the extent context of use plays a role). When we question whether a source is reliable, we don't ask whether it is not previously published, or a synthesis of other material, or reaches a new conclusion not present in the source it, itself, may draw from (i.e. the questions we ask when considering whether material added here is original research). Rather, we ask whether it is the type of source that has a reputation for fact checking and accuracy; whether the manner it's material is drawn from comes from a scholarly source or a self-published blog, and so on.

To provide an example, if you wanted to add to an article the statement that: "Colorless green ideas sleep furiously", we would ask whether a reliable source out in the world had already published that concept and whether, even if some had, the way we were presenting the material presented a conclusion not explicitly held by any of the sources. That would be the original research query. However, if we were questioning (quite separately), reliability of sources, and were presented with "Colorless green ideas sleep furiously<ref>External Source 1</ref><ref>External Source 2</ref> we would ask whether external sources 1 and 2 have a reputation for fact checking an accuracy. Does that help? Did I miss your point?--Fuhghettaboutit (talk) 23:10, 19 April 2012 (UTC)

Each publication decides what kind of material it wishes to publish. Golf Digest does not print articles about seismology. Lancet does not publish articles about knitting (unless knitters have an inordinate tendency to develop a medical disorder). Wikipedia is not the first publisher of original research. Jc3s5h (talk) 23:46, 19 April 2012 (UTC)

Re:Fuhghettaboutit: You are pretty well spot-on about what I was asking... Basically, you are saying "Don't make a statement here unless you can back it up", right? Similarly to how you will automatically will be flunked on a research paper if you do not supply what sources you used in writing the paper (so as to prove you are presenting factual statements and are not committing plagiarism), an article you either write here, or contribute to here, will not be considered acceptable if you do not mention on what basis you are making your statements... correct? Blozier2006 (talk) 00:01, 20 April 2012 (UTC)

yes and no. It is not sufficient to simply show on what basis your are making your statements, but we impose some (strong) criteria on the "basis" as well and in a way stronger it is for academic publication as they are allowed (or even expected) to contain OR. Even decent OR article backs up his claims as well, but his claims are new conclusions/claims/theories, which is something we do not allow in WP. We do not allow "backed up" new conclusions in WP, they need to be published (and reviewed) elsewhere first.--Kmhkmh (talk) 12:55, 20 April 2012 (UTC)

To try to clarify... WP:NOR goes a step beyond saying "Don't make a statement here unless you can back it up"... it also specifies how we have to back it up. We must to be able to point to reliable sources that (effectively) say the same thing that we say.

Our job, as editors of a tertiary source, is to summarize (in our own words) and report on what others (reliable sources) have said about a topic. The sources might say something original... but we are not being original when we summarize and report what they said. Blueboar (talk) 15:15, 28 April 2012 (UTC)

Dictionary as a source

A dispute has arisen at Talk:Welfare about the main meaning of the word "welfare" and whether the meaning of "government aid" is primarily North American. Three dictionary sources have been produced as follows

The Oxford Dictionary

"health, happiness, and fortunes of a person or group"
"chiefly North American financial support given to those who are unemployed or otherwise in need

The Longman dictionary

"health and happiness"
"help for people with personal or social problems" (welfare benefits, welfare services, welfare programs etc)
American English. "money paid by government to the poor or unemployed"

The Macmillan dictionary

the health and happiness of people or good care
1. good care and living conditions for animals
care provided by the state or another organization for people in need
1. Mainly American money given to people who do not have work or who are in need. (The usual British word is benefit)

As far as I can tell all three dictionaries say that the meaning of "money or financial aid from the government" is mainly North American.

Another editor has claimed that these dictionary references are PRIMARY SOURCES and that we need other secondary sources before we accept that the meaning of "government aid" is primarily North American usage. He claims that my producing these as sources is OR or SYN. Is that correct? I would have thought that dictionary compilers refer to a wide variety of other primary sources before accepting an entry and that these dictionary entries are therefore a scholarly and reliable impeccable secondary or tertiary sources for Wikipedia. 84.250.230.158 (talk) 17:47, 27 April 2012 (UTC)

See WP:DISAMBIG: articles are about topics, not words. Where a word nay have more than one meaning, we create more than one article. TFD (talk) 19:32, 27 April 2012 (UTC)

I agree. It has been suggested and I am in favor of that. The issue is about whether the word WELFARE equates to the topic in the article (which, as it stands is about "government aid"). It seems that this meaning is used primarily in the U.S. and therefore a dispute has been about whether the title can be used without further clarification given that the main meaning in countries outside the U.S. would seem to be "well-being". Another editor seemed reluctant to accept that the meaning of the term "welfare" as "government aid" was primarily North American. It was suggested that I needed a secondary source to prove this before the other editor would accept it as a valid argument. I produced the dictionary definitions but he said I was engaging in OR or breaking a rule about not using a primary source. But as I see it, this is not OR and the source is perfectly valid. A dictionary is surely a secondary sources as the dictionary compilers use many other sources). I don't understand how I could be engaging in SYN (another accusation put to me). Now he seems to be accusing me of coming here to WP:CANVAS. All I am doing is trying to establish whether a dictionary is an acceptable source for something being asserted as factual. It would be helpful to have a yes or no answer to that question.84.250.230.158 (talk) 21:05, 27 April 2012 (UTC)

An article is about a topic. Does the first sentence of the article 'Welfare refers to a broad discourse which may hold certain implications regarding the provision of a minimal level of wellbeing and social supportfor all citizens.' describe the topic of the article? That is really all that matters. The dictionary is not very relevant. There is no need for a separate article for the same topic in another country just because a different term is used for it. There are various terms like social security for instance but welfare is perfectly okay. We don't need separate articles on cootie catcher and Salt cellar (origami) and paper fortune teller, the first two simply redirects to the third. Dmcq (talk) 22:52, 27 April 2012 (UTC)

The first line may capture both concepts, "well-being" and "social support", but the bulk of the article is about financial support, and in most countries that is NOT called "welfare". It is Welfare Support, or Welfare aid, or Social Protection, or Social Security or Income Support etc. etc. but not "Welfare" on its own except perhaps in the USA. Another editor has said that America is 1/7th of the World's English speakers and so WELFARE (as financial aid) is alright. I disagree. Isn't Football a good model here? Here we have a topic that means different things to different people in different cultures. So the article talks about the wide variety of football games without ever assuming that one meaning is the right meaning. I think we need to do the same with Welfare. It too means different things and we should not assume that one is more correct than the other. 84.250.230.158 (talk) 23:30, 27 April 2012 (UTC)

Well to start with, the "social support" meaning is what the article is about and probably the only meaning that would merit an article vs. just a dictionary definition. As far as where you go from there, there are several possible routes. My suggestion (amongst all of them) would be to continue to have it be about social support, but acknowledge and briefly discuss the other meaning. This would do a good service to readers. I have run across other major words/topics where the meaning is different in different places in the world, and it was a painfully slow process learning this on the article talk page. Clarifying it in the article will save readers all of that work/pain and present very useful world view level enlightenment. North8000 (talk) 23:46, 27 April 2012 (UTC)

A child living with wealthy parents that offer him/her no love or affection is not lacking "government support services" but is certainly is lacking in "welfare". If the article is about welfare as "social support " then the article title should be Welfare (social support and Welfare should be the disambiguation page that it was for many years. It would be wrong to make the football article solely about rugby football (even if the lede said that this was the topic of the article) when most of the world cannot be sure what "football" means when talking to an international audience. The argument at Talk:Welfare is now about the title of the article. The issue I had in coming here was to determine whether using a dictionary definition to prove the regional bias of the minor definition amounts to WP:OR. So far nobody has said it does.84.250.230.158 (talk) 09:08, 28 April 2012 (UTC)

I think calling this a WP:OR issue is stretching things. It is really a typical WP:Article title debate... one that requires a multi-step analysis to resolve... First we need to determine the scope of the article in question (Is it limited to the system of government support for the poor that is specific to the US, or is it covering the systems of government support for the poor that exist in multiple nations?).

Once that is determined, we can then move to the next step... If the scope is specific to the US, then the discussion becomes whether to use the proper name of that system ("Welfare") as the title, or whether to use a descriptive phrase (such as "Welfare system in the United States") as the title.

If the scope is international, then we can not use a proper name as the title (since various nations have different names for their systems of support)... we must entitle the article with a descriptive phrase (such as "Governmental support for the poor"]]).

The final step is to determine if the chosen title requires disambiguation. That is determined by whether there is an existing article that uses the title. If there is not, then there is no need to disambiguate... or rather there is no need to do so at this time (this may change in the future ... once another article that wishes to use the title is written). Blueboar (talk) 13:50, 28 April 2012 (UTC)

Thanks. I think this is a very sensible approach.84.250.230.158 (talk) 22:51, 29 April 2012 (UTC)

Scientific papers in peer reviewed publications as supporting citations

I have been told that no scientific articles published in peer reviewed publications are suitable for inclusion in wikipedia as references because they violate the Original Research policy. Could you take a look at Talk:Bacteriological water analysis and let me know if, in fact, no scientific papers should be used as sources on Wikipedia? Thank you. --User:Ceyockey (talk to me) 16:33, 6 May 2012 (UTC)

Just to declare my interest as I reverted the addition. If Wikipedia were to permit the use of individual peer reviewed scientific publications as references then even such an obscure topic as Bacteriological water analysis could be overwhelmed by references - I might be able to find 1000+ without getting out of my chair. There is also the real risk that scientists anxious to ensure that their paper gains a wide readership would insert highly tangential references just to get a mention. I believe that the current guidance is appropriate and reasonable and would advocate maintenance of the status quo. Velella ^{Velella Talk} 17:14, 6 May 2012 (UTC)

→ Just to be clear - there wasn't any removal of content by Velella in this case; I suggested the consideration of a scientific paper at the article in question on the talk page. That suggestion led to the present conversation. --User:Ceyockey (talk to me) 20:38, 6 May 2012 (UTC)

That slippery slope argument isn't going anywhere. If you can find 1000+ different peer-reviewed papers to cite the same statement than that statement probably doesn't need a reference or can be referenced to a textbook. But in many other cases, peer-reviewed articles are the best option. Not everything that appears in print need appear on Wikipedia though and your concern about tangential subjects and tangential references should be handled using guidelines on due weight and notability. Citations of peer-reviewed journal are so central to Wikipedia that User:Citation bot which is entirely devoted to this issue has tens of thousands if not hundreds of thousands of edits. Pichpich (talk) 17:40, 6 May 2012 (UTC)

Citation bot works with all citations and with any sources. The fact that Citation Bot exists and operates in Wikipedia is not an endorsement of the use of primary sources. I have made no reference to a"slippery slope" nor should my comments be used to draw such an inference. I stated an opinion that if primary sources were routinely accepted as reliable sources then there is a risk that some articles would be overwhelmed by references. A cursory glance at any scientific parer will usually show that for that paper alone very many other quoted sources are listed. Wikipedia articles are (thankfully) generally more wide ranging than scientific papers and the number of references is potentially very much greater. The readership of Wikipedia articles will also be poorly served by an interminable list of references which are likely to dwell on research minutiae rather than provide enlightenment on the general topic. Velella ^{Velella Talk} 19:16, 6 May 2012 (UTC)

You're arguing that citing any individual peer reviewed paper is against Wikipedia policy. This is nonsense and again I think it's obvious that citation bot would have never been created or allowed to perform the tasks it currently performs if a significant number of people shared that view. Pichpich (talk) 20:43, 6 May 2012 (UTC)

→I think that this is largely a misunderstanding around initially the definition of "original research" and second around whether or not scientific publications are acceptable sources. As I noted on the talk page which prompted my coming here, scientific publications in general are 'secondary sources', which are allowed by policy for inclusion as supporting citations and on which Wikipedia relies heavily. Granted, one cannot generally support the retention of an article which relies solely on secondary sources; a consequence of this is that massively obscure topics which are either too new, fringe or abstruse to be considered by the tertiary sources which meet minimum Wikipedia inclusion guidelines would not survive here as stand-alone articles, something which addresses in part Velella's slippery-slope concern. --User:Ceyockey (talk to me) 03:08, 7 May 2012 (UTC)

Different disciplines use the term "primary source" and "secondary source" in different ways. The usage in this policy is necessarily a compromise among the fields. A peer-reviewed scientific paper might be considered a primary source if it reports new experimental results, rather than discusses previously reported results. A secondary source would not report new experimental results, but might report new conclusions based on analysis of previous reports. Tertiary sources are typically encyclopedias and textbooks. The notability policy requires that articles contain citations to secondary or tertiary sources. The lack of tertiary sources is not grounds to call for an article to be deleted. Indeed, tertiary sources are usually considered less desirable than secondary sources. Jc3s5h (talk) 04:12, 7 May 2012 (UTC)

Please note... OR is about how we use sources, not about their classification... Note that the policy explicitly allows us to cite primary sources (as long as we do so appropriately and with caution). "Primary" does not equal "bad" ... nor does citing a primary source automatically make a statement OR. So... even if we classify a peer-reviewed scientific paper as "primary", citing it is not OR unless we go beyond what the paper actually says (drawing our own conclusions, interpretation or analysis from it).

Also, citing a secondary source does not automatically make a statement "not OR". It is quite possible to misuse a secondary source in ways that are considered OR (using it to support your own analysis, interpretation or conclusion). In short... Original research is really about whether the statements we write are directly supported by our sources, and not about the classification of those sources. Blueboar (talk) 12:42, 7 May 2012 (UTC)

There are 2 slightly different things that often get confused somewhat here. a) OR by wp editors (which imho the policy primarily deals with b) OR as in brand new (external) research that has not really been accepted or confirmed by the scientific community (at large). You can do b) without violating a), however in many cases we don't want b) in WP either (at least not in the form of reporting it as a "fact"), but wait until it is confirmed or invalidated by other sources.--Kmhkmh (talk) 13:28, 7 May 2012 (UTC)

Any given field has hundreds and thousands of papers, with varying results. Choosing one of them over the others is original research. That's why you use critical articles, reviews, meta-reviews, and scholar books that analyze the field. Those sources will tell you which papers were significant and why. --Enric Naval (talk) 15:09, 7 May 2012 (UTC)

No, choosing among reliable published sources is source-based research, which is what Wikipedia editors are supposed to do. Jc3s5h (talk) 15:14, 7 May 2012 (UTC)

Choosing between many different sources is what we expect our editors to do. Research papers have always been acceptable sources from the point of view of the NOR policy. If there is a specific issue with some particular article and citation, that has to be resolved on the talk page of that article. — Carl (CBM · talk) 00:59, 8 May 2012 (UTC)

Wikipedia aims to reflect the consensus of the papers. The best method of reflecting the consensus of the thousands of papers is by looking at the reviews, meta-reviews and the scholarly books. Choosing a single original research paper and saying it represents the consensus is problematic. IRWolfie- (talk) 14:19, 11 May 2012 (UTC)

Quick points:

WP:MEDRS is relevant.
This is a content dispute, not a proposal to change this page. Therefore, the question needs to be taken elsewhere.
WT:MED is a relevant "elsewhere" that you can take it. (Read MEDRS first. People experienced with medicine-related articles are generally not sympathetic to primary [using the hard-science definition, which claims all original experimental work as primary] sources.) WhatamIdoing (talk) 00:09, 15 May 2012 (UTC)

suggest addition to point out that primary sources such as interviews do count towards notability

Note that an interview with a person about themselves or something notable they have achieved, such as a book they have published or a film they started in, does counts towards notability of them or their works.

To avoid confusion that comes up from time to time in AFDs, even with editors who have been around for years now, I suggest this be added to the article, or something similar. [3] Or just a note saying that nothing on this page had anything to do with notability. Dream Focus 00:15, 13 May 2012 (UTC)

Not always, as it also depends who is doing the interview, in what context, and the like. A fanblog without established reliability but managing to get a good interview is not going to count for notability. On the other hand, a reliable third-party source that has taken the time to interview someone is an indication of notability. --MASEM (t) 00:30, 13 May 2012 (UTC)

This has come up before, but it seems that the problem has yet to be corrected. It is the attention given by the interviewer to the interviewee that shows that the topic attracts attention. The interviewee may be an expert in a subject matter about which the comments are secondary. It is the interviewee speaking about themselves that is primary. I added the words "(depending on context)" to the Project Page. Unscintillating (talk) 06:57, 13 May 2012 (UTC)

Instruction creep... notability issues have nothing to do with Original Research. This is an issue for discussion at WP:Notability. That said... Masem and Unscintillating are correct in noting that the issue is more complex... an in-depth interview with the local dog catcher in a small town newspaper does not "count towards" notability. An interview with that same dog catcher in the New York Times might... depending on why the dog catcher was being interviewed. Blueboar (talk) 14:09, 13 May 2012 (UTC)
Actually, I think it's the other way around: it's the interviewer, not the interviewee, that matters. If Barbara Walters publishes a long interview with you about your career, then that counts towards notability as a sign that the mainstream media is paying attention to you ("attention from the world at large", to use WP:N's old language). If a student newspaper publishes an interview with you, it doesn't.
The reason this comes up here is that the GNG demands that only secondary sources be counted. Naturally, since most news stories are primary sources, and since this page defines all interviews (without exception) as being primary, that is awkward for demonstrating notability of current events and the like. I think, though, that the solution needs to be re-defining GNG to accept that we should give some (perhaps only a little) weight to sources that are technically primary, but are also high-quality, independent, and non-self-published. WhatamIdoing (talk) 00:14, 15 May 2012 (UTC)

This page contains some footnotes that mention interviews as examples of possible primary sources, but it really depends on who published the interview (both author and publisher) and what else is in the source. A book written by a historian who, in 1990, wrote a comprehensive history of a World War II battle, and included an interview with an officer that the author conducted personally, would have to be considered a secondary source, including the material about the battle contained in the interview. This is because, after reviewing a multitude of primary and secondary sources, the historian decided the interview was worth putting in the book. However, the interview could still be considered a primary source for personal details about the officer that were not verified by the author. Jc3s5h (talk) 13:51, 15 May 2012 (UTC)

I asked this question in relation to the work of Wendy Lowenstein, a widely recognised social historian of Australia who used oral history methodology and published her results as extensive sourcebooks of interviews. RS/N's opinion was that relying on material selected by a historian and contained in a sourcebook is unreliable and original research. However, I would agree that a historian who quoted extensively in a work under their own name (a monograph) would lend their credence and authority to that quote (as long as they didn't rip it to shreds or impugn its account while quoting, ie: as the lying liar Liar Liarson lied, "…"). Fifelfoo (talk) 22:33, 20 May 2012 (UTC)

Sounds like a good idea but in the wrong place. That would be a good change where it belongs, wp:notability. North8000 (talk) 00:32, 15 May 2012 (UTC)

Masem hits the nail on the head. If the article is published in the New York Times, it's evidence of notability. If an article is published by a fanzine, that's not evidence of notability. A Quest For Knowledge (talk) 00:51, 15 May 2012 (UTC)

Exceptions for primary sources

The policy states that "Do not base an entire article on primary sources, and be cautious about basing large passages on them." Can an article be majorly based on primary sources which have received coverage in tons of news reports from highly reputed newspapers i.e. do reports in loads of reliable secondary sources over-rule the fact that the site is a primary source and can the policy be flexed for that reason? Secret of success (talk) 12:06, 20 May 2012 (UTC)

Could you give us more specifics? A lot depends on what the article is about... For example, the "plot summary" section of an article about a book, would be appropriately based entirely on the book itself... but other sections of the article (such as a section on how the book was received) would not. Blueboar (talk) 14:00, 20 May 2012 (UTC)

The policy doesn't really distinguish between primary sources used a lot secondary and those which are used less. But if a primary source is evaluated/analyzed in some secondary source, there is usually no reason to cite primary source at all, but you simply use the secondary source instead.--Kmhkmh (talk) 22:09, 20 May 2012 (UTC)

I was specifically referring to box office figures for films, which are without doubt, one of the most sensitive issues across Wikipedia. I have seen articles like this being sourced mostly to primary sources. It has had two AfD's in the past and both of them have resulted in a "keep". Now, I do not want to discuss that specifically, my comment above was more general.

If content from a primary source is put up in a secondary source with attribution and some amount of analysis, does it mean that the secondary source endorses the content? I don't think that is the case, right. Secret of success (talk) 09:13, 21 May 2012 (UTC)

The mere use of a primary source does not necessarily mean the secondary source "approves it", you need to read secondary source to see what it does with the primary source. I was saying if you have secondary sources using a primary source for specific information, then usually you get that specific information you want from the secondary source anyway, hence you don't the primary source anymore and arguing whether you could use or not is a bit of a moot point. As far as box office figures are concerned, I see no problem with citing (standard) primary sources for it, but they should be clearly attributed. However these hinges a bit on the question as how reliable you assess a primary source in the first place and you could consider its general use in secondary sources as one factor (of many) to assess the reliability of the primary source. If BoxOffice.com figures are widely used in media and literature, than you consider that as hint that their data is generally accepted/considered reliable. -Kmhkmh (talk) 09:40, 21 May 2012 (UTC)

Academic question: Does recognizing something count as OR?

There's a discussion going on at Talk:Blackwater (Game of Thrones) about whether or not certain types of content from primary sources count as original research. I refer to the discussion as academic because there does not seem to be any disagreement (at the moment) about what the text of the article should read. In these cases, the episodes themselves are considered primary sources and off-Wikipedia articles about the episodes are considered secondary sources.

If a work of fiction has a character say, "To be or not to be. That is the question," is it original research for the Wikiarticle to say, "Character Bob Smith quoted part of Shakespeare's Hamlet"? If so, which part of WP:OR covers this?

If a work of fiction has a character hum or sing a song, is it original research for the Wikiarticle to say, "Foghorn Leghorn can be heard humming 'Camptown Races' in the episode 'Fearsome Foghorn'"?

Does it make any difference if the content is well known or obscure? The statement that set off the debate in question was, "Tyrion Lannister can be heard whistling 'The Rains of Castamere' in a previous episode." (This statement was later backed up with a secondary source.)

There seems to be an unspoken consensus that no secondary source is needed to identify characters by their faces ("Bob Smith appears in a scene with John Jones") or to identify words by sound ("Bob Smith says 'Hello, I'm Bobby' to John Jones in the second episode") or to describe events ("Bob Smith can be seen to smack John Jones in the shoulder in the third episode"), and most plot summary sections consist primarily of these kinds of statements. If identifying melodies is different in some way, then which part of WP:OR covers this? Darkfrog24 (talk) 05:45, 10 June 2012 (UTC)

It isn't OR. It is a simple descriptive statement about something that occurs in the work of fiction that a reader can verify by looking at the work of fiction itself.

That said, the fact that the statement isn't OR does not mean we necessarily must (or should) include it. A good plot summary will not include every tiny trivial action made by every character, it will only mention the important ones. So, the next question is... is the fact that the character sings that particular song at that particular point in the plot really worth mentioning in our article? Is this action central to the plot, or is it trivial? Blueboar (talk) 11:43, 10 June 2012 (UTC)

I agree that just because it's not OR doesn't mean it merits inclusion; that's a separate issue. The article in question has a small section about music. The song "Rains of Castamere" features rather heavily in the episode, and the article says where the lyrics came from, who wrote the song, and who performs it in the credits. I figured it was relevant to mention that this was not the first time that it appeared in the series. Feel free to see for yourself. The section is quite short.

The guy I've been talking to was pretty adamant that listening to the melody and concluding that it was the same song as another string of pitches was OR. Darkfrog24 (talk) 14:10, 10 June 2012 (UTC)

Doesn't WP:MOSFICT#Plot summaries describe about plots and refer to some essay and help about them? I think it describes the general consensus about this area saying what people agree is about okay as a straight factual description. Dmcq (talk) 13:31, 10 June 2012 (UTC)

Original Research from reliable primary sources for exclusion, not inclusion

Please discuss -

“When a secondary source is in conflict with a reliable primary source, then the reliable primary sources can be used to exclude in information, but not include it.”

An example is when a line stating living person’s age has a reliable mainstream newspaper as the source, but the very reliable primary source of government birth records indicates the newspaper made an error, then the information about age should be removed from the article based on the reliable primary source, but not included from the primary source which is original research. PPdd (talk) 13:47, 4 April 2012 (UTC)

When reliable sources conflict, the simplest solution is for the article to say "source A says X, whereas source B says Y". Deciding whether A or B is more likely to be wrong cannot be reduced to a simple rule like one being primary and the other secondary. jnestorius^(talk) 16:50, 4 April 2012 (UTC)

PPdd your assumption that government records (like birth certificates) are inherently more accurate than newspapers is flawed. This is not always the case. Let me present an example: If you were writing an article about me (assuming I was notable), and you looked at my birth certificate, you would read that my mother was born in Manilla, RI (ie in Rhode Island)... in fact, she was born in Manilla, PI (ie in the Philippine Islands). What obviously occurred was that somewhere in the process of filling out my birth certificate, someone swapped an R for a P (easy enough to do). Thus, my birth certificate contains erroneous information. Now, if I were to be interviewed by the newspaper, and they asked me where my mother was born, I would say "in the Philippines". The newspaper account would be accurate.

Of course, there is no way for you (the person writing an article for Wikipedia) to know that the newspaper is accurate and my birth certificate is not. All you know is that there is a discrepancy. So you have two options a) Look for for further sources that will clarify the facts, or b) present the discrepancy by mentioning what both sources say. Blueboar (talk) 18:16, 4 April 2012 (UTC)

To Blueboar, to make sure I am understanding your explanation correctly; in the specific example of confusion over a living person's age or birthplace (or something similar), an outright statement by the living person in question, intentionally given to debunk any false or misleading statements about the issue in dispute (DOB, birthplace, etc.), would trump any inaccurate legal documentation, incorrectly sourced media statements, deliberately-made slanderous or libelous statements, or any other source that conflicts with their own from-the-horse's-mouth statement? (Assuming, of course, that the Wikipedia author/editor was the person this statement was made to) Blozier2006 (talk) 22:31, 19 April 2012 (UTC)

Not necessarily. I certainly tend to give a lot of weight to published statements by the person in question, but to know comparative weight between sources (ie which source outweighs or "trumps" the other) I would have to look at the specifics of each case separately. Most of the time, neither source "trumps" the other... so we present what both have to say. In my example, I would say something like: "Blueboar's birth certificate gives his Mother's birth place as Manila, RI (Road Island) <cite birth cirtificate>. In an interview given in the Boston Tribune, Blueboar claimed that his birth certificate contains a clerical error and that his Mother was actually born in Manila, PI (Philippine Islands) <cite Tribune interview>. Blueboar (talk) 23:25, 19 April 2012 (UTC)

OK, thank you for the clarification. Blozier2006 (talk) 23:50, 19 April 2012 (UTC)

And just to complicate things... we also need to keep in mind that there might be other sources that confirm (or debunk) either source. This too impacts how much weight to give each source. Blueboar (talk) 00:21, 20 April 2012 (UTC)

A birth certificate is not a reliable source unless it has been published, as for example the American president's birth certificate. But that is rare. If a newspaper adds ten years to your age, you should get the paper to issue a correction, rather than to correct it here. The typical conflict between a primary and secondary source comes when a secondary source misrepresents a primary source, in which case we could just use the primary source. For example, a newspaper reported that x said y, while a published transcript says that he said z. TFD (talk) 00:38, 20 April 2012 (UTC)

A birth certificate is published if it is available to any member of the public who wishes to pay the fee; this is the case in some US states. A birth certificate that is only available to certain select persons is not published. Jc3s5h (talk) 00:49, 20 April 2012 (UTC)

Blueboar is correct that birth certificates often contain errors, but newspapers contain errors vastly more often. So, if we prefer Wikipedia to be right more often than wrong, while maintaining our ban on original research, we ought to allow birth certificates that are publicly available and not allow those which are not publicly available. Of course, if a secondary source has examined the birth certificate and declared it mistaken, we have to go with that. Zero^talk 14:31, 28 April 2012 (UTC)

We do allow birth certificates and other such publicly available documents ... however, we should also remember that such documents are primary sources, and as such are subject to the limitations and cautions of all primary sources. They can be used, but should be used with caution and care. Blueboar (talk) 14:47, 28 April 2012 (UTC)

I believe the point is whether a source explicitly states what is claimed by wikipedia. The problem with primary sources is that they usually need to be synthesised to be able to draw useful conclusions. If the primary source says: the Mona Lisa has black hair and there is no discrepancy between sources, than you can state that she has black hair. If however one source says grass is green when it's alive and a second source says that it's brown when it's dead than you may not conclude that it turns brown when it dies unless you reference a secondary source which makes that conclusion for you. That being said: a birth certificate is often not a primary source, I don't know how it works in the USA, but if the data has to be reported to an official by a parent or a doctor, it's a secondary source (the primary source being the report from the doctor or the parent). PinkShinyRose (talk) 00:10, 27 June 2012 (UTC)

Which is why we don't allow synthetic OR from primary sources—encyclopaedists lack the capacity to make expert and original judgements about primary material. If you want to publish biographic compendiums, then the free online publishing industry is that way (but our reliable sourcing standards are high). Fifelfoo (talk) 00:46, 27 June 2012 (UTC)

A birth certificate is not a secondary source.[4] [5][6][7] It's not a matter of counting up links in the chain. WhatamIdoing (talk) 03:06, 27 June 2012 (UTC)

Fringe, established facts and synthesis

The WP guideline on synthesis may have a major weakness when it comes to attacks by fringe authors on established facts. Image the following: some recent author A makes the obviously ludicrous assertion that Tower Bridge was constructed by 15th century Chinese and he is cited so in the WP article. No other reliable author ever cared to address and refute this statement, hence it remains undisputed' so to speak. Now some editor takes an older book B which states that the bridge was built in the 19th century by the British and adds this statement as a rebuttal. He is, however, reverted on the grounds that this would be synthesis because B did not refer to A explicitly. Consequently, the statement of B is removed as refutation, even though it is an established fact, while A's statement remains in the text as undisputed. Was this right? This case does not seem to have been covered by WP:SYN properly, a kind of loop-hole which favours outlandish and more recent claims. Gun Powder Ma (talk) 15:25, 6 June 2012 (UTC)

That isn't synthesis. That is a source simply being wrong and there being a lack of other sources which is a different problem. An example of synthesis would be 'The Chinese made the earliest iron bridges. Tower Bridge is one of the most famous iron bridges made before the twentieth century'. Dmcq (talk) 17:04, 6 June 2012 (UTC)

So even though the source isn't discussing Menzies, we can use it? At Where Troy Once Stood this edit[8] of mine was, I think correctly, removed. Can I reinstate it? The actual edit that Gun Power Ma is discussing is [9]. Dougweller (talk) 18:11, 6 June 2012 (UTC)

Who are you replying to? If you could state what you see as the problem in a straightforward way it would help. Personally I can't see why anyone would worry about some fantasy pseudohistory like those but if people have gone out of their way to write refutations I suppose they should be in too. Dmcq (talk) 18:27, 6 June 2012 (UTC)

Oh no no MEnzies stuff please:-).--Kmhkmh (talk) 18:30, 6 June 2012 (UTC)

The point is that none of these sources were written as refutations, they are simply about the same subject. The question seems to be should our policy allow us to use them, even though in these two examples they weren't written in response to the authors' books. Dougweller (talk) 20:27, 6 June 2012 (UTC)

Synthesis is using two or more sources to advance a position not advanced by any of the sources. If all the sources are being considered to support an article, there is no problem with ignoring a fringe source and not mentioning it in the article. It's a different story if the article is about the fringe source. Setting out "facts" purported by the fringe source and listing sources that contradict the fringe source could be viewed as a Wikipedia editor creating a novel review of the fringe source. If no reliable sources can be found to review the fringe source, the article about the fringe source should probably be deleted. Jc3s5h (talk) 20:48, 6 June 2012 (UTC)

SYNTH would be something like "Source A says the Tower Bridge was built by Chinese laborers. Source B says it was built by the British. Therefore (here's the SYNTH part), China and the UK are the same place." When different views exist indifferent sources, simply stating what the different views are is not a SYNTH violation, because yu're not adding up the views to get something entirely new that can't be found in any of the sources. WhatamIdoing (talk) 14:16, 7 June 2012 (UTC)

You don't have to even say the conclusion, just imply it as in my example near the start implying the Chinese built Tower Bridge using two perfectly okay and true statements. Dmcq (talk) 14:29, 7 June 2012 (UTC)

Does it matter that these are book reviews? Dougweller (talk) 16:40, 7 June 2012 (UTC)

If they are published in a reliable source like a newspaper of journal I cant see why not and they'd be good secondary sources. Dmcq (talk) 22:37, 7 June 2012 (UTC)

No, it is the articles that are book reviews. The authors of the sources may not even have ever heard of the books in question and they certainly aren't commenting on them. Dougweller (talk) 13:01, 8 June 2012 (UTC)

Sounds very strange, but if a book review has been commented on in a number of other ways in some notable way then I guess so. I suppose someone could write a notable book review if they were very good authors themselves or particularly biting critics who attracted attention or something like that. Dmcq (talk) 13:13, 8 June 2012 (UTC)

Let me try again. There are no book review involved here. We are talking about Wikipedia articles about books, or sections of a biography that discuss a book, where sources are being used that do not mention the book or the author or even the subject of the book and are being used to refute the book. As I understand it, this is not acceptable according to our NOR policy. Dougweller (talk) 14:47, 9 June 2012 (UTC)

Perhaps you should just point at the discussion on the point that concerns you. Dmcq (talk) 20:30, 9 June 2012 (UTC)

I think my answer to the original question is that author B should not be mentioned. However the article should not state what A says as fact, it should only be stated as what A asserted. It is not up to us to write refutations but we don't have to write rubbish up as truth. Dmcq (talk) 20:37, 9 June 2012 (UTC)

Thanks, I agree entirely. Dougweller (talk) 06:02, 10 June 2012 (UTC)

Well... I don't believe it's necessary for the "refuting" source to mention the book by name. Let me give an example: some <insert disparaging term here> in Hawaii have come up with this silly theory that breast cancer is caused by wearing bras. They wrote a book about it. Every single mainstream source that addresses the question says that their idea is wrong. Almost none of those sources mention their names or the title of their book. They just say "Hey, if you heard this rumor that clothing causes breast cancer, it's wrong". But we use those sources to present information about their silly idea, because there's really no other way to meet NPOV and FRINGE requirements to present a disproven, tiny-minority position as being a silly, disproven idea.

So I'd say that if New Book says that the Tower Bridge was built by 15th century Chinese, and mainstream experts have a very different idea, then you can safely present the mainstream position as being the mainstream view. You can't say "Alice Expert rejects this book's claim", but you can say "Mainstream expert opinion holds that the bridge was built by British laborers" (or whatever). WhatamIdoing (talk) 19:30, 16 June 2012 (UTC)

New edits to WP:CALC

A recent edit was made in a good faith attempt to clarify WP:CALC (see this diff). I have reverted because I think the edit needs some discussion, especially the idea that pulling numbers from multiple sources is in all cases an improper synthesis - and therefor the numbers need to come from a single source. I am not sure this idea is completely accurate. Suppose I want to calculate the basic land area of the State of Virgina is 1850. We know that, at that time, Virginia included what is today West Virgina, but otherwise its boarders were the same as today. Now, suppose I have a source that gives me the modern area of Virginia, and another source that gives me the area of modern West Virginia... simple addition will give me the area of 1850 Virgina. This is exactly the sort of basic calculation that CALC was intended to allow. However, the edit would disallow it. Blueboar (talk) 11:48, 16 June 2012 (UTC)

I think that as long as its clear what the units/scale are and how absolute/unquestionable they are. Take an example where I attempt to add 1 million homes with Tivo boxes and 2 million homes with DVRs (from two different sources) and attempt to say that a total of 3 million homes have some type of digital recorder. The problem is that Tivo is a type of DVR and unless source 2 is clear that they didn't count Tivo in their DVR #s, then that 3 million may be double counting. --MASEM (t) 12:51, 16 June 2012 (UTC)

My guess is this is to counter the business where editors agree a calculation is okay in consensus but the calculation is not some standard one for the application. There was quite a row over various editor own calculations in the articles on usage share of web browsers and operating systems putting in a median of the various figures published by outside agencies. The articles no longer have the median figures they calculated so it doesn't look like the change is necessary for that purpose, and I'm not altogether sure the ones pushing the calculation in would have agreed with the interpretation even so anyway as they were arguing that what they did was useful and a fairly standard technique. Dmcq (talk) 12:58, 16 June 2012 (UTC)

The edit would often disallow presenting material in the same units of measure so that numbers in the article could easily be compared. For example, if an article discussed energy usage, home-heating sources might give energy usage in therms and barrels of oil, and electricity-related sources might use kilowatt hours. To make these diverse units comparable, the article might use megajoules. The conversion factors might be found in a different source from the sources that give the energy usage figures. So under the proposed change the unit conversions would be disallowed. Jc3s5h (talk) 13:40, 16 June 2012 (UTC)

I think we all agree that it certainly is possible to create an improper calculation by taking numbers from multiple sources... I hope that we all agree that it is also possible to create an absolutely appropriate calculation by taking numbers from multiple sources. My concern about the edit was that it went too far... phrasing what probably should be a "sometimes" caution as if it were an "always" rule. WP:CALC is one of those things where we simply can not make an "always" rule ... Ultimately, we have to examine each calculation on a case-by-case basis and determine whether that specific calculation is proper or improper. Blueboar (talk) 13:52, 16 June 2012 (UTC)

I would go along with Blueboar. In essence, the paragraph that was reverted didn't add anything to the article. A note describing the pitfalls of using numeric data from multiple sources might be more appropriate, but as an essay, not as part of this partiuclar article. Martinvl (talk) 14:23, 16 June 2012 (UTC)

I am the one who made the bold edit. It was an attempt to avoid situations like the one Dmcq was referring to. I do not agree that the edit would in any way disallow presenting material in the same units of measure as other sources. Converting units of measure of one source is clearly just a single source calculation and does not involve multiple sources, even if the goal of the calculation is to converts numbers from multiple sources into comparable units. I agree that there are some calculations which are relevant across multiple sources. But if they are relevant, surely the editors should be able to find a source which supports that calculation. That was the reason why I put in the phrase about a source directly supporting the calculation involving the sources. Useerup (talk) 16:34, 16 June 2012 (UTC)

Well, let me give you an example, and you tell me whether this is improper OR in your mind:

Source 1 says that 100,000 white Americans will get Scary Disease this year.
Source 2 says that 100,000 non-white Americans will get Scary Disease this year.
I add it up and say that 200,000 Americans will get Scary Disease this year.

Is that okay in your mind? WhatamIdoing (talk) 19:34, 16 June 2012 (UTC)

The proposed change disallows unit conversions unless the conversion factor is contained in the same source as the unit to be converted. I reject Useerup's claim to the contrary. Jc3s5h (talk) 20:03, 16 June 2012 (UTC)

@WhatamIdoing: Yes, that is OR. Even though you used perfectly complementary concepts (white and non-whites) it is not at all clear that the sources could be used that way. Rarely are multiple sources directly comparable that way. The sources could use different standards for accepting a diagnose or use different standards for "Americans" (native, immigrants, natural born or naturalized?, North Americans or just US citizens). In those situations you do *not* add up the numbers, you report on each of the sources. Useerup (talk) 01:01, 17 June 2012 (UTC)

Would you feel the same if the example was "males" and "females" or "adults" and "children" rather than race? What if the sources were actually papers published by the same people, just split up so they could get two publications on their CVs?

Also, do you think that the typical editor would object to my example? WhatamIdoing (talk) 04:20, 17 June 2012 (UTC)

If the papers were published by the same people I would assume that they would reference one paper from the other. That would actually open the door for combining over multiple sources. If the 1st paper was referenced from the 2nd paper in such a way as to compare them or include the conclusions it opens the door for calculations. But think about the opposite: What if two papers were written by the same people, but did not reference each other. Would you then be comfortable drawing conclusions by calculating numbers across the papers, given that it would have been obvious to the authors? Useerup (talk) 06:56, 17 June 2012 (UTC)

Yes, I'd feel perfectly comfortable combining these sources. I'd feel comfortable doing this even if the sources were written by different people. I am, for example, happy to take three solid sources on epidemiology, each of which provide an incidence for one of the three countries in North America, and do a simple addition to produce a statement about "The incidence in North America" rather than "The incidence in Mexico... in the US... in Canada".

You have not answered my second question: Do you think that the typical Wikipedia editor would object to adding up such numbers? WhatamIdoing (talk) 16:45, 25 June 2012 (UTC)

@Jc3s5h: Can you propose a change which would in your opinion allow unit conversions? I certainly agree that unit conversions should be considered allowed under the policy; they are (usually) non-controversial and straightforward. Useerup (talk) 01:01, 17 June 2012 (UTC)

I think the policy was fine before the edit. When doing calculations involving different sources (or even different pages within the same source) it is important to make sure one is comparing apples to apples, but I don't think there is any concise formulation that will disallow the incorrect calculations and allow the correct calculation. But a service that an encyclopedia should perform for its readers is to collect information from diverse specialized sources and present an overview that is easier to quickly understand, and this requires activities such as unit conversion and other calculations so that data from diverse sources is more readily comparable. Of course there will be times when the exact nature of the data in various sources is not clearly enough stated to decide if data may be combined, in which case some other approach will be needed. But I think this should be looked at on a case-by-case basis on each article talk page. Jc3s5h (talk) 12:55, 17 June 2012 (UTC)

I think that the bold changes that Useerup made were absolutely appropriate. Those edits are just an explicit restatement of the principles in WP:SYNTH - specifically "Do not combine material from multiple sources to reach or imply a conclusion not explicitly stated by any of the sources". The principles in WP:CALC are extremely useful because they empower users to restate quantitative claims made by a source into the form demanded by the article. The use of routine calculation is analogous to indirect quotation - you are simply restating verifiable claims in the matter most appropriate to the text. However the principles of WP:SYNTH should always be observed to avoid novel claims not made by any single source; and calculations made with data from multiple sources to generate new data not directly attributable to any given source is clearly such a violation. I think, as is pointed out in the discussion above, it certainly is possible to correctly do calculations from multiple sources to reach a true conclusion; however whats at issue is not whether the calculations are correct or the whether the new number is true, its whether it is verifiable and directly attributable to a reliable source. Solid State Survivor (talk) 07:47, 26 June 2012 (UTC)

You then run into the problems that exist with the present text of the verifiability policy page that are as of yet not settled. One of the problems is that you can have verifiable facts that in the relevant field are entirely uncontroversial to state, yet they cannot be attributed to a single source. E.g., in physics and chemistry, it is routine to combine information from sources like reference books to do computations. Then if e.g. we have some list of physical properties of materials here and we want to present that in some way, we should be free to do the relevant computations to extract the correct figures. As long as what we do doesn't deviate from how every professional in the field uses the sources, there shouldn't be a problem here. Count Iblis (talk) 16:21, 26 June 2012 (UTC)

RfC on the verifiability policy lede

Wikipedia:Verifiability/2012 RfC

Hello all. I'd like to draw your attention to an RfC about the lede of the verifiability policy. We have been drafting this RfC for some time as part of a MedCab mediation, and it is finally ready for comment. In the RfC we have included a few specific drafts of the policy lede for you to comment on, and we have twelve general questions to find editors' views about how the lede should look. All editors are warmly invited to join the discussion at the RfC page. Thanks! — Mr. Stradivarius on tour ^{(have a chat)} 02:04, 29 June 2012 (UTC)

Newspapers (again...)

I just want to make sure I've not missed something that may have changed in regards to the primary/secondary nature of newspapers.

There's been discussion about the appropriateness of "X On Twitter" articles (This is not to clear that issue up). A discussion I started is at WP:VPP#Appropriateness of "X on Twitter" (or similar) articles. One protracted issue that has come up is the nature of newspapers articles as primary or secondary sources. Last I checked, and this is consistent with the wording of the policy page and other places on WP that talk about primary/secondary sources, newspaper sources are generally considered primary if they are reporting on a topic and not providing any type of analysis or critical assessment; another user (Alanscottwalker (talk · contribs)) is insisting because the newspaper is not the original source and had editorial control to chose to publish it, they're secondary sources.

As the case in point, I found an example of an article that is talking about one of the celebs with a Twitter account, and the article quotes one of the celebs tweets - and says nothing else about the Twitter account [10]. In considering the application of that article to a WP article specifically on the celeb's Twitter account (and not on the celeb himself), that, by all standards that I've read and checked is flat out a primary source. (It may be secondary for the celeb, but certainly not for the Twitter account). Alanscottwalker insists that its secondary because the news-reporting group used editorial control to select that quote and thus implicitly added an evaluative aspect to it, thus making is secondary.

I want to make sure I'm not mistaken in that Alan's view here specifically on this example is completely at odds with standard consideration of primary and secondary sources. --MASEM (t) 12:43, 4 July 2012 (UTC)

You're right. The term secondary was used for so long as a synonym for any positive quality (independent, fact-checked, peer-reviewed, non-self-published, etc.) that some editors are pretty confused. See WP:PRIMARYNEWS for a summary. WhatamIdoing (talk) 02:30, 16 July 2012 (UTC)

IUPAC names for chemicals, especially for drugs

The International Union of Pure and Applied Chemistry (IUPAC) has a nomenclature for chemical substances, so that one can derive a name if the chemical structure is given. (Resulting in names such as "2-Amino-4-(ethylcarbamoyl)butyric acid" for theanine.) For the sake of completeness: sometimes there is more than one way to derive a name for a structure. Both {{chembox}} and {{drugbox}} have fields for IUPAC names, and one editor has started fact-tagging these a while ago. Now the problem is that the obvious sources (PubChem, Chemspider) aren't reliable, and articles in medical journals generally don't bother with chemical nomenclature. Even articles in chemical journals often have the structures but not the names. Deriving IUPAC names by ourselves is OR. Any way to solve this? --ἀνυπόδητος (talk) 11:19, 16 July 2012 (UTC)

It should be noted here, that in principle any creation of those names is 'original research' in a way (also outside Wikipedia) - the IUPAC has schemes (an algorithm) with which you can generate the name of a compound based on structure - they do not check or validate the names that one is giving to a compound - anyone can generate a name, publish that in a source and it is to the discretion of the peer review systems whether the names are actually checked (I have never seen that happen, and the reviewers would follow the same schemes anyway). --Dirk Beetstra ^{T C} 11:41, 16 July 2012 (UTC)

In violation of WP:BEANS, I'd like to mention here that SMILES and InChI (both having fields in chembox and drugbox) are also text representations of chemical formulae that can (and generally have to be) computed from the structures, so the mentioned problem applies to these as well. (Thanks to Dirk Beetstra for the tip.) --ἀνυπόδητος (talk) 13:49, 16 July 2012 (UTC)

The core of the problem is that generating systematic names from anything but simple chemical structures is complex. The set of rules (or more accurately, guidelines, I believe) described by IUPAC are not sufficiently comprehensive to make unambiguous IUPAC name assignments for complex multi-functional chemical compounds. There are plenty of computer algorithms out there that do a decent job, but they do not produce the same results in all cases. And since IUPAC does not produce or endorse any official algorithm (or other method), as far as I know, Wikipedia and other databases must work with what we have. If there is a way to determine a database that has the most reliable IUPAC names, we should go with that; but I'm not confident that could determined. In the absence of any "official" IUPAC name authority, perhaps we should just dispense with IUPAC per se, and simply go with a more general "systematic name", meaning any name from which one can unambiguously derive the correct chemical structure. In the mean time, we shouldn't be putting {{citation needed}} on any IUPAC name if we do not recognize any sufficiently reliable source for such a name. -- Ed (Edgar181) 17:31, 16 July 2012 (UTC)

But even "non-IUPAC systematic names" will be OR unless we find a reliable database, which is unlikely given the existence of computer algorithms. --ἀνυπόδητος (talk) 17:34, 16 July 2012 (UTC)

Yes, but I'm not too concerned about the OR aspect. OR concerns are important where subjectivity is an issue, but systematic chemical names can be objectively and readily verified by widely available chemistry drawing software or at various websites. -- Ed (Edgar181) 17:41, 16 July 2012 (UTC)

I'm concerned about OR because dozens (or more) systematic names have been fact-tagged, and neither PubChem nor ChemSpider are reliable sources. So maybe the question is whether using chemistry drawing software for verification counts as OR. --ἀνυπόδητος (talk) 17:46, 16 July 2012 (UTC)

Let's put a real example into your question: If I take the IUPAC name "2-Amino-4-(ethylcarbamoyl)butyric acid" which is currently in theanine and tagged with {{citation needed}} and I plug it into this webpage from Univ Cambridge which parses IUPAC names, it returns a chemical structure that matches what is in our article. Is that original research by Wikipedia standards, or not? And is that sufficient verification to remove the "citation needed" tag? -- Ed (Edgar181) 19:11, 16 July 2012 (UTC)

I came here from a note at WT:WikiProject Pharmacology. I don't really see a problem with OR in this case. It's not any more an OR violation than would be computing the molecular weight of a compound from the periodic table masses of the individual atoms, or converting a measurement to or from metric. The principle of OR is to prevent editors from passing off opinions and pet theories as facts. To reply to Edgar's specific example, I think it would be reasonable to remove the cn tag, while best practice would be, instead, to provide an inline cite to a book that contains the structure, which I bet exists somewhere. --Tryptofish (talk) 21:57, 16 July 2012 (UTC)

If no reliable publications on a chemical use its IUPAC name, why do we need to give one? Someguy1221 (talk) 22:01, 16 July 2012 (UTC)

It's not really the case that no reliable publications use IUPAC names - most chemical databases will list IUPAC names. The problem is that reliable sources can differ in the ways they determine the IUPAC name and Wikipedia chemists have not determined which otherwise-reliable sources are sufficiently reliable for IUPAC names. -- Ed (Edgar181) 23:27, 16 July 2012 (UTC)

I hope you all remember how the whole situation started. After this we are back full circle. --Dirk Beetstra ^{T C} 04:18, 17 July 2012 (UTC)

I was contributing to the section above and I lost my place and started reading your section. As a complete ignoramus of chemistry and pharmacology, may I give two-cents. First, NOR only applies when you think something is wrong and the first step is to request reference(s). Do you think that "2-Amino-4-(ethylcarbamoyl)butyric acid" is wrong for theanine or that it's one of many IUPAC names? If the latter then why not add all valid names in a list, with an inline citation, if anyone requests it? Isn't this what WP:WEIGHT is all about?

This is just an idea and I freely admit that my lack of knowledge about the subject may make my opinion moot. --RoyGoldsmith (talk) 15:27, 17 July 2012 (UTC)

We've no idea whether Plasmic Physics thinks this name is wrong, he just goes around and adds fact tags to chemical names. Everything else (there being several possible names and all) isn't a problem in my opinion. --ἀνυπόδητος (talk) 17:14, 17 July 2012 (UTC)

Actually, I think that RoyGoldsmith has it right. If, hypothetically, someone has a real factual doubt about a particular chemical name (apparently not the case here), and they discuss it on the article talk page, then it should be sorted out. But if someone is just going around adding cn tags, it's probably worth asking them why, but if they don't have a specific reason, it's reasonable to revert them and request that they stop.

(Another example occurred to me. If you look at the end of Theanine, there are a bunch of interwiki links to Wikipedias in languages other than English. In all likelihood, those links got there from editors translating the title into/from other languages, but it would be a stretch to call that translation original research as it's defined here. We don't expect someone to cite a dictionary that such-and-such is the (language) word for theanine, and it would be silly to put a cn tag on such a link. If someone realizes that the link is incorrect, then correct it, but there's no need to worry about OR.) --Tryptofish (talk) 20:39, 17 July 2012 (UTC)

As a technical matter, if any reliable source in the world has ever claimed that this is the (or "an") IUPAC name for a given chemical, then there is no NOR violation. Providing a citation proves there is no violation, but if such a source exists, then there is no NOR violation even if there is no citation.

I would like to know why this person believes that his tagging is helpful. WhatamIdoing (talk) 00:42, 18 July 2012 (UTC)

I agree that there is no NOR violation here. For any given chemical structure, there are a number of valid IUPAC names that can be unambiguously used to re-derive the structure (many to one relationship). There is also something called the Preferred IUPAC name, which is supposed to be a unique, canonical name for the structure (one to one relationship). The fields in the chembox (IUPACName) and drugbox (IUPAC_name) are for any valid IUPAC name and does not mandate a preferred name. As long as an chemical software package can regenerate the structure from the supplied IUPAC name, that should be sufficient justification for including the name in an infobox and no citation is required. Boghog (talk) 05:49, 18 July 2012 (UTC)

Do note, that the rules for the Preferred IUPAC name did not finish yet (I am 'following' the discussions that should lead to them), the official guidelines/rules for that have not been published in a finished form. --Dirk Beetstra ^{T C} 05:53, 18 July 2012 (UTC)

As I said, we are going full circle. This situation started a long time ago (I'll try and find the discussions), where self-generated names were put on chemicals, where there were serious concerns that those names were not correct and many were reverted. IIRC, that self-generation of names was deemed original research, and hence, people were asking for references for those self-generated names ... --Dirk Beetstra ^{T C} 04:01, 18 July 2012 (UTC)

Oh, it is too much to read - many threads in Wikipedia_talk:WikiProject_Chemicals/Archive_2010

moving into 2011

(I may finish this list later, but there are more in the archives of WT:CHEM). I've marked some with a <-, which are the ones which are of interest. --Dirk Beetstra ^{T C} 04:16, 18 July 2012 (UTC)

Responding to Tryptofish 20:39, 17 July: A minor point but according to WP:V, it's not legit to revert CN-tags (see WP:BURDEN):

The burden of evidence lies with the editor who adds or restores material. [In this case, the IUPAC name.] You may remove any material lacking an inline citation to a reliable source.

In other words, the person who objects doesn't really have to insert CN's -- that's just a courtesy. They can delete any statement without an in-cite immediately, just like BLPs. Then it's up to the editors who believe in that IUPAC name to find a reliable reference.

Reponding to Dirk Beetstra 04:16, 18 July: OK. Then what would you like as a resolution to this section? Has ἀνυπόδητος gotten his answer? --RoyGoldsmith (talk) 00:43, 19 July 2012 (UTC)

Yes, it's a minor point, but sometimes BURDEN becomes too much of a burden, the way it gets interpreted. I agree that it's bad form simply to revert the tag when a cite is legitimately needed, but it seems to me to be perfectly reasonable to revert one that was placed either in error or stupidly. Please note that I did specify that one should attempt to consult with editor placing the tag, before considering reverting it. --Tryptofish (talk) 20:16, 19 July 2012 (UTC)

Sometimes you really should revert fact tags. The rules about not rewarding vandals and pointy behavior still apply. WhatamIdoing (talk) 20:33, 19 July 2012 (UTC)

I love that diff! It's on my list of the best one, two, three, four.... --Tryptofish (talk) 21:19, 19 July 2012 (UTC)

Eyes

I'd appreciate it if some more people kept an eye on Template:Content policies for a little while. It seems to be attracting links to pages that aren't content policies. WhatamIdoing (talk) 16:33, 23 July 2012 (UTC)

WP:WAF conflict with WP:NOR?

I am confused. Recently I added this talk section (please read) for the TV series The Big Bang Theory. In it, I warned that I would delete a section in the article that contained about 200 lines. My rationale was that this section was either (a) almost entirely based on the episodes of The Big Bang Theory and that is regarded (see WP:TVPLOT) as a primary source, which allows no interpretation (see WP:PSTS), or that (b) the majority of statements in the section should have inline citations to secondary sources.

However, the WP:WAF guideline, which covers classic plays as well as comic books, talks about in-universe perspective, based on the work itself. It implies that we can use in-universe perspective for fictional works, with limitations. This is even more egregious on the talk page (for example, see this). WP:NOR says nothing about in-universe perspectives or their effect on primary sources.

The policy under WP:NOR seems to be:

Any interpretation of primary source material requires a reliable secondary source for that interpretation. A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the source but without further, specialized knowledge. For example, an article about a novel may cite passages to describe the plot, but any interpretation needs a secondary source. Do not analyze, synthesize, interpret, or evaluate material found in a primary source yourself; instead, refer to reliable secondary sources that do so.

Normally, I would say that policies trump guidelines but the history of WP:WAF gives me pause. Allowing some in-universe perspective is not a new idea. The majority of articles about fiction do use in-universe perspective for things other than the plot.

Does a "straightforward, descriptive statements of facts" include fictional "in-universe perspectives", which, of course, must be based on the literary text, a primary source? I would like a clear statement in WP:NOR about using the work itself for fiction.

--RoyGoldsmith (talk) 17:05, 13 July 2012 (UTC)

The plot guideline seems to be well accepted. According to that one can put in a short neutral description of a plot. The in-universe bit only extends to describing things and one shouldn't assume too much in-universe background. You can say 'he erected a psychic shield' without worrying about the real world existence of such things, but one should be careful to avoid interpretations like 'in a satire of Soviet bureaucracy' unless they have been made very obvious or there is a secondary source. This is interpreted as being in line with only using primary source for facts. Dmcq (talk) 18:53, 13 July 2012 (UTC)

First, where would I find the "plot guideline" you refer to? Second, I would like to get your opinion. Which of the following article sections would need secondary sources:

Characters, see Hawaii_Five-0#Main_cast or Alphas#Main_characters;
Settings, see General_Hospital#Setting;
Themes, see The_Big_Bang_Theory#Elements_of_the_show;
Season summaries (in addition to episode summaries), see Private_Practice_(TV_series)#Series_Overview

I've taken all the examples from TV series; believe me, comic books would be far worse.

As a comparison, look at The Simpsons featured article. It has secondary source inline citations for Characters, Setting and Themes, as well as Hallmarks and Culture. --RoyGoldsmith (talk) 03:31, 14 July 2012 (UTC)

Non-interpretive summaries of primary works are completely acceptable. If you can find secondary sources to back it up, great, but it's generally not required if you're summarizing a single work. But as you move towards series-wide themes for things like TV shows or comics, broader claims may become a bit more interpretative depending on how they are taken, and secondary sources should start being used more. But again, if the claim of a broad scope is obvious to any reasonable person that would have access to the primary source (eg, the BBTheory has a lot of science-related humor) it doesn't need sourcing. --MASEM (t) 03:54, 14 July 2012 (UTC)

See WP:PLOTSUM for the plot summary guidelines. Dmcq (talk) 04:27, 14 July 2012 (UTC)

As I said before, I only want to bring NOR policy in line with current consensus. In my opinion, all statements within Wikipedia guidelines and essays should be based on some policy (or, failing that, the lack of any applicable policy). How about the following change (in boldface) to one sentence of WP:NOR that I quoted above:

Any interpretation of primary source material requires a reliable secondary source for that interpretation. A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the source but without further, specialized knowledge. For example, an article about a novel may use plot elements to describe the characters, themes, etc., but any interpretation (such as literary criticism) needs a secondary source. Do not analyze, synthesize, interpret, or evaluate material found in a primary source yourself; instead, refer to reliable secondary sources that do so.

--RoyGoldsmith (talk) 13:33, 15 July 2012 (UTC)

That's just unnecessary cruft. More is less with cruft. Dmcq (talk) 01:22, 16 July 2012 (UTC)

Actually about 95% of Wikipedia (everything except direct quotes) violates a rigorous literal interpretation of wp:nor/ wp:synth. But the de-facto policy is to bend it a little, allowing straightforward neutral summarization. North8000 (talk) 17:36, 15 July 2012 (UTC)

You've said that for years, but nobody agrees with you that everything eccept quotations and copyvios is a NOR violation. In fact, the third paragraph of the lead to this policy directly tells you that you are supposed to summarize sources in your own words, while retaining the meaning of the sources. WhatamIdoing (talk) 02:43, 16 July 2012 (UTC)

I understand what North8000 is saying. He is talking about an extreme interpretation where every editorial action should be based on secondary sources. The answer is that the level of synthesis that is forbidden is that which would be considered creative. In other words, WP:NOR says that editors are not allowed to do themselves the sort of transformation that converts primary source material to secondary source material. --SmokeyJoe (talk) 03:47, 16 July 2012 (UTC)

Actually, it is my impression that North believes that "everything except direct quotes" is a violation of a strict reading of NOR. Like I said, nobody else agrees with him, but he does seem to sincerely believe that summarizing a multi-page source into a single sentence or paragraph, even if your summary is obviously a plain, neutral summary of the source, "violates a rigorous literal interpretation of wp:nor", to use his words. WhatamIdoing (talk) 00:26, 18 July 2012 (UTC)

Given that WP:NOR, despite being "core policy", is flexible and not prescriptively written, in the phrase "is a violation of a strict reading of NOR", I'd equate "strict" to "mis-". WP:NOR is not intended to be read that way. I'd ask North8000 if many editors on the ground are reading it that way. --SmokeyJoe (talk) 00:55, 18 July 2012 (UTC)

I think that Roy has merely rediscovered that applying the advice of WP:WAF causes violations of WP:NOR to become obvious. Go back to the previous in-universe form and look carefully; it would have contained the same OR only better disguised. --SmokeyJoe (talk) 03:53, 16 July 2012 (UTC)
- It's not obvious; NOR doesn't depend on if its in- or out-of-universe. (that's a matter of writing style than the correctness of the information.) You need to be a more explicit to say what the problem is, as as give, WAF follows standard practice for using primary sources to summary primary works. --MASEM (t) 05:33, 16 July 2012 (UTC)
  - When something is written in-universe, it is very easy to include multiple references from the work of fiction itself to support an analysis. When writing out-of-universe, the work-of-fiction references no longer suit, and then the lack of references makes it obvious that some synthesis is being presented. --SmokeyJoe (talk) 06:06, 16 July 2012 (UTC)
    - That's not what "in-universe"/"out-of-universe" distinction is. OOU basically means to discuss the plot/characters as the viewer sees them, not as the characters see them. Say a character holds a secret that the viewer is obviously aware of, but no other character in the work knows this until the end of the work. In our summary of it, we can expose this secret, describing it without being coy about it. OOU writing most commonly manifests itself in being reasonably creating in untangling simultaneous-running plot lines into a single one, putting events in a order that is easiest and more concise to explain than the scene-by-scene of the work itself. But in-universe/OOU has zero to do with anything about sourcing. --MASEM (t) 12:59, 16 July 2012 (UTC)
      - You seem to be disagreeing with me, but I'm not sure with what. What you say is correct. I didn't mean to repeat any of the "in-universe"/"out-of-universe" distinction, WP:WAF explains it well. I think what I was saying is that in-universe writing makes pseudo-sourcing easier. --SmokeyJoe (talk) 03:00, 17 July 2012 (UTC)
        Yes, I am , as you're implying that to write out-of-universe requires extra sourcing beyond the work itself. This is not true. A properly written out-of-universe summary is sufficiently sourced to the primary work, period. We only encourage this because when writing in-universe, it encourage original research from that perspective, not because we are requiring more sourcing. --MASEM (t) 03:51, 17 July 2012 (UTC)
        Every fiction article should have sourcing beyond the work itself. The plot summary is an obvious exception. I didn't think we were talking about plot summaries. --SmokeyJoe (talk) 00:50, 18 July 2012 (UTC)
        There is no question, per NOT#PLOT, that we need sourcing beyond just a plot summary. But that is not what we talk about when we talk out-of-universe; out-of-universe describes the type of tone one takes when writing plot summaries, or more particularly, when writing on characters and larger meta-plots. We write as a distanced all-knowing viewer, not as invested character. Both tones still can rely on the primary work and require no additional secondary sourcing for supporting the plot summary. Everything else in a fiction article, heck yes, we need better sourcing but that's not the original question poised here. --MASEM (t) 01:08, 18 July 2012 (UTC)

It says 'but any interpretation needs a secondary source.' Nothing about in universe there. Dmcq (talk) 08:40, 16 July 2012 (UTC)

"Interpretation with a secondary source" will get you to the right place nicely. Few reliable independent secondary sources are in-universe. The in-universe type advice can be helpful in parallel. If in doubt, note that one is policy. --SmokeyJoe (talk) 03:00, 17 July 2012 (UTC)

Responding to WhatamIdoing, I stand by what I said, at least with respect to the nuts-and-bolts operative clauses of wp:nor/wp:ver. It's also true that the "intention" clause that you cite says that summarization is OK. That, plus common sense is what makes this work....most of the time. But once it degenerates into e.g conducting a POV war via wiki-lawyering, the nuts and bolts operative clauses always win out over the "intention" ones, and the warrior can use the fact that summarization in some respect always violates those clauses to knock out any material that they wish. Think about what summarization actually is. It is synthesis combined with selectively leaving out material. (The latter is fully legal despite some rumors otherwise) Like most policies it work most of the time but is prone to misuse, also a frequent occurrence. Tweaks would reduce the latter. Sincerely, North8000 (talk) 12:12, 16 July 2012 (UTC)

It's the cases where people say 'Oh I see now' what it means where one should be thinking about rephrasing to cope with misunderstanding. If they argue after having it explained a number of times then explaining better is probably not a solution. Dmcq (talk) 13:29, 16 July 2012 (UTC)

Not sure that I understand. But the problematic cases for most policies aren't cases where someone doesn't understand the overall policy. These are people who misuse operative clauses to suit their own agenda. Violations of policies are quickly corrected. Mis-uses are more common because they don't get corrected and just keep going. North8000 (talk) 14:10, 16 July 2012 (UTC)

Dmcq, I agree with you: Wikipedia has gotten far too bureaucratic for my taste. The primary lesson I've gotten from all this is that you have to judge truth along with verifiability before you challenge anything. If you believe something is true, it probably doesn't need for you to insert a CN-tag.

This is different than BLP and most articles on current politics (see US debt-ceiling crisis, Tea Party movement or Abortion). Unlike fiction, in these articles practically every statement requires a source. Once you have a source (or sources), you may spend endless hours discussing reliablity and weight before you get consensus of the active contributers to add one sentence. In fiction, my guess is that we have far fewer editor eyes to review and much less contentious opinions to referee.

I'm mostly from this world of writing about politics in Wikipedia. I'm still wondering whether we can do something in WP:NOR that makes it less likely that someone like me may interpret the definition of primary sources too legalistically. How about adding one word:

"Any creative interpretation of primary source material requires a reliable secondary source for that interpretation."?

The rest would remain the same as it is now. Other words we could use in place of creative might be: imaginative interpretation, original interpretation, inventive interpretation, novel interpretation or something else. What's your idea? If I had seen something like this in the NOR definition of primary sources, I'd have been a lot less prone to starting this whole rigmarole in the first place. --RoyGoldsmith (talk) 00:37, 17 July 2012 (UTC)

I just had a good look through the guideline writing about fiction and it seemed okay to me. However when I looked at one of its exemplars Gulliver's Travels it did seem to be badly lacking in citations in some important places. The section Gulliver's Travels#Major themes has no citations at all even though it is pure interpretation and there must be loads of people who have analyzed the work. My guess is the problem is with getting editors to implement the guideline. Dmcq (talk) 09:35, 17 July 2012 (UTC)

I don't think that adding the word "creative" is going to do anything except cause problems from people who are obviously going too far, but who say that their interpretations are so obvious that they're not creative and are therefore okay.

What we really need is for a few people to understand that "interpretation" is about assigning a meaning to something, not simply describing it. For example: Georgia O'Keeffe made a famous painting of a cow's skull on a red, white, and blue background. You may describe the painting, using the painting itself as your source: rectangular painting containing one cow's skull, centered, on a tricolor background. You may not "interpret" the painting: you may not talk about even the most "obvious" (to you) and "uncreative" meanings of the painting's contents. There might well be patriotic symbolism the choice of the colors, but you need a secondary source to write that. The skull may make you think of your own mortality, but you need a secondary source to write about that.

Description is not interpretation and is not banned. WhatamIdoing (talk) 00:35, 18 July 2012 (UTC)

I agree with WhatamIdoing. Non-creative interpretation is so narrow as to not be worth mentioning, and the problems introduced outweighs benefit. "Interpretation" already implies a degree of copyrightable creativity. Is it OK me for to copy your interpretation and present it elsewhere, without attribution, as if it is my interpretation? I think not. While there is a low-end to interpretation, which is not creative, it is probably better described as description. Someone might interpret the cental thing in Cow's Skull: Red, White, and Blue as looking like an animal skull. This would not be creative, but I think that in this case "interpret" is an exaggeration of "describe". --SmokeyJoe (talk) 01:12, 18 July 2012 (UTC)

I think that a good tweak be to more deeply embed "summarization is OK" into the operative parts of wp:nor & wp:ver, especially when the content of the summarization is not questioned. North8000 (talk) 01:14, 18 July 2012 (UTC)

I think Potter Stewart's comment regarding pornography applies to "interpretation" as well. It's very difficult to codify precisely where the line is overstepped between "describing what happens in the plot" and "analysing what happens in the plot" sometimes, but it's rarely difficult to tell the difference in a given bit of prose. (Or at least it is to me. YMMV, and obviously this is a major problem in almost all non-stub articles on fiction on en-WP and always has been.) It may be best tweaking the guideline so as to note this effect and to advise editors who disagree on whether a given summary strays too far into personal interpretation to take it to talk. Chris Cunningham (user:thumperward) (talk) 12:54, 18 July 2012 (UTC)

I tend to think that when people are sincerely questioning the content itself as overstepping summarization, there usually is really an overstepping problem. When they don't really question the material itself but just quote rules to knock out the material, now we've gotten into where the inconsistency that I pointed out above becomes problematic. That's why I think that this rule (and others) would work better if a part of the process of invoking the rule is to question the material. This is just procedural, not a criteria for the outcome. So, the person raising the question does not need to debate the question nor win a debate on it, they just need to just raise it. North8000 (talk) 20:21, 19 July 2012 (UTC)

Getting back to my original problem (what should we do about The_Big_Bang_Theory#Elements_of_the_show), I still have trouble with NOR policy being understood one way for fiction and a totally different way for, say, current politics. I think the difference is that a work of fiction is the primary source itself while, in other articles, you're talking about primary sources that describe the article's subject but are not the article's subject itself.

I examined various other, non-fictional works that were themselves the subject of their article; for example, the Bible and Common Sense (pamphlet), among others. Almost all described their subject based on the work itself. (For example, see this section in Common Sense.)

How do we allow this difference to be expressed in the NOR section that I quoted up above? If for no other reason than preventing anal-retentives like me from starting this discussion all over again. --RoyGoldsmith (talk) 01:57, 20 July 2012 (UTC)

On regards to the specific example of that section of the Big Bang Theory, you're absolute right that most of its problems in NOR. One station "The show centers around physics" (paraphrasing) is actually ok - you only have to watch a few episodes of the show to come to this obvious conclusion. But most everything else in that section is saying "The show has theme X because of these examples." without sourcing. As we now talking about the show and not the narrative itself, that steps out of the primary sources and requires secondary sourcing. Contrast this to the Themes section of Inception, which does supply appropriate sourcing to discuss the meta-details. So to your point, when talk about the broader scene of a work and not just the events as they happen on screen, we do expect better sourcing than just the primary work itself. But when you're just talking the narrow aspect of the events of one episode or a faith summary of the plot of a movie, the primary work is sufficient. --MASEM (t) 02:23, 20 July 2012 (UTC)

I'd agree that the material which you are discussing falls within the 95% of Wikipedia that is in violation of a strict/literal interpretation of some operative clauses of wp:nor/wp:ver. Assuming we aren't going to start deleting that 95% the question becomes a reasonable interpretation per the norms of Wikipedia and the "intent" and other portions of those policies. This one is a bit more out there than the norm because much of it is is a (IMHO excellent) summary from primary sources. My opinion/advice is that if you dispute the summary (or parts of it) itself, then do so and bring out the rule book; if not don't. North8000 (talk) 11:28, 20 July 2012 (UTC)

What I'm looking for is a (hopefully small) change to WP:NOR's policy on primary sources. That is the paragraph under Primary sources that begins with Policy.

I think we've all agreed that this policy doesn't state explicitly the use of primary sources in fiction and other, written works. (If you think it does, please, show me where.) Talking about common sense doesn't work because my common sense may be different than yours. (For example, I would still basically delete the entire The_Big_Bang_Theory#Elements_of_the_show section as well as Common Sense (pamphlet)#Paine's arguments against British rule, given the current NOR policy.) I believe that the policy should be changed but my belief doesn't make it policy. --RoyGoldsmith (talk) 13:22, 20 July 2012 (UTC)

Arguably what you are looking for is already under WP:SYNTH - we can't come to conclusions from facts from primary sources that are otherwise not obvious. We can add the entirity of the Big Bang Theory episodes to conclude "hey, they make a lot of jokes about physics", that's a non-SYNTH, obvious conclusion - but also one I suspect that it would take a trivial amount of time to search and verify with secondary sources. However, there are plenty of other statements that add 2 or three episodes together and claim long-running themes and that may be more difficult to source; that would be SYNTH. SYNTH as it is is fine, what probably is needed is clarification in WAF that points back to SYNTH to say that whenever you start to generalize a work of fiction (moving beyond the plot summary), secondary sources start to become required to do that.

Also, as a point: we prefer not to delete material if it doesn't immediately fail core problems like BLP or NPOV, particularly if the material is likely true and just lacks sourcing. The section in Big Bang Theory screams for the need to sourcing, and if I were taking it to Good Article or better, I'd either find sources or remove that section before GA, but as it is developing, it makes no sense yet to remove; the claims made all appear to be true, but just presently are technical NOR violations. --MASEM (t) 13:51, 20 July 2012 (UTC)

The policy on primary sources states (emphasis mine) "Do not analyze, synthesize, interpret, or evaluate material found in a primary source yourself; instead, refer to reliable secondary sources that do so." I therefore conclude that the examples in WP:SYNTH refer to reliable secondary sources.

My problem is that primary sources are used differently in articles that have written works (including fiction) as their subject. In that case, the written work is treated (and, IMO, deserves to be treated) as the most important source -- the canon, if you will. But nowhere in NOR or the other core policies does it say so, either explicitly or implicitly.

Let's say that you found an article in a reliable secondary source (bear with me) interviewing a writer of The Big Bang Theory and he said "I know many people think that the show focuses on science, especially physics, but they are wrong. We could have used accountants or computer programmers or even lawyers...". Would that source, simply by being secondary, make the current Wikipedia material (based on a primary source) wrong? If anyone thinks no or thinks they could weasel around it (for example, by saying that the writer really meant to say that "the show must focus") then I think they're engaged in the original synthesis: that they think that their opinion is as good as the writer's.

And I would agree with them. That's if we were just having a conversation. But I don't think anyone should have a discernible opinion when acting as a Wikipedia editor. Even if just "summarizing".

On your second point, I agree: we ought to move slowly, especially about deleting material. I am a firm inclusionist- I hate removing anything. Why else would I be having this involved discussion? But do you believe that the section in the BBT we've been discussing is based on anything other than its primary source or that, given time, the vast majority of the material can be sourced? The warnings about have been posted for a year and a half. If they were going to find secondary sources, wouldn't they have done so already? --RoyGoldsmith (talk) 22:25, 23 July 2012 (UTC)

The short answer is that we already have plenty of advice in NOR, WAF, NOT#PLOT, and elsewhere on putting too much reliance on primary sources, and using them to make claims outside of summarizing a work and for simple, obvious WP:SYNTH, in the absence of more authoritative sources to say otherwise. As in your example, if a writer confirms something that is completely against the obvious interpretation, we go with the authority, not the interpretation.

As to the BBT section in question, I don't know enough about the show to know if those are accurate but they do seem to be reasonably truthful statements that otherwise aren't being put there to create bias or falsify info. I wouldn't be making an article a Good Article with that section as is, and would ultimately require better secondary or third-party sourcing to affirm that or otherwise remove the claims; but as a currently airing/produced TV show, I would expect one shouldn't have a problem to find sources, so I wouldn't be quick to remove it. A better example I can vouch for is Mythology of Fringe which I created, and does have blocks of text without direct sourcing and relies on the primary. I've not spent a lot of time with it, but I do know that if someone came and begged for improvements from it from, I know I can source every element give from secondary third-party sources that talk about the show, as well as information from the producers/directors. Now, if we were suddenly talking about an old TV show, say, "I Love Lucy" and someone started throwing up similar types of thematic elements based only on the primary source, I'd beg to question if that is synthesis and if they would ever be able to find sources for it since the show's long off the air and production and many of the original people working on it have passed away. It's basically a matter of giving editors the benefit of the doubt. Basically, if it doesn't look like the material is a total fabrication or otherwise harmful to WP, but only begs for better sourcing, it is better to leave it in place and nudge those editing it to get the improvements in place. --MASEM (t) 23:50, 23 July 2012 (UTC)

OK, you've convinced me. Temporarily and tentatively but I'm convinced. For now. If you can get Mythology of Fringe to survive for more than a year without being shortened by 75% I guess I have to agree. For the time being, I'll only add or replace the episode templates in the Elements of the show section with non-primary source needed tags.

In "A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts...", my interpretation of "facts" was that fiction is not and could not ever be, by definition, facts. So that any synthesis beyond the sequential plot elements would have to be "interpretation". This was such an ingrained aspect of my psyche that I didn't realize it before now. I guess what the policy means is that "facts" can include fictional, in-universe facts as well.

This was acerbated by having the bulk of statements in the Elements of the show section rely on too many episodes. You can't specify episode numbers for "Much of the show focuses on science, particularly physics" because (IMO) two-thirds of the episodes support that (in-universe) "fact".

There is another, related point I want to make but, since it's another point, I'll make it in another talk section below. Thanks again. --RoyGoldsmith (talk) 15:15, 29 July 2012 (UTC)

Too many primary sources?

The sections below are all comparisons of films to their sources (the novels or plays that the movies were based on). These sections, which present significantly different styles, have no, or very little, inline citations.

The theory has been advanced that these sections (and other ones like them) should be removed in their entirety because, since they were based on at least two primary sources (the work itself and its antecedents), they are obviously interpretations of those two sources and thus, original research and/or synthesis.

Do you think that these sections should be totally removed because of WP:NOR? Why? Would you change them? How? Or would you indicate by tags that they should be changed? If so, for how long? --RoyGoldsmith (talk) 16:51, 29 July 2012 (UTC)

This is probably covered by WP:PLOT... briefly noting differences in plot between two versions of the same story (purely descriptive statements, involving no analysis, interpretation or conclusion) is not OR. Blueboar (talk) 13:32, 1 August 2012 (UTC)

Technically true, but I do find that it is very easy for readers to either slip in every tiny change, or start engaging in subtle OR here. My opinion here is that such sections border on trivia, and what differences to include should be guided by third-party sources. --MASEM (t) 13:42, 1 August 2012 (UTC)

Synthesis example

The "more complex example of original synthesis" is not a good one. It claims that "the second paragraph is original research because it expresses a Wikipedia editor's opinion that, given the Harvard manual's definition of plagiarism, Jones did not commit it." But it doesn't: it expresses the simple logic that IF Jones did such-and-such (copying references without consulting them), then such-and-such would be the case (this would contravene the Harvard manual, but not fall under its definition of plagiarism)." Certainly this expresses the editor's understanding of what the Harvard manual says, but understanding is a part of writing about anything, is always open to correction, and is not by itself OR. A reference to the manual itself may be sufficient. If the manual is not so explicit, then reference to a reliable source that comments on the manual and its definition of plagiarism may indeed be desirable, but such a reference need not comment specifically on Smith and Jones as stated in the article. [PS. I write the preceding without knowledge of the "actual Wikipedia article" concerned. If the latter clarifies the issue, then someone who knows what it is should link to it.] — Preceding unsigned comment added by 163.1.246.64 (talk) 13:54, 4 July 2012 (UTC)

Although Smith accused Jones of plagiarism, he did not accuse him of violating the practice recommended in the Harvard manual. If Jones were formally accused of plagiarism by Harvard, they may present another manual whose rules he violated. Or an author may sue Jones based on state or federal laws. Perhaps Smith meant that Jones' actions were plagiarism because he used another author's ideas [that author's interpretation of the sources] without citing that author. These are determinations that we cannot make. TFD (talk) 16:32, 7 July 2012 (UTC)

Indeed, but I don't think they're relevant here. The supposed OR is not commenting on the intentions of Smith or on the actual actions of Jones. It discusses how a hypothetical action ("If Jones ...") relates to the actual manual. If interpretation of the manual is difficult, then a reference for that is desirable, but no specific reference to the actual Smith or Jones is required, which is the main claim being made. — Preceding unsigned comment added by 82.10.109.67 (talk) 14:14, 5 August 2012 (UTC)

Replaced with the correct text from the consensus version of WP:V?

I still think this is premature. Certainly the closing statement did not say that "VnotT" would be removed. It says "there remain open questions about the exact status and place of the phrase "Verifiability, not truth", and recommend that the community continue discussion on these points. In particular, jc37 considers that there is consensus for option D only insofar as it forms a basis for such continued community discussion." Having said that, as this is just one small paragraph, perhaps there is no problem. Dougweller (talk) 15:43, 31 July 2012 (UTC)

We should definitely think about this carefully and go slowly. I don't object to rewording in light of the recent RFC at WP:V, but we do need to remember that the koan "The threshold for inclusion in Wikipedia is Verifiability, not Truth" directly relates to this policy... Indeed that specific wording originated here in WP:NOR. It was stated here before it was added to WP:V.

If we think of VNT as a statement about Original Research (as opposed to it being a statement about Verifiability), I think the koan is still conceptually valid (an unverifiable truth is excluded - on the grounds that it is OR.) Blueboar (talk) 12:34, 1 August 2012 (UTC)

The closing statement was to use option "D". Wp:ver and wp:nor are a 90% overlap of each other and wp:nor certainly needs to be updated. Further, the item updated here is a statement of what is IN wp:ver. It certainly must be a correct (and thus updated) statement of what is in there. Sincerely, North8000 (talk) 12:51, 1 August 2012 (UTC)

Which is why I am not opposed to discussing a rewording... what I am suggesting is that the koan (or something very like it) might still have a place in this policy... not as a restatement of what is stated at wp:ver, but as it originally was - a statement to clarify an important aspect of NOR. Blueboar (talk) 13:14, 1 August 2012 (UTC)

I have to admit to being a little surprised at my own reaction to this issue. After all, I'm a long-time supporter of VnT, and I'm a big believer in taking things slowly and discussing them fully. But nonetheless, I agree with S Marshall's removal of the sentence! The purpose of the "other policies" section is to point the reader to what those other policies currently say. And that phrase has now been deprecated to a footnote, so it shouldn't be where things start off here. But taking this discussion a step further, I wonder why we even have a lengthy section about V and NPOV here at all. Why not state briefly, as in the V lead section, that the policies work together, and link to them, and leave it at that? --Tryptofish (talk) 13:40, 1 August 2012 (UTC)

Truth be told, these policies are a 90% duplicate of each other and should probably be merged. The "10%" is basically emphasizing the aspect of it's title, which then, as it inevitably should, goes to verifiability. The more realistic (and possibly better) alternative is to start slowly taking out the 90% of it that is a duplication of wp:ver. North8000 (talk) 14:01, 1 August 2012 (UTC)

VNT separate from the "Other Policies" section

I am going to start a new sub-thread here... because my comments above were not really discussing a specific edit, or even the language of the "other policies" section... what I was asking is a slightly different question: Is there a place for the VNT koan (or something like it) somewhere in the NOR policy? (Not in the "other policies" section).

Remember... VNT originated in this policy. It was originally crafted to explain an aspect of Original Research. It was subsequently added to WP:V (in the context of being a restatement of something that this policy said). My question is, regardless of what WP:V now says... does it make sense for this policy to include the VNT koan (or something like it)? If the answer to that question is "yes", then we can move on to discussing the specifics of exactly where and how to include it. Blueboar (talk) 14:31, 1 August 2012 (UTC)

The answer is no. Surely it's fairly obvious that the correct response to the resolution of such an epic battle as has just been had is not "so is it OK if I put it in a different policy, then?". Formerip (talk) 14:33, 1 August 2012 (UTC)

Agree. (except I'd say "decision" rather than "battle") Blueboar, unless I am misunderstanding, this seems very uncharacteristic of you. Sincerely, North8000 (talk) 14:54, 1 August 2012 (UTC)

It isn't a question of moving it to a different policy... as it has always been part of this policy (even before it was added to the WP:V policy). I suppose what I am asking now is whether its depreciation over at WP:V means we should remove it from this policy, or whether we should return it to its former status as a stand alone element of this policy - separate from "other sections"?

As to the personal comment... I am not sure why you find my asking about this uncharacteristic ... I have been arguing all along that VNT is important in terms of understanding NOR. That's why I was initially so reluctant to depreciate it at WP:V. Sure, I (eventually) came to understand why it was problematic at WP:V... but I still think it is important to state here at WP:NOR. I am very flexible as to how we state it (I am not locked into any specific wording)... but as a concept I think it is important to state in this policy and would strongly oppose removing it completely. Blueboar (talk) 15:31, 1 August 2012 (UTC)

(added later) I didn't mean that the viewpoint was uncharacteristic, I meant that the angle you are taking on this is, respectfully, pretty weak and far-fetched, and that is uncharacteristic of you. North8000 (talk) 16:11, 1 August 2012 (UTC)

Please reassure me that you're joking.—S Marshall T/C 15:45, 1 August 2012 (UTC)

No Joke. Blueboar (talk) 15:52, 1 August 2012 (UTC)

Okay, I'm not sure if I have the heart for this discussion but let's just try it out.

The consensus version that appears at WP:V and now in WP:NOR reads "Even if you're sure something is true, it must be verifiable before you can add it." What's missing from this explanation?—S Marshall T/C 15:58, 1 August 2012 (UTC)

It's even more no-brainer than that. The section that it was in was the "other policies" section, and the wp:verifiability subsection, and it was and is a statement of what is at wp:verifiability, as such. North8000 (talk) 16:08, 1 August 2012 (UTC)

Blueboar, when you say "it has always been part of this policy", do you mean to say that at some point in the dim past WP:V was a section in WP:NPOV, which was then spun out and became a page of its own? (I have no idea, this is just a hunch). Formerip (talk) 16:14, 1 August 2012 (UTC)

No... I am saying that both the concept behind VNT, and the specific phrasing "The threshold for inclusion in Wikipedia is Verifiability, not Truth" started out as part of the WP:No Original Research policy... and has remained part of this policy since it was created. The language was originally intended as an explanation of an important policy point (that we should not add unverifiable material, no matter how true it might be... because doing so is a form of OR). It was not created for WP:V and then added here... it was created here and subsequently added to WP:V (as a repeat of what this policy already said)... and it was only after it was repeated at WP:V that it took on a life of its own and became problematic.

To reply to S.Marshal's question... I suppose what is missing is a statement that ties that sentence, conceptually, into NOR... something to clearly explain the reason why we should not add unverifiable material - even if that material is true (the reason being that doing so is form of Original Research). This is the core concept that the VNT koan was written to express.

Thinking further... I suppose I am expressing my concern that we might be tossing the baby out with the bathwater... I am asking people to pause before we rush around deleting every instance of the "toxic triad"... and to question whether, in the context of NOR, the koan is really all that toxic (and to consider that perhaps it is significantly less toxic in this context than it was over at WP:V). While VNT has been appropriately depreciated at WP:V, does that necessarily mean that it should be depreciated here - retained in its original context. I am asking people to consider the possibility that VNT continues to be appropriate here, even though it was not appropriate at WP:V.

Now, if we can get the concept across without using the koan, I am quite willing to consider options. I am not insisting on retaining the language of the koan (what I insist on is retaining the original concept behind the koan)... I merely asking whether there remains a legitimate place for the koan in this policy. Blueboar (talk) 17:11, 1 August 2012 (UTC)

Your response doesn't seem very clear. What I'm suggesting in that VnT originated on the NOR page because there was at the time no V page. If that's the case, then it's even less of a big deal that it would otherwise be.Formerip (talk) 17:35, 1 August 2012 (UTC)

The earliest incarnation of "Verifiability, not truth" that I've been able to trace is an edit to a page that's now called Wikipedia:No original research (draft rewrite 5th December 2004 to 5th February 2005) (the edit being this one of 8th November 2004). The phrase appears fully-formed: "The threshold for inclusion is verifiability, not truth". (If anyone knows of an earlier version that explains more about where it actually came from or why, then I would be interested to see it.) WP:V existed then, having first been created on 2 August 2003, but there seems to have been quite a bit of reorganisation of the two policies with text moving between them. VnT got moved to WP:V in this edit by Uncle G.

Examining the history is instructive. The original version seems much less objectionable to me, because the phrase is completely surrounded by context and explanation. By the time I encountered it, all the thought and reasoning had been cut and the key words emphasized in bold, and it had been moved into the first sentence of the lede of Wikipedia's only positive content policy. (What I mean by "only positive content policy" is that all the other content policies are negatively phrased: they describe things you shouldn't do. WP:V is the only policy that describes what you should do. To my mind this is an important distinction, and it's why I think it's a good thing that V and NOR are separate.)

In any case, now that I understand Blueboar's objection better, it seems easily-overcome to me. Blueboar says: "I suppose what is missing is a statement that ties that sentence, conceptually, into NOR... something to clearly explain the reason why we should not add unverifiable material - even if that material is true". I agree that this would be a beneficial thing to explain. I would like to point out that the phrase I removed does not, in any way, explain it; the sentence "The threshold for inclusion on Wikipedia is verifiability, not truth" contains exactly zero reasoning. But I would welcome discussion about how we could explain it to Blueboar's satisfaction.—S Marshall T/C 21:15, 1 August 2012 (UTC)

First of all, I strongly believe that Blueboar's raising of the question is both good faith and thoughtful, and I hope that no one else is really calling that into question. Blueboar's question, whether the possibility that VnT arose here historically, instead of the policy page that was just discussed in the RfC, should mean that we evaluate it differently here, is an entirely reasonable question to ask.

So here is how I would answer it. Everything on Wikipedia is dynamic, such that it may always be changing (and maybe even improving). The purpose of the threshold-VnT sentence, no matter how it arose originally, was to explain the concept of verifiability. Not to explain the concept of original research! Therefore, the recent consensus to downgrade that sentence's status at WP:V means that it's entirely appropriate to do likewise here. The only reason not to would be one that I'm not seeing in this discussion: that VnT somehow clarifies one's understanding of original research in a way that it does not clarify one's understanding of verifiability. Obviously, that's not the case. (I recognize that it originally communicated the point that "truth" is a form of OR. But that's not the same thing. One can fully understand what OR is, without that sentence. As long as we continue to tell readers of the policy page that it works together with V (and NPOV), no one is going to think that NOR allows VnT's concept of "truth".) So I (who have long been a stick in the mud for VnT, let's not forget!) am all in favor of deleting it here. --Tryptofish (talk) 00:31, 2 August 2012 (UTC)

Blueboar wrote: "Now, if we can get the concept across without using the koan, I am quite willing to consider options." S.Marshall wrote: "I would welcome discussion about how we could explain it to Blueboar's satisfaction." So how about this? "Truth without verifiability is OR." Kalidasa 777 (talk) 08:42, 2 August 2012 (UTC)

Thanks to S. Marshall for the links to the early history of the VNT koan... I encourage everyone to read them. Understanding the original intent of the koan, and how it evolved over time, will better inform our discussions.

The type of evaluation and discussion that is now going on is exactly what I was asking us to do.

As for Kalidasa's suggestion... hmmm... that might do it. I'll think about it. Blueboar (talk) 14:04, 2 August 2012 (UTC)

OR, truth

When I think of my experiences looking at others' edit summaries and talk page discussions for articles, when someone tries to put OR into an article, they are simply told that it is OR, and usually the material is successfully deleted, without any discussion of truth. In some cases the claim that something is OR is disputed, but the discussion does not turn to whether or not it is true, but whether or not it is OR. --Bob K31416 (talk) 09:42, 2 August 2012 (UTC)

"OR" has become Wikipedia shorthand for "does not comply with wp"ver", and "unsourced". WP:ver/nor creates a (sourcability) requirement for inclusion of material. And say that there are no exceptions to meeting that requirement (e.g. by claiming that it is true.) Striving for accuracy of material is a related but different topic, and a good goal. North8000 (talk) 23:09, 2 August 2012 (UTC)

Re “ "OR" has become Wikipedia shorthand for "does not comply with wp"ver", and "unsourced".” — Not for me. OR means unpublished info that has originated with the editor that put it in, i.e. the editor’s original research. If one interprets OR to mean something more general, it would lose its usefulness of specifying a particular type of information that isn’t allowed in Wikipedia.

But that is a digression. I began this section to share my personal experience and implicitly asking other editors to consider their own experiences, not alleged experiences of others, regarding the situations that I mentioned in my message. --Bob K31416 (talk) 11:53, 3 August 2012 (UTC)

Sidebar/digress comments: Whether it's right or wrong, I think that it is useful to point out what the common use/meaning of the term "OR" is. And the range of unsourced/unsourcable material can range from new creative personal research (which is what the term brings to mind) to someone putting in a wrong fact (e.g. Babe Ruth's date of birth) from memory/personal knowledge. IMHO, from a structural/policy standpoint, they are both the same to Wikipedia.

On the main topic of your comment, I really didn't understand what you were getting at, but was pointing out a related distinction anyway. That the following are two different things:

WP:Ver/NOR establishes a requirement that must be met for inclusion. Meeting it is a requirement for inclusion, not a force for inclusion.
A quest for accuracy.

In some ways #1 helps #2, but they are two different things.

In other different ways, #1 also inadvertently hindered #2, but that has been fixed.

In cases where objective accuracy exists, the operative mechanics boils down to 4 cases:

Meets wp:ver/nor and is false: WP:ver/nor has no effect. Other Wikipedian processes are free to determine inclusion/exclusion
Meets wp:ver/nor and is true: WP:ver/nor has no effect. Other Wikipedian processes are free to determine inclusion/exclusion
Fails wp:ver/nor and is false: WP:ver/nor prevents the material from being in Wikipedia.
Fails wp:ver/nor and is true: WP:ver/nor prevents the material from being in Wikipedia.

Sincerely, North8000 (talk) 12:27, 3 August 2012 (UTC)

This actually raises some interesting side questions... are WP:V and WP:OR distinct concepts or simply different ways of expressing the same concept? Is it possible for something fail WP:V and yet still pass WP:OR? And is it possible for something pass WP:V and yet still fail WP:OR? Blueboar (talk) 13:05, 3 August 2012 (UTC)

IMHO they are the same concept. The only difference (the 10% that is not overlap) is wp:nor places extra emphasis on and provides extra discussion on one type of material that is excluded by wp:ver. North8000 (talk) 13:18, 3 August 2012 (UTC)

WP:NOR and WP:V are similar, but are distinct concepts. WP:V states that a reliable source must exist that supports the statement; WP:NOR adds that sources must not be improperly used to create original research. It is possible for something to meet WP:V but fail WP:NOR. The classic example is the one given in WP:Synthesis, reliable sources exist for the definition of plagiarism in the Harvard Writing with Sources manual, but inclusion of that definition would be OR unless a published source refers to it in the context of the dispute. LK (talk) 13:43, 3 August 2012 (UTC)

I think that puts a better point on it, addresses an abstract variant more directly, and explains it better. But I think that that is also a violation of wp:ver. North8000 (talk) 14:52, 5 August 2012 (UTC)

OR, Verifiability, Accuracy

North, you keep saying that verifiability is "not a force for inclusion". If you mean that verifiability is not a guarantee of inclusion, then I agree. However, if you mean that verifiability is not, in any circumstances, an argument for inclusion, then I can't agree.

A comparatively simply example from the WP:Inaccuracy essay -- report about road closures in Oslo after 2011 attacks, extent of road closures questioned by an Oslo resident, on the grounds that "I was there". My own thoughts about this...

It is clear that we cannot simply add a previously unpublished eye-witness report into mainspace, as that would be OR.
Is someone saying "I was there", sufficient grounds, by itself, to remove the information that was found in the verifiable published report? I don't think it is.
On the other hand, I'd agree that in such a case it is a very good idea to check what other published sources have to say about how much of Oslo was closed. In fact WP editors did consult further sources in this instance.

I think it is a very important principle that, as the RfC-approved lede of WP:V says, "content is determined by previously published information rather than by the personal beliefs or experiences of its editors." Kalidasa 777 (talk) 23:45, 4 August 2012 (UTC)

Well.... OR can be discussed on the talk page, and while it should not influence what information we present in the article, the discussion can sometimes influence how we choose to present to information. For example, suppose we are writing about a road that closed (not talking about the ones in Oslo now), and based on a source we say: "The road was closed in August of 2011"... then a Wikipedian says "that's incorrect... I live there and drove on the road every day until it closed... it did not actually close until September", we can reach a consensus to add attribution: "According to <source> the road closed in August of 2011" ... or we could compromise and say: "The road closed during the late summer of 2011" (which effectively resolves the conflict between the OR assertion and the source). What we can not do is say the road actually closed in September, based purely on the OR assertion. Blueboar (talk) 02:30, 5 August 2012 (UTC)

I don't think you can call that an OR assertion, it rather a primary source. Writing WP content based on it, that's usually OR.--Kmhkmh (talk) 03:15, 5 August 2012 (UTC)

Kalidasa 777, one/curse/gift of mine is that I can see the logical/structural underpinnings of policies, conversations and think along those lines. I mentioned "curse" because, when I try to communicate it on a more complex ones it often comes off sounding like Greek. Your example is one of those so here goes. The short answer is that the fact that the material meets wp:ver means that wp:ver allows ("allows" means only "does not prevent") the material to be included, and that's the extent of wp:ver for your example. So it will be up to other Wikipedia policies, guidelines and processes to determine whether or not the material stays. Who knows what those could be on any particular example. A few examples of how it might flow for your given example could be for the editors at the talk page to look at it and say and decide:

A "RS" source says that it is true, and a credible sounding editor/argument says that it is false. That means that we should look for more sources on this question.
"That information was sort of off topic and not useful anyway. Before this question came up we were on the fence anyway on whether to keep/remove it. Now, with this additional new question, let's just leave it out."
They make one of the decisions that Blueboar described.

Now on your:

"Is someone saying "I was there", sufficient grounds, by itself, to remove the information that was found in the verifiable published report? I don't think it is."

I would have to respectfully say that you have accidentally misheard what I said into a straw man version. While I agree with your conclusion "I don't think it is", I must note that nobody is arguing that a personal opinion is sufficient grounds to force removal of material, and so that is not relevant to what I said. Your conclusion does not conflict with or relate to my argument. Again, I say this respectfully; the bluntness is only for brevity& clarity.

Finally, when I said that meeting wp:ver is a requirement for inclusion, not a force for inclusion, I should clarify that I was talking about the policy, not the sourcing that made it meet the policy. The sourcing that enabled it to meet wp:ver would certainly matter in the discussion. Sincerely, North8000 (talk) 11:31, 5 August 2012 (UTC)

Original mathematical proofs are not interpretations

I would like to request that the policy be clarified concerning mathematical proofs. Almost all mathematics articles on Wikipedia are based on a specific standard axiomatic system that only a small minority reject. And with the exception of articles that deal specifically with axiomatic systems, the rest do not explicitly state this set of assumptions, but under these assumptions, all mathematical proofs are absolutely true regardless of whether they are original research or whether they have been published before. Therefore editors have always been freely writing new proofs or variants of them because they can be systematically and unquestionably checked to be correct. In short, there is no room for interpretation, so original research is perfectly fine. Of course, it might actually be a good thing if every single proof had a link to the axiomatic system within which it is established, but the typical reader would not care, so omitting it as has always been the case is fine too. Lim Wei Quan (talk) 05:26, 6 August 2012 (UTC)

New proofs that have not been reviewed by reliable sources may have errors. We as WP editors cannot rely on other editors finding and correcting them. Ergo, they are original research and cannot be included. --MASEM (t) 05:39, 6 August 2012 (UTC)

First of all most math articles do not contain proofs (see Wikipedia:Manual_of_Style_(mathematics)#Proofs). As far as describing proofs in WP are concerned "straight forward"/"obvious" or minor variations of known proofs are up to an editor discretion, but new proofs are considered WP:OR for the reason already stated above. WP is not a publication venue for new material, but it merely collects and describes what was already published elsewhere (in reliable, reputable sources).--Kmhkmh (talk) 06:59, 6 August 2012 (UTC)

Wikipedia is not a place for "proofs", but it is a place for "explanations". In the article IBAN, I wrote "It can be shown that^ref ..." and in the reference, I stated "This is an undergraduate level problem". In other words, if you have a degree in maths, by all means comment, if you don't, please leave comments to somebody who can assess this statement. That article might well serve as a point of reference for this thread. Martinvl (talk) 10:29, 6 August 2012 (UTC)

The content of Wiki articles needs to be verifiable from reliable sources by any reader, not just the ones able to prove it for themselves, so a source to enable them to confirm that "It can be shown that..." would be requiered in the example you give. Canepa (talk) 12:16, 6 August 2012 (UTC)

I totally disagree with Canepa. Many readers cannot obtain/understand citations in articles - such citations might be in a foreign language or they might be in an obscure journal or a hard-to-get book. The statement that I made comes under the unbrella of "simple calculations".

Since I volunteered information about the article IBAN here, I regard your immediate rejection of sources as being in bad faith, especially as the article gets 90,000 hits a month and that statement has been in place for over a year. Maybe other editors might like to make their views known, meanwhile would Canepa please hold back from reinserting his changes to that article until he has had input from other editors. Martinvl (talk) 12:56, 6 August 2012 (UTC)

WP:V does not require every reader to be able to get the source; foreign-language or difficult-to-understand sources are completely acceptable. But that said, if the proof has not be printed, and its anything beyond simple, obvious calculations per WP:SYNTH, it's original research. If it is a proof that one could expect a college undergraduate to complete, it's not obvious; "obvious" is things like "2+2=4", simple arthimatic, not logical proofs. --MASEM (t) 13:07, 6 August 2012 (UTC)

First of all, Masem is confusing the words "proof" and "explanation". If you look at the article in question, you will see that I was explaining (not proving) how a piece of computer code works. One needs a certain level of maths to understand what I was writing, and if you could understand that, then what I wrote is obvious. Martinvl (talk) 14:40, 6 August 2012 (UTC)

As far as I can see Masem was not contradicting you, but rather stating a somewhat similar opinion, i.e. he disagreed with canepa. Possibly he was even replying to him. Unfortunately in WP the ident is often not used as in the usenet (mimicking a a tree structure), where the ident indicates who you reply to. In WP some people floow that convention, but others don't or are simply not aware of it.--Kmhkmh (talk) 15:18, 6 August 2012 (UTC)

The requirement is that statements can be verified from reliable sources, and that the sources must be cited if asked for. We cannot assume that all readers have the math abilities required of a mathematics undergraduate. I didn't reject any sources or change the article, what I did was request that a source be cited to make the math claim verifiable. If you cannot supply an appropriate source, then the statement will have to be removed from the article. Canepa (talk) 15:05, 6 August 2012 (UTC)

If an fellow editor cannot really verify a source due to a lack domain knowledge, he simply should leave the verifications to others who have that domain knowledge. If he has the suspicion that something fishy is going on, he simply should report it to editors or project sites (like special subject portals) that possess the required domain knowledge. The policy you've cited above is not to be understood as "can be verified by everybody" but as "can be verified by some/many editors or readers".--Kmhkmh (talk) 15:18, 6 August 2012 (UTC)

Policies aside, Canepa, do you have a question / concern regarding the material?

Policies not aside, IMHO I think that Canepa's tagging is proper. Hopefully it is because there is a sincere question rather than as a way of conducting battle. That "cite" that is not a cite that is/was earlier in that sentence raises an additional yellow flag. North8000 (talk) 15:14, 6 August 2012 (UTC)

What we have here is a problem of where do we draw the line between someone stating that "1 + 1 = 2" and User:AEinstein stating that "e=mc². Following up on what User:Kmhkmh wrote, if the reader does not understand what is being written, then citations for a purely mechanical section will not help. Let us assume that the citration referenced a book that was not availalbe to Canepa, would (s)he understand what was being written. If so, it is a trivial exercise to make the link in question, if not then no number of citations will clarify things. Martinvl (talk) 16:12, 6 August 2012 (UTC)

The policies demand a reference to a reliable source in these circumstances. Without such a source to support it, the content may be deleted. Canepa (talk) 08:26, 7 August 2012 (UTC)

In reply to the comments from North8000, I too have a serious concern about the use of fake references. You might assume from the appearannce of the reference number that the statement can be verified from a reliable source, when in reality no reference actually exists, just an arrogant assertion that if you are clever enough you might be able to work it out for yourself. This isn't good enough for Wikipedia. This needs a reliable source. Canepa (talk) 08:26, 7 August 2012 (UTC)

No, I already told you above, what you should do if you suspect abuse or something fishy, involve other people with domain knowledge and WP has special project sites to contact them (hiere for instance Wikipedia_talk:WikiProject_Mathematics). However expecting a sourcing that will anybody allow by mere text comparison whether any content is correct/sufficiently sourced is out the question. Editors won't work under such a nonsensical restriction.--Kmhkmh (talk) 11:49, 7 August 2012 (UTC)

Wow I did not expect such a lively conversation. Glad to see that there is one going! I have a reasonable mathematical background, and my opinion is that mathematics should not require citation, because generally cited mathematical articles are unavailable to the general public, making them unverifiable even by mathematically knowledgeable editors, thus defeating the purpose of citation! When I come to Wikipedia to find out about some mathematical theorem, a complete proof of it would be perfectly in line with Wikipedia's encyclopedic nature, since it can easily be hidden by default (as I have seen in some articles). If instead Wikipedia only gives a link to some article requiring subscription or, worse still, printed material, will surely get me to leave immediately. Mathematically trained editors will correct mistakes, so better, I think, to have proofs than nothing at all. Proofs are not opinions and thus do not fall under the original intended scope of this policy in the first place. Lim Wei Quan (talk) 10:29, 7 August 2012 (UTC)

Well no new proofs do fall under under WP:OR and this subject has been been discussed in the math portal /under mathematician) several times, always with the same result: That is proofs are normally not a part of articles, but only in special cases and new proofs are WP:OR. I'm a bit puzzled from where you get the idea that math articles are unavailable to the general public, math articles are as available as any other academic articles. Yes, some journals are too expensive for individuals, but you can get them in university libraries and math institutes and many WP editors are directly connected to them, moreover often freely accessible online are available (preprints on arxiv.org for instance). That aside most math content of WP or an encyclopedia can be found in textbooks or monographs, which are often easier to access. Math is not different from any other academic subject in that regard (availability/accessability of sources). Take a look at other larger math encylopedia like MathWorld (Wolfram) or the Encylopedia of Math (Springer) and you will see they do not contain proofs either. If you want to provide proofs as an additonal service readers to readers, the way to go is provide an external link containing the proof (link to existing proofs on university or expert page, or use the book of proofs in Wikibooks (or initiate a similar project on wikibooks) or use the proofwiki). If you feel further need to discuss that subject or have questions, I suggest to continue this discussion at Wikipedia_talk:WikiProject_Mathematics, but as I mentioned earlier that subject has been discussed many times before, so don't expect a different answer/result.--Kmhkmh (talk) 12:01, 7 August 2012 (UTC)

There's clearly a need for Wikipedians to produce original calculations where they concern articles within the scope of WikiProject Mathematics. For example, Frobenius endomorphism contains at the bottom an example which is completely uncited, and seems to consist of calculations performed by User:Gene Ward Smith. Does the article benefit from a worked example? Absolutely; it enhances the reader's understanding of the subject. Should Gene Ward Smith have copied a calculation directly from a textbook? Not necessarily, because although a mathematical formula couldn't be copyright, a sample calculation could be. If it's undergraduate-level maths then it's no problem for other Wikipedians to check, which means it's verifiable. (If you don't have the maths to check it, ask a Wikipedian who does.) And it can't possibly be original research if Évariste Galois proved it in the 1820s!
However, I'm sceptical of the idea that a Wikipedian could produce an original proof (as opposed to a worked example) using undergraduate level maths. I think a new proof would have to be published in a reliable source before it would be usable here.—S Marshall T/C 01:34, 8 August 2012 (UTC)

I think you are on the right track, though the boundary between "worked example" and "proof" doesn't strictly exist. Perhaps the difference is that what we think of as a "worked example" consists only of a sequence of steps that are considered routine, whereas a "proof" involves an additional degree of creativity. My experience is that editors can tell the difference fairly consistently. Zero^talk 12:31, 8 August 2012 (UTC)

This is not 100% on-topic, but I have pointed out once before that a large fraction of mathematical articles (not to mention theoretical physics, etc) are in violation of the following sentence from the NOR policy: "A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the source but without further, specialized knowledge." Since research articles in mathematical journals are primary sources, and a typical "educated person" would have no chance of understanding them enough to verify that a statement is correctly cited to them, a strict application of the policy would mean that they can't be used. A large number of excellent technical articles owe their excellence to the fact that editors ignore this rule. Once before I proposed a change of wording to fix this anomaly but I didn't follow through on it. I'll try again, below soon. Zero^talk 12:31, 8 August 2012 (UTC)

It's one thing to repeat a proof that is given in a primary source where there has been no change to how the proof is presented (an educated reader with access to that proof would be able to validate that the proof as printed in WP matches what is printed in the primary source), and another to create a new proof that hasn't been published anywhere else first. The statement you are pointing to doesn't require the "educated reader" to understand how the conclusions in the primary source were arrived at, only that if given the primary source, they could understand how the summary in WP follows from it. --MASEM (t) 12:39, 8 August 2012 (UTC)

No, even if a proof is copied verbatim from a mathematical research paper, specialised knowledge is required to verify that the premises and assumptions of the proof are correctly represented. And anyway we are supposed to paraphrase for copyright reasons and only someone with mathematical training is able to check that mathematics is correctly paraphrased. Zero^talk 13:16, 8 August 2012 (UTC)

No, that's not our verification policy. We do not expect the reader to be able to confirm if the information is correct or not, only that if they were given the original source, they would require no special knowledge to see how that is summarized in an article. (This is why reliable sourcing is appropriate so that we don't have to worry about putting the onus on "truth" or correctness on the reader). We do assume that an average reader can comprehend the symbols used in proof (either as standard symbols for logic, or as reported in the work) but we don't expect them to understand the step-by-step approaches in the proof. You can't copyright math statements (the language around it, yes, but not the math), but even with summation of the proof, the average readers - able to comprehend the proof as written - should be able to come to the same conclusions made in the WP's summary of it. --MASEM (t) 13:23, 8 August 2012 (UTC)

Of course symbolic mathematics is copyrighted, I don't know where you got that idea from. But copyright is a side issue. The real problem is that mathematics is a language which few understand. Suppose I want to mention this result in Wikipedia (from a recent research paper): Let P and X be closed oriented topological n-manifolds, and assume that P is simply connected. If there exists a map

f:P\to X

such that

f^{\ast }:H^{n}(X;Z)\to H^{n}(P;Z)

is an isomorphism, then

\pi _{1}(X)=1.

As a good editor I would explain the meanings

\pi _{1}

and

H^{n}

and change the notation so that it matches the notation of the existing Wikipedia article. Will you be able to verify that my version correctly reflects the source? I would also read the paper carefully to check if there are additional requirements for the claim to hold, which is very common. Indeed, I see global assumptions like "For the manifolds satisfying our assumptions, without loss of generality, we assume that the

S^{1}

action is effective." If I don't mention that on Wikipedia, will you be able to verify that it isn't needed? Zero^talk 14:01, 8 August 2012 (UTC)

Per WP:V and WP:OR: restating that statement without the consistency changes along with the citation is completely fine. Now, assuming that the authors of the paper have stated what pi and H are prior to this use, and in the article on WP, you have stated what your terms that are equivalent to it are, then it is not original research, and within verification policy, to make that translation (you would likely want to note that in the running text to help though). Now, here's the gotcha: I took a modern algebra course maybe 15 years ago, so I understand the symbols being used and the general terminology, but I have zero comprehension of what that statement is trying to say. But that by no means does that mean this fails OR or V; I can follow from the source to WP just fine, and that's what we would expect the average reader to do. What we would expect you as the WP editor to do is not only to make that non-original, verified restatement of the proof but to provide some rudimentary english prose statement of what that means and implies.

(and on the subject of copyright: you cannot copyright facts, which is what proofs are. But the presentation - including any supplementary texts around the proofs, can be copyrighted.) --MASEM (t) 14:16, 8 August 2012 (UTC)

I do mathematical research for a living, but I deliberately selected a paper in a specialty that I don't understand much. So actually I'm not sure whether "the

S^{1}

action is effective" is required for the correctness of the lemma I quoted. But I'll be perfectly happy if someone who can understand it properly comes along and summarizes it in an article. About copyright, facts are not copyrightable, but proofs are not facts. The mathematical truth that the proof establishes is a fact, but the proof itself is a creative presentation. I can't copyright the fact that London is the capital of the UK, but I can copyright a paragraph explaining exactly how we know that London is the capital of the UK; it's the same thing. Zero^talk 15:26, 8 August 2012 (UTC)

Actually individual steps/part of the proof including math formulas are facts. The proof as a whole package is under copyright as any other text, but if you retell it in your own words then there is no copyright violation. The literal text is under copyright not the content.--Kmhkmh (talk) 21:00, 8 August 2012 (UTC)

There is one thing missing from the discussion which affects the answer. Without clarification there, trying to answer that question "what about putting proofs in an article" is like saying "what is our policy about putting nouns in the article. : And that question is,, in what way (structurally) is the proof put into the article. Is it:

Using the proof to create / support final statement/result? And saying that the final statement does not need to be sourcable because it has been derived in the article? (the section title implies this situaiotn, but then the conversation went elsewhere)
A statement of the final result/statement, and using a proof done outside of Wikipedia and using that proof as a source for the final statement/result?
The article is discussing the proof itself....I.e. discussing that someone did it, what is in it etc.

North8000 (talk) 14:31, 8 August 2012 (UTC)

I agree, In my view for your topic (1) this would be original research as it is synthesis inside Wikipedia (2) The proof outside of Wikipedia should be references and should be a reliable source (per WP:RS), although this will likely be a mathematical article thus primary, (3) In this case the reference should definitely be secondary, otherwise the Wikipedia interpretation of someone doing the proof would be the original research. Arnoutf (talk) 15:23, 8 August 2012 (UTC)

I would argue that unless the proof (or equation derivation, or the like), it is almost never necessary to include the individual steps of a proof within a WP article, unless such steps have been highlighted by secondary sources as being relevant to understanding the proof. We shouldn't even have fully worked examples here. That's what Wikiversity is for. --MASEM (t) 15:28, 8 August 2012 (UTC)

The fundamental problem for Wikipedia here is the fact that the writers of graduate level textbooks assume that you passed your exams at university, so I do see a tension with our policies here. If we don't allow editors who know a lot about some subject some lattitude with presenting complicated stuff so that it becomes understandable, all you can do is present this sort of stuff in a way that it is only understandable to readers who passd all their undergraduate exams at university and perhaps also took some graduate level courses. But then those readers would likely already know more about the topic than is in the Wiki article.

A good example of such a wiki article is Yang–Baxter equation, completely and utterly useless. To get something that is of any use, you have to consider what the audience is we want to write for. We should aim for people with the least background to whom you could still explain the stuff in a few pages. That would then limit you to undergraduate physics and math students. To them you could explain the way some models can be solved using tranfer matrices, that if you have a some suitable family of commuting transfer matrices you can get to the Yang-Baxter equation in some cases and how that leads to solution of the model. This is doable and would not constitute "original research" as we in the field would call it, but you really do need to include examples and extra explanations that you do not find in typical textbooks.

A good example of a wiki article that does this right is Methods of contour integration, it explains in detail how one can use contour integration methods to compute real integrals, this is exactly what a typical Wikipedia reader who would want to look up this stuff would want to read about. Simply quoting a few theorems from a textbook would not achieve this goal. Nothing in this article is really "original research" in the way that a mathematician would define it. Count Iblis (talk) 16:20, 8 August 2012 (UTC)

I'd actually argue that the contour integration is a bad example, simply because it is primarily composed of examples and how-tos, which we explicitly do not include per WP:NOT. This is what our sister project Wikiversity should have, and what we should be linking to. (eg the contribution of those examples don't have to be lost). That article should be at Contour integration, and should explain what it is and what purpose it serves in mathematics (which is partially does) but should not be explaining here how to actually work with it. --MASEM (t) 16:46, 8 August 2012 (UTC)

I'd say in many cases Wikipedia can and should do both. Yes, Wikipedia is not Wikiversity, but it is also used as "hands on reference" by many (students in particular). Meaning we do not only explain what a theorem means and what its general applications are, but we often do provide concrete examples how to use it as well. WP articles can and should serve various audiences and there can be sections targeting different groups/different interests. After Wikipedia is not paper.--Kmhkmh (talk) 17:48, 8 August 2012 (UTC)

But that's exactly why we have Wikiversity, where many more examples and worked problems can be provided. We're a reference work, not a teaching guide. If anything, first and foremost these articles need to bring the concept into some outside frame of reference to explain its importance, where it's used in board terms, etc.; once those basis are covered, then maybe more detailed examples could be provided, but I still argue that the one-click-away Wikiversity approach is better and more consistent for all. --MASEM (t) 17:57, 8 August 2012 (UTC)

A "hands on" reference book is not a textbook. There is no argument that the primary goal to explain importance and use in broad terms, but that is not inconflict in provided concrete examples/apllication as well. Again WP is not not paper. Wikiversity is not structured for quick lookups but rather for learning subjects thoroughly and from the scratch (which is different thing). Also while Wikiversity is just click away, the state of some/many of its sites is not good enough that I personally would recommend them. Currently you often find better free material elsewhere.--Kmhkmh (talk) 21:29, 8 August 2012 (UTC)

I think the focus here on WP:NOTHOWTO is misguided as an approach to mathematical proofs, though it can provide some guidance in the use of examples. An encyclopedia article is, fundamentally, an explanation of a topic, and so is a proof. It explains why something is true, and can in some circumstances also explain the history of a line of thought. So, high level overviews of proofs already present in the literature are, to me, unproblematic, especially when (but not only when) the proof itself is covered in a survey article or textbook rather than only in research articles [rewording here to sidestep the separate debate over whether peer-reviewed research articles in mathematics are primary or secondary]. The problem comes when proofs are made up directly by Wikipedia editors and presented without any source for the proof. In many cases these kinds of proofs (e.g. proof by induction of simple mathematical identities) are straightforward to verify, but they are still original research. As for relegating all examples to some other web site: no. Examples are an essential part of explanation. Like anything else they can be done to an extreme, and an example worked in tedious step-by-step fashion is usually going too far, but forcing mathematics articles to avoid mentioning examples would be like forcing book and movie articles to avoid any mention of plot (something that is also often overdone but would be nonsensical to eliminate altogether). —David Eppstein (talk) 18:43, 8 August 2012 (UTC)

Here's an example of a page that I know I can talk about (familiar with the subject) but recognize as a problem: Fundamental thermodynamic relation. Ok, perhaps the derivation in the first section is useful to show how it derives from the first and second law of thermo, ergo universally consistent, so maybe that derivation (equivalent of a math proof) could stay. But with the article's focus so much on the derivations, it tells the reader nothing about the equation beyond that it is tied to the laws of thermo. The article needs context, in this case, noting that one can now take any equation of state or experimental data and generate thermo equations and relationships from that, among other uses. In that light, it absolutely doesn't need the proof from statistical thermo, though this is certainly useful as a Wikiversity element. Math articles that get proof-heavy often tend to ignore this factor, that we need to related this down to the common reader first and foremost before filling in proofs that appeal to the mathematician. --MASEM (t) 19:57, 8 August 2012 (UTC)

I think we are getting a bit off the track here. Nobody is arguing against providing context and this is not one of the most important task of a good article. Nobody (aside from Zero maybe) is arguing for proofs as a standard components of math articles. As I understand it there is a long standing consensus of the math project, that proofs are not a standard component. So nobody is arguing for proof heavy (in particularly not in a textbook or wikiversity sense), but there cases where providing a proof or derivation or parts of them is somewhat natural and useful for an article (as for instance described by David Eppstein above), similarly for examples. Another thing has to be kept in mind here, we are arguing here about the formulation of a policy, which also can be seen as minimal standard that is required by all WP entries. That includes "bad" and "mediocre" ones as well and not just what we might consider good or excellent articles. Meaning the policy needs to be workable for mediocre or even bad articles.--Kmhkmh (talk) 21:50, 8 August 2012 (UTC)

RfC about WP:BLPPRIMARY

WP:BLPPRIMARY currently forbids the citation of any public record that contains personal details. It currently reads:

Do not use public records that include personal details, such as date of birth, home value, traffic citations, vehicle registrations, and home or business addresses.

Should this sentence in be kept as is, removed, or changed? RfC is here:[11] LK (talk) 15:28, 10 August 2012 (UTC)

List articles and NOR

Upon reading WP:SALAT I came to a conclusion that there is some contradiction between those guideline are the NOR policy. The guideline say: "The potential for creating lists is infinite. The number of possible lists is limited only by our collective imagination. To keep the system of lists useful, we must limit the size and topic of lists." The guidelines contain no reference to the NOR policy, and the restrictions applied to the size and topic are dictated only by the usefulness of lists.
However, I see a serious potential of original research here. By arbitrarily choosing the list's subject, a user may create new topics. By expanding the annotations and the lede of the list article users can create de facto full articles about the topics that do not exist in literature, because the difference between the article and extended lists is not obvious.
In connection to that, I propose to add to WP:SALAT a link to WP:NOR, and to add to the policy a notion that, when no sources exist that draw more than a trivial connection between the items listed in the list article, such a list must contain just a very brief description of each item (not more than one short sentence), and the lede of such list must be totally descriptive.--Paul Siebert (talk) 21:43, 26 June 2012 (UTC)

I agree. We should only have finite lists of clear populations, e.g., the list of current members of the US Senate. A list of famous graduates of Oxford for example would be a nightmare, but we could mention in the article - if sourced - that many PMs were graduated, then have an article that lists them. TFD (talk) 22:36, 26 June 2012 (UTC)

We've long had lists of "People of X" (where X may be from a certain city, country, college, etc. etc.) Almost always these lists are strictly limited to those people with non-redirect bluelinks - aka notable people, and generally stated in the lede of the list. Are these lists OR? Not really - we know people must originate from some city, or graduate from some college, or the like, and we do fully expect these claims to be sourced.

I think what we need to be careful on is when the list is construed in a way to push a point. "List of people believed to be homosexual" would never fly regardless of the quality of the sourcing, for example, though on the other hand "List of homosexual people" (where a criteria requires self-assertion of this fact in a reliable source interview) would be ok. This is a whole issue we had discussed in trying to describe the notability of lists at WP:N a while back, but we never came to a clear answer to clearly saying when lists topics are notable, and I think the same problem extends to NOR and NPOV. Basically it is a point we all should be aware of in constructing the theme of the list, but it is difficult to nail down to a line since some novelty in organization may actually be nature and appropriate. --MASEM (t) 22:47, 26 June 2012 (UTC)

Lists need notability like any other article, one needs to show someone is interested in such a list by having something similar in a reliable source. The only special exception for lists is that one is allowed to set up a article on a section of a list, e.g. composers from A to E. Dmcq (talk) 23:49, 26 June 2012 (UTC)

The problem is that WP:SALAT does not tell about that. Per WP:SALAT, it is quite possible to create, for example, a List of Jewish mass-murderers, List of black homosexuals, etc. Not only such list will create non-existent topics, they would violate WP:NPOV also. We need to link list-related MOS to the NOR policy and to explain clearly that notability criteria for lists and for articles should be essentially the same.--Paul Siebert (talk) 01:03, 27 June 2012 (UTC)

Dmcq, when we developed WP:LISTN, we never came to that conclusion because there's just too many exceptions. We strongly recommend lists be notable themselves, or be a grouping that is notable, but we could not say lists that didn't fall into either case were or were not notable. And yes, there's definitely a need to link NOR/NPOV in discussion of list topics, but we just need to be aware that we can create novel groupings of elements that have never been published before as long as its not POV pushing or an unique way that doesn't fall naturally out of regular coverage. --MASEM (t) 01:37, 27 June 2012 (UTC)

Disagree. What Paul Siebert and TFD are essentially proposing is that it should only be possible to create a List of fictional dogs if there exists a single source that lists all fictional dogs contained in that list. He is proposing that the use of multiple sources to create such a list is deemed to be synthesis and thus such a list be prohibited. Clearly that is too restrictive and is contrary to the whole purpose of a list which is to aid navigation through the organisation of articles. --Nug (talk) 02:18, 27 June 2012 (UTC)
I'm with Nug. This comes up every now and again. NOR does not prohibit you from making a list merely because your sources don't include every item, or because your sources deal with it in prose. If you've got an entire book on The Top 100 Causes of Fever, and it uses paragraphs throughout, you may (1) present that information in a list format if you so choose and (2) expand the information citing other sources) to include items not in that one source. WhatamIdoing (talk) 03:10, 27 June 2012 (UTC)

Yes, NOR do not prohibit me from making a list if the sources don't include every item. However, and I explained that to Nug elsewhere, the topic per se must exist, i.e., some source must exists that groups majority of the list's items according some non-trivial criterion. Thus, you may create a list of WWII battles based on some WWII handbook, and add to that list some battles not mentioned in this concrete source. However, you cannot create a list of black rapists based on the book about mass rapes during the WWII if no non-fringe source exists that defines such a topic (rapes by Afro-american, or Caucasian, or Jewish, etc military personnel during the WWII).

In actuality, the WP:SAL lede says: "Being articles, stand-alone lists are subject to Wikipedia's content policies, such as verifiability, no original research,neutral point of view, and what Wikipedia is not, as well as notability guidelines." The problem is that that thesis has not been properly explained in the guidelines' body. In addition, no strict dividing line exists between the lists and the articles (for example, compare this ans that), so it is quite possible to create a piece of original research by gradual expansion of the list created based on the arbitrary chosen criterion.

In my opinion, the problem may be resolved if we add to the policy that strictness of the notability, NOR and NPOV criteria directly correlate with the amount of prose in the list: if the list consists of a very brief and descriptive lede followed by short descriptive single sentence items, the NOR criteria should be much milder.--Paul Siebert (talk) 03:23, 27 June 2012 (UTC)

Belated clarification: we have had editors claim in the past that you must be able to find one source that includes every item on your list, and it is that false claim that I reject. So, for example, if you have 100 causes for fever, it would be impermissible to have more than 100 items on the list, no matter how stellar your sources were for the 101st cause of fever. The example may be in the archives; the editor asserted that Barack Obama could not be added to List of US Presidents (after his election to that office) unless and until a reliable source made a list of US Presidents that included his name. WhatamIdoing (talk) 04:26, 22 August 2012 (UTC)

I don't think this restriction would solve any problems. Also, WP:N plainly states that notability does not affect what or how much you can put on the page, just whether we should have that page at all. WhatamIdoing (talk) 18:09, 28 June 2012 (UTC)

In any event, WP:SALAT currently says: "The potential for creating lists is infinite. The number of possible lists is limited only by our collective imagination. To keep the system of lists useful, we must limit the size and topic of lists." Some users may interpret that as a carte blanche for creation of any list they want. In my opinion, the sentence

"The number of possible lists is limited only by our collective imagination."

should be modified as follows:

"The number of possible lists is limited only by our collective imagination, and by our content policy WP:NOR, WP:NPOV and WP:V."

--Paul Siebert (talk) 19:52, 28 June 2012 (UTC)

Lists are properly seen as navigation aids, that is why they are discussed in Wikipedia:Categories, lists, and navigation templates. Your proposal is still too restrictive. --Nug (talk) 20:06, 28 June 2012 (UTC)

My proposal is by no means restrictive. Without this reservation, the WP:SALAT section simply contradicts to what the lede of the same page says ("Being articles, stand-alone lists are subject to Wikipedia's content policies"). As this discussion demonstrates NOR NPOV and V are applicable to stand alone lists, no matter if they are seen as navigation aids or not.--Paul Siebert (talk) 20:23, 28 June 2012 (UTC)

Support change to WP:SALAT as detailed by Paul Siebert, i.e. to say WP:NOR, WP:NPOV andWP:V apply. People seem to have got the idea that the manual of style overrides the content policies and guidelines. Dmcq (talk) 22:16, 28 June 2012 (UTC)

WP:NOR applies in what way with respect to lists? That all items in a list should be sourced to a single publication? This is in essence what Paul Siebert is proposing. --Nug (talk) 22:27, 28 June 2012 (UTC)

Nug, you again misinterpreted my words: I didn't say all items need to be listed in the same source. My point is that some reliable non-fringe source is necessary to define the topic. Do you see a difference? I believe my explanation is crystal clear. Please, do not misinterpret my words any more.--Paul Siebert (talk) 05:04, 29 June 2012 (UTC)

Comment. The background to Paul Siebert's request here is this discussion at Talk:Soviet_occupations#List_Article, his contention being that there is no one single source that mentions all the Soviet occupations mentioned in the list. Although all individual occupations mentioned in the list are cited to reliable sources, he contents that aggregating them into a list is synthesis and thus should be prohibted. --Nug (talk) 22:22, 28 June 2012 (UTC)
- Yea, Paul's view is completely counter to how list articles can be constructed. No single source has to discuss all the occupations in whole, but the grouping of the occupations (presumably each well sourced) is completely fine particular since its a neutral, non-POV state ("here are the occupations of the Soviet Union"). And because the individual occupations are notable themselves, there's no issue there as well. --MASEM (t) 22:27, 28 June 2012 (UTC)

First of all, you Masem seems to argue not with what I propose, but with Nug's interpretation of my words: " What Paul Siebert and TFD are essentially proposing is that it should only be possible to create a List of fictional dogs if there exists a single source that lists all fictional dogs contained in that list." In actuality, what I say is the following: since per WP:SAL, list articles, similar to other articles, are subject to our content policy, we can add new items in the list that belongs to some existent topic, but the list cannot create a non-existent topic. If the Nobel prize winners A, B, C, and D are Caucasian males, and each of them is notable enough to have his own article, it would be hardly correct to create a list "List of Caucasian males who got a Nobel prize". We can create such a list only if some mainstream reliable source exists that discusses a category "Caucasian male Nobel prize winners. Otherwise, we open a Pandora box: we create a topic, it may grow to a full article and eventually we get a piece of non-neutral original research.

With regard to "occupation of the Soviet Union", have you noticed that that article (which, according to Nug is a list, but it is meets all criteria of the full article) is not about the occupations of the Soviet Union, but the article about occupations by the Soviet Union. I asked, for several times, to provide a source that links such different events as annexation of the disputed province Bessarabia (a former Russian province, which was annexed by Romania, the act, which had never been recognized by the USSR and many other states), Soviet occupation of Nazi Germany after the Allied victory in the WWII, and the war in Afghanistan (which is not described as Soviet occupation in the main article, so it is simply a POV-fork). I have been waiting for almost three months, but no sources have been provided to link all those events. Meanwhile, when I type "Soviet occupations" google gives mainly the links to Wikipedia itself, and to its mirrors, thereby confirming that Wikipedia has created a new topic. --Paul Siebert (talk) 05:01, 29 June 2012 (UTC)

PS. By writing that, I by no means argue against a disambiguation page, similar to the Japanese occupation page. --Paul Siebert (talk) 05:10, 29 June 2012 (UTC)

Paul, if you first find a source that links Akamaru, Kiba Inuzuka's dog in the television series Naruto with Chief from The Fox and the Hound, two dogs listed in List of fictional dogs, then I will find a source that links the Soviet military occupation of Afghanistan with the Soviet military occupation of Germany. According to Paul "List of fictional dogs" can only exist only if some mainstream reliable source exists that discusses a category "fictional dogs". --Nug (talk) 05:40, 29 June 2012 (UTC)

There are plenty of lists that are "Lists of X" where "X" is not a notable topic in of itself, but is one that derives from natural or logical extension of other notable topics. Yes, care must be taken that that grouping is not biased or created to push a point; NOR and NPOV apply to lists in as much as they apply to articles. But you do not need to have one single source that link all those events to X, as long as individual sources for each event declare it part of that arena. The Soviet Occupation article would likely be considered a list because it is simply grouping several already-existing articles as well as some non-notable events into a navigation aid. --MASEM (t) 05:50, 29 June 2012 (UTC)

The Soviet occupations article is definitely not a single topic but a list type article and should be treated as such. It looks like original research to me trying to make an integrated article out of them, that would require a source talking about the topic as a subject rather than just a list. Dmcq (talk) 10:49, 29 June 2012 (UTC)

Well obviously it is a list of military occupations by the Soviet Union. The lede describes a clear criteria for inclusion and each section has a link to the respective article and a summary of that article. --Nug (talk) 13:24, 29 June 2012 (UTC)

I agree it is a list article in its approach, but again, that doesn't require that we have one single source that links them all together. Is there any question that the former Soviet Union has occupied other nations? If not, this is a natural grouping of such occupations as a list, it doesn't matter that no source explicitly lists each one in a single shot. --MASEM (t) 13:35, 29 June 2012 (UTC)

Again, we do not need a single source that links all of them together. However, we do need a source that sets the topic: is there any question that some white men got Nobel prize? However, we can create a list of all Nobel prize winners, a list of American Nobel prize winners, but I don't think a list or white male Nobel prize winners would be a legitimate subject of the list. Similarly, it is possible be to create a list of countries under Soviet occupation as a result of the WWII (that is a part of a single story), however, a list of countries that are believed to be occupied by the USSR during different periods of its history is hardly a correct subject of the list: I asked, for several times, for a source that links WWII, post-war and Cold War Soviet occupation within the frames of the same concept, however, no sources have been provided.--Paul Siebert (talk) 14:31, 29 June 2012 (UTC)

--Paul Siebert (talk) 14:05, 29 June 2012 (UTC)

"believed to be occupied by the USSR during different periods of its history" - as long as 1) they are being called "soviet occupation" or a near-equivalent phrase, and 2) backed by reliable sources for each entry, there's no NOR/NPOV going on here in combining all Soviet occupations. I do recognize the concern when you say "believe", as I could see the possibly of someone in a politically/ethnically-charged manner assert that some action by the Soviet Union was an "occupation" but no RS ever back that up; I have not reviewed every statement on that list to confirm this, but AGF that this is not the case. We use a similar logic on List of commercial video game failures in that we require sources to call out the entries as commercial failures instead of just making assumptions without sources. --MASEM (t) 14:22, 29 June 2012 (UTC)

Using the same logic, as long as some white males are believed to get a Nobel prize, and that is confirmed by reliable sources, there is no NOR/NPOV issues with combining them together in the List of white male Nobel prize winners.

Well, I'll ask the same question is a different way: is it correct to create a full article that discusses all Soviet (American, British, etc) occupations (or white male Nobel prize winners) in a single article devoted specifically to this subject? I believe, in the absence of RSs that confirms that such a topic does exist in literature the answer would be "no". However, for the lists, which must comply with the same content policies, the answer is, according to you, "yes". Don't you see any inconsistencies here?--Paul Siebert (talk) 14:39, 29 June 2012 (UTC)

The reasoning on the white male Nobel prize list being a problem is that there is implicit gender and racial biases when we talk about a large grouping of people; to classify any list out by gender, race, religion, sexual preference, etc. is going to be a non-neutral action. To demonstrate that we (the Wikipedian editors) are not introducing that bias ourselves, we'd need reliable sources that calls out this grouping as novel, at which point its not our OR or POV that we're putting into text. You'd very likely never see any type of "white male Nobel prize" (doing a quick glance through Google) but there's plenty of reasons to have "female Nobel prize winners" (as this group has been called out) or similar "minority" classifications. But that just doesn't seem to exist for white males, so while assembling the list is not violating NOR, it is a problem with NPOV unless sources defer that claim for us.

And no, "List of X" does not depend on "X" being a valid topic for an article. I doubt there is any specific source that describes, in generic terms, the idea of a "Soviet occupation" without ever referring to any specific occupation. But there have been several Soviet occupations stated by reliable sources over the years, so there's no doubt they happened - before, during, and after WWII It's clear that the concept is related to the notable topic of the USSR, and more likely under History of the Soviet Union; its even better that its limited to the lifetime of the Soviet Union and not of the previous Russian empire or the post-dissolution. There is thus no NPOV connecting these together, and as long as each occupation is clearly identified as such by RSs, no OR in putting them into a single list. --MASEM (t) 15:17, 29 June 2012 (UTC)

And to add (after seeing a response below), a list that describes all occupations of other countries by another country (extending outside of the Soviet Union), would not be OR or NPOV as well, as long as each occupation is sourced to be described that way in RSes. There may be organization and size issues (I figure there's a LOT of those in human history) but the discussion of all known country occupations is a completely fair list topic. --MASEM (t) 15:20, 29 June 2012 (UTC)

You are talking about "discussion", but the List of fictional dogs does not discuss anything, and the List of military occupations does not discuss either. By allowing stand-alone lists to expand we allow users to create the topics that do not fit notability criteria applied to normal articles. In connection to that, see my comment below.--Paul Siebert (talk) 15:46, 29 June 2012 (UTC)

Oppose to change. There is nothing wrong with creating pages like List of Japanese occupations, List of military occupations by the United States, etc. This does not contradict NPOV or anything. The objects included in these lists are notable and obviously belong to the same category. Such lists can have such format and include such amount of information about every item as decided by WP:Consensus at talk page of the corresponding list article, not here. The items of the list can be taken from multiple sources that are not necessarily consistent with each other. They are even required to be taken from multiple sources per WP:NPOV. My very best wishes (talk) 14:00, 29 June 2012 (UTC)

Firstly, Japanese occupation dab page already exists. However, it combines the events that have been combined together (by many sources) according to a very concrete criterion: all of them occurred during the WWII (and occupation of Japan was an immediate result of the WWII). With regard to List of military occupations by the United States, I have a double feeling. Initially, I had no objections against the List of military occupations by the USA, USSR, or similar articles. However, upon realizing that stand alone lists must comply with the same NOR and NPOV criteria as other articles do I started to doubt. Our NOR policy says: "Do not combine material from multiple sources to reach or imply a conclusion not explicitly stated by any of the sources." By combining all white male Nobel prize winners in the same list we imply the idea of white male supremacy. By combining all military occupations by some concrete big country we inevitably imply it was conducting aggressive and expansionist policy during its whole history. If such a conclusion does not explicitly exists in some mainstream reliable source, it is synthesis. If lists comply with the same content policy, such an approach is hardly legitimate.--Paul Siebert (talk) 14:17, 29 June 2012 (UTC)

No, I do not think that such lists actually imply anything beyond the existence of many occupations. For example, a list of software would hardly imply anything except the existence of numerous software. Of course if a list of occupations by nation X includes a direct statement that nation X is a "nation of occupiers" (in fact such statement can frequently be well sourced), that might be a matter of concern related to specific list, but not to the general policy. My very best wishes (talk) 14:45, 29 June 2012 (UTC)

By combining all US (Soviet, British, etc) occupations together we imply a conclusion that those countries conducted expansionist policy during whole their history. By combining all white male Nobel prize winners we imply the idea of white male supremacy. All of that may be right or wrong, however, we need a source that tell that explicitly.

And, by the way, you seem not to understand the main idea of this thread: I propose no changes to policy, I propose to bring the guidelines in accordance to our policy.--Paul Siebert (talk) 14:48, 29 June 2012 (UTC)

A reader (like you or me) has every right to personally interpret the list of Microsoft software as a proof that Microsoft was conducting aggressive and expansionist policy [on the market] during its whole history, but it does not make such list problematic. And if the claim about "aggressive and expansionist policy" can be reliably sourced, there is nothing wrong with including such claims in the introduction.My very best wishes (talk) 17:03, 29 June 2012 (UTC)

Oppose change - a cure in search of a disease, a solution in search of a problem, and a great example of instruction creep. Cheers. Collect (talk) 14:43, 29 June 2012 (UTC)

Could you please elaborate on that? Your post contains no concrete statement, so it is just a universal "IDONTLIKEIT".--Paul Siebert (talk) 14:50, 29 June 2012 (UTC)

1. You have not demonstrated that a problem exists.

2. You have not demonstrated that this would solve an existing problem.

3. This is a "new instruction" hence clearly is "instruction creep."

To wit: Instruction creep occurs when Wikipedia guideline or policy pages contain statements that expound upon unimportant details and minutiae.

Thus my post is not an IDONTLIKEIT post - it is grounded in the facts and substance of what has been presented. Collect (talk) 18:52, 29 June 2012 (UTC)

Upon looking at List of fictional dogs I partially understood Masem's point. Yes, such a list has a right to exist, because it implies no conclusions and is used for navigational purposes only. The close analog of such list would be List of Nobel laureates and List of military occupations. These lists contain no original research, imply no conclusions and are in full accordance with out core content policies. However, a List of white male Nobel laureates would violate our policies if no mainstream sources exist that define such a topic. Similarly, list of military occupations by country, during whole history of this concrete country implies some conclusion, and if no sources exist that supports such a conclusion, then we can speak about a violation of our policy.
In addition, in some lists, the prose has a tendency to expand, and such a list may become a full scale article. In connection to that, I am asking again, what our policy tells about that, and, if different criteria are applied to the lists and full articles, what are the criteria that allow us to discriminate the former from the latter?--Paul Siebert (talk) 15:33, 29 June 2012 (UTC)

You're missing the NPOV aspect. "White male Nobel laureates" is immediately a biased statement, simply because of the weight of gender/racial equality concerns. "Occupations by the Soviet Union" is not biased at all; the Union did occupy countries, they were called "occupations", etc. There's no conclusion to be made from that beyond listing it out. If it was more the case of said occupations of being of what individual editors made of it (hypothetical example: a Czech calling the 1968 action as an "occupation" why a Russian calling the same thing "a nice friendly visit") so that there was a large disagreement in RSs of what really was an occupation or not, ok, there may be some question, but this doesn't appear to be the case; the majority of RS agree each case can be defined as "occupation". So there's zero POV pushing happening here as best I can tell. --MASEM (t) 15:49, 29 June 2012 (UTC)

I am not missing anything. If something is not biased, that is easy to demonstrate. We do not outline a limited set of statements that are intrinsically biased, and we do not provide a carte blanche for other topics.

Regarding your "There's no conclusion to be made from that beyond listing it out", that depends on the form of the list. If the list has the same style as the List of fictional dogs, yes, there is probably no problems with such a list, for example, a List of countries occupied by the USSR from 1939 to 1991. However, when we are talking about an expanded list, I see almost no difference between such a "list" and a full article. In connection to that, I repeat my question: what is the difference between the "expanded list" and full article, and why the NOR/NPOV criteria should be milder to the former than to the latter?

Regarding your second point, yes, if different terminology can be applied to some item of the list, NPOV should be taken into account. In your example, "a nice friendly visit" to Czechoslovakia would be a totally fringe views, however, "occupation of Afghanistan" is described as a "proxy war" in the main article, not an occupation. Therefore, a list should not give undue weight to one existing viewpoint. However, that is a different story.--Paul Siebert (talk) 16:14, 29 June 2012 (UTC)

There is no difference in the application of NOR/NPOV between a list and an article, and certainly not in this case. All pages in mainspace, short of disambiguation pages, are expected to meet all content policy and guidelines, and the only place where lists and articles really come up is if you're trying to go to Featured content, where we decide if its more a list or more an article. You're trying to argue that some exception is being made on the Soviet occupation page but there isn't. It complies fully with NOR and NPOV. And as for the Afghanistan piece, there's plenty of sources search on "Soviet occupation of Afghanistan", so just because WP calls it a proxy war doesn't mean that RS never called it an occupation. --MASEM (t) 17:19, 29 June 2012 (UTC)

Well, is there is no difference, then explain me why do you oppose to addition of that explanation to WP:SALAT? I see the need in such an explanation, because during some discussion I saw the argument that, since the article is in actuality a list, then we are not restricted with NOR/NORN/V policies here. Since that argument was put forward by an experienced user, I think the WP:SALAT is ambiguous.

With regard to "Soviet occupations", I am not trying to argue about that here. You probably noticed that I didn't bring this argument here. I am perfectly comfortable to discuss any other example.--Paul Siebert (talk) 17:59, 29 June 2012 (UTC)

I personally never said that NOR/NPOV shouldn't be added here. But what does seem to be in play is that you're trying to argue for their addition here to nullify the Soviet occupation (and other similar articles) article, starting with the false premise that a topic needs to be discussed in full by a source to be an appropriate topic for WP, which is a fallacy. Again we should be applying NOR and NPOV to list contents, but remember that NOR does allow for non-biased synthesis of information which is what we're talking about. --MASEM (t) 18:06, 29 June 2012 (UTC)

If you never said that NOR/NPOV shouldn't be added to SALAT, then I see no reason to argue with you about that. We need just to implement these changes in SALAT and forget about that.

With regard to Soviet occupation, are you talking about my real intentions, or about Nug's interpretation of them? My views of the "Soviet occupations" page evolved with time. Initially, I asked the users who worked on this article to provide sources that demonstrate that such topic exists. In response, they agreed to change the lead and declated that the article is a list now (although some of them still maintain that this is an article about some single subject). Upon changing the lede they declared that the issue had been resolved, because, since the article is a list, no sources that prove the existence of the article's topic in general are needed.

I didn't mind to keep this list, although, I insisted on its renaiming into the "List of Soviet occupations and on bringing it into accordance with List of military occupation style, because the article in the current form is the article, or summary style article (similar to the WWII article), but by no means it is a list. However, upon meditation, I realized the we have a general issue, and I decided to discuss it here. I believe this my explanation should resolve the issue. In future, please, do not accuse me in attempts to modify policy to push some local POV.--Paul Siebert (talk) 18:19, 29 June 2012 (UTC)

The problem is that you started with a statement like "By arbitrarily choosing the list's subject, a user may create new topics. By expanding the annotations and the lede of the list article users can create de facto full articles about the topics that do not exist in literature, because the difference between the article and extended lists is not obvious." which has been pointed out is not true. It didn't matter if it was specifically the Soviet occupation article, just that that statement was a fallacy. NOR/NPOV still apply to lists and list content (so I do support their addition here), but neither of them prevent the type of list your original statements suggested.

And no, renaming doesn't have to be done either. List articles, expanded or not, do not require them to start "List of". I've found its generally more the case as the list becomes more a navigational aid that "List of" is used more often but certainly it is not required. As long as it is clear what the inclusion rules are (in the case of Soviet occupation, its right there, first sentence), the title doesn't have to indicate that. --MASEM (t) 18:32, 29 June 2012 (UTC)

All of that is of secondary importance. I see you have no objections to updating of WP:SALAT, and that is good. With regard to the rest, see my questions below. If you believe that the same criteria are applied to SALs, to summary style articles and to articles, please, explain, can I take several items, each of them are notable per se, combine them into a list based on some nominal linkage between them, expand the list, and eventually get a summary style article about the topic that does not exist in literature?--Paul Siebert (talk) 18:38, 29 June 2012 (UTC)

To your question: yes, but only if that linkage is not something that is POV-loaded or biased, or that would require that POV/biasing would be necessary to shoehorn in topics. "Soviet occupations" is fine as the country did perform several - albeit different reasons, but it was still the same country, so that's a non-biased link. "White male Nobel prize winners" is a problem since you're introducing the bias of race and gender. The same problem would apply to "Female Nobel prize winners" but here while there is the bias, we can find sources that describe it out that way, thus removing the bias from WP's editors and to the RS, which apparently see it this way. --MASEM (t) 18:42, 29 June 2012 (UTC)

Do you mean NPOV are applicable, but NOR is not?--Paul Siebert (talk) 18:45, 29 June 2012 (UTC)

NOR applies too, but keeping in mind that NOR allows for reasonably simple, obvious synthesis, of which, say, gathering together the Soviet occupations. --MASEM (t) 18:48, 29 June 2012 (UTC)

Then it would be useful to define the limits of simple and obvious synthesis that are allowed per NOR. For example, let's consider the following case. Some reliable sources describe behaviour of some Western powers towards some small states or nations as "betrayal" ("western betrayal at Munich", "western betrayal in Yalta", "western betrayal of Arabs" etc.), so each of those topics is notable per se. In connection to that, I have following questions. What of the following can be considered as "simple and obvious synthesis that is allowed per NOR":

Creation of the "List of western betrayals";
Development of such a list into an extended list;
Creation (based on this extended list) of a summary style article "Western betrayal" that starts with "Western betrayal is the term which refers to some aspects of the foreign policy of several Western countries during the period from the Treaty of Versailles in 1919 through World War II and to the Cold War".

As this search demonstrates, each separate topic is notable and verifiable.--Paul Siebert (talk) 20:37, 29 June 2012 (UTC)

The only thing that would prevent this is the quality of the RS's in how they define "betrayal", as it is a potentially more opinionated word than "occupation". If the only sources for these are, say, for the Yalta pieces are coming from Yalta sources, and so on, their reliability is in question, and there's like NPOV going on. But if dozens of historians all agree they're all individually betrayals, that's fine. --MASEM (t) 22:46, 29 June 2012 (UTC)

I don't think this interpretation of NOR policy is correct. Many instances of different "betrayals" by one or another Western power do not form a single concept "Western betrayal". Just compare this discussion with the discussion in the previous section: much more innocent things are not allowed per NOR.--Paul Siebert (talk) 23:44, 29 June 2012 (UTC)

The problem is that you have a team that is "loaded", that being the word "betrayal". Yes, everyone knows what it means, but how it is used in the geopolitical stage? I have a feeling if you looked through highly reliable sources of a purely academic nature, you'd find several different definitions. On the other hand, the geopolitical meaning of "occupation" is pretty dang clear. Again, there are fringe takes on the word, but in academic circles, it seems to be a pretty sound, concrete topic. It would be very difficult to contest any of the Soviet occupation as "occupations" but it would rather easy to contest "betrayals". But you can always be very explicit in what "betrayal" means - you must have such-and-such a source confirm this, or it must be X doing Y to Z, or whatever, and avoid the fringe interpretations.

The end result: there is absolutely nothing wrong with grouping numerous elements that share common aspect X as long as the aspect of X can be reliably sourced (WP:V), does not require "creative" interpretation or correlation (NOR) and does not create a bias (NPOV). --MASEM (t) 00:06, 30 June 2012 (UTC)

Masem, you will be surprised, but the term "occupation" is also not as clear as you think, and some events that are described as "occupation" by one author are described otherwise by another one. However, I suggest you to forget about occupation. I came to this page for quite a different reason. (Regarding "Soviet occupations" I initially was pretty comfortable with the idea to convert it to the classical list, similar to List of fictional dogs). Therefore, again, please stop your allusions: I am here not because I want to change policy to push some POV.

With regard to betrayal, that is a part of a real dispute (I just changed some arguments for simplicity): if one author speak about "western betrayal" of Czechoslovakia by Britain and France in 1937, and another author speaks about "Western betrayal" of Poland by the US and the UK in 1945, can we speak about some general "western betrayal" as a general concept that is used to describe policy of western powers in XX century?

This is a very interesting question, and I am even contemplating to start an RfC on that account: can Wikipedians create new topic by themselves based on well sourced and notable examples that may be (or may be not) combined within some more general concept, but which have not been combined so far in reliable mainstream sources? I cannot predict the result of such a RfC, but, in my opinion, such a question deserves to be asked.--Paul Siebert (talk) 00:33, 30 June 2012 (UTC)

The selection metric you're using is actually very important here. "Occupation" might be a loaded term, but in reference to everything on the Soviet occupation page, those examples are far from it. "Betrayal" is a loaded term - is it as simple as a broken verbal promise? - and thus if we group betrayal together we've got to make sure that the reliable sources are talking the same thing. I can't recollect the exact cases above, but if one case was where country A was betrayed by B causing significant economic damage to A, and another case of country C being shunned because country D didn't hold their promise to hold an event these but otherwise causing no other issues, there's no way those two would be connected. But this is because "betrayal" has a wide breadth of meaning, particularly into the fringe. A list would be possible, but its got to have a strong definition of what betrayal is for each item added to it to prevent trivial/fringe considerations. Again, I bring up List of commercial failures in video gaming in which we don't include any "flop" but ones specifically called out by the industry press as commercials failures; Similarly, List of Internet phenomena requires a highly-reliable source to claim the material is viral or the like, despite lots of ways in the fringe you could says something's popular on the Internet. Making that type of definition (explicitly in the article text, or in the talk page) to limit what elements are included or the like is ok, as long as we're still following NPOV and WP:V.

There's also a flip way of looking at this. Imaging you could write an article that would encapsulate the list, talking about the subject that's at the next highest level; in the case of Soviet occupation, it may be "Military actions of the Soviet Union". (and lets say that's a true article). Does it make sense that there would be a section "Soviet occupations" in that article? Or in the case of betrayals, lets say its an article on Intra-country politics, and would a section on "betrayals" make sense? If that section makes sense to have because its one way to organize the larger information, then its probably okay as a standalone list. This is not a test that works well in all cases but works quite often. --MASEM (t) 00:50, 30 June 2012 (UTC)

Re your "A list would be possible". Whereas I tend to intuitively agree that for some topics lists are acceptable, and articles are not, I still cannot understand that, taking into account that, according to WP:SAL, the same criteria are applied to lists and to articles.

Re your " Does it make sense that there would be a section "Soviet occupations" in that article? ", no. Different events currently described in the "Soviet occupations" article are seen quite differently in literature: Bessarabia (a former Russian province annexed by Romania) had been annexed by the USSR, and this fact is not a subject of debates in literature (except, probably, Romania). The events in the Baltic states are also described differently, from "annexation/incorporation" through "occupation quasi-bellica' to "military occupation" (depending on the views of one or another author). Occupation of Czecholovakia was a typical military occupation. Occupation of Germany was a very strange case (some authors even use the occupatio sui generis formula). Occupation of Afghanistan was a proxy war similar to Vietnam, and so on. As a result, those events will never be analyzed together in the hypothetical "Military actions of the Soviet Union" article; each of them will be discussed in the proper historical context.

However, all of that is just details. I am more interested to analyze a general question: can we create new topics by themselves?--Paul Siebert (talk) 01:42, 30 June 2012 (UTC)

Within limits/reason we can, we not just have list but articles as well the exact naming of which exact naming does not exist in sources but the subject as such is still considered notable. In these cases usually a purely descriptive term is used. But note it may only be "new" in the sense that there is no established name or definition in the sources, it's actual content however is not new at all.--Kmhkmh (talk) 09:18, 30 June 2012 (UTC)

Oppose I have a problem with how the suggested change might be used. I see no problem with example of the "list" of soviet occupations. I think we need to keep in mind the goal of our guidelines is to ban "real" (novel) OR and "real" (novel) synthesis, but not to ban informative and useful articles/lists which could be seen as possible OR or SYNTH violation if you try really hard. I have hard time seeing anything novel in that example. Summarizing/categorizing/listing/combining material from different sources is standard task in any encyclopedic writing.--Kmhkmh (talk) 16:17, 29 June 2012 (UTC)

I am not sure you understand what we are talking about. The discussion is primarily about bringing WP:SALAT (a WP:SAL's section) in accordance with the WP:SAL itself. If some topic contains no original research or it does not violate the neutrality policy, it can be easily demonstrated. However, to say that "The potential for creating lists is infinite. The number of possible lists is limited only by our collective imagination. To keep the system of lists useful, we must limit the size and topic of lists." (as WP:SALAT currently says) would mean to give a green light to various violation of NOR, V and NPOV policy.--Paul Siebert (talk) 16:48, 29 June 2012 (UTC)

I can't the see the (alleged) violation you're talking about, in particular not in that line yu've cited. Also if you're primary aim is to change the text of WP:SALAT you should discuss that there rather than here.--Kmhkmh (talk) 17:58, 29 June 2012 (UTC)

See my responce to Masem. It was not me who started a discussion about "Soviet occupations". I am discussing a policy in general. If we assume that the same criteria are applied to the articles and to the lists, then the opposite is equally correct: we can apply the same criteria to the articles as to the lists. If we assume that the lists can combine several items that are notable per se according to some arbitrarily chosen criteria, then the articles can do that too. In other words, we are allowed to write WP articles about non-existent topics by combining several existing notable topics, for which needed reliable sources exist. That is not how I saw our policy.--Paul Siebert (talk) 18:04, 29 June 2012 (UTC)

Why would you assume an identical treatment of articles and lists to begin with?--Kmhkmh (talk) 18:23, 29 June 2012 (UTC)

I do not assume that. It is what WP:SAL (its lede) says. In any event, my questions here are general: (i) are the same criteria applied to the articles and to the lists (as you can see, some participants of this discussion believe that they are); (ii) if the criteria are different, then what is the concrete difference; (iii) if different criteria are applied to the lists and to the articles, what is a borderline between SALs, summary style articles and full articles, and which parts of NOR policy are applied to each of them? I see some gray zone here, which may serve as a serious source of original research.--Paul Siebert (talk) 18:28, 29 June 2012 (UTC)

The concrete difference as far as I'm concerned is that a list does not have to cover a topic that has been covered in detail, it just has to be based on lists that people have thought worth making. For instance one would have trouble with an article on universities in the US as there isn't really a topic that people go into in any detail like Europe even though Europe is composed of countries. However people do make lists of things like American universities. Dmcq (talk) 18:37, 29 June 2012 (UTC)

American universities is a legitimate subtopic of the Education in the United States article. I see no problem with V, NOR and NPOV here.--Paul Siebert (talk) 18:40, 29 June 2012 (UTC)

List of London railway stations then. Dmcq (talk) 23:55, 29 June 2012 (UTC)

That is obviously a quite legitimate subject for the list (as a supplement to the London metro article). I am not sure I understand the idea you are trying to convey.--Paul Siebert (talk) 00:37, 30 June 2012 (UTC)

What about List of fictional canines? What article does that supplement? It is subsidiary to the list of fictional animals. In just the same way Soviet occupations is subsidiary to the List of military occupations. --Nug (talk) 01:02, 30 June 2012 (UTC)

That is an interesting question. But that is a part of a bigger question. See our discussion with Masem.--Paul Siebert (talk) 01:42, 30 June 2012 (UTC)

Re what I was trying to convey, you asked what was the concrete difference between an article topic and a list topic. I answered that lists do not have to be dealt with in depth as a subject, they just have to be things that people have thought worthwhile to list, i.e. they have list notability not topic notability. Dmcq (talk) 12:54, 30 June 2012 (UTC)

The problem I have with that is it means I could add my niece to the list of notable people from Denver. Or a friend's film company to a list of film companies. Dougweller (talk) 13:45, 30 June 2012 (UTC)

That is a question of what can be included in a list which is covered by what the list inclusion criteria and verifiability. Editors may or may not require that the entries are notable for instance after checking what the list is about. If all that is required is verifiability then it is possible your friend's film company may be included even if nobody is particularly interested in it. Dmcq (talk) 16:16, 30 June 2012 (UTC)

Invitation to discussion

There is a discussion on the Beitar Illit about the removal of part of a generic text which is sourced to general sources, not specific to Beitar Illit. The discussion mentions WP:OR as an argument. Please weigh in. Debresser (talk) 15:16, 21 August 2012 (UTC)

Is simple logic a SYNTH ?

1. If one reliable source say A:X->Y and an other reliable source say B:Y->Z, is the obvious conclusion C:X->Y->Z or C:X->Z a synth?

2. If one reliable source say A:Y happen when X is happen, and an other reliable source say B:Z happen when Y happen, is the obvious conclusion C:Z happen when X happen a synth?

I saw "SYNTH is not explanation - SYNTH is when two or more reliably-sourced statements are combined to produce a new thesis that isn't verifiable from the sources. If you're just explaining the same material in a different way, there's no new thesis".

79.182.215.205 (talk) 23:58, 18 June 2012 (UTC)

Very very few sources meet the specification requirements of the logic you're using. It is usually SYNTHetic and Original Research for an editor to determine that Y(a) = Y(b). Fifelfoo (talk) 05:56, 19 June 2012 (UTC)

In A:X->Y, in B:Y->Z you only have Y in common. Then you derive X->Z which does not mention Y. Aside from it probably being synthesis I'd wonder what the topic was, it would need to cover at least two of X, Y and Z. And why hasn't somebody else pointed out this obvious bit about the topic - so what would be the weight? And without weight why would you be putting it in? Dmcq (talk) 07:13, 19 June 2012 (UTC)

Fifelfoo, Y(first source)=Y(second source)

Dmcq, because there are many (things)->Y, where X is only one such (thing). Thus researchers of Y->Z are not interested in what (thing) caused Y. They are just interested in investigating if Y->Z.

79.182.215.205 (talk) 08:07, 19 June 2012 (UTC)

I guess from that the topic involves X and X implies Y but that Y implies Z does not explicitly involve X. Then yes it is synthesis. And if it was any sort of mentionable weight they would have figured that out without needing editors on Wikipedia to do that for them and we'd have a citation pointing out the connection between X and Z. Dmcq (talk) 08:12, 19 June 2012 (UTC)

There are very few reliable sources about Y->Z, maybe because it is a lengthy experiment. One primary source though did state the conclusion X->Z, however, I understand that primary sources are not accepted.

It is true that Y->Z does not explicitly involve X. Suppose there are X->Y, X2->Y, X3->Y. So Y could be the result of X,X2, or X3. But why does it matter?

79.182.215.205 (talk) 08:19, 19 June 2012 (UTC)

It matters because Wikipedia is an encyclopaedia and is in the business of summarizing what people have said. They haven't said X implies Z. That is original research. That's not Wikipedia's job. We should be approaching the inclusion of stuff in Wikipedia with a neutral point of view and putting in the stuff that has some weight. Reliable sources are what give weight. Editors own conclusions have no weight. If an editor wants to popularize their ideas then they should do it outside Wikipedia and then eventually perhaps they might gain enough weight to be included. Dmcq (talk) 09:02, 19 June 2012 (UTC)

IP User 79.182 Y(a:Communism) almost never equals Y(b:Communism). You'll find that in actual writing, almost no terms are defined consistently between sources; such that any attempt to use syllogistic logic in the real world will fail, as sources do not use identical definitions of terms. Fifelfoo (talk) 08:24, 19 June 2012 (UTC)

This editor (79.182) is engaged in a debate about X-ray computed tomography, and has been forum shopping on WT:MED and now here about how to circumvent rules about original research. JFW | T@lk 08:34, 19 June 2012 (UTC)

Medicine is an area where we have to be especially careful not to start putting in synthesis. Apart from the general policy medicine is just rife with examples where one needs to check things are really so rather than applying logic, never mind anything about possible real life consequences. Dmcq (talk) 09:11, 19 June 2012 (UTC)

Dmcq,JFW, your reference to other discussions is off topic here in this theoretical discussion, which may or may not be related. Please remove these off topic comments.

Fifelfoo, In this case lets assume it is possible it is evident that both Ys are the same.

Dmcq, I understand that Wikipedia summarize what other people say, and other people say X->Y and Y->Z (and even X->Z in a primary source). In the article there is a discussion about stuff=(?) such that Y->(?) and stating there (?)=Z based on sources indicating Y->Z seem to me appropriate, and on topic.

79.182.215.205 (talk) 10:21, 19 June 2012 (UTC)

What's the problem with the X->Z primary source? Dmcq (talk) 10:29, 19 June 2012 (UTC)

Dmcq, other editors indicated that they want only a review from the best of journals, I think.

79.182.215.205 (talk) 10:39, 19 June 2012 (UTC)

I'd guess because the other citations in the article are of about that WP:WEIGHT. That's more of judgement call but mixing things that are properly reviewed along with results in a primary source that people haven't commented on can certainly be problematic. User's own conclusions with SYNTH have much lower weight than that!, so the primary source would be a far better one to look at, maybe someone has talked about it? Dmcq (talk) 10:52, 19 June 2012 (UTC)

Why is this generating this much text? It's clearly synthesis. Taemyr (talk) 11:44, 19 June 2012 (UTC)

I didn't get a specific comment about that primary source yet. I asked in the article's talk page. I did write it in the format of 'A 2004 cohort study concluded ....', so one can see it is a study. In my view it is very important to include it in the article.

79.182.215.205 (talk) 11:54, 19 June 2012 (UTC)

Consider the case - Source A says that the area of a piece of land is 347 ha. Source B says that there are 100 ha in one square kilometre. It is not synthesis to state that the area of the land is 3.47 km², yet this is what various people are saying (or is it synthesis to make this assumption?). Moral of the story - use common sense. Martinvl (talk) 13:35, 19 June 2012 (UTC)

Martinvl asks us to "consider the case". No. This question has been asked within the last few weeks, perhaps by the same person. There is a good chance the original poster is forum shopping. In any case, it almost always turns out that the premise "one reliable source say A:X->Y and an other reliable source say B:Y->Z" is false because Y₁ is not identical to Y₂, or "->", which in math is spelled out as "implies", is not what either source actually says. In math, "X->Y" means that in every single instance throughout the past, present and future, from the big bang to the heat death of the universe, whenever X is true Y is true. When an everyday non-mathematical source makes a statement using the English word "implies" or similar wording, it is hardly ever meant in such absolute terms. Jc3s5h (talk) 14:27, 19 June 2012 (UTC)

Jc3s5h, this is the first time I ask this question here. Thus, if the question has been asked before, I guess that means that other editors have also received comments, that they make a SYNTH due to this simple logic, and they too thought that it doesn't make sense, that obvious logic can't be used.

Jc3s5h, I have already wrote above "In this case lets assume that it is evident that both Ys are the same.". I.e. I am asking only about the case in which Y₁ is identical to Y₂, and both sources do say and mean imply.

Martinvl, in my actual situation, the other editors are saying something like - the second source doesn't say 'land' anywhere, so you can't use it in an article about land.

79.182.215.205 (talk) 14:45, 19 June 2012 (UTC)

79... request "In this case lets assume that it is evident that both Ys are the same." No. Frankly, I don't believe you. This is because many editors have come here with such over-simplified descriptions of the situation, and I have never seen a case where the situation was actually as simple as described. Jc3s5h (talk) 14:53, 19 June 2012 (UTC)

Jc3s5h, What I ask about, is of such a simple case.

79.182.199.172 (talk) 21:09, 19 June 2012 (UTC)

Jc3s5h, not believing violate WP:AGF.79.182.199.172 (talk) 23:07, 19 June 2012 (UTC)

I believe you have had a consensus answer to the question you have posed. It is synthesis as far as Wikipedia is concerned and should not be done. If you want to discuss some other case or divulge more details that's fine but as far as the question is posed just arguing isn't going to change peoples minds about it as it seems a pretty strong consensus. Dmcq (talk) 00:33, 20 June 2012 (UTC)

Yes, very strong consensus, two editors vs one. --Nenpog (talk) 13:45, 4 July 2012 (UTC)

Hypothetical case - Source1:"Kabusha drink include 40% alcohol", Source2:"Drinking 40mL alcohol can cause a hangover", Contested text:"Drinking 100mL of Kabusha drink can cause a hangover". Synth? OR? --Nenpog (talk) 15:11, 4 July 2012 (UTC)

Definitely should not be included as it is synthesis. We should not be making up our own deductions and sticking them into articles. We are supposed to be summarizing what other people have thought worth writing about. Dmcq (talk) 15:27, 4 July 2012 (UTC)

Do the Wikipedia policy writers expect that researches will think that it is worth writing, again and again, about every alcoholic drink, that it too can cause a hangover? --Nenpog (talk) 16:29, 4 July 2012 (UTC)

Just don't try sticking it into an article if it isn't something people think is worth writing about the topic. How much simpler can anyone put it to you? If researchers don't think it is worth writing again and again it is not worth us writing again and again. Wikipedia summarizes so it should have less not more than what is written about in reliable sources. Dmcq (talk) 17:21, 4 July 2012 (UTC)

But researchers don't write again and again about each alcoholic drink, that it too can cause a hangover, because when they write that 40mL alcohol can cause a hangover, they know that everyone can make the above simple deduction, and realize that if the certain drink contain 40mL alcohol, then that drink can cause a hangover. You see researchers like to summarize too, and when they write 40mL alcohol can cause a hangover, they mean it for every drink that contain 40mL alcohol. The fact that they didn't write a research paper for each drink doesn't imply that the fact is not important enough to be mentioned, but rather that fact was already pointed out in a generic form. If the researchers would have thought that their audience is so stupid that it can't perform a simple deduction, I think that they would have kept a list, and update it with every alcoholic drink. Do Wikipedia policy writers expect researchers to write this important fact again and again explicitly for each drink, just for people who are too thick to understand a simple deduction? --Nenpog (talk) 18:16, 4 July 2012 (UTC)

It's no more appropriate to go around to every article on an alcoholic beverage and say that it might cause hangovers than it is to go to every article on something that causes dsDNA breaks and say that it might cause cancer. Even if it's not original synthesis, it's undue weight. The outcome is the same using either policy: If you don't have sources for the claim you're making, you are not aloud to make it. Someguy1221 (talk) 04:10, 5 July 2012 (UTC)

Did you read the page of undue weight? Please read/reread it. It is a bit confusing, because in the Wikipedia definition the term undue weight refers only to the weight that is given in the Wikipedia article, e.g. by depth of detail, quantity of text, prominence of placement, and juxtaposition of statements. The Wikipedia definition of weight doesn't refer to the weight given in the sources. Instead in the sources what is important is the degree of acceptance of that view e.g. a majority view, a minority view, a tiny minority view or fringe. The due weight policy calls for a correspondence between the weight of the text in the Wikipedia article, and the degree of acceptance of that text in reliable sources. Thus, according to the Wikipedia definition, if the view, that dsDNA break might cause cancer is the majority view (and it is), then that view deserves a high weight in Wikipedia articles. Likewise, in this hypothetical example, if it is the majority view that 40mL alcohol can cause a hangover, then that view deserve high weight in Wikipedia articles. Thus inserting that view into the Kabusha drink article, would not be of undue weight, as it doesn't give prominence to a minority view, as it isn't a minority view, but the mainstream view. Please see also this discussion at the NOR talk page, in which it was suggested, that determining the degree of acceptance of a view is performed by confronting supporting vs. objecting reliable sources and evaluating from that which of the views is more accepted. At this case that would be reliable sources supporting the view that 40mL alcohol cause a hangover, vs. reliable sources objecting that view. At that case there would be only supporting reliable sources, hence the view is the majority view, hence WP:WEIGHT can not be used in order to exclude that view from any Wikipedia article. --Nenpog (talk) 06:15, 5 July 2012 (UTC)

This started as a content dispute on Talk:X-ray computed tomography where Nenpog faced a lack of consensus (every other editor opposed the changes he wished to make.)
He was then blocked for edit-warring and tendentious editing.[12][13][14]
He then started Wikipedia:Forum shopping, taking his dispute to:
Wikipedia talk:WikiProject Medicine,
Talk:Ionizing radiation,
Wikipedia talk:No original research,
Wikipedia:Dispute resolution noticeboard,
User talk:Elen of the Roads,
Wikipedia talk:Neutral point of view
Wikipedia:Conflict of interest/Noticeboard,
And now he has come back to Wikipedia talk:No original research. --Guy Macon (talk) 05:29, 5 July 2012 (UTC)

And Guy Macon has been with this off topic message in all (or most, didn't bother counting) of these forums, so that everyone would be able to avoid a theoretical discussion, and discuss the situation only with the prejudice of a specific content dispute in mind. --Nenpog (talk) 06:15, 5 July 2012 (UTC)

Please join us at WQA and share your view there, if the above disclaimer from Guy Macon is related to the discussion or is off topic. --Nenpog (talk) 13:57, 6 July 2012 (UTC)

And the point that was made is that undue weight is not a valid argument here, and that the interpretation of the WP:SYNTH that was written above is like expecting that after researchers have determined that 40mL alcohol can cause a hangover, the researchers will continue to publish papers about each drink that include alcohol, and state if it too cause a hangover. That just don't make any sense, and thus that WP:SYNTH interpertation doesn't make sense either. --Nenpog (talk) 15:40, 6 July 2012 (UTC)

Fine that's your opinion. However it is not in line with the consensus on Wikipedia as acceptable grounds for inserting a reference which does not mention the topic into an article or for putting a conclusion based on that reasoning into an article. If you wanted to change that you'd need to raise a WP:RfC and to be frank I believe you're more likely to win the lottery than get your idea accepted. For far less money than the lottery you could set up your own website or publish your own book pushing your idea, and who knows you might eventually become a recognized authority and so make them acceptable references. Dmcq (talk) 11:44, 7 July 2012 (UTC)

This is now at Wikipedia:Administrators' noticeboard/Incidents#User:Nenpog. --Guy Macon (talk) 01:40, 10 July 2012 (UTC)

This is an important point. I was recently involved in an article which I thought justified it's inclusion in a section though the law of Law of Syllogism but it was claimed by another editor this wasn't true since it required the use of synthesis. I think this is the same argument as used in point 2 at the top of this section. If Wikipedia is excluding information which can't even be classed correctly through logic, then perhaps it needs to rethink this rule? This may also be inconsistent with the rule that Wikipedia allows simple calculations providing they are an obvious, correct, and meaningful reflection of the sources.--Andromedean (talk) 12:04, 1 September 2012 (UTC)

If your simple logic resulted in an idea that had never been published in a reliable source, then it definitely violates this policy. WhatamIdoing (talk) 21:31, 1 September 2012 (UTC)

It isn't my logic it is an established truth. The following is an example:

1.If Larry is sick, then he will be absent from school.
2.If Larry is absent, then he will miss his classwork.
3.If Larry is sick, then he will miss his classwork

Say if there was a reliable source for 1, and a reliable source for 2, then Wikipedia rules means that you can't state 3) through the law of Law of Syllogism, unless you can find a common source stating 3 in its entirety, even if you try to justify it's inclusion through this rule! If so, this is actually quite profound, because it censors defined truths which a reputable encyclopedia needs to cover. I suspect when the rule of synth was dreamt up this wasn't intended or realised.--Andromedean (talk) 06:12, 2 September 2012 (UTC)

In fact the rules of synth are so ill defined they could be used to censor almost anything other than a defined quote. This is because summarising a statement requires language intepretation as well, therefore this is combining two bits of information to create another, a synth.--Andromedean (talk) 06:12, 2 September 2012 (UTC)