Wikipedia talk:No original research/Archive 56

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Archive 50 Archive 54 Archive 55 Archive 56 Archive 57 Archive 58 Archive 60


WP:NOR should stay on topic

The following is mostly from WP:TOPIC:

The most readable articles [or policies] contain no irrelevant (nor only loosely relevant) information. While writing an article [or policy], you might find yourself digressing into a side subject. If you find yourself wandering off-topic, consider placing the additional information into a different article, where it will fit more closely with the topic. If you provide a link to the other article, readers who are interested in the side topic have the option of digging into it, but readers who are not interested will not be distracted by it. Due to the way in which Wikipedia has grown, many articles contain such redundant texts. Please be bold in deleting them. (talk) 13:57, 7 March 2011 (UTC)

Could you be a bit more specific about which portions of this policy you feel are off topic? TenOfAllTrades(talk) 14:14, 7 March 2011 (UTC)
Not wishing to speak for the IP, but we did discuss this not long ago, and people seemed to agree that (if this page is to remain separate from V) then the detailed section about sources is off-topic, and we should confine ourselves here to saying what is and what isn't original research. That's largely the same point I'm making in the thread above this.--Kotniski (talk) 14:53, 7 March 2011 (UTC)
The section on sources is not off topic... it just takes a long time to get around to the topic ("don't use primary/secondary/tertiary sources to support OR"). Blueboar (talk) 15:06, 7 March 2011 (UTC)
All right, that much is on topic... but this shows that all the detailed definitions of the various types of sources are not on topic (since for the purposes of OR, the distinctions turn out to be pretty much irrelevant).--Kotniski (talk) 15:42, 7 March 2011 (UTC)
One example is the section Primary, secondary and tertiary sources. An already existing sentence in the previous section sufficiently and appropriately states the relevant NOR concept, “Drawing conclusions not evident in the reference is original research regardless of the type of source.” The questionable section should be moved to another page and there should be a link to it. (talk) 18:02, 7 March 2011 (UTC)

At this point, it will probably be instructive to look at the history of the PSTS section, and see how it developed to be the way it did. A few key dates:

  • Starting in Feb 2004 the entirety of the discussion of source types in the NOR policy consisted of one line: "Wikipedia is not a primary source". (added with this edit) This remained stable in the policy for just over a year. This idea (that Wikipedia should not itself be a primary source) was directly tied to the concept of Original Research. However, it was felt that editors might not understand what was meant by the term "primary source", so...
  • In March of 2005 - a definition was added (and, to help clarify further, a definition of what a secondary source is was also added). Unfortunately, at the same time the key phrase "Wikipedia is not a primary source" was removed (with [this edit). Personally I feel this was a mistake, as that phrase went to the heart of this policy. However, the policy at least retained a loose tie to the concept of OR (by stating that OR produces a primary source).
  • July 2005 - the discussion of primary and secondary sources was, for the first time, given its own section - around this time the concept that Wikipedia should not itself become a primary source had been restated as: "Original research that creates primary sources is not allowed". This language remained stable in the policy for more than a year.
  • Then, in October 2006 - a restructuring of section expand the definitions (adding tertiary sources)... but, sadly, the edit also removes the phrase "Original research that creates primary sources is not allowed".

As I see it, we ended up with a detailed definition of terms, but along the way we cut the initial reason why we included that definition in the first place. Blueboar (talk) 18:46, 7 March 2011 (UTC)

For consideration

The following is not a proposal ... just some thoughts for consideration and discussion. It stems from my belief that the PSTS section overly focuses on defining terms and does not do a good job of explaining how the terms and definitions relate to the concept of "No original research". Not sure where a statement like this would go, or how it would integrate with the rest of the section... and the wording is very preliminary... but I do think something like it would improve the policy.

  • Wikipedia is not a primary source. Primary sources present information for the first time. Do not add unsourced material from your personal experience, because that would turn Wikipedia into the primary source of that material. All information on Wikipedia must be published elsewhere before it can be mentioned or discussed on Wikipedia.
  • Wikipedia is not a secondary source. Secondary sources compile, synthesize, analyze, interpret, and evaluate material found in primary source, and draw conclusions from them. Do not analyze, synthesize, interpret, or evaluate material found in a primary source yourself, or state your own conclusions based on primary sources; instead, refer to reliable secondary sources that do so.
  • Wikipedia is a tertiary source.. Tertiary sources summarize the information, analysis, interpretations, evaluations and conclusions found in reliable, published secondary sources (and, to a lesser extent, in other tertiary sources). This is what we should do when building and editing our articles.

Your thoughts? Blueboar (talk) 20:21, 7 March 2011 (UTC)

WP is sometimes a secondary source, in that we do allow the use of primary sources. SlimVirgin TALK|CONTRIBS 06:30, 9 March 2011 (UTC)
But we don't compile primary source material, though we may cite such material occasionally, and we don't (or at least we shouldn't) synthesize, analyze, interpret, and/or evaluate it. Wtmitchell (talk) (earlier Boracay Bill) 06:57, 9 March 2011 (UTC)
We cite it more than occasionally, and there's no requirement that we compile it before becoming a secondary source (not even sure what that would mean here). When we include primary source material, we become a secondary source for it. So it's just not correct that we're invariably a tertiary source, and it varies a lot between articles. SlimVirgin TALK|CONTRIBS 07:01, 9 March 2011 (UTC)
I'm so used to coming across once-controversial debates that are cut and dried by the time I get to them, I almost didn't realize I could actually reply here! Yes, I think these categories (or some set like them) would be helpful to draw the Wikipedia Is/Isn't distinction across; given my own understanding of the terms, to ask for synthesis-free restatements of reliable sources that are not direct quotes, and yet do not plagiarize, is paradoxical. This more nuanced explanation, that Wikipedia publishes second-order analyses (etc.) of original research, seems far clearer - and attainable. Mercenary42 (talk) 04:11, 8 March 2011 (UTC)
Yes, looks good. I.e. we use this page to talk about primary/secondary/tertiary sources in terms of what kind of source Wikipedia is. The information that addresses the issue in terms of what kinds of sources our sources are (and what consequences that has) should be moved away to the relevant places in V and IRS. This is starting to look almost logical - well done Blueboar!--Kotniski (talk) 10:36, 8 March 2011 (UTC)
I am not at all sure that I agree with the idea of moving the current PSTS material into WP:IRS or WP:V (I have trouble seeing how the material relates to the twin concepts of Verifiability and Reliability). I am not going to knee-jerk object to the idea, but I would need a lot of convincing. Perhaps we should save that for a later discussion. I realize that I am suggesting a new and different way of thinking about the topic of primary/secondary/tertiary (or perhaps a return to an older way of thinking about it) ... but that does not mean I want to throw the baby out with the bath water either. While I am sure that my new approach will impact the current language, I do like a lot of the current language, and would like to see it retained if we can. So... if people like where I am going with this, let's think of it (at least initially) in terms of being a clarification/addition, rather than as an outright replacement. Blueboar (talk) 13:55, 8 March 2011 (UTC)
Very clear and illucidating language. PPdd (talk) 14:37, 8 March 2011 (UTC)
I'm not saying the current language should be replaced; just moved to the appropriate page. Whether a source is primary, secondary or so on is an issue so closely related to the reliability of that source (for Wikipedia's purposes) that it just cries out to be moved to WP:IRS. I certainly can't see how it belongs on this page (if we're assuming that there really is a topic called "original research" that's distinct from verifiability).--Kotniski (talk) 16:21, 8 March 2011 (UTC)

I agree with some of the motivation, but I have a couple concerns. The first is that I'm not sure "Wikipedia is not a secondary source" is completely accurate. For example, consider the list of Pokemon characters. Is that article really a tertiary source? I don't think it is; the main sources are all primary, because there aren't secondary sources about Pokemon characters as there would be, say, for the French revolution. The deeper issue is that "Wikipedia articles must be based on secondary sources" isn't completely accurate: many of our articles are based on primary sources. But taking that idea one step farther seems to increase the difference between this policy and reality. — Carl (CBM · talk) 15:25, 8 March 2011 (UTC)

Yeah... Despite the fact that multiple policies and guidelines (including the current version of NOR) repeatedly say that articles should be based on secondary sources, there are some topics (especially on Pop-culture) where we routinely Ignore all these policy and guideline statements and rely on primary sources. I don't think that changes anything. I still think Wikipedia is (or, at least, is supposed to be) a tertiary source, even if we occasionally forget (or choose to ignore) that fact. Blueboar (talk) 16:05, 8 March 2011 (UTC)
(As an aside... I have long felt that most of our "list of" articles are more appropriate for an almanac than an encyclopedia. If I were running things, I would create a WikiAlmanac sister project as a place for such material) Blueboar (talk) 16:37, 8 March 2011 (UTC)
I tend to approach it from the opposite side: I look at the collection of articles we have, and they tell me what Wikipedia is supposed to be. When I came to WP, I was attracted by the goal of providing access to a large portion of human knowledge – which it has made some progress towards doing. It's true that some policies have been written that conflict somewhat with that basic goal, but I tend to view that situation as saying that the policies aren't completely right, not as saying that we have ignored the policies when writing articles.
It's hard to write policies that accurately describe our best practices, which include articles based on many different types of sources. For the French revolution, we use secondary sources because that's the right thing to do; for the list of Pokemon characters we do what's right for that article. In many cases the articles predate the policy wording, which makes me think the policy is ignoring the articles rather than the other way around. If the policy said, "use the best, most appropriate sources for each article", that would represent our best practices better than "every article should be based on secondary sources". — Carl (CBM · talk) 17:10, 8 March 2011 (UTC)
Source typing and the prohibition on synthesis both came into the policy (t seems to me)as tools (weapons) against some editing on some topics that was disturbing to some editors. Wikipedia is an encyclopedia and ought to, as a first approximation, include everything a good encyclopedia can include. Because of the specific nature of Wikipedia there is the possibility that some editors, either through zeal for a particular point of view or through ignorance, will introduce material that violates the basic principle: Wikipedia is not a primary source and is not to be used as one. Original research, no matter it's merit or appeal (to whatever audience it may appeal) is not to be included ion Wikipedia. A lot of time and effort have been expended on source typing and synthesis, on the grounds that use of some sources might introduce original research or the use of synthesis might introduce original research. Simple declarative sentences might be used to introduce original research but (for which we can be thankful) no prohibition on simple declarative sentences has yet been proposed or added to the policy. Should not the emphasis be on quality (of Wikipedia as an encyclopedia?) If some editor or faction of editors is aware of how some source (kind of source) or some technique may be improperly used to introduce OR surely those editors will be doing a service by correcting the error, probably by removing the offending material. To perpetually wrangle over the policy seems not to serve the goal of quality.
Does anyone else notice that the place in Wikipedia where OR most often occurs is in this very page, the discussion of WP:NOR?
I like Blueboar's first point (Wikipedia is not a primary source.) That's the basic policy: no original research. Attempts to go further will inevitably get mired in discussion. I suggest that the reason for this is that the goal (having some specific prohibitions in the policy) isn't a good goal. Clarifications have to be made, modifying clauses have to be added - and soon the community is once again engaged in an OR frenzy in the discussion of WP:NOR.
I find it plausible that the EB doesn't have the same prohibition on OR that Wikipedia does. I think the EB is a fine encyclopedia. It has a different nature from Wikipedia and that different nature can perhaps allow OR, at least in some limited case. (After all, many major articles are written by true experts in the field of the article, persons qualified to do OR in that field.)
The underlying polices of the EB and Wikipedia can be largely the same - but in the case of Wikipedia there is a prohibition against OR. Wouldn't it seem reasonable that if the policies of both were codified that the only major difference between them would be the sentence in the Wikipedia policy that says OR is forbidden? OK, editors need guidance. That guidance has to be clear, it has to be accurate (meaning, for instance, that the definition of terms such as "primary source" cannot be floating, ought not to be Wikipedia-specific misshapen variations of what the term means outside Wikipedia), has to be motivated by a desire to guide (not by a desire to create a quickly-citable reason for summarily edits.)
It might be the best thing to do would be to back up by several years or to simply start over. Blueboar has pretty much attempted something of that sort. Why not rethink? Minasbeede (talk) 05:19, 9 March 2011 (UTC)
EB differs from WP regarding all three content policies. They allow OR, especially in named expert articles. They don't require sourcing. And they don't require neutrality. So there's really no point of comparison. SlimVirgin TALK|CONTRIBS 06:29, 9 March 2011 (UTC)
And the online version of EB solicits suggested revisions from readers. Collect (talk) 12:52, 9 March 2011 (UTC)
well, OK, then the Wikipedia policy and the EB policy differ in three ways, not one. The comment isn't aimed at attacking or weakening the Wikipedia policies, it's an attempt to direct policy-writing focus to what it is that makes for a good article in a good encyclopedia. I still maintain that there have to be broad areas of agreement between the two simply because of the huge overlap between the purposes of the two. My plea is for the policies, particularly when they are intended to provide guidance, to serve to favor the creation of good articles. EB and wikipedia are both universal encyclopedias. I say they are very similar in that regard. The implied comparison is focused on the factors that make for a good article and I cannot see how that is invalid. [The only neutrality point I can think of with respect to the EB is their having two articles on the Great War (later called the 1st World War) written by experts from each of the major sides in the conflict because the editors recognized that single-article neutrality that close in time to the end of the war was impossible. Now, I think, they no longer do that.]Minasbeede (talk) 01:50, 11 March 2011 (UTC)

Blueboar, this type of 30,000' view discussion is what we need more of, so nice work! Now on to the discussion. The premise of your statement is that it is useful to define WP's place in it's own PST source classification system. The question is: why? Sincerely, North8000 (talk) 13:04, 9 March 2011 (UTC)

Well, that wasn't the intent of my statement... but to answer the question: Because I think doing so will better tie our source typing system to the concept of "No original research". I think something along the lines of what I am suggesting helps to explain to our editors why we discuss source types in this policy in the first place. The current PSTS section does an excellent job of telling editors "don't do this"... but it does not clearly explain why they should not "do this". Blueboar (talk) 14:23, 9 March 2011 (UTC)
Well... I thought we'd established before that it doesn't do an excellent job at all, because it keeps making statements about the particular kinds of sources that apply equally well to all the other kinds of sources. I still don't see why you think all the information in that section belongs on this page (accepting for the moment that there's a reason for this page to exist, i.e. that original research is something distinct from just breaches of WP:V). --Kotniski (talk) 14:40, 9 March 2011 (UTC)
To be honest, I don't think we have really "established" much of anything (except, perhaps, that each of us have different, sometime overlapping, views about what is good and bad about the PSTS section). Do I think the current language has flaws?... absolutely. Do I never-the-less like and agree with a lot of the current language?... yup. Is this contradictory? ... perhaps. Do I think that approaching the entire issue from a completely different angle could improve the policy? Yes, that is why I started this thread. Do I have a firm grasp of exactly what that new angle should be, or how it will impact the policy? Nope... I am groping towards it, and asking the rest of you to grope along with me. Blueboar (talk) 18:44, 9 March 2011 (UTC)
PSTS doesn't add anything to policy that isn't already covered in the rest of WP:NOR. It strays off topic when it defines various types of sources. That information is better placed in Wikipedia:Identifying reliable sources. A previous comment about staying on topic may be useful for guidance. (talk) 23:39, 9 March 2011 (UTC)
I understand (and to some extent agree with) the concern that the definitions in PSTS are "off topic" in this policy ... but I am concerned that moving them to IRS would simply shift the problem to a new page. Can you explain why you think the material is "on topic" at that page? What do the definitions have to do with the concept of identifying reliable sources? Blueboar (talk) 00:16, 10 March 2011 (UTC)
Well, reliable sources (except for the purposes of just saying what's in the source) need to be secondary or tertiary. This seems to be effectively one of the criteria for identifying a reliable source. The main reason we find ourselves debating whether a particular source is primary/secondary and so on is that primary sources are not generally considered reliable for facts other than the facts of their own content. Also for the purpose of establishing notability, we expect mentions in reliable, secondary sources. The two concepts (though not identical, of course) are so closely tied together that it seems obvious to me that they ought to be dealt with on the same page.--Kotniski (talk) 11:11, 10 March 2011 (UTC)
For secondary sources, can we say things that are not in the source? I don't understand the first sentence. Reliability seems to be unrelated to the nature of the source. — Carl (CBM · talk) 13:24, 10 March 2011 (UTC)
We can't say things that are not in the source, but we can say things other than that something is in the source. This keeps coming up - I'm sure there must be some jargon for it, any ideas? If a source X contains statement P, then if X is a totally reliable source then we can say P, but if its reliability is limited to its own content then we can no longer say P, but we can still say "X says P". Pretty much any (genuine) sources are reliable as to their own content, but to be reliable in the former, stronger sense, we expect more - that the source be published, reputable, etc., and generally speaking that it be a secondary (or at least not primary) source. That might be an oversimplification, but it still remains clear to me that the primary/secondary/.. distinction relates very closely to reliability of sources, and not at all to originality of research (if you use a primary source as you would a secondary one, it might constitute a breach of our standards of neutrality or verifiability, but it can hardly be called original - we're assuming here that this page has the right to exist at all, i.e. that it relates specifically to something called "original research" that doesn't just mean any breach of WP:V).--Kotniski (talk) 14:13, 10 March 2011 (UTC)
The "X says P" way of writing is (1) not required, in general, for every reference of a primary source and (2) is sometimes required even for secondary sources. — Carl (CBM · talk) 16:03, 10 March 2011 (UTC)
I'm sorry that I don't have time to participate to the extent that I'd like to in this discussion, but I do just want to say that Carl has my proxy. I've read through the thread, and agree with everything he's said here, so pretty much anytime he writes something, please mentally add a "What he said", afterwards, followed by my sig. :-)  – OhioStandard (talk) 16:38, 10 March 2011 (UTC)
(This was discussed many years ago, too.) For scientific material the criteria are different. You don't go to a secondary source for an important breakthrough, you go to the primary source. So this creates the Wikipedia-specific idiosyncratic definition of "primary" and of "secondary" (at least that's how this was handled way back then.) If it is fresh research you may wait (or need to wait) until the new material has been verified before incorporating it in an encyclopedia but when you incorporate it you cite the work of the discovery/developer, not later commentators.
The point here is that the discussion about PSTS centers on non-science articles but the policy is a general policy and is wrong for science articles. Sure, the proponents of PSTS don't concern themselves with scientific articles and won't bother editors who (quite properly) cite primary sources but the policy isn't appropriate for those articles and there is always the danger that some zealot will start ripping up articles for violating a principle that is wrong for an important part of Wikipedia. Minasbeede (talk) 01:34, 11 March 2011 (UTC)

There are a lot of good and interesting points being made here, but on the question of which page these issues ought to be discussed on, I'm still not seeing anyone making a link between the issue of distinguishing P/S/T sources and the subject we propose to call "original research". Does anyone want to make such a link? If not, I still don't see any reason to keep the material on this page rather than V or IRS.--Kotniski (talk) 12:44, 11 March 2011 (UTC)

Re: "I'm still not seeing anyone making a link between the issue of distinguishing P/S/T sources and the subject we propose to call "original research". Well, that is what I was trying to do in opening this thread... by shifting the focus of our discussion away from "what type of source can be used in Wikipedia" and towards "what type of source is Wikipedia itself". To me, the fundamental reason why we don't allow OR is that Wikipedia is not a primary source, and the inclusion of OR would make it so. Blueboar (talk) 15:18, 11 March 2011 (UTC)
It would make it a primary source or a secondary source, depending on what type of OR it was (presentation of completely new material vs. analysis of existing primary sources). But I don't think this is the fundamental reason we don't allow OR (after all, as long as we're just reporting on the content of primary sources, we're quite happy to be a secondary source) - and even if it were, that wouldn't justify a lot of detailed analysis about what kind of sources count as primary, secondary and tertiary.--Kotniski (talk) 16:15, 11 March 2011 (UTC)
... and for moving the definitions material, consider the articles Primary source, Secondary source, and Tertiary source if the material isn't already there. Also note that the material for removal would include most of the material in the Notes section. In any case the material doesn't belong in WP:NOR. (talk) 15:05, 11 March 2011 (UTC)
Additionally, the terms are unclear (a single source can be classified as more than one type) and unnecessary for WP:NOR. (talk) 04:52, 19 March 2011 (UTC)

Multi article SYNTH

Does it count as original research via WP:SYNTH if an article has information that does not apply for what links to the article?


  • Bob adds a sourced sentence in the article "Swastika": Bill said that everyone wearing a swastika deserves to be punished.
  • Bob then adds links to the "Swastika" article in another article that states The symbol used to represent Jainism is a Swastika with the dots present.
  • These articles, together imply that Bill has problems with Jainism, when in reality Bill has nothing against them. (talk) 00:20, 16 March 2011 (UTC)

Um, I don't see the implication in the links. So I would have to say, no, it isn't WP:SYNTH. Blueboar (talk) 22:23, 16 March 2011 (UTC)
Adding Easter-egg links can indeed amount to SYNTH or just straightforward OR. SlimVirgin TALK|CONTRIBS 22:29, 16 March 2011 (UTC)
Yes, but this does not seem (to me) to be an Easter egg situation. I would assume that "Bill" is objecting to the swastika due to its association with Nazis, not its association with Jainism. If so, the situation can be easily fixed by editing the "Bill said" statement so that this "anti-nazi" context is clear. Blueboar (talk) 22:42, 16 March 2011 (UTC)
My understanding of the example was that Bob said everyone wearing a swastika (implication: Nazi) deserves to be punished. But someone had snuck in an Easter-egg link that linked swastika to Jainism. Ergo, everyone involved with Jainism deserves to be punished = a form of SYNTH, or just plain OR. SlimVirgin TALK|CONTRIBS 23:50, 16 March 2011 (UTC)

I didn't really meant to imply original research, just whether or not something should be changed to avoid misleading, and more specifically, what could possibly be change to avoid the misleading. (talk) 00:20, 17 March 2011 (UTC)

Easter egg links aren't a good idea if the link is in any way contentious or misleading, so that's one thing you can ask be stopped. It definitely can be a form of OR. SlimVirgin TALK|CONTRIBS 01:14, 17 March 2011 (UTC)
Presumably the article swastika makes clear that not all swastikas are Nazi ones. If we are sure (for some reason) that what Bill had in mind was Nazi swastikas, then that quote should presumably appear in the article in the context of specifically Nazi swastikas; so I don't see how anyone could really be misled. (Of course, Bill might indeed think that all people wearing swastikas, Nazi or otherwise, should be punished - in which case he presumably really does have something against the Jains' use of the symbol.)--Kotniski (talk) 07:16, 17 March 2011 (UTC)
(I do agree that "easter egg links" are often a problem, but from my understanding of the situation as described here, no easter eggs are involved this time.)--Kotniski (talk) 07:19, 17 March 2011 (UTC)
I agree, see Swastika--Kmhkmh (talk) 22:56, 17 March 2011 (UTC)

A few points here:

  • The example seems rather artificial.
  • I can't see how the proclaimed conclusions necessarily follows.
  • You cannot treat separate articles as blocks in a logic puzzle, since the formal logic usually doesn't capture the complexity of articles. What might appear as mandatory logic conclusion at first glance to some, is usually none when you take a closer look.

--Kmhkmh (talk) 10:22, 17 March 2011 (UTC)

What was probably had in mind with the "Easter egg link" thing would be a sentence like, say, "Jains us the swastika as a symbol", where the word "swastika" is linked to the article Nazism. That would certainly be inappropriate. Or even the same sentence with "swastika" linked to swastika, but only if the article swastika were exclusively about the Nazi symbol (which in fact it isn't). I wouldn't put this practice under the heading of original synthesis, but they're something to be avoided whatever they're called. Sometimes they're just mistakes, of course, as editors don't always check the subject matter of the articles they link to.--Kotniski (talk) 10:43, 17 March 2011 (UTC)

Agree. Blueboar (talk) 13:05, 17 March 2011 (UTC)
Okay, I see now what the IP's saying; sorry for being dense. I don't see what we could do about the kind of example's he offering. I had a similar thing (assuming I've understood him properly now) with an article about an alleged miscarriage of justice. The accused's lawyer is appealing the conviction, and he has a WP BLP, which I'd normally link to. But the BLP is problematic, accusing the lawyer of shady practices—reliably sourced, but written in an unnecessarily breathless tone. I decided not to link to it, but someone else did, and I could hardly object. The problem is that the link to the BLP arguably undermines neutrality in the miscarriage of justice article, if you look at the two articles as one unit. It's not SYNTH or OR, but it's awkward. SlimVirgin TALK|CONTRIBS 22:27, 17 March 2011 (UTC)
However in that case I'd see it simply a case of (malicious) false linking and a not a case of OR. You cannot combine different articles to create OR.--Kmhkmh (talk) 22:54, 17 March 2011 (UTC)
This case wasn't even false linking; the link was accurate. It was just that the lawyer's BLP was very negative, and it therefore undermined the article about the lawyer's case. I accept it's not really OR. It just had the flavor of it, slightly, in that none of the sources writing about the case had mentioned the lawyer's negative reputation, so it was an interesting form of (sort of) SYN-by-linking. SlimVirgin TALK|CONTRIBS 23:39, 17 March 2011 (UTC)
Sounds as if the problem here was not so much the link, but the writing of the target article. But I think we certainly can create OR by linking; for example, if a word has two senses on which we have two different articles, and we give a quote from someone which includes that word, and link that word to one of the two articles - it needs to be very plainly clear that it was that meaning and not the other that was intended, otherwise I would say it was a form of OR (i.e. making our own interpretation of the source).--Kotniski (talk) 08:00, 18 March 2011 (UTC)

Table of Contents structure, Mind maps

There is discussion of use of mind maps at WikiProject Medicine[1]. The issue is more broad than use in medical articles. They may be considered OR. Defects cannot be edited out, and they cannot be updated. On the other hand, they are roughly equivalent to organizational structure created in a table of contents, which is usually not considered OR, and they may supplement a table of contents and improve WP for readers.

  • (1) Is organizational structure in a table of contents OR when the organizational structure itself is not verifiable with RS?
  • (2) If organizational structure in a table of contents is not verfiable with RS, why is a mind map OR but not the table of contents?
  • (3) If a mind map is not OR, does it sometimes improve WP by helping some readers by supplementing a table of contents?
  • (4) Is there a way to make mind maps editable and not static?
  • (5) Is there a way to make mind maps link to sections and subsections as does a table of contents, or more strongly, to link to specific content in the article body?
  • (6) Should an article have one or more mind maps?
  • (7) Should an article have a "mind maps" section at the bottom, similar to the "See also" section, referring to a collection of various mind maps associated with article X, which may be located in a separate "X (mind maps)" article, similar to a "List of X's" associated with article X?

This matter has been discussed before: here, here and here. PPdd (talk) 14:21, 17 March 2011 (UTC)

It's really a style issue and it's not really a question of "function" it's just a question of "form" and while there are potential WP:UNDUE questions as to what's included or not, I doubt that the reliability is really an issue. It's really more a question of utility, which is an article-by-article decision. Honestly, if we have a table of contents why do we need a mind map? It seems like it's just an alternate way of presenting very similar information, with the additional problem of barriers to editing. It'd be an interesting way of formatting an article, incorporating the see alsos and the external links as well, but in general I don't see the value. SDY (talk) 15:15, 17 March 2011 (UTC)

tertiary sources

Can we expand this sentence...

"Reliably published tertiary sources can be helpful in providing broad summaries of topics that involve many primary and secondary sources."

with this clause...

...especially when those sources contradict each other.

I do a lot of editing on pages where editors hold deeply felt minority viewpoints (Jesus, evolutionary psychology). They have an easy time finding secondary sources that support their viewpoint but have a substantially harder time finding tertiary sources. Naturally, they denigrate the whole concept of tertiary sources, saying (indeed rightly) that WP should be based mostly on secondary sources. Tertiary sources, however, have a special role in providing an overview of a topic where the primary and secondary sources are in conflict. It's exactly this broader view that defenders of minority viewpoints want to hold at bay. In order to find out that tertiary sources are especially good at identifying a majority viewpoint or describing a scholarly disagreement, you have to go to WP:NPOV. If we could add this clause here, it would help editors see what tertiary sources are especially good for, and it would help settle a lot of disputes over contentious topics.

While we're at it, my English professor father always taught me to write actively, so I'd change the sentence to "Use reliably published tertiary sources to provide broad summaries of topics that involve many primary and secondary sources, especially when those sources contradict each other." Leadwind (talk) 15:15, 8 March 2011 (UTC)

This page is clear. Tertiary sources are not as solid a base as secondary sources are for Wikipedia articles. Using a tertiary source to invalidate a reliable secondary source is problematic at best - as it is a truism that folks who write them frequently do not use the WP NPOV standards. Collect (talk) 12:51, 9 March 2011 (UTC)
I agree that we shouldn't in general use a tertiary source to contradict a reliable secondary source. The clause I propose adding would emphasize that it's when the secondary sources themselves disagree that a tertiary source can be especially useful. The clause would make it less likely that an editor would use a tertiary source to contradict secondary sources that agree with each other.
Presumably you agree with the policy that tertiary sources are useful for providing broad summaries. Presumably you agree with the policy that commonly accepted reference texts are useful for determining the majority viewpoint. And you agree that secondary and tertiary texts that describe a scholarly disagreement from a disinterested viewpoint are useful for determining proper balance among contending viewpoints. That's all WP policy already. But the policy that explains when tertiary sources are especially useful is on WP:NPOV, not here. Why not make a reference to that policy here on this page? If you don't want to add a simple clause ("especially when those sources contradict each other"), then just a hyperlink to the majority viewpoint and balance sections of WP:NPOV? Leadwind (talk) 15:39, 10 March 2011 (UTC)
I support the proposed wording, it is helpful and appropriate. In economics for example, there are sometimes several contradictory viewpoints that are supported by reliable secondary sources. The most reliable way to distinguish the majority mainstream viewpoint is to refer to tertiary sources – advanced level university textbooks, academic reference books, other encyclopedias. Tertiary sources are useful for determining academic consensus on an issue. LK (talk) 04:55, 14 March 2011 (UTC)
Tertiary sources are not less reliable than secondary sources per se. Whether they are or not depends on the particular case. Also keep in mind that many WP articles are actually using tertiary sources, since standard textbooks and (standard) special subject encyclopedias are tertiary sources for the most part.--Kmhkmh (talk) 16:01, 14 March 2011 (UTC)

[2] (talk) 13:57, 15 March 2011 (UTC)

Thanks, LK and Kmh. I find that it takes a little extra work to tell people something they don't want to hear, and editors who support minority views don't like tertiary sources because tertiary sources are more likely to support majority viewpoints. Leadwind (talk) 14:57, 24 March 2011 (UTC)

"Common sense is not original research"

i personally think that this should be added to the policy. For example if someone, in good faith, adds to the meat substitute page that "Meat substitutes are commonly made of tofu" and someone reverts it because "it is original research". that is completely ridiculous, again, this is just my personal opinion, this doesn't HAVE to be done. (talk) 15:02, 22 March 2011 (UTC)

The specific statement isn't really an original research issue, it's just incorrect (i.e. tempeh is not tofu, nor is seitan). Honestly, the common sense exception already exists, it's called WP:IAR, and it's basically a question of if there's no chance anyone would bother to challenge it. SDY (talk) 15:35, 22 March 2011 (UTC)
This statement along with SDY's refutation, in fact, is perfect evidence for why we shouldn't allow "common sense" exceptions--because very often, common sense is wrong. Qwyrxian (talk) 21:56, 22 March 2011 (UTC)
I wouldn't remove something like that unless I thought it was wrong. If I was just wondering and couldn't find a reference quickly I'd put {{citation needed}} on. If you see a citation needed that's been there for some time and can't find a confirmatory citation then of course deletion is in order even if you're not altogether sure it is wrong. Dmcq (talk) 17:00, 24 March 2011 (UTC)

This... sucks

I would like to use my freedom of speech to criticize this policy. If an claim can be proven than it should be used. The goal of Wikipedia shouldn't be to re-phase stuff said by reliable sources but it has become that with WP:OR. --SomeDudeWithAUserName (talk with me!) 00:45, 24 March 2011 (UTC)

To clear up your first misconception, you have no free speech right on Wikipedia. However everyone is free to participate if they don't deviate too far from acceptable use - which isn't fully formalised but WP:5P is a good indicator.
Secondly summarizing what other people have written in reliable sources is practically exactly what editors should be doing on Wikipedia. WP:5P is a good summary of that too. Dmcq (talk) 00:58, 24 March 2011 (UTC)
No, just because something is made for something doesn't mean that is how it should be used. That is like saying the Ku Klux Klan(modern version, not the original) should being hating on African-Americans and do bad things.
The Ku Klux Klan should be a group for people being proud for being white but not attack non-whites.
That is exactly what I fell like with regards to Wikipedia. IT SHOULD BE CHANGED! I think the biggest problem with Wikipedia is that it usually ignores that there is a community and that it is not a paper encyclopedia. Thanks, Steve T. R.!
WP:NOR is what keeps Wikipedia from being just another site on the internet with people making stuff up. If you want to write about your great ideas, there are many other Web 2.0 options. SDY (talk) 06:18, 24 March 2011 (UTC)
My great ideas, the only one that I explain clearly is Stevecracy. I guess I was wrong on that idea, but I still think Wikipedia should get reformed. Thanks, Steve T. R.! --SomeDudeWithAUserName (talk with me!) 18:04, 24 March 2011 (UTC)

Ithaca college library link

I have some reservations about the contents of the Ithaca college library link (currently footnote 4 in the text of the policy):

The Ithaca College Library compares research articles (primary sources) to review articles (secondary sources).

This draws a bright line between "review articles" being secondary sources and "research articles" being primary sources. This may be more true in some of the sciences, but it is certainly less true in others. In many sciences (for instance, mathematics, physics, and economics), "research articles" include a substantial review, since they build on earlier work by others. So a research article may be either a secondary source or a primary source, depending on how it is used. Is it helpful to endorse a bright line between primary and secondary sources, even if it goes against established practice in some areas? Sławomir Biały (talk) 15:06, 17 March 2011 (UTC)

I have rewritten the footnote to more accurately (as I see it) reflect the contents of the link. Sławomir Biały (talk) 14:58, 18 March 2011 (UTC)
Actually, in pretty much any scientific discipline a research article is has a substantial portion which is a secondary source and reviews the earlier literature. But it is not surprising that the confusion extends to some less-informed librarians as well. II | (t - c) 17:52, 18 March 2011 (UTC)
Isn't it also the case that all such review material indicates a reference from which it comes? If a particular fact (etc.) is going to be cited wouldn't it seem the better course in most cases to consult the original source cited and refer to that rather than to cite the review section of an article describing other new research inasmuch as the review is there as background for the currently reported work and is not conceived as or intended to be anything remotely resembling a comprehensive review? If, though, the author Jones references the work of Smith in a particularly cogent way then perhaps in that case the Jones article might be the preferred source to rely on and to cite for Smith's information for the purposes of creating an encyclopedia article of quality. The goal is to honestly edify. Minasbeede (talk) 23:23, 22 March 2011 (UTC)
Not necessarily (see WP:PSTS, which we are discussing here). But given the substantial confusion about this point of policy, it might be a good idea (as others have suggested) to have a separate guideline (or essay) on PSTS which is informed by concrete examples. Sławomir Biały (talk) 00:31, 25 March 2011 (UTC)

Research literature as primary and secondary source

Coming from a background in mathematics editing, it's often the case that research literature can be used as either a primary source (about the conclusions of the authors) and as a secondary source (as a review of the literature). I'm also aware that in other fields (medicine, for instance) the reviews contained in primary research literature tend to be less reliable as secondary sources. Is there some way we can harmonize these differences? I feel like the policy, as it is now written, excludes all research literature as a "primary source" (except for review articles), and I don't think that's helpful as a blanket stance. At the very least, it might reinforce a misunderstanding of the notion of primary and secondary sources. Sławomir Biały (talk) 16:30, 18 March 2011 (UTC)

I recall this being mentioned elsewhere, perhaps WP:IRS. Do you happen to know what it says there about reviews of other work in papers that are primary sources? It might be best left to discuss that there. Or if we introduce it here it should be the same wording to reduce confusion. SlimVirgin TALK|CONTRIBS 16:37, 18 March 2011 (UTC)
I somewhat agree, however even in math many primary sources (original publications) should be treated with care, since they not firmly established yet. Original papers which are not that high profile may have a good chance containing errors. So even in math writing articles based on original publications only is usually not a good idea (there might be notability issue there as well).--Kmhkmh (talk) 16:39, 18 March 2011 (UTC)
That's a good point. There is an additional confounding issue of source quality separate from the primary/secondary distinction. Sławomir Biały (talk) 16:52, 18 March 2011 (UTC)
This is partly why this whole section ought to be moved to WP:IRS, as we were talking about in one (or more) of the preceding threads. It seems to be a historical accident that this information has found itself on this page - it has very little specifically to do with the topic of "original research", and everything to do with the reliability of sources.--Kotniski (talk) 12:34, 19 March 2011 (UTC)
I am still not happy about simply moving the section to IRS (I am beginning to see that IRS might be a better location for the material than NOR is... but... I think the section would be an awkward fit there as well.)
That said. It is clear that the issue of using primary vs secondary sources is one that relates to multiple policies and guidelines. So... I think it is time to revive an old suggestion... create a separate PSTS guideline page. This would give us a single place to discuss all the complex issues that relate to the uses of primary vs secondary sources ... and allow us to summarize in other policies and guidelines (such as here)... which would help to keep those other policies and guidelines more focused. Blueboar (talk) 14:02, 19 March 2011 (UTC)
Maybe not a bad idea - but why do you think it would be an awkward fit at IRS? (From what I can tell, when the issue arises of whether a particular source is primary or secondary, the subtext is usually "if it's primary, it's not reliable". I think PSTS would actually fit at IRS pretty well.) --Kotniski (talk) 14:37, 19 March 2011 (UTC)
BB, you can't downgrade part of policy to guideline status without very wide consensus, especially when it's a core content policy, and this is an integral part of it that lots of people rely on. SlimVirgin TALK|CONTRIBS 18:29, 19 March 2011 (UTC)
Slim... Of course we would need wide consensus for such a change... I never said otherwise. But you can't get consensus (or even find out what consensus is) if you don't suggest the idea. And if people like the general idea of splitting the section off and expanding it, but want the resulting page to have policy status, then we can do that... I have no problem with crafting a PSTS policy.
My concern has always been that within the context of the WP:NOR policy page the PSTS section amounted to confusing instruction creep. I think it distracts our readers from fully understanding the concept of "No original research" (and I think it would distract the reader from understanding the concept of "Identifying a Reliable Source" if we moved it there). But I agree that it contains useful and important concepts on its own. I like PSTS and agree with what it says (although I can quibble with how it says it). That is why I would like to see the section moved out of this page, and given its own page. One where we could expand on the topic, and better explain how it fits into several policies and guidelines, without distracting the reader from understanding the policy/guideline concepts presented on those other pages. I suggest this idea not as an attempt to "demote" anything... My intent is simply to suggest a different way of presenting it. Blueboar (talk) 19:08, 19 March 2011 (UTC)
You say "of course" but we've had people (not you) on this page for several weeks trying to make major changes without that wide consensus, and reverting when challenged, so it's worth stressing that any substantive change can't be made by people on this page acting alone.
I can't see the point of hosting it as a separate policy. First, it's closely tied to the concept of NOR, and you'd lose that on a separate page. More importantly, there are multiple discussions going on at the moment, including with the Foundation, about the increase in WP's instruction creep, the number of guidelines, the attempts to impose guidelines as though they were policies. I don't think we should do anything to add to that, whether it's adding more rules/less clarity to this page, or splitting off parts of it to create yet more policies. SlimVirgin TALK|CONTRIBS 21:59, 20 March 2011 (UTC)
This is a reason why I think it would be better moved to IRS, rather than start another page. But I think you're wrong in saying that [the distinction between primary/secondary/... sources] is closely tied to the concept of NOR - as discussed previously, it's really tied closely to other concepts and only loosely to that of OR. The section's being placed on this page (and worded to try to make it fit here it isn't even worded to make it fit here) encourages the fallacious thinking that it's more tied to OR than it really is (and more practically, removes this important information from the main guideline on reliability of sources, where it should occupy a central place).--Kotniski (talk) 09:34, 21 March 2011 (UTC)
It's closely tied to the concept of OR because OR makes WP a primary source; and using primary sources to interpret and draw conclusions makes WP a secondary source —whereas for the most part (not always), WP should be a tertiary source. That's the relationship between types of source and the OR policy.
I know you want to see V and NOR on one page. So did I. But we tried and failed, and the community, including Jimbo, showed that it wanted the two ideas (V and NOR) to remain separate, both conceptually and in terms of presentation. There's no point pretending that consensus doesn't exist. SlimVirgin TALK|CONTRIBS 00:50, 25 March 2011 (UTC)
I was under the impression that a significant majority actually supported the proposal, so it would be wrong to say that consensus existed, even at that time, for keeping things the way they are (and of course consensus can change over time). But I don't mind that much having two separate pages as long as there's a clear division of scope and they don't keep trying to duplicate each other. About the relationship - I don't understand how what you say implies any kind of close relationship between this information and OR. Original research is most easily identified on its own terms - trying to define it in terms of what kind of source it makes Wikipedia just confuses the issue totally (in fact I don't think it can be done, since OR can be based either on no sources - making WP a primary source - or on overinterpretation of n-ary sources - making WP an (n+1)-ary source - it's still OR whatever the value of n).--Kotniski (talk) 11:04, 25 March 2011 (UTC)
We did get support; see Wikipedia:Attribution/Poll. We got 424 supporting, 354 opposing, and 102 neutral. So not enough. But it was huge support given that Jimbo had come out very publicly against the merger, and in those days people were more inclined to go along with that. So I felt it was a good show for the merger. I don't know whether you know the story, but the merged page was actually policy for a few weeks, and people did seem pleased with it. Then someone added a "deprecated" tag to the NOR policy, Jimbo saw it, and that was that. Even though he'd been told about the merger, he strongly objected to the "deprecated" tag. And so the whole thing unravelled, and we were asked to hold a wiki-wide poll, which took ten years off my life, with polls about which question to ask in the poll, an RfC about which tags to put on the poll page, then polls about that too. And of course when you get that many people involved, quite a few are voting without knowing what the issues are, so there were oppose comments from people who'd misunderstood what the proposal was. All very depressing.
But the only reason we got the support we did is that the merger didn't actually change the policies. We just spliced the pages together; redirected IRS; and we got rid of the idea of "verification" and replaced it with "attribution." So it was very simple:
  • Unattributable = material for which no reliable source exists = original research = never allowed.
  • Unattributed = material for which no RS has been added, but it exists = sometimes allowed (if not challenged or likely to be challenged, etc).
So material on WP has to be attributable, but not necessarily attributed. Two policies in one sentence. Nice and clean. SlimVirgin TALK|CONTRIBS 20:13, 25 March 2011 (UTC)
Thanks for the background; I'd kind of heard the details in bits and pieces, but not the full story. So perhaps we could have another shot at trying to persuade people one day; maybe consensus can be achieved if it's handled a bit differently. But on what we were talking about above, are you persuaded yet that PSTS really doesn't belong in this policy? --Kotniski (talk) 11:34, 26 March 2011 (UTC)
Apologies for the lengthy reponse, but I want to explain where I'm coming from, because I think you've developed a view of me as intransigent for no reason, or for no good one. Since the ATT poll, my own preference has been to respect the strong consensus we saw (for no policy change, regardless of whether on one or two pages), and not to mess around with the policy pages anymore. Instead, I watch the discussions carefully to see which issues keep arising, and who keep bringing them up, to check that it's not always the same people. Then one day, my hope is to write a new version of ATT, but a different one: something lean and mean that gets to the core of the objections. But we'll only get one more shot at it in the foreseeable future, so I don't want to waste another attempt.
The problem is that there are a small number people who hang around these pages who don't understand them; who have difficulty with writing skills; who will object no matter how sensible the proposal is, or will agree no matter how many problems it would cause, because they see change where there is none, and don't see change when it's real. This was a big issue during the ATT proposals. I don't mean by that that there were no valid objections, because there were. But there were others who were almost trolling, and they can be incredibly wearing. It's because I can't face dealing with those kinds of objections that I've never given it a second try.
So to answer your question: I can see the relationship between PSTS and V, of course. But there's also a relationship between PSTS and NOR; indeed the concept of NOR evolved, in part, around the misuse of primary sources. So I don't see that there's any clear benefit in moving it. In addition some of the people who want to move it would like to see PSTS done away with entirely, or downgraded to a guideline. So that concerns me too, in case the move offers people an opportunity to weaken it. So instead of fiddling around with that, I want to keep my eye on the ball, which is the eventual—and incredibly careful—rewriting and merger of these two policies. SlimVirgin TALK|CONTRIBS 00:00, 27 March 2011 (UTC)
Well, glad we seem to share the same long-term aim, anyway;) But I don't see that this is incompatible with making improvements at a more detailed level in the meantime. I still don't see, for example, any particular connection between (most of) PSTS and OR - the latter may have developed as a result of misuse of primary sources, but that doesn't imply that the ability to distinguish between primary and secondary sources is of any great importance for the question of OR (given that secondary sources can be abused in just the same way, and it's still OR). Anyway, discussion seems to be emerging over at V about the proper division (if any) between these two pages, so maybe this matter will be settled as a result of that.--Kotniski (talk) 10:26, 27 March 2011 (UTC)
──────────────────────────────────────────────────────────────────────────────────────────────────── SlimVirgin wrote, "The problem is that there are a small number people who hang around these pages who don't understand them; who have difficulty with writing skills". — I agree that there are these people, but I'm sure I disagree with SlimVirgin regarding who some of them are. Note that being an expert at maneuvering in the Wikipedia system and gaining control by any means, does not mean that one is a good writer and organizer of policy pages. (talk) 13:30, 27 March 2011 (UTC)

WP:MEDRS#Biomedical journals addresses this issue as follows:
"Research papers that describe original experiments are primary sources; however, they normally contain previous-work sections that are secondary sources (these sections are often incomplete and typically less useful or reliable than reviews or other sources, such as textbooks, which are intended to be reasonably comprehensive)."
This covers the important points: Yes, it's a secondary source, but, no, you shouldn't rely very heavily on these, especially if proper sources are actually available (e.g., it's not a newly desscribed or rare disease). In particular, these sections are often criticized for cherry-picking sources to make a case for the authors' own research (which we have rendered diplomatically as "incomplete"). WhatamIdoing (talk) 20:09, 19 March 2011 (UTC)
That does seem to cover the important points. The point should be made that one can use research papers as secondary sources, and that one must also use caution in doing so to avoid the authors' potential biases. I think this is likely more of a problem in empirical research than in other more analytical kinds of research. Sławomir Biały (talk) 22:26, 19 March 2011 (UTC)

This Talk page section is off the topic NOR. It is only being discussed here because of the inappropriate inclusion of the excessive discussion of primary, secondary and tertiary sources in PSTS and the Notes. (talk) 19:44, 21 March 2011 (UTC)

This policy is being abused, badly!

My comment is a variation on "This Sucks". I understand that Wikipedia is not a peer-reviewed publication and needs a policy that allows arbitration of conflicting viewpoints among editors who are essentially hobbyist volunteers. So I understand the motivation for "No original research".

However, that being said, something is wrong with the whole shebang when practically every article I happen to come across has a multitude of headers complaining about inadequate sources and inline citation needed comments when it is fairly easy to see from the text and accompanying citations where the writer got his information, and that information obviously came from reliable sources and is not being criticized for its accuracy.

I should note in particular, that in many fields, primary sources are the ultimate, indisputable authority. In an article describing how a piece of computer software or hardware is supposed to work, the appropriate manual for the product is the indisputable authority. No secondary or tertiary source even approaches that level of authority. Similarly for any proprietary product, no matter what the field. The manufacturer is the indisputable authority. I should not be seeing any complaints about the adequacy of the sources because they happen to all be from the manufacturer.

'Nuff said, I think.

Dlw20070716 (talk) 00:35, 25 March 2011 (UTC)

Not quite 'Nuff... This policy does not ban primary sources... it bans the inappropriate use of primary sources. Thus, in situations where a primary source is the "ultimate authority" you can (and should) cite it. Blueboar (talk) 00:39, 25 March 2011 (UTC)
Also, I think it's important to note that the manufacturer is not in any way an "indisputable authority", particularly in the area of software these days. Even in the area of manuals and specifications of hardware and software, there is variance between what is described therein and the implementation of same in the real world, and outside of those narrow areas, one sometimes finds oneself in a spinning wonderland of the not-quite-true and outrageous misrepresentation as one gazes upon product descriptions. --Nuujinn (talk) 22:11, 28 March 2011 (UTC)


For the purposes of WP:NOR, the policy of the whole section PSTS can be covered without the use of the problematic terms primary, secondary, and tertiary by simply using the following sentence.

”Do not add to Wikipedia articles your own analysis, synthesis, or evaluation of material found in sources, nor add your own ideas, opinions or experiences that have not been published in a reliable source.” (talk) 10:50, 20 March 2011 (UTC)

No. It is definitely important to emphasize not to use primary sources, since it is much harder to avoid committing original research if one relies too much on them. The point of the above thread is that the policy should convey an accurate understanding of primary and secondary sources. Sławomir Biały (talk) 12:15, 20 March 2011 (UTC)
I have to quibble... the policy does not "emphasize not to use primary sources"... In fact, it explicitly says we can use primary sources... but we include a strong caution about how to use them. The point of that caution is (or should be) to emphasize how it is easy to (inadvertently) form Original research when relying on primary sources. But the ban is on the OR... not the use of primary sources. Blueboar (talk) 16:01, 20 March 2011 (UTC)
I should have written more carefully. My point was the same as you are making, though, that original research can be harder to avoid when dealing with primary sources. This bears emphasizing in policy. Sławomir Biały (talk) 20:19, 20 March 2011 (UTC)
Ah... ok... then we agree. Blueboar (talk) 21:39, 20 March 2011 (UTC)
But that's still not the main reason we distinguish primary from other sources. The distinction is far more significant for notability and verifiability than it is for "original research" (accepting for the moment that original research really is a distinct topic that justifies having its own page). It may be most common for original research to be committed by over-analysis of primary sources, but it isn't necessary to know that the source is primary in order to identify that over-analysis as original research. And of course we can still emphasize on this page that people have to be particularly careful with primary sources, without actually including the definition of primary source here (we can link to a definition that appears in a more suitable place). --Kotniski (talk) 09:27, 21 March 2011 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── I like the guideline:

  • WP:V is about what sources are, and which sources are reliable, while
  • WP:NOR is about how to use sources properly and how to avoid original research

Given the above I think PSTS should remain in NOR. LK (talk) 01:15, 29 March 2011 (UTC)

That sounds a bit muddled to me. "Which sources are reliable" is an essential part of "how to use sources properly" (i.e. you don't use a source to support a statement for which it is not reliable). And the only real concrete consequence of PSTS is that primary sources are not generally reliable for statements other than about themselves, so it would seem to belong in V rather than NOR even on the basis of the above split.--Kotniski (talk) 09:29, 29 March 2011 (UTC)
A while ago I proposed that the WP:NOR should concentrate on the cautioning against overanalysis of primary sources. I suggested that the definition of them and other advice relating to them should go in a seperate guideline. I even went so far as to draft such a guideline. See WP:PSTSPROP. It was shot down. Mostly by some editors who thought that it was weakening the OR policy. I still don't get how concentrating on OR and not getting distracted by unrelated matters is weakening the OR policy. But there you go.
Yaris678 (talk) 14:19, 29 March 2011 (UTC)
Also note that material not specific to policy, about the subjects primary, secondary and tertiary sources, can be put in the corresponding wikipedia articles and referred to, instead of cluttering up policy and guideline pages.
Furthermore, the terms are problematic and unnecessary. For example, a peer reviewed journal article is put in the same category of primary as are works of literature or a speech; the terms are relative to how the source is used; and a single source can be partly one type and partly another. NOR does not need to be discussed using these problematic terms. (talk) 18:42, 29 March 2011 (UTC)

Routine interpretations

May I suggest that the sections "Translations and transcriptions" "Routine calculations" be merged into one section with the name by "Routine interpretations". The proposed text for the replacement section is:

Routine interpretation of data that follows standard rules which are easily understood by well-informed non-specialists is not considered original research provided editors agree that the information presented and its application correctly reflect the sources. Such interpretations include, but are not restricted to:
  • Faithfully translating sourced material into English, or transcribing spoken words from audio or video sources. For information on how to handle sources that require translation, see Wikipedia:Verifiability#Non-English sources.
  • Making routine arithmetical calculations, such as adding numbers, taking averages, converting units, or calculating a person's age. See also Category:Conversion templates.
  • Reading data off maps such as measuring distances, simple interpretation of spot features or contour lines and the like

The reason for this proposal is to incorporate guidance on the use of maps as a reliable source. In addition, this change will consolidate permitted interpretation into one section. Martinvl (talk) 09:31, 28 March 2011 (UTC)

Slight change in wording. Martinvl (talk) 08:49, 29 March 2011 (UTC)
I am concerned that editors will read the first sentence of your proposal and misconstrue it. The three items you list are specific exceptions to a firm rule against interpretations. (indeed the entire point of the exceptions is that these things don't require interpretation). Better to deal with them separately. Blueboar (talk) 18:53, 29 March 2011 (UTC)
I don't want to see the math part (routine calculations) lumped with other stuff, since this is under discussion (and from some perspective a special case) elsewhere (also from some perspective a special case) and the wording might get slightly adapted to suite the needs for math articles in the future anyhow.--Kmhkmh (talk) 19:55, 29 March 2011 (UTC)
How about a new section "Use of maps". The text would be:
An interpretation of information contained in a map is not considered original research provided that such intepretation does not draw on specialist knowledge that is not on the map itself. This includes but is not restricted to calculating distances, directions, orientations and interpretation of contours. Thus, stating that a valley is "U-shaped" on the basis of contour information is permissable, but stating that it is "glacial in origin" will require further evidence.
Martinvl (talk) 20:53, 29 March 2011 (UTC)
I agree with your assessment of what is allowed and what not (in geography), but i'm getting a bit wary regarding the level of detail. Imho such things are matter of common sense. There seem to be an increasing (somewhat nonsensical) notion of some editors, that anything which isn't a almost literal replication represents OR. You cannot write intelligible ona subject without a certain degree of interpretation and actually understanding of the (secondary) sources and you need some common sense and careful consideration to see when an interpretation is appropriate and when it turns into OR. The problem that I see here now, is that if we are trying to detail this fine line between required interpretation and unwanted interpretation (OR) for every area, then we get an ever increasing policy. Also we'll get more editors who will be attempted apply the policy formalistic manner without using their brains.--Kmhkmh (talk) 13:41, 30 March 2011 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── I agree that the policy sohuld be given as examples rather than a list of "authorized uses." The test that I always use is the same as the one for sourcing: is it reasonably likely to be challenged? If it's taking two numbers from separate points in a source to extrapolate a third number, that's likely to be challenged (e.g. a plane's average speed and the distance travelled to calculate a time), but there are situations where it's appropriate. SDY (talk) 17:45, 31 March 2011 (UTC)

The consensus suggests that the boundary that I have defined is correct, though there is ambivalnce as to whether or not it should appear in the form that I proposed. I have decided to be bold and have added the text as described above. Lets see if this stimulates more discussion leading to a more general guidance, possibly using the "U-shaped valley/Glacial valley" as a case study. Martinvl (talk) 09:32, 1 April 2011 (UTC)
I've proposed some other language just for discussion, it's in the edit history of the policy. Other than that I've reverted the addition about maps. It smacks of WP:CREEP and WP:BEANS. Honestly, even the stuff I'm considering adding seems redundant. WP:IAR allows for common sense "violations" of policy, and explaining every possible situation just makes our policies longer, clumsier, and more intimidating to new users. SDY (talk) 15:45, 1 April 2011 (UTC)
(edit conflict) The problem it that this allows interpretation to be be allowed as fact, something that goes against the cetnral principle of NOR. An experienced reader of maps can see from contours that a valley is U shaped by virtue of it having a flat bottom. However, that depends on experience and judgment. How flat is flat exactly? What if (as in many real life examples) the water course that created the original valley has cut a new, smaller, V shaped valley in the bottom of the U shaped one in the post-glacial period. These are essentially issues of judgment and interpretation rather that simple conversion of existing data into textual form.
Even things like direction can be troublesome since it can depend on the projection used in the map. An extreme situation would be something like travel between New York City and London. Your initial bearing on leaving New York would be approximately north east as the map would tell you. Going in the reverse direction you would set off from London in a direction north of west. This is a seeming paradox that can't be resolved from the map - how can both cities be north of each other? A globe would tell you the truth and easily show the great circle which indeed is initially northerly in both directions, but not a map.
By encoding what potentially could be complex issues of interpretation here in NOR we give carte blanche to use maps to prove anything we want, whether it is true or not, with only a map that does not explicitly make the point being made as a reference. Debates over what is routine or clear can easily become troublesome and usually boils down to "I'm smarter than you in this area and understand this stuff. You don't so your opinion doesn't count", regardless of the merits or accuracy of either position.
Simple matters of fact (Germany has a land border with France to its west) that are essentially re-presenting data on the map in textual form are fine, but those are never going to be challenged by a reasonable editor. Any element of judgment or inference is not a routine calculation as has historically been permitted, and shakes NOR to its very core. We demand references for inferences everywhere else and maps should be no different. The current drafting does not even include a presumption against the inference so this policy can be used to justify a statement even after a challenge. Again, that goes against the previous presumption against a re-presentation of details presenting in a different form where that re-presentation proves contentious. Crispmuncher (talk) 16:10, 1 April 2011 (UTC)
(I have indented the previous editor's comments). What is the difference between "translating" a map or "translating" foreign text. May I propose that the first sentence of the of the section "Translation and Transcription" be replaced by "Faithfully translating or transforming sourced material into English text using standard techniques is not considered original research. Sourced material can include, but is not restricted to foreign language text, spoken words from audio or video sources, maps and reputable computer programs. Ideally computer programs should be available on-line as web pages." One would expect standard techniques to be described somewhere in Wikipedia.
The extreme example cited by the previous editor can easily be resolved by the fact that great circles cannot easily be identified from flat maps. Furthermore, if two editors argue "I am smarter than you", then they should call for mediation of one sort or another. Martinvl (talk) 16:40, 1 April 2011 (UTC)
I strongly agreen with Martinvl. Calculating a slope on a topographic map is about the same as saying "France is west of Germany." If you give a young earth creationist and a geophysicist the same map they will come to the same conclusion about what the slope is (though they'll disagree on pretty much everything else). There's no interpretation involved, just reading data. It's like the joke "there are 10 kinds of people in the world, those who understand binary and those who don't." If you don't, someone can explain it to you because they're "translating." SDY (talk) 17:32, 1 April 2011 (UTC)
In my view calculating a gradient is fine as a routine computation. This problem is in interpreting the result. Looking at a map and stating "The gradient between these two points is 1 in x" is uncontroversial. The difficultly comes in determining whether that amounts to flatness - "Since this gradient is less than 1 in y the valley has a flat bottom". Who chooses y exactly? How do they arrive at that figure? What factual basis does that conclusion have with respect the the published source, i.e. the map? There are other problems: presumably we would need the measured gradient to be some proportion of the overall width of the valley (and repeated over its length) so that purely local elements do not affect the consideration of the valley as a whole. Who determines those proportions? The same considerations apply as to the maximum gradient.
The comparison with translated sources is a false one in my view: policy on translations represents an additional requirement rather than an easement since it begins from the master principle and builds on it. Statements must be sourced. If that citation is in a foreign language then a translation must also be provided. The authoritative citation remains the foreign language source. The translation is nothing more than a convenience to the reader.
Since when (and where) do require translations? I just recall conviency translations and translations upon request.--Kmhkmh (talk) 09:54, 3 April 2011 (UTC)
As for the problem of identifying great circles on a map - that is precisely my point. You can't easily do it on a map. A simplistic analysis of a map (drawing a line between two points) will therefore give a misleading result. A misleading result than can't be bounced out of the article as uncited since it is the result of a simple analysis of the map that would be specifically authorised here - remember there is no provision for challenging dubious or unlikely interpretations. Crispmuncher (talk) 18:00, 1 April 2011 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── In the editorial process there will always be decisions, and honestly it comes down to consensus among editors if there's a disagreement. In all cases, I strongly agree that if there is a calculation, the raw data must be available, but does that mean the map has to be copied to the article? The map-specific issue is partially why I proposed the alternate language - this is still all about editorial consensus, and giving specific rules for specific situations is only going to cause problems because the policy might explicitly authorize something that, as Crisp is pointing out, might be inappropriate. SDY (talk) 18:21, 1 April 2011 (UTC)

I've been trying to get my hands around something like this. Suppose the "routine calculations" are, never the else, extensive. For example, a table providing historical budget vs expenditure numbers for a city. They can easily be derived from primary sources (city budget and end of year financial statements), but may include adding up a bunch of numbers. For aesthetic reasons, you don't want to put the details in the article, but it makes verification easier to say how the numbers were derived. My read of the policy today, is that one would be discouraged from including such a table (even if it has a NPV) because the act of synthesis is "original research". The current NOR article isn't very clear on how complex but routine calculations like that might be viewed. Thoughts? Thengeveld (talk) 23:54, 1 April 2011 (UTC)
In response to User:Crispmuncher - there will always be a degree of interpretation. As an example compare Driver location signs and this]. Both were written under the same conditions and share at least one references.
I have updated the article with my last proposal. Martinvl (talk) 06:56, 2 April 2011 (UTC)
After reading [this reference] I changed the text "taking measurements from ..." to "extracting information from ...". For the benefit of non-UK readers, this reference is aimed at 16-year-old students in conventional English schools. I assert that anything described in the reference is "routine". Martinvl (talk) 06:49, 3 April 2011 (UTC)

Without radical reform of this policy, Wikipedia will be doomed

Google is about to implement a new page ranking policy which will favor webpages that contain original content. Webpages that don't contain content that can't be found elsewhere, like many Wiki articles, won't show up prominently in Google searches. So, if we don't implement radical reform, we'll see Wikipedia losing its prominent status, which in turn will lead to many editors to stop editing here.

Of course, we can't just abandon the NOR policy. But we should reform it in the direction of allowing more nontrivial analysis of existing content. Wikipedia will then still be depreciated by Google somewhat, but it may then survive. Particularly the Wiki-articles on political subjects will be hard hit, as there is little nontrivial analysis to dothere. But we can save the articles on hard sciences here.

The very ridiculous clause in the current version of the NOR policy allowing only "routine calculations" should go. Almost no one editing math and physics pages sticks to this anyway. Instead, we should move in the opposite direction, invite more expert editors here who are able to produce good quality original educational content, which is however 100% based on what can be found in textbooks and the scientific literature. So, this would not be "original reseach" as scientists define it. Of course, the way Wikipedia defines "original research" is ideosyncratic to Wikipedia, we should not confuse real "original research" with what we currently call "OR" here. Count Iblis (talk) 14:36, 1 April 2011 (UTC)

I don't quite get why she should worry about Google. Being ranked top in Google searches is hardly any of hour project goals.
As far as "routine calculations" are concerned I agree however. WP has the problem of becoming overly bureaucratic and to develop a quality control insisting on a literal application of guidelines, nevermind how braindead their application in an individual might be. I've noticed that some math articles often get missing or improve references tags when they don't use inline references for basic middle or highschool algebra--Kmhkmh (talk) 22:04, 1 April 2011 (UTC)
What is Google's new policy exactly? If it's just demoting pages that have copied text from some pre-existing page (and that's all I can imagine it doing), then Wikipedia should be unaffected - we generally produce original text, even if it's not supposed to contain original ideas or reasoning.--Kotniski (talk) 14:42, 1 April 2011 (UTC)
I suspect you may be confusing "original research" with "original content". The two are very different concepts. Kansan (talk) 15:02, 1 April 2011 (UTC)
  • My only question here is whether we will be hurt by the existence of mirror sites. Kansan (talk) 14:47, 1 April 2011 (UTC)
  • Um... why should we care where Wikipedia appears on Google search results? tfeilS (talk) 14:57, 1 April 2011 (UTC)
  • Motion to PAnic Wikipedia is surely doomed! Brilliant excerise in logic Count! The Resident Anthropologist (talk)•(contribs) 15:04, 1 April 2011 (UTC)
  • Any chance of getting that information sourced please, Count Iblis, so we can see what it involves for both them and us please? Motion not to panic until we see what's going down. No More April Fools. 15:19, 1 April 2011 (UTC)
  • Not a problem for WP: we do not allow duplication of content from other sources except in small, appropriately attributed quotations. If we wholesale taking the content of other pages and replicating them here, completely ignoring WP:COPYVIO, that might be of concern, but since we have copyright violation partoling, its not an issue. It certainly isn't an WP:OR issue. (And if Google actually has a linguistic engine that can compared differently-stated sections of text and call them equal -- well, that's an amazing human achievement...) --MASEM (t) 15:21, 1 April 2011 (UTC)

I am not sure the proposed solution is the right one. Changing something just because it doesn't work is never the right solution. Instead, we need to add a new layer process. I suggest we just create WikiDYK as a project for creating large amounts of original research on topics that are sufficiently boring so that nobody cares whether it's true or not, original or not. This is tricky, but should be doable. Then every article will be allocated a random tidbit that will be placed there in a sidebar. For example the article Newfoundland and Labrador might have the following sidebar: "Did you know? In 1758, there were 5 weddings in Massenheim, Germany – more than in any other year of the same decade." We already have a lot of editors who concentrate on writing BLP stubs for little known people or DYK articles on ships, so the motivation for such a project is certainly there. Hans Adler 15:32, 1 April 2011 (UTC)

This seems to me to be worrying over nothing. Google's ranking system is fairly opaque and the recent changes are essentially fiddling at the margins in most cases. Most Wikipedia pages will still float up the the top. To be honest, downgrading Wikipedia in general would be no bad thing: if there is an article for a given search term in invariably gets a top or high ranking. If it is a well-developed article then fine. If it is little more than a dictionary-style stub then it is often little more than a distraction that reflects poorly on the project. In any case, it seems premature to be worrying about what might happen as a result of something, and altering core policies seems precisely the wrong reaction even if it does turn into something. Crispmuncher (talk) 16:42, 1 April 2011 (UTC)

  • (I originally wrote this with the lead-in 'Is this an April Fool's prank?' and I still don't know if it's serious or the greatest April Fool's hoax of all, but I'm putting back what I wrote now with a few small edits) If this is true, I have hope that the Google execs are smart enough to demote the Wikipedia clones, not us. Even the now-dead Cuil understood that. THis is assuming, by the way, that by "original content" you mean exact copies. If you meant statements in Wikipedia that are merely cited to another source are still considered not original, even though our wording is in most cases completely different, then I think this either is a prank by someone at Google or a misunderstanding. My apologies if Im spoiling a joke, but it seems this has already caused significant confusion. Soap 15:11, 1 April 2011 (UTC)
  • Could you link to a page describing Google's new page rank policy? Soap 22:43, 2 April 2011 (UTC)
Hmm...interesting. I didn't think of April Fools Day when reading the first posting, however at least on the replies seems to carry a clear 1st of April mark :)--Kmhkmh (talk) 02:21, 2 April 2011 (UTC)

Most of Wikipedia is built upon "violations" of wp:nor, if it is taken literally. First, just about anything except a direct quote violates a literal reading of wp:nor. (e.g. summarization is a violation of wp:nor-taken-literally.) Second, the way that most of Wikipedia is actually written is from consensused knowledge (knowledge = an integration of hundreds of sources, with sources at hand in mind) and THEN sourced per wp:ver & wp:nor. Most of the time this disparity between the rules and reality works fine. But on contentious articles, any wiki-lawyer warrior can roll out the oft-"violated" rules to exclude the other side's content Using this they can also shut down the above process which works elsewhere in Wikipedia, and there is no mechanism to get that process back in place because it is technically "illegal". That is a big part of why most or all contentious articles in Wikipedia are eternal messes and eternally unstable. Sincerely , North8000 (talk) 13:52, 3 April 2011 (UTC)

  • What exactly is the rumored change at Google envisioned to entail? If all they are doing is deprecating mirror sites and link farms then this is nothing to panic about. Is anybody going to link to the actual rumor, or should this discussion be closed as a bit of April folly? ~ Ningauble (talk) 15:15, 3 April 2011 (UTC)

Reputable computer programs, and maps

I removed this from the Translations and transcriptions section because I couldn't work out what it meant: "Sourced material can include ... information returned by reputable computer programs." [3] What is being referred to here? SlimVirgin TALK|CONTRIBS 13:32, 3 April 2011 (UTC)

I had in mind the use of Google Earth to compute distances, conversion between the Gregorian calendar and the French Revolutionary Calendar here, the use of computer-based dictionaries for simple text translation (as used in Kilometres per hour and similar such programs. Martinvl (talk) 16:43, 3 April 2011 (UTC)
Yes, I have no problem with that. So why was my edit to your section reverted twice? SlimVirgin TALK|CONTRIBS 16:45, 3 April 2011 (UTC)
Wow, three times. Martin, you're out of order. You're adding poorly written material to a core content policy, and reverting efforts to fix it. When people object to new material, you discuss; don't keep reverting. SlimVirgin TALK|CONTRIBS 16:47, 3 April 2011 (UTC)
I edited the new map edition, so that it said: "The policy also allows information to be extracted from maps so long as there is nothing contentious about the material." This was reverted, so I've removed the new map material entirely, because maps are actually highly contentious in some areas, so I'm not quite sure what the aim of the new language is. SlimVirgin TALK|CONTRIBS 16:39, 3 April 2011 (UTC)
This is what two editors want to add, but it strikes me as almost meaningless. SlimVirgin TALK|CONTRIBS 16:45, 3 April 2011 (UTC)
"...also not clear what the new words even mean" was said and they seemed understandable to me. May I recommend Wiktionary? Banaticus (talk) 16:49, 3 April 2011 (UTC)
I agree with Slim on this... maps are extremely problematic. In some ways using a map appropriately is as difficult as using a primary source appropriately. While there are ways to use maps without ending up with OR, these are few and far between. Blueboar (talk) 16:52, 3 April 2011 (UTC)
(ec) The writing is poor, to the point where it's unclear what's meant: "Sourced material can include, but is not restricted to foreign language text, spoken words from audio or video sources, extraction of information from maps and information returned by reputable computer programs."
What does it mean to say "sourced material can include .." What is the point of "not restricted to"? What does "information returned by computer programs" mean? What does "extract of information from maps" refer to, and are there no restrictions when the maps are contentious? SlimVirgin TALK|CONTRIBS 16:54, 3 April 2011 (UTC)
I believe it's because the section in question is regarding computer-added or computer-translated information. Source material written in English is easily understood -- no computer intermediary is required -- you and I can just read it. It's only foreign language material that requires some form of translation (perhaps computer translation in the absence of anything else). I agree with Martinvl, although I agree that his phrasing may be a little tough to understand. That should result in a rewrite, though, not an excision of those statements. Banaticus (talk) 17:11, 3 April 2011 (UTC)
We can't rewrite it, because we can't understand it. It therefore should be removed for now. You can't just turn up to a policy page and add whatever words you want, even while acknowledging they're hard to understand, then engage in serial reverting and templating of people who restore the previous wording. That's a disruptive waste of time. SlimVirgin TALK|CONTRIBS 17:15, 3 April 2011 (UTC)
I'll see if I can't rephrase it and I'm sorry that you don't understand it. Did you see my post just before yours where I explained what I believed was meant? Banaticus (talk) 17:22, 3 April 2011 (UTC)
There are two issues: (1) maps, which are often highly contentious, so that has to be fixed; and (2) the computer-generated thing, which I don't understand as written. SlimVirgin TALK|CONTRIBS 17:26, 3 April 2011 (UTC)
I included the words "but not restricted to" as a get-out to allow for other sources that are equally reputable and verifiable, but which were not in the list. Such sources could include great-circle calculation programs etc. If I were to quote the great circle distnace between London and New York and cite internet-based application XYZ, then anybody could verify the calculation by follow the program's operation.
I agree that maps can be contentious, on the other hand they might be quite clear. The same goes for written text. That is why I used the word "standard" in the first sentence. Martinvl (talk) 17:33, 3 April 2011 (UTC)
How's the current version, "Faithfully translating or transforming sourced material into English text using standard techniques is not considered original research. Sourced material may include foreign language text, spoken words from audio or video sources and extraction of information from maps. For information on how to handle sources that require translation, see Wikipedia:Verifiability#Non-English sources." Should I change it to say, "Sourced material translated or transformed into English may include... (added text emphasized)" How are maps sometimes highly contentious? Banaticus (talk) 17:37, 3 April 2011 (UTC)
Maps are contentious where the status and ownership of land is disputed. That's why I wrote: "The policy allows information to be extracted from maps so long as there is nothing contentious about the material." But you reverted me.
I can't see the point of "sourced material may include ..." What do you mean by that? SlimVirgin TALK|CONTRIBS 17:41, 3 April 2011 (UTC)
I see nothing wrong in including contentious material, provided that it is flagged as contentious. I nothing wrong with the text "British mapsRef 1 show the Falkland Islands as being a British dependency, but Argentime mapsRef 2 show them as being part of an Argentine province" even though it refers to contentious maps.
The words "source material" refer to the previous sentence "Faithfully translating or transforming sourced material into ...". Martinvl (talk) 17:51, 3 April 2011 (UTC)
This is what it said: "Faithfully translating sourced material into English, or transcribing spoken words from audio or video sources, is not considered original research. For information on how to handle sources that require translation, see Wikipedia:Verifiability#Non-English sources." The new additions to that section are very mixed up, and should be removed for now. SlimVirgin TALK|CONTRIBS 18:14, 3 April 2011 (UTC)

Editors may wish to be aware of discussion on a similar and relevant topic here, where information from a map has been questioned as OR. Pfainuk talk 18:23, 3 April 2011 (UTC)

Indeed. They're questioned constantly in contentious areas of the world, and a frequent source of disputes. So we definitely can't include the statement that they're fine to extract information from, without making clear that it has to be entirely non-contentious -- in which case why bother saying it? SlimVirgin TALK|CONTRIBS 18:26, 3 April 2011 (UTC)
The matter in question in the Falkland Islands article is width and depth of the northern entrance to the Falkland Sound. These questions can be answered totally and unambiguously by a reference to the relevant Admiraly Chart using standard techniques as taught in English schools to 16-year-olds. I am quite sure that during the Falklands War, every Argentine ship's captain had this information to hand (if not, what were they doing captaining a ship?????) and maybe even the charts (or an older version) to hand. Thus, by definition this is not "original research" - if I were to submit this to a learned journal as a piece of original research, they would roll around on the floor with laughter (metaphorically speaking of course). Martinvl (talk) 18:44, 3 April 2011 (UTC)
Of course maps can be (and often are) used for OR; it happens all the time, particularly when there are territorial disputes, which also happen all the time. What you are describing is not "routine mathematical calculations, such as adding numbers, converting units, or calculating a person's age" - not by a long shot. For that matter, neither is "information returned by reputable computer programs", an incredibly vague phrase. This is an on-going problem; people want to do OR in an article but find they cannot, so they attempt to loosen policy to allow it. Please stop modifying the policy to allow giant loopholes in it. Jayjg (talk) 20:23, 3 April 2011 (UTC)

Idea for a way to allow indirect original research

Couldn't the Wikimedia foundation develop something like a "Wikipeer" site where people can produce research on subjects and allow it to be mass reviewed? I've seen Knol by google and they work under a very similar model as that. It's a shame when someone produces a lot of work and they're not allowed to use their content. Maybe if there was a proving ground for that sort of thing then such data could be used for Wikipedia. Or we could give people a chance to collect their notes and take them to Knol as a place for it, and that way if the research wasn't originally backed up it would be safe somehow. WikiRefundAdjuster (talk) 15:21, 3 April 2011 (UTC)

Examine Wikiversity where such appears to be allowed within reason. Collect (talk) 15:27, 3 April 2011 (UTC)

What is a routine mathematical calculation?

An interesting point has been raised here My point being that as we do not know all the variables we cannot come up with a definitive answer but (at best) a range of possible answers, thus her year of birth would be OR. The counter point is that this still does not constitute OR because it’s a mathematical calculation. Which is the correct interpretation?Slatersteven (talk) 18:35, 3 April 2011 (UTC)

We should err on the side of caution, particularly with a BLP. It's not a routine calculation if we don't have sufficient information; it's guesswork and that's OR. SlimVirgin TALK|CONTRIBS 20:27, 3 April 2011 (UTC)
It's not guess work but straightforward calculation and definitely not OR (if you are 50 in 2010-4-27, then you've turned 50 between 2009-4-28 and 2010-4-27 and hence you were born between 1959-4-28 and 1960-4-27). So giving ca 1960, 1959 or 1960 or 1960 or earlier are all fine. Even 1960 with the inline reference would imho be fine, if we don't take our readers for completely braindead. Because they can notice that no exact date but just a year is given (meaning there is an obvious uncertainty regarding the exact date to begin with) and they can actually read the explanatory footnote/inline reference. I agree in terms of BLP we should be careful in general but we are hardly talking about defamatory information here, so I see no reason for concern.--Kmhkmh (talk) 20:57, 3 April 2011 (UTC)
I see no problem to give 1960 or better ca 1960 with an inline reference to the Daily Express article as year of birth. And yes to compute the year of birth is routine calculations in the sense of this policy (under the assumption the information 50 in 2010 stems from valid source). The argument "we don't know all the variables" is to put it plainly hogwash. Strictly (or philosophivally) speaking we almost never know "all variables" on anything. There needs to be some common sense here. As long as there is no additional controversy ca 1960 with the Daily Reference is fine. Now if I understand that correctly the "controversy" arises from some accounts claing different dates. However something doesn't become really "controversial" because other editor is makimg claim. It becomes "controversial" when some other editor makes a claim that comes with some evidence/sources to back it up rather than somewhat arbitrary speculations (maybe the dail express got the date from WP, maybe thompson asked the daily express to beautify her age, maybe Thompson was lying to the daily express, etc.). You can always make up some (pseudo)rational explanations why some source might have gotten it wrong, if we'd allow such idle speculations to render sources invalid, we'd hardly be able to use any sources at all.--Kmhkmh (talk) 20:57, 3 April 2011 (UTC)
I agree that it's fine to say circa 1960. SlimVirgin TALK|CONTRIBS 21:14, 3 April 2011 (UTC)
Agree, if the Age 50 source is reliable and not rounding, then "circa 1960" works. --SarekOfVulcan (talk) 14:58, 4 April 2011 (UTC)

Editing policy for this policy page

A general question. When an editor makes a change of any kind to this policy page, and another editor reverts it because it isn't an improvement, in that editor's opinion, who has the burden of gaining consensus on the talk page for their action, the reverter or the editor who made the initial change? Can the initial editor or someone else simply re-revert the change back into policy, claiming that it is not a change in meaning so the change must be allowed on that basis? Is it the understanding here that any new change in wording that is not a change in policy or meaning, can only be reverted by consensus? In other words, does restoring the previous wording require consensus? (talk) 12:15, 4 April 2011 (UTC)

No, changing the original wording requires consensus, but people ought to go into the process in good faith - too often on these key policy pages people oppose and revert any change as a knee-jerk reaction, without really thinking or explaining what their objection is.--Kotniski (talk) 12:26, 4 April 2011 (UTC)
When making edits to policy pages that change meaning, best practice is to discuss first. If you make the edit first and you're reverted, then start a discussion, but don't keep reverting back. If you can't gain consensus on this page, open a neutrally worded RfC. Even if you can get consensus on this page, if the change is a substantive one it's better to open an RfC anyway to bring in fresh eyes.
As for copy edits that don't change meaning, it's okay to go ahead and make those so long as you're sure you're improving the writing. But again, if there's an objection to the change it's best not to keep reverting. The default with policy pages has to be a degree of stability because lots of people rely on them. SlimVirgin TALK|CONTRIBS 12:30, 4 April 2011 (UTC)
SlimVirgin is correct... The editor who makes the initial change should not re-revert. Doing so is the start of edit warring. So... I suppose you could say that the burden to start a discussion does fall on the person who desires to change the policy. However, once that discussion has started is up to both editors to engage in discussion, and to consult with other editors so that a consensus can be determined.
The changing editor must realize that while his/her intent may not have been to the change the meaning of the policy, others may disagree. If there is any question, we default to the previous version and discuss. Even small shifts in wording that we agree do not change the meaning can have unintended results. Sadly there are a lot of Wiki-lawyers out there who parse our words and give them meanings we never intended. There are POV pushers who will exploit any loop-hole we give them. So we do need to be conservative about changing policy pages. That does not mean changes are impossible... but it does mean that even small changes can require a lot of discussion before a consensus as to whether to accept them forms. Blueboar (talk) 13:11, 4 April 2011 (UTC)
This is why experience with policy editing matters, because the more you edit them, the more you're likely to recognize the threads that join them, and how an apparently innocuous tweak can have consequences. SlimVirgin TALK|CONTRIBS 16:04, 4 April 2011 (UTC)
There is the other side of the coin, of course - since there are people out there giving our words meanings we never intended (sometimes in perfectly good faith, of course - how are they supposed to know what we intended if we don't write clearly?), it's quite important to to plug the holes as we find them - preferring the version that makes the intended meaning most clear, regardless of whether it happens to be the "new" or the "old" version. So too much conservatism is harmful too. (As you may have noticed, I think that on these particular "core policy" pages, there is most definitely too much conservatism.) --Kotniski (talk) 13:28, 4 April 2011 (UTC)
Yes, I would agree that we can be too conservative at times... there are editors who will reject a change simply because it is a change. My point was that instead of getting into revert wars, we need to fully discuss. We need to fully examine the change and think about how it will be interpreted, misinterpreted, and abused. And, sometimes, that discussion can take a long time. Editors on both sides need to be patient. A consensus will form... eventually. Blueboar (talk) 13:50, 4 April 2011 (UTC)
<hollow laugh /> Yes, and when it does, and you try to put it into effect, the resident page-owner will STILL revert you, and the process has to start all over again... --Kotniski (talk) 14:37, 4 April 2011 (UTC)
All part of the process. Perhaps it was not yet time to try to put the change into effect; that a deep enough consensus had not yet been reached. I know it can be frustrating to be patient and to sit through what can seem to be endless discussions... but that is how it works. Blueboar (talk) 14:56, 4 April 2011 (UTC)
No, it's how it fails to work. Consensus should not require unanimity; yet if we all sit by and say "ah, that's just how it is", then one intransigent editor who has learnt to apply the "undo" button without feeling shame can effectively ensure that nothing gets done without her say-so. Result: endless rehashing of the same discussion points, policies remain in the same poor overall state, editors who could improve them are driven away from the pages as they feel it's not worth the effort. And Wikipedia is left with a vastly bloated mass of largely incomprehensible or inaccurate "policies" which misinform as to what our practices actually are, and help to deter potential new editors from joining or staying with the project.--Kotniski (talk) 15:23, 4 April 2011 (UTC)
You are correct, consensus should not (and does not) require unanimity... but on the other hand, it should (and does) require an indication that more than small group of editors support the change. Policy needs to have broad support from "the community" (ie a large group of people). Unfortunately, the "community" does not pay close attention to the day by day arguments on policy talk pages. Where it gets most frustrating for everyone is when a consensus is emerging, but is not yet completely clear: when several editors are expressing support for a change, but there are still some hold-outs who either object to it or are not yet certain. Both sides tend to claim "consensus" (either for or against the change) when in fact no consensus has (yet) been reached. Blueboar (talk) 15:37, 4 April 2011 (UTC)
That's just another way of saying that consensus does (in your view) require unanimity.--Kotniski (talk) 16:04, 4 April 2011 (UTC)
Not unanimity, but it has to be strong, and more importantly it has to be informed. If 20 new editors turn up to insist on a change that will help them push their POV, they can't be allowed to drown out two established editors resisting that change. So it's almost never about numbers alone. SlimVirgin TALK|CONTRIBS 16:07, 4 April 2011 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────An example of what I was asking just occurred on the policy page. SlimVirgin changed some wording, and I reverted it.[4] Then SlimVirgin reverted my reversion.[5]

Perhaps SlimVirgin could explain how her action is consistent with what she wrote in her previous comment here,

"As for copy edits that don't change meaning, it's okay to go ahead and make those so long as you're sure you're improving the writing. But again, if there's an objection to the change it's best not to keep reverting. The default with policy pages has to be a degree of stability because lots of people rely on them." (talk) 15:58, 4 April 2011 (UTC)

The words you removed were already in the lead, 75. So I've restored them, just in a different position to avoid repetition. SlimVirgin TALK|CONTRIBS 16:01, 4 April 2011 (UTC)
If there's an objection to the change it's best not to keep reverting. Apparently objected to the change. It's important for established editors to model the type of editing we wish other editors would use on pages like this. — Carl (CBM · talk) 16:05, 4 April 2011 (UTC)
He apparently didn't read what he was editing. The sentence he thinks is new was already there. SlimVirgin TALK|CONTRIBS 16:09, 4 April 2011 (UTC)
Did you raise that on the talkpage and get a consensus? According to Blueboar above, "The editor who makes the initial change should not re-revert. Doing so is the start of edit warring." According to you, "If you make the edit first and you're reverted, then start a discussion, but don't keep reverting back." But you violated that very principle today [6] [7]. I'm pointing this out because, in the event that an RFC were started in the future, it's necessary to have diffs where the problem was openly stated. If you wish other people would not re-insert their own unilateral changes to this policy page, how can you re-insert your own unilateral changes? — Carl (CBM · talk) 16:14, 4 April 2011 (UTC)
That edit stems back to the edit to the lead suggested months ago by you, Carl, and yes it got consensus from you and several others. You hailed it as a great improvement. But I included inadvertent repetition in it, which I've just removed. SlimVirgin TALK|CONTRIBS 16:16, 4 April 2011 (UTC)
That's the sort of thing you could point out in a talk page thread. How is supposed to realize your motivations if you simply re-revert your own edit? It has a bad appearance, and it leads to complaints like the ones here when other people feel there are inequities. Having pointed that out, I am now going to drop this thread for the time being. — Carl (CBM · talk) 16:20, 4 April 2011 (UTC)
I did point it out, Carl, and you're aware of it too, because you were the one who initiated the change. So you could also have explained it to the anon. SlimVirgin TALK|CONTRIBS 16:30, 4 April 2011 (UTC)
The point is that when anyone else does what you did (make an edit, see it reverted, and then just do it again), you start shouting at us for being male aggressors and so on. You seem to apply different standards to yourself than you demand of others.--Kotniski (talk) 16:38, 4 April 2011 (UTC)
Please, everyone, focus on the content. If anyone has a serious and credible complaint about the conduct of an editor, the etiquette noticeboard would be the place to bring it up. --Nuujinn (talk) 20:50, 4 April 2011 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────Note that the topic of this section pertains to the editing process, not the content of the example edit. The issue wasn't whether the content of the example edit was right or wrong. The issue was whether or not reverting a revert is proper without discussion on the Talk page. In general, editors here seem to have the opinion that there should be discussion on the Talk page, instead of a revert of a revert. But that didn't occur in actual practice with the example, because the editor was sure that the edit was an improvement. I certainly don't think it is and I gave my reason below.

So is the lesson here, that when an editor makes a change to this policy page, and it is reverted, the editor is justified in reverting again, instead of going to the Talk page for discussion, if the editor is sure that the change is an improvement? (talk) 05:45, 6 April 2011 (UTC)


Since it is possible to slant presentation of information on any subject, by selectively removing some contributed information obtained from verifiable cited reliable secondary and tertiary sources, which other editors or readers reject (for whatever reasons they have), something should be said about producing a "Synthesis-by-omission", i.e. removing or excluding information for the purpose of producing a particular conclusion that could misinform the reader. (For example, removing or excluding information from an article on Abraham Lincoln relevant to Lincoln's opinions on slavery before 1860, or from an article on George Washington, or on Benjamin Franklin, relevant to their particular opinions on what some called the "necessity of slavery" to the economic stability of the new nation among the community of nations on earth, in order to avoid or promote or reach a conclusion about these men's views and policies that would be false.) I am convinced that the policy articles should explicitly say something about avoiding a "Synthesis-by-omission", and should point out that merely saying "this is not relevant" is not good enough reason for a reverting or excluding edit. --Michael Paul Heart (talk) 17:11, 9 April 2011 (UTC)

That's more of an NPOV issue, Michael, where it says we should represent the balance of what the sources say, rather than cherry-picking from one source, or from the literature in general. SlimVirgin TALK|CONTRIBS 18:17, 9 April 2011 (UTC)
I concur with your reasonable NPOV issue assessment. I really do believe that in some way the issue of "Synthesis-by-omission"/"Synthesis-by-selective-exclusion" editting (sometimes it could be a form of vandalism) should be explicitly addressed and expressed in policy statements by the community under that heading. It is presently addressed in an indirect way. And I do acknowledge that. Thank you for your response. --Michael Paul Heart (talk) 03:06, 10 April 2011 (UTC)

Original research and patent rubbish

Is it original research to point out patent rubbish? The example that I have in mind invovles the digging of a road tunnel. The tunnel is 1830 m from end to end. One of the references states that two tunnelling machines, one at each end, moved forward at one metre a day. Another reference in the same article states that tunnelling started in February 2008 and was completed in February 2009. Is it OR to note the contradiction that each machine would have had to dig 915 metres, but that the digging period was about 365 days? Martinvl (talk) 19:37, 9 April 2011 (UTC)

Yes, it would be OR to note the contradiction (without a source that does so). The better approach is to look for additional sources that would allow you to resolve the contradiction. Perhaps another source will indicate that the pace of tunnelling was faster than the first source says... or different start-end dates... or a different length for the tunnel. Blueboar (talk) 15:02, 10 April 2011 (UTC)
I don't think our technical terms are very helpful in such situations. This kind of thing is tricky. We should not normally say that there is a contradiction unless some reliable source has remarked on it. But it's a red flag and as such cause to look very closely at the available sources and then to possibly supress one of the statements, or the other, or both, as a matter of editorial judgement if the contradiction cannot be explained away. Hans Adler 16:46, 10 April 2011 (UTC)


An edit was recently made to this policy page.[8][9] Part of the edit was not an improvement, in my opinion. The part of the edit in question concerns the last part of the following sentence as it appeared on the policy page before the edit, and I have underlined that part for discussion.

The term "original research" (OR) is used on Wikipedia to refer to material—such as facts, allegations, ideas, and stories—not already published by reliable sources.

The last part was changed to the following (also underlined here for discussion).

The term "original research" (OR) is used on Wikipedia to refer to material—such as facts, allegations, ideas, and stories—for which no reliable source exists.

I believe the original is superior because it explicitly includes the “published” aspect which is important. For example, an editor could personally interview a reliable source, such as the spokesperson for some major corporation, and if unpublished that interview would be considered “original research” for the purposes of this policy. That idea is no longer expressed in the sentence. (talk) 23:17, 4 April 2011 (UTC)

I agree. The version with the word "published" was better. On closer look... the "no reliable source exists" verison is better at that point... it is explained further later (where it talks about the need for being published) Blueboar (talk) 23:31, 4 April 2011 (UTC)
I think they're both wrong. When people were defending having OR and V as separate pages, the claim was that something can fail to be original even if it's not supported by RS (taking the word "original" literally, presumably - if something's supported only by unreliable sources, then it's not properly verifiable, but it can hardly be called original, since the unreliable sources have precedence). So again, we must make up our tiny minds - if original research is to be defined as anything that isn't supportable by reliable sources, then "no original research" really does reduce to exactly the same thing as "verifiability". But if we want to maintain NOR as a separate topic, we must define OR in its own terms - as ideas that don't come from any source.--Kotniski (talk) 07:16, 5 April 2011 (UTC)
Blueboar, The problem with that is, using the previous example, the editor who interviewed the spokesperson could come to WP:NOR to check if the interview is OR, come across the sentence that has the part, "...for which no reliable source exists", and conclude that the info in his interview is not OR because a reliable source exists for it, and go back to editing. (talk) 11:50, 5 April 2011 (UTC)
I understand your concern 75... but "reliable source" is a defined term on Wikipedia, one that requires the source to be published. When we say "no reliable source exists" we mean "no published reliable source exists" (if it isn't published, we don't consider it reliable). So in your example... Since the interview is unpublished it is by definition unreliable.
Kotniski's concern is a more valid (and more subtle) point... It is possible to find interpretations, analysis, or conclusions in unreliable (but published) sources. Summarizing these in an article would not violate NOR (because it does not originate with the editor who is summarizing). That said, the fact that there is no NOR violation does not mean that there isn't another policy or guideline being violated. Passing NOR does not necessarily make the material acceptable. Blueboar (talk) 13:49, 5 April 2011 (UTC)
(edit conflict)Your comment, "I understand your concern 75... but "reliable source" is a defined term on Wikipedia, one that requires the source to be published."— I appreciate that you're trying, but I think you're looking at the sentence through the eyes of an editor, such as yourself, who already knows the ins and outs of policy. The original version of the phrase contains the same number of words as the newer version, so I don't understand why the objection to the original version that explicitly mentions published, up front, where it can do the most good. Anyhow, I don't think there's going to be any progress towards agreement between us on this matter, so I'll leave the discussion to any other editors who might want to continue. (talk) 15:29, 5 April 2011 (UTC)
Supplying a non-reliable source (within our definition), when that's all there is, amounts to OR, as a matter of definition. There's no point trying to force our meaning of OR into everyone else's. What we mean by it is that no reliable source exists for it. The policy inherited the term from various discussions, and now that we have it we can't change it, because people are very attached to it. So we simply define it. We could just as easily say, "Blah is adding material for which no reliable source exists. Don't do blah." SlimVirgin TALK|CONTRIBS 15:23, 5 April 2011 (UTC)
So by this definition, "OR" and (breaches of) V seem to be exactly the same thing. So why bother with two pages? (Not so much a question to Slim, but to anyone who thinks that the two pages ought to be kept separate.) --Kotniski (talk) 17:20, 5 April 2011 (UTC)

(I also suggest that we shouldn't lazily carry on using words in misleading, jargonny ways just because insiders have become attached to them. What we call "original research" should correspond as far as possible to what people in the outside world would understand by that term - otherwise we build up further unnecessary barriers to communication, particularly with new editors.)--Kotniski (talk) 17:33, 5 April 2011 (UTC)

I disagree with that. Jargon is not inherently misleading, and part of the process here is to define terms to facilitate development of consensus. We need some anchors, and the terms used serve that purpose. People do get wound up in the jargon, especially if they only edit policy pages, but it serves a purpose. That we could do a better job with new editors is a valid concern, but being as precise as one can be is not lazy.
Well, if we're defining completely new concepts, then we need new terms for them, sure. But I'm not sure that Wikipedia's concept of "original research" is a completely new one; and even if to some extent it is (probably no-one's gone to the same effort to define it as we have; at least, not in the sense of something bad) the term(s) we use for it shouldn't be at odds with the way those terms are used in natural language. Calling it "blah" (or "non-att" or something) would in some ways be better, since at least then people will know it's jargon. But calling it with a phrase that already has a different meaning in the English language is misleading, and quite at odds with your stated goal of "being as precise as one can be" - this practice leads precisely to imprecision, confusion, and indeed laziness (people don't bother to say what they really mean, they just write "OR" and expect readers to guess).--Kotniski (talk) 08:01, 6 April 2011 (UTC)
The phrase "original research" does not have a precise meaning in standard english. I've been a student/instructor/staff member at a uni since 1979, and I can honestly say I have never heard the phrase used in an academic setting ("do your own research" is the usual phrase). There's nothing wrong with a term being defined in a particular way in a particular context. We don't ask people to guess, generally we point users unfamiliar with our use of terms to the appropriate page where the usage is explained. And I just plain disagree with your assertion regarding "precision, confusion, and indeed laziness", and ask that you please assume good faith on the part of other editors, rather than assume the worst of them. --Nuujinn (talk) 11:49, 6 April 2011 (UTC)
?? I'm not saying anyone is acting in bad faith. But even if it were true that the phrase "original research" were never used in an academic setting (in fact Google tells us otherwise), the words "original" and "research" are still standard English, and readers would be justified in assuming, on being told that something is original research, that what was meant was research that's original. --Kotniski (talk) 12:38, 6 April 2011 (UTC)
If I had been involved in the discussions that led to the creation of this policy page, I would have suggested "No Original Interpretation or Analysis" (NOIA) as the title of this policy... as being a more accurate description of the policy's salient points. But this is all spilt milk under the bridge after the horse has bolted. "Original research" is what we call the concept. "NOR" may be Wiki-jargon, but we have used it long enough that it would be disruptive to change it at this late date. So the question is: Do we explain what we mean by this bit of Wiki-jargon clearly enough? If we can avoid using other bits of Wiki-jargon in explaining it, I think it would help. Blueboar (talk) 13:25, 6 April 2011 (UTC)
Maybe so, although I'm not so sure everyone on Wikipedia understands Original Research to mean what we define it as here (rather counterintuitively) on this page. (Last time I asked people what they thought the difference between NOR and V was, I recall one response along the lines of "if it's just something my professor told me, that's not original research, yet it doesn't satisfy verifiability".) But if we do accept this definition as a done deal, then what's the remaining difference between this concept and that of V?--Kotniski (talk) 13:51, 6 April 2011 (UTC)
Blueboar, Except for the NOIA part, which could still be criticised in the same way as the term NOR, I wholeheartedly agree with the rest of your statement. (talk) 14:17, 6 April 2011 (UTC)
75, are disagreeing with the idea that we should not include our own original interpretations or analysis? Or just my choice of words for the title of the policy? Blueboar (talk) 15:40, 6 April 2011 (UTC)
I agree with Blueboar, and will also point out that the google hits on "original research" are mostly from the UK. So it's apparently used there, but no so much elsewhere. And I see no advantage to trying to merge OR and V, since they are distinct topics (at least they seem so to me). I can engage in OR with verifiable sources.... --Nuujinn (talk) 15:55, 6 April 2011 (UTC)
In my view, it is regretable that Her Majesty's Government does not see fit to publish information that most other civilised countries would publish. This might explain why a certain dergee of what appears to be OR is done in the UK. For example, driver's dates of birth are encoded into their driving licence number. About thirty years ago, a driving licence was printed onto a sheet of A4 paper and the date of birth was printed in un-encoded form on the bottom right-hand corner. The driver could cut that corner off if they wished their date of birth to be secret - or so they thought - it was jumbled up in their licence number, but the decodng was simple enough that a policeman could do it without having to call for help. In these circumstances it was easy enough to reverse-engineer the encoding by comparing the licences of half a dozen drivers. This raises the questions - "Would this reverse-engineering be OR? - after all the information has been published - every police station has a copy of it. How should I have gone about getting that information? Martinvl (talk) 16:12, 6 April 2011 (UTC)
Yes, it would be OR. Also, I would suggest that if that fact is not covered in reliable secondary sources such as journal, magazine, or newspaper articles, we have no need for it in any article. It's an interesting fact, but not everything interesting belongs on WP. I know, for example, why a particular uni has the id number as a bar code. Interesting story, but not worth including here. Now, if I can convince Time magazine to interview me, maybe I can get that datum published. --Nuujinn (talk) 20:43, 6 April 2011 (UTC)
You say above you think OR and V are distinct topics, and that you can "engage in OR with verifiable sources". Can you explain what you have in mind?--Kotniski (talk) 10:09, 7 April 2011 (UTC)
I'm confused by the first half of this conversation. Both versions require "reliable sources". Neither version permits unreliable sources.
The only difference is whether we're specifically saying that an unpublished "reliable source" is insufficient. For example:
  • Is a hand-written, never-published letter from an eyewitness sufficient to avoid NOR violations? If I cite a letter that I found in the closed archive of the local history museum, is that a NOR violation? It exists, but it is unpublished.
  • Is an unpublished manuscript for a book sufficient? If I'm describing the plot of a new novel from the pre-publication manuscript, is that a NOR violation? The book exists, but it is unpublished. Isn't that a little too "Proof by forward reference" for NOR?
If publication matters, shouldn't we say so? WhatamIdoing (talk) 21:45, 6 April 2011 (UTC)
I'm not sure it does, really - at least, not so much that it needs emphasizing (to the extent that publication is required, it's already included within the meaning of "reliable sources" - along with all the other features we expect reliable sources to have). For me the problem (apart from the overriding one that we are reinventing the English language by giving the ordinary word "original" a new meaning) is - the same as where a similar phrase appears in WP:V - that the actual "material" does not need to have been published. It's better in this case than at V, since here we define a narrower sense of "material", but it's still not completely accurate - the actual "facts" we write do not always need to have been published; they need to be supportable by something that's been published (i.e. we should say we have a reliable source "for" the facts, not that the exact same facts appear in a reliable source, which can fail in the cases of (i) trivial deductions (ii) reporting of opinions/content of primary sources).--Kotniski (talk) 10:18, 7 April 2011 (UTC)
Nowhere in WP:NOR is it said that "reliable source" means the same thing as "published reliable source". So without going to the wikilink, the reader wouldn't know that important point from just reading WP:NOR. Furthermore, including "published" when mentioning reliable sources is done over and over elsewhere in WP:NOR, suggesting that editors of this policy page feel it is needed, and that the term "reliable source" alone is not sufficient for clearly expressing the idea that it must be published. The recent edit that excluded the important word "published" from the second sentence of the lead was incorrect and "published" should be restored there. WhatamIdoing is perfectly correct. Also to everyone, please stay on the topic of this section. (talk) 12:16, 7 April 2011 (UTC)
Would the hybrid version "for which no published reliable source exists" satisfy you?--Kotniski (talk) 12:44, 7 April 2011 (UTC)
It says the same thing as the version before the recent change and is OK in that respect. However, for the purpose of stability of this policy page, the previous version should be restored. One shouldn't change this policy page by whim. (talk) 12:54, 7 April 2011 (UTC)
It was not changed on a whim... SV has told us that she changed it to avoid what she thought was unnecessary repetition. We can debate whether her change was for the best or not, but let's at least acknowledge her intent. Blueboar (talk) 13:19, 7 April 2011 (UTC)
Adding to my previous messages, the wikilink reliable sources does not even link to a place that says "reliable source" is the same as "published reliable source". That place at WP:V says, "All quotations and any material challenged or likely to be challenged must be attributed to a reliable published source using an inline citation." There too the editors felt that it was necessary to explicitly mention "published". (talk) 14:54, 7 April 2011 (UTC)
Interesting... the link in question used to point to WP:IRS... the first sentence of which is:
  • "Wikipedia articles should be based on reliable, published sources, making sure that all majority and significant minority views that have appeared in reliable, published sources are covered (see Wikipedia:Neutral point of view)." <note... bolding changed so as to be relevant to this discussion>
I have added a statement to WP:V making it clear that our sources do need to be published to be considered reliable. Blueboar (talk) 15:30, 7 April 2011 (UTC)
(edit conflict)I think that additional "published" are definitely an improvement. The WP content must be verifiable, which means that everyone can theoretically be able to find the cited source by themselves and to compare the WP text with what the source says. If the source has not been published, such an opportunity does not exists even in theory. I support this proposal.--Paul Siebert (talk) 16:02, 7 April 2011 (UTC)
You might want to reread the quote in your last message since the editors of the quote found it necessary to use "published" in two instances of the phrase "reliable, published sources", rather than using "reliable sources" alone, as well as later in that policy page. Your quote only supports the point made in my previous message, "There too the editors felt that it was necessary to explicitly mention 'published'." (talk) 15:59, 7 April 2011 (UTC)
P.S. So far there isn't a consensus for the recent change[10][11] that removed "published" from the second sentence of the policy page. (talk) 16:11, 7 April 2011 (UTC)

Update: The editor that removed "published" has just restored it to the sentence.[12] With that, there is no remaining given reason for changing the wording of the rest of the phrase. Furthermore, the previous wording is more concise than what is presently there, and it is best not to change wording without good reason, so I am restoring the original wording. (talk) 17:00, 7 April 2011 (UTC)

Well, I've given a reason (I don't know if it was the same reason that editor had in mind) - material in Wikipedia may well not have "been published" before; what we mean is that it must be supportable by something that has been published - i.e. the requirement is that there be a reliable source "for" the material, not that the material itself literally be in a reliable source. (All seems obvious to us - but remember this is giong to be read by people for whom none of these ideas will be obvious.)--Kotniski (talk) 17:11, 7 April 2011 (UTC)
I think your point pertains to the other part of the sentence in both versions, rather than the phrase at issue. (talk) 17:21, 7 April 2011 (UTC)
No, I don't think so - aren't we discussing whether to say "for which no reliable published source exists" or "not already published by reliable sources"? I certainly prefer the first, not for reasons of phrasing, but because what it actually says is closer to the truth.--Kotniski (talk) 17:25, 7 April 2011 (UTC)

(edit conflict)Update: Blueboar has reverted my restoration of the rest of the original phrase.[13] The reason in his edit summary for changing the original wording, "It's simply a matter of sentence flow... but I prefer this version" is not true regarding flow, in my opinion, and does not observe the principle suggested in my last update message that wording should not be changed without good reason. (talk) 17:37, 7 April 2011 (UTC)

75, Improving sentence flow is a very good reason for a change. You are absolutely entitled to have an opinion as to whether SV's wording does improve the sentence flow (or not)... but please don't claim that my revert was "without good reason". Blueboar (talk) 18:11, 7 April 2011 (UTC)
It wasn't a good reason because you didn't explain why you thought it was good flow, when you could have explained it in this section. One could carry your reasoning even farther by simply saying it was changed because the new version is better. Using your reasoning, one could say "better" is a good reason, but I don't think that just saying "better" without explanation is a good reason. However, I'm interested in why you thought it was better flow, especially since there is a phrase before it, set off by dashes, that breaks the flow for both versions. (talk) 19:21, 7 April 2011 (UTC)
Kotniski, Your comments may or may not be true that "not already published by reliable sources" is more easily misunderstood as meaning verbatim than "for which no reliable published source exists". In my opinion, it's not clear when one considers the context of the rest of the sentence. Fortunately, the important issue about "published" has been corrected. This subsidiary issue is not so clear. If this was an article instead of a policy page, I wouldn't care whether it was one way or another. But I think we should strive for stability and I don't agree that it's worth making the change that you, SlimVirgin and Blueboar prefer, although I have to defer to consensus. (talk) 17:59, 7 April 2011 (UTC)
Whatever the merits of "stability", it must certainly take a lower priority than telling the truth. We all seem to have different reasons for preferring the "for which" wording, but mine is that it is actually incorrect to say that "all...facts...must already have been published..." Sometimes we give facts which don't themselves appear in the source, but which result from our examination of the source, or are uncontroversially deducible from the source.--Kotniski (talk) 11:03, 8 April 2011 (UTC)
Aside from whether your point is right or wrong, how does it apply to one version and not the other?
Here's the actual parts that you are referring to, and they mean the same thing, IMO.
such as facts, allegations, ideas, and stories—not already published by reliable sources.
such as facts, allegations, ideas, and stories—for which no published reliable source exists. (talk) 14:44, 8 April 2011 (UTC)
A published reliable source can exist "for" a fact without that fact having been "published by" reliable sources. Simplest examples: (a) source says 5 miles, we say 8 km; (b) source (Smith) says P, we say "Smith says P". Here the sources support our statements without containing them, thus it's not true (and I've seen this be a genuine source of confusion) that our statements must already have been published by reliable sources.--Kotniski (talk) 10:44, 9 April 2011 (UTC)
You are suggesting that the phrase "not already published in a reliable source" is associated with verbatim but the phrase "no published reliable source exists" isn't. Is that basically your point? I don't see that either necessarily implies verbatim. The meaning of both phrases is the same. (talk) 14:22, 9 April 2011 (UTC)
Not verbatim necessarily (though someone might indeed interpret the first sentence as having such a meaning). But to me the first sentence implies that the fact we state must itself appear in the source, while the second version implies it need only be supportable by the source. With my first example I suppose you might claim that "T is 5 miles north of U" is the "same fact" as "T is 8 km north of U" (though readers might not share this interpretation), but with the second example, "Smith says P" is most definitely not the same fact as "P" (P might not even be a fact).--Kotniski (talk) 14:45, 9 April 2011 (UTC)
Please note for example, that a "fact"[14] is not the same as a "statement of the fact". Both phrases that we have been discussing are in a sentence that is referring to facts, rather than statements of the facts. Does that help at all? (talk) 18:00, 9 April 2011 (UTC)
I don't see how it changes anything. "P" and "X says P" are not just different statements of a fact - they're different facts (if they are facts in the first place).--Kotniski (talk) 19:05, 10 April 2011 (UTC)
────────────────────────────────────────────────────────────────────────────────────────────────────Yes, "P" and "X says P" are different facts. But that is not what I was referring to.
Consider the example of an object that is 12 inches, i.e. 1 foot, long. This is a fact about the length of the object. A statement of this fact is, "The object is 12 inches long." Another statement of this fact is, "The object is 1 foot long." These are two different statements of the same fact about the length of the object.
From the same first part of both sentences that we have been discussing, the subject of both sentences pertains to facts, not statements of facts. The phrase "not already published in reliable sources" is about the "fact" not being published, rather than a particular "statement of the fact" not being published. Similarly, the phrase "for which no published reliable source exists" is about the "fact" not being in an existing published reliable source, rather than a particular "statement of the fact" not being in an existing published reliable source. The two phrases mean the same thing.
I don't think I can make it much clearer than that, and I also suspect that my explanation isn't sufficient for you. So this may be my last message on this topic. (talk) 21:06, 10 April 2011 (UTC)
If you're making a point about logic, there is only "statements about facts", i.e. propositions: "Object X is 12 inches long," "object X is a foot long," etc.
If you're making an editorial point, namely that we need a source that supports the material, but not necessarily one that includes the exact words we want to use, yes, you're right. SlimVirgin TALK|CONTRIBS 23:29, 10 April 2011 (UTC)
(to 75:) The explanation remains insufficient because you haven't addressed the second (and more clear-cut example), where the statement in the source (written by X) is "P", but the statement we make (for one reason or another) is "X says P". In this case the difference between the published words and our words is not just that we've reworded a particular fact - we've stated a fact which is supported by the source but not contained in it.--Kotniski (talk) 06:46, 11 April 2011 (UTC)
Is the fact that "X says P" in the source? If not, you haven't shown that there exists a published reliable source for the fact that "X says P".
Or are you saying that X is the source? In that case, "X says P" is simply an in-text attribution to the source that is allowed by both phrases, just like inline citations are allowed by both phrases. (talk) 13:01, 11 April 2011 (UTC)
I mean X is the (author of the) source, the source says "P", and Wikipedia says "X says P". The first phrasing would imply that this isn't allowed (since the fact "X says P" has not been published by a reliable source), whereas we know that of course this is allowed (as you say, it's simply in-text attribution), hence the first phrasing is wrong. The second phrasing seems OK, since "source for (something)" is a vaguer expression which doesn't require that the source actually contain that something.--Kotniski (talk) 13:23, 11 April 2011 (UTC)
Whether you call X the work that is a source or X the author of the work, "X says P" is an in-text attribution that is allowed by both phrases, just like inline citations are allowed by both phrases. The phrase "not already published in reliable sources" means the same thing as "for which no published reliable source exists" because they are both referring to a fact and because they both allow in-text attribution for a fact P from a source X, either by specifying the work or the author of the work, e.g. "X says P". (talk) 13:52, 11 April 2011 (UTC)
All right, I think we're inaccessible to each other's light, and since the matter seems to have been settled on different grounds, we may as well stop discussing it (until sometime when I decide to bring up the issue at WP:V, which has a similarly problematic sentence...)--Kotniski (talk) 14:13, 11 April 2011 (UTC)
Just IMO, but I prefer "for which no reliable published source exists" to the original wording. In particular, I think the "exists" language correctly emphasizes the fact that NOR violations depend on whether a proper source exists (anywhere in the world), not merely whether the source is currently cited in the Wikipedia article (a significant misunderstanding by some less-experienced editors). WhatamIdoing (talk) 18:25, 7 April 2011 (UTC)
Although I disagree because I think the basic idea of your remarks should apply to both versions, with appropriate rewording, I found your reference to "the 'exists' language" useful. Since "exist" is used in the second paragraph on the policy page, use of it in the sentence under discussion would tie them together. On the other hand, someone may feel that it is better style not use the same word in both places. Either way seems alright, and if there is a difference in worth, it's very small, IMO. (talk) 19:37, 7 April 2011 (UTC)

Do original images require a source?

A dispute over the use of an image as an illustration at Fan service is leading some editors to argue whether any original image requires a source to support an image's use as an illustration. The outcome of this discussion could have a substantial affect on all other original images. —Farix (t | c) 16:30, 10 April 2011 (UTC)

The original image doesn't require a reliable source and none can be given since it is the product of an editor. However, the image may not be allowed.
"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."[15]
Whether or not it illustrates or introduces unpublished ideas or arguments may need to be determined on the Talk page. (talk) 17:39, 10 April 2011 (UTC)
Also... while the original image may be allowable, very often it's the caption accompanying the image that introduces the OR. Blueboar (talk) 17:46, 10 April 2011 (UTC)
Where an image is contentious, in the sense that editors dispute what it represents, you would need a reliable source saying "this is an image of X," or words to that effect. SlimVirgin TALK|CONTRIBS 20:13, 10 April 2011 (UTC)
How would one source that an original image is X without taking the creator's own comments? Or is the creator's comments sufficient enough of a source per WP:AGF? —Farix (t | c) 00:14, 11 April 2011 (UTC)
It depends on the image. If I have a source that says "Barak Obama often wears neckties", I can use my own original photograph of Obama wearing a necktie (captioned: "Obama wearing a necktie") to illustrate that sourced comment. The image and caption are not OR since it clearly illustrates a sourced fact. Blueboar (talk) 00:36, 11 April 2011 (UTC)
(ec) If there's no dispute, or where no reasonable person could deny that it's an image of X, then we can take the creator's word for it; or rather, we don't need the creator's word. But where there's a good-faith dispute about whether the image is an example of X (and people are being reasonable and not DICKish), then we need a reliable source that discusses the image as an example of X, or in terms of X. SlimVirgin TALK|CONTRIBS 00:37, 11 April 2011 (UTC)
The claim is that the image's creator is not an expert in X, therefore he is not a reliable source that the image they created is an example of X. But this is just the latest tangent used in order to have the image remove. Previous arrangements have involved WP:SELF, self-promotion, child porn, "cheesecake", drives off female editors, and etc. —Farix (t | c) 01:05, 11 April 2011 (UTC)
That's not the claim (that the image's creator is not an expert), the issue is that there isn't a reliable source linking the image as an illustration of the articles topic. It's a small but important difference. RxS (talk) 02:38, 11 April 2011 (UTC)
No, I've already show were the image's creator states that the image is fan service (which is the subject of the article the image is used on).[16] But you claim that additional sources are required because the creator is not an expert. —Farix (t | c) 02:47, 11 April 2011 (UTC)
You can't just proclaim yourself an expert on so[mething without reliable, published sources backing your claim up. RxS (talk) 02:55, 11 April 2011 (UTC)
One doesn't need to be an expert in X in order to create an illustration of X, especially when X is a simple topic. —Farix (t | c) 11:15, 11 April 2011 (UTC)
I think that Wikipedia:Images#Pertinence_and_encyclopedic_nature is what you want to be looking at. If editors can agree that the image in question looks like (not "is", but "looks like") what the well-sourced article is describing, then the image is acceptable as far as WP:NOR/WP:V/WP:RS are concerned (you might still reject it for other reasons). And if it doesn't "look like" the thing, then it's not an appropriate image choice (even if it is unquestionably authentic), because an image of X that looks like something else entirely won't help the reader understand what X looks like. WhatamIdoing (talk) 00:54, 11 April 2011 (UTC)

For anyone wondering what this is all about, it's whether the image [17] belongs in Fan Service as an illustration of the subject. RxS (talk) 02:41, 11 April 2011 (UTC)

There's also been some who question the image because the character who appears there hasn't appeared in any related media production (that anyone knows about) and is thus claims that since that is the case cannot be used. The others claim that is irrelevant; the creator used the style of art which itself does not appear to be contested (knock on wood) and labeled it as a fan service image to represent what (one type of) fan service looks like.Jinnai 03:20, 11 April 2011 (UTC)
For the purpose of discussion, can anyone give a link to a similar image that is known to be an example of fan service? (talk) 12:38, 11 April 2011 (UTC)
Some examples for the fairly often used "unwilling bathing intruder" scene, which usually is also related to ecchi theme: [18], [19], [20], [21], [22], [23]. --Niabot (talk) 13:04, 11 April 2011 (UTC)
If those are valid examples of fan service, then the question is whether the original image in the Fan Service article illustrates or introduces ideas or arguments that are not in those examples. If it doesn't, then the original image is allowed. (talk) 13:28, 11 April 2011 (UTC)
I should add that my remarks are with respect to WP:NOR and the image may or may not be excluded for other reasons. (talk) 14:34, 11 April 2011 (UTC)
The article states "The typical, but not only, variety of fan service is racy or sexual content, such as nudity or other forms of eye candy. Shower scenes are very common in movies, and in anime of the 1980s and 1990s, while many more recent TV series use trips to onsen (Japanese hot springs) or trips to tropical locales (or in some cases a swimming pool), in order to showcase the characters in bathing suits." An anime image of a girl in a bathing suite (with the caption: "A bathing suit is typical 'fan service'") seems to me to be an acceptable illustration of this text (There might be a better illustration, but the one given does acceptably illustrate the text.) I don't see any OR here.
Now, that said... there is another issue to consider... the text that the image is illustrating is supported by several sources which are tagged as being "self-published" and thus (potentially) unreliable. That question needs to be resolved. If the text is based on unreliable sources, then the text should be removed. And if the text is removed, the image would no longer be illustrating something stated in the text (and thus would also have to be removed). Blueboar (talk) 14:03, 11 April 2011 (UTC)
That problem could be fixed with the info at which is from Anime News Network. (talk) 22:01, 11 April 2011 (UTC)

Anyway I agree with some who think OI needs to be rewritten to be less vague with regards to non-photos.Jinnai 16:21, 12 April 2011 (UTC)

Removal of material using OR

From the lead, "This means that all material added to articles must be attributable to a reliable published source, even if not actually attributed."

This policy mentions OR that adds material to articles. However, an editor might use OR to remove reliably sourced material from an article. In some cases this may also leave behind information that is OR because it is incomplete. This aspect of OR doesn't seem to be clearly mentioned here, since only adding OR is clearly discussed. In my opinion, such a situation is happening at Operation Odyssey Dawn where OR is being used to remove a reliably sourced list of UK forces from the Deployed forces section of the article.[24] See Talk:Operation Odyssey Dawn#UK forces in Operation Odyssey Dawn. The messages beginning with the one at 06:21, 18 April 2011 are where the OR is mostly discussed. (talk) 17:09, 18 April 2011 (UTC)

how long?

If a page is "tagged" - though nothing specific in the article is tagged; how long until the claim of OR is removed? I asked the question on the talk page of the article and have received no answer. The article is listed as: "missing citations or footnotes" & "may contain original research or unverifiable claims" both "Tagged since January 2010." However, there is nothing listed as "needing citation" or listed as "unverified" - who decides when to remove these tags? (talk) 20:08, 18 April 2011 (UTC)

Inline tagging is not required for these problems. Fifelfoo (talk) 22:22, 18 April 2011 (UTC)
There is no time limit on a tag... The point of a tag is to inspire someone to fix the problem. A tag should stay until the problem is fixed... and can be removed once it is fixed.
Looking at the edit history, a lot of work has been done on the article in question since the tags were added. I would say the problems have been fixed (or at least there has been a good faith effort to fix it). It may be that there are still specific problems... individual statements that are unverified or OR (I don't know the topic well enough to say yeah or nay on that), but the article has been improved to the point where a generalized tag is no longer appropriate, and where more specific (in line) tagging or talk page discussion is called for. I have removed the tags. If they are replaced, note who replaced them and ask that editor to be more specific. Blueboar (talk) 23:58, 18 April 2011 (UTC)

Thanks! Free Patriot press (talk) 01:15, 19 April 2011 (UTC)

Well the tag stays as long as it applies. If it does not apply anymore (or never did in the first place), you may remove it. The decision to remove a tag is made by individual editors according to their best judgement or in case of disputes a review/opinion of larger group of editors might be required. In other words tagging is handled like regular editing more or less.--Kmhkmh (talk) 03:46, 19 April 2011 (UTC)

A request for thoughtful opinions

I have a policy question that I believe belongs here, though it also relates to "Verifiability" and "Reliability." I've been searching on WP for an explicit statement one way or the other for several days, without much success. The question concerns the use of information published on a website as the basis for statements in a WP article. (I should perhaps say that I'm a retired librarian/archivist/historian and I've been doing "close research" in both primary & secondary sources for forty-plus years, so I believe I understand the distinctions made in the various policy discussions.) And I'm perfectly aware that mere publication doesn't make a statement "reliable," whether it's on paper or on a website. It comes down to the motives and expertise of the author, case-by-case, and whether assertions are sourced.

In this case, the WP article on William Jennings Gardner has great gaps in it, as well as misinformation and unnecessary assumptions, all of which can be demonstrated and corrected by reference to primary sources (most of which are available online these days). And use of primary sources implies "original research" on WP (though there are exceptions, which I have also made note of). However, there is a published article on Mr. Gardner on the WeRelate website, which includes a good deal of information taken from primary sources and skillfully interpreted by an expert researcher with whom I am well acquainted. Yes, WeRelate is a genealogical site (using MediaWiki, actually), and I'm aware that there's a bias on WP against genealogical research (as well as frequent wince-producing ignorance and misstatements regarding what "genealogy" consists of and the quality of available research). But WeRelate works hard to maintain high standards for what's published on the site, and most of the admins there are professionals. So, in this case at least, the information provided is reliable and entirely verifiable, every substantive statement being sourced on the page. I.e., the WeRelate article is, in my opinion, the equivalent in quality of an academic journal article. And the fact that the information from primary sources has been published there makes the WeRelate article legitimately a "secondary" source -- does it not? That's how I interpret WP policy, anyway. And because there has never been a published-on-paper biography of Gardner (only one chapter in one book, that I'm aware of, plus brief mentions in a couple of others, all of which also will be used as sources), this website is really the only available resource for extended, verifiable information.

I would like to fill out the William Gardner article on WP with reliable, sourced information, but I would prefer to avoid having material I include be summarily deleted because of knee-jerk reactions to its origins. Would a couple of you folks with combat experience care to give me your thoughts on this? --Michael K SmithTalk 14:43, 13 April 2011 (UTC)

I would go ahead and use the primary sources, Michael; see the policy on using primary sources at WP:PSTS. There's no BLP issue here. The Werelate page doesn't actually say much, so there'd be no point using it as a secondary source (plus, it uses WP as a source). It doesn't make sense to use that kind of a site as a source anyway, when you can go straight to the sources they used, if they're good. Just be careful to use high-quality primary sources, and don't interpret them. SlimVirgin TALK|CONTRIBS 21:46, 13 April 2011 (UTC)
Question... if there are no secondary sources on the subject... can we say the subject is notable enough for a stand alone article? Blueboar (talk) 03:17, 14 April 2011 (UTC)
Yes, there would have to be sufficient secondary sources to make the subject notable, so that you're just filling in details with primary sources, not basing the article on them. SlimVirgin TALK|CONTRIBS 04:14, 14 April 2011 (UTC)
I would not go ahead and use the Primary Sources. It very strongly constitutes original research. I would not use the WeRelate page given, as the page does not meet Wikipedia's reliable sourcing criteria. The "United States Department of Veterans Affairs. Nationwide Gravesite Locator. " secondary or tertiary source given on that page and Sally Jenkins. The Real All Americans (New York: Doubleday, 2007), pg 233, 306. are citeable in wikipedia. Most of the data given at WeRelate, apart from an unreliable primary source interview, is not unique, it is sourced to the USDVA data block. The marriage dates aren't cited to anything, and the persons married to is cited to an oral history interview which lacks ethics clearance and demonstration of intact and inviolate transmission (and would be a primary source unless edited and commented by a recognised oral historian in a peer reviewed process). We can't cite to WeRelate as WeRelate lacks an editorial oversight. And we can't use the primary materials as this would constitute original research. I most certainly would not compare the standard of the indicated WeRelate page to academic research, as it has not been subject to an academic peer review process prior to publication. What you really need is an _edited_ genealogy site for the publication of original genealogical research which wikipedia could cite. Fifelfoo (talk) 03:42, 14 April 2011 (UTC)
Michael K. Smith,
1) From WP:NOR - Using sources - Reliable sources third paragraph,
”Drawing conclusions not evident in the reference is original research regardless of the type of source.”
Is this occurring in the contrbution? Then it’s a violation of WP:NOR.
2) From the beginning of WP:V.
”The threshold for inclusion in Wikipedia is verifiability, not truth: whether readers can check that material in Wikipedia has already been published by a reliable source, not whether editors think it is true.”
Is a reliable source being cited? If not then it’s a violation of WP:V. (talk) 14:15, 15 April 2011 (UTC)

Well, not all of these replies are quite coherent, but anyway: The "oral history interview" was carried out by a widely experienced and expert interviewer and Mr. Woody is Gardner's stepson. I know both people quite well. I don't know what you mean by "ethics clearance" in this regard; Mr. Woody has given permission to have his info posted or it wouldn't be there, since WeRelate follows the same Creative Commons standards as WP.

All of WeRelate is "peer reviewed" in the same way that Wikipedia is -- and, in my experience, with a lot more expertise. As to citations for marriages and such: This page (and the associated Family pages) are not at all complete. There is a great deal more information in the process of being posted, from Indian censuses, public records, family letters, and so on. My inquiry was looking to the future, trying to sort out the policy underpinnings as they relate to WP.

The reference on Gardner's page to Wikipedia, by the way, is the result of a small group of enthusiasts who think anything biographical that appears on WP ought automatically to be sucked into the relevant articles on WeRelate. I don't agree, especially when the info on WR is better than that on WP.

It appears from some of the replies that the feeling is that inaccurate and inadequate information from out-of-date secondary sources is to be preferred to more recent, accurate, sourced information that hasn't been published on paper. I would call that bad scholarship. The information is what it is. I'm aware that "The threshold for inclusion in Wikipedia is verifiability, not truth." But I would think that verifiable good information from primary sources would trump verifiable bad information from inadequately researched paper books. --Michael K SmithTalk 10:21, 20 April 2011 (UTC)

WeRelate is not a reliable source. If they cited reliable sources, primary or secondary, you can use those yourself. But your article must be based on reliable secondary sources, with primary sources filling in details, if needed. SlimVirgin TALK|CONTRIBS 05:02, 21 April 2011 (UTC)
Exactly... we don't "ban" primary sources... but we do limit how they are used, and what they are used for. Blueboar (talk) 14:00, 21 April 2011 (UTC)


Bit of an issue I've run into in several places, most recently as talk:pseudoscience (twice), with what I tend to think of as 'quote-diving' (the textual equivalent of pearl diving, where one dives deep and digs through a whole lot of clams to find the one little bitty thing that's sellable). The problem arises becase when I point out that this extreme form of cherry-picking is un-scholarly and misrepresents the intentions of the source and the context of the quote, I get accused of original research for trying to take the intentions and context of the material into consideration. The two examples from the pseudoscience page:

Illusions of causality at the heart of pseudoscience
This source is used to support the statement that "pseudoscience is a threat to public health" in the article. I point out repeatedly that (i) this statement is based on the first line of the abstract of the journal article, (ii) this statement appears nowhere in the article itself, (iii) the journal article is actually talking about cognitive distortions that might might lead people to believe in quackery, and (iv) the journal is a psychology journal and the authors psychologists. These contextual factors make it clear to me that the authors were not trying to discuss pseudoscience as a whole (or even quackery - they were trying to talk about cognitive distortions) and that the quote being used is nothing more than a toss-off line to quickly frame the article abstract. However, when I point this out, I get accused of engaging in original research. just one example, there are dozens of them on the page if you need more.
this document from the NSF
Here we have a document that says explicitly (on page 5) that belief in pseudoscience is being used as one of three indicators of scientific literacy, and where the authors are clearly glossing over terms - they make a reference to 'ten pseudoscientific beliefs' but point to a Gallup poll where the ten items are related to 'belief in the paranormal'. The authors are talking about scientific literacy, and say that they are talking about scientific literacy, but the quote is being leveraged to assert that these 'ten things' are overtly pseudoscience (even against common sense - 'witches' and 'ghosts', for instance, are just superstition; there's little to nothing pseudoscientific about them). Yet again, when I point this out I get accused of OR for addressing the context of the quotes: this edit, and this

Now this would never fly in academia; I've given students C's and D's on papers for this kind of thing. In academia, you just don't take some off-hand, out-of-scope comment that an author made and use it as though it were his/her main point. However, NOR is silent on this in both directions. Nothing in it prevents this kind of misuse of sources to make them say things they never intended to say, and nothing seems to prevent these counter-accusations of OR for trying to put quotes back in the proper context of the source. So how to deal with this? I can recommend a revision or two to policy that would patch this up, but I want to make sure we're all agreed on the problem itself, first. --Ludwigs2 18:04, 26 April 2011 (UTC)

I gave my rationale in the talk page here. --Enric Naval (talk) 09:22, 27 April 2011 (UTC)
Also the NSF report doesn't call them pseudoscience, and neither does our article, it calls them "pseudoscientific beliefs". Ludwigs2 says that "there's little to nothing pseudoscientific about them", but I can find RS that relate ghosts and witchcraft to pseudoscience, specifically, how witchcraft relates to experimental science in the 17th Century, for example:
  • Robert Cogan (1998), Critical thinking: step by step, University Press of America, p. 205, ISBN 0761810676, 9780761810674 Check |isbn= value (help), Section 14.1: the nature of pseudoscience and the paranormal. Subjects in which pseudoscientific claims are made include (...) ghosts and chanelling, (...), satanism and witchcraft, (...). The "paranormal" is that area of pseudoscience which relies on dualistic pseudo explanations. The paranormal includes (...) ghosts (...).  Unknown parameter | ignored (help)
  • The Skeptic Encyclopedia of Pseudoscience, chapter "Witchcraft and the Origins of Science" analyzes the relationship between "belief in ghosts, witches and demons", and attitudes toward science, as well as the scientific claims made to prove scientifically that they existed, see subchapter "Joseph Glanvill and the scientific defense of witches".
  • Science and Pseudoscience (transcript of a talk by Imre Lakatos, who was interested in philosophy of science) "But what exactly is 'experimental' reasoning? [If we look at the vast seventeenth-century literature on witchcraft, it is full of reports of careful observations and sworn evidence - even of experiments. Glanvill, the house philosopher of the early Royal Society, regarded witchcraft as the paradigm of experimental reasoning. We have to define experimental reasoning before we start Humean book burning.]"
  • Why astrology is a pseudoscience (talk by Paul Thagard) "I think that discussion would show that the [demarcation of pseudoscience] criterion marks as pseudoscientific such practices as witchcraft and pyramidology, (...)"
Due to the demarcation problem, there is no clear line separating pseudoscience. There will be RS including ghosts and witchcraft as pseudoscientific, and others not including it.
Note that Ludwigs2 has cited no RS to support any of his points, he is basing all his argument in personal opinion and personal inference. His edits have been reverted because other editors believe that it's him who is doing original research. --Enric Naval (talk) 13:18, 27 April 2011 (UTC)
Normally I'm a bit laissez-faire myself and in the interest of getting a wide view would not object to that like Ludwigs2 has. However in this instance I think one needs to exercise more than the usual amount of care that the references stand up as pseudoscience has strong medical implications, I'd hate to see the article having easy grounds for sniping attacks. The article is quite big too so there is not a burning need to collect citations. Just my 2¢ rather than a strong view on the matter. Dmcq (talk) 14:20, 27 April 2011 (UTC)
I'd ask that we please keep focus on the question I asked. This is not a content question about what can and cannot be called pseudoscience, nor is it a philosophical question about the demarcation problem. This is a sourcing guideline question about whether sources can and should be used in a particular way. to be clear:
  • Yes, Enric and I disagree on the content point in a number of pseudoscience cases
  • Yes, that disagreement can and should be resolved by properly appealing to sources - I may even be wrong, who knows?
  • However, improper use of sources does not help resolve the issue, it just gums up the discussion
If someone wants to find sources that specifically and intentionally make a given statement, that's fine. For instance, note that in the last source Enric presents above - "Why astrology is a pseudoscience" - we see a author who has sat down specifically to explain that astrology is pseudoscience. This is good: we know for sure that anything Thagard says about astrology and pseudoscience is what he intended to say and any quotes to that effect are in line with the context of the work. However, the NSF report and the 'illusions of causality' article are not like that: the first is intending to talk about scientific literacy, the second is intending to discuss cognitive distortions from a psychologist's perspective, and in both cases discussions of pseudoscience are introduced solely to frame or support those other positions. Neither source is intending to talk about pseudoscience directly or specifically; neither source goes into any great detail about pseudoscience; quotes about pseudoscience cannot be taken in literal form from these sources without removing them from their context in a way that significantly changes their meaning. This is bad.
The content discussion will work out however it works out. My only concern here is to establish whether we can use sources in this (to my mind, thoroughly unscholarly) manner. What this amounts to is the following:
  • Source X writes statement Q (verifiable fact). Do we:
    • Treat Q as a definitive, intentional statement of X without regard to its context in the document or its relationship to the topic that X was actually discussing?
    • Weight Q appropriately according to its context in the larger document and the primary focus of that document, so that apparently unintended/off-topic comments are not exaggerated or misapplied?
The second seems like the obvious choice to me (I know that I would hate to have someone take some out-of-context quote from something I've published and use it to say something I never intended to say), but I keep getting countered by people who aver the first. --Ludwigs2 16:34, 27 April 2011 (UTC)
I would tend more to the first and the context only provides a check on whether the meaning is being used as intended. Using the whole document means that in a large work all the bits get to have less and less weight and we'd only be able to use sources which are practically soundbites to start with. Dmcq (talk) 17:06, 27 April 2011 (UTC)
So according to you it would be fine if I added the following sentence to Controlling interest, based on Chapter 6, footnote 21 of the NSF's Science and Engineering Indicators 2006 document?
"According to the National Science Foundation, a controlling interest is defined as 10 % or more of a company's voting stock, or for unincorporated businesses an equivalent interest.[25]"
Our article on the topic currently does not have any specific numbers, so surely that would be a very welcome addition – unless, of course, this kind of quote-mining produces nonsense statements that only make our articles inaccurate. Hans Adler 17:45, 27 April 2011 (UTC)
@Dmcq: it's the "meaning is being used as intended" part that's significant here. I don't think we can even begin to address that without taking the source's intentions and the quote's context into consideration. I mean, there's obviously common sense limitations - a book length work probably has multiple themes and chapters, and a given quote may only be relevant to one of the themes in one of the chapters, but should still be usable regardless. The point here is that I don't think we want off-hand, out-of-scope comments to get reified as uber-reliable quotes just because. When scholars sit down to write a work, they are usually very careful about the topic they are intending to discuss (that's part of being a scholar), but being human they often toss in extraneous background material or interesting associations that are not really pertinent to the topic, don't really receive the same careful thought and attention as their main points, may not be material they are really qualified to talk about at all, and etc. Generally speaking, these kinds of things are graciously kept in perspective by the rest of the scholarly community (scholars usually like that kind of trivia - enquiring minds, you know - but don't confuse interesting tidbits with analytical arguments). I mean seriously: if a world-reknowned apiologist happens to write in some publication - in a fit of humor - that he never liked the song A Taste of Honey, would we add this in as a scientific critique of the The Beatles?
@ Hans - did they really say that? This is one of those times when I wish wikipedia had a good icon for eye-rolling. --Ludwigs2 18:52, 27 April 2011 (UTC)
The NFS is a RS. The document has a definition of "what is pseudoscience" in the same section where the poll is cited. The NFS not only asserts that the poll lists pseudoscientific beliefs, but then goes and mentions two specific items in the poll as pseudoscientifc beliefs that have gone down since past years. The article accurately reflects the nuances of the NFS assertion. The article doesn't treat the items as "overt pseudoscience" anywhere. I provided like four RS that support that the NFS assertion is reasonable (like, in the second comment of this very same thread). Two of them are books and the other two are talks by scholars interested in philosophy of science (the discipline that examines the demarcation problem of science, among other things). I think that this satisfies the WP:BURDEN of proof. Ludwigs2 is appealing to common sense and to our personal definitions of what should be considered pseudoscience. Ludwigs2 is also trying to mind-read and second-guess the intentions of the drafters of the NFS document. --> I think that Ludwigs2 is performing OR in order to reject perfectly good RS that don't agree with his personal ideas. --Enric Naval (talk) 11:24, 28 April 2011 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── You seem to be talking about the specific situation at Pseudoscience rather than the general principle. OK, you can have that. Apparently you think the main point is the distinction between "pseudoscience" and "pseudoscientific beliefs". You are mistaken. Here is what that article currently says:

An example of characterization as pseudoscience by a national scientific body is provided by the US National Science Foundation (NSF), whose statements are generally recognized to harmonize with the scientific consensus in the United States.[1] In 2006 the NSF issued an executive summary of a paper on science and engineering which briefly discussed the prevalence of pseudoscience in modern times. It said that "belief in pseudoscience is widespread" and, referencing a Gallup Poll,[2] stated that belief in the ten commonly believed examples of paranormal phenomena listed in the poll were "pseudoscientific beliefs". The ten items were: "extrasensory perception (ESP), that houses can be haunted, ghosts, telepathy, clairvoyance, astrology, that people can communicate mentally with someone who has died, witches, reincarnation, and channelling."[1]

Anyone who bothers to follow the references can see that this is a bizarre overstatement of what the source actually says. Not to mention that "whose statements are generally recognized to harmonize with the scientific consensus in the United States" is a blatant case of WP:SYN that tries to get around WP:RS/AC. (An earlier version violated that explicitly by claiming: "The scientific consensus, as expressed by the National Science Foundation, has identified and described ten subjects [...]". [26])

This has been going on for over a year now, and I am beginning to wonder if it's necessary to go to Arbcom with it.

What's going on here? Maybe some editors are just not very good at reading technical texts and distinguishing between the points that a text makes and their own conclusions which they draw from the text. And maybe these editors don't realise that they have this problem. Add to this a general feeling of moral superiority on the side of the clueless (because they are fighting for the right thing), and conflicts are inevitable. But we can't have our content decided by those who don't understand what they are doing, just because they feel more strongly about it than those who can work with texts properly, to scholarly standards.

For easy reference, this is the claim that is plain wrong: "[In 2006, the NSF] stated that belief in the ten commonly believed examples of paranormal phenomena listed in the poll were 'pseudoscientific beliefs'." Stated is a pretty strong word which implies that they explicitly made this connection. However, all we have is the fourth paragraph in this section, which in abbreviated form (for clarity) says this:

Belief in pseudoscience increased significantly [...] and then fell somewhat between [..]. The largest declines were in the number of people who believe in ESP, clairvoyance, ghosts, mentally communicating with the dead, and channeling. Nevertheless, about three-fourths of Americans hold at least one pseudoscientific belief; i.e., they believed in at least 1 of the 10 survey items [...].[Footnote defining those ten items without mentioning pseudoscience] In addition, 22% believed in five or more of the items, 32% believed in four, and 57% believed in two. However, only 1% believed in all 10. [citation to the Gallup poll]

This paragraph is not about defining pseudoscience, it's about how belief in it changed over time. It's reasonable to assume that this follows the same patterns as the closely related belief in paranormal. (Belief in paranormal as a proxy for belief in pseudoscience.) As only data about paranormal was available, the used that. In the original context this is proper because for the point that the NSF document wanted to make it doesn't really matter. But for an encyclopedia article that happens to be the article about pseudoscience and not the separate article about paranormal, the difference obviously matters.

For easy reference, the most important policy violations that occurred here:

  • Original research: Explicit claim that the NSF document provides "an example of characterization as pseudoscience".
  • Improper synthesis: "whose statements are generally recognized to harmonize with the scientific consensus in the United States" (an unsourced claim, btw) added for no reason other than to create the impression that the claim represents scientific consensus.
  • Misrepresenting a source: A source does not "state" that A is B, merely because it talks about B and in the next sentence draws examples from A.
  • Cherry-picking/overstating, and ignoring that the source itself contradicts the overstated claim: In the previous paragraph the NSF document quotes the following correct definition of pseudoscience: "claims presented so that they appear [to be] scientific even though they lack supporting evidence and plausibility". This is obviously inconsistent with a claim that every [belief in] ghosts is [belief in] pseudoscience, since belief in ghosts has a priori nothing to do with appearance as scientific.

It is true that in common parlance the words pseudoscience/pseudoscientific are sometimes used more broadly. But here the broadest possible dictionary definition is not appropriate for an article about a topic from philosophy of science, especially not without making it explicit where such an overbroad definition (never used by the scholars who work in the field) is used. If we want to base our articles on this definition, we will have to merge pseudoscience and paranormal, because then the two terms are basically synonyms. (As in the German book title My Paranormal Bicyle that makes fun about number mysticism pseudoscience.) Hans Adler 12:55, 28 April 2011 (UTC)

I've come across a few things like that 'controlling interest' and they can be quite difficult to deal with properly. I believe the best thing is to just produce better sources with explanations to show there's a problem, though some seem to really get a bee in their bonnet about the wrong stuff. I do not believe though we can say it is not reliably sourced or that it is original research.
I think Hans is going around this the better way rather than trying to ban 'quote-diving'. I tried to find out who wrote or checked those 'Science and Engineering Indicators 2006' by clicking 'About Science and Engineering Indicators' on [27] but the page does not exists, which is a little worrying, but the 2010 one [28] says they are subject to extensive review, and the page that was used about pseudoscience doesn't mention that at all in the 2010 version [29]. Perhaps we should just abandon the 2006 ones and go to 2010 instead of trying to get outdated stuff into the pseudoscience article, in which case there's nothing there which should be included. Dmcq (talk) 14:23, 28 April 2011 (UTC)
I in fact believe this case is OR with synthesis, I was thinking there of the case of the 10% controlling interest an a couple of things I'd had to contend with. Dmcq (talk) 09:46, 29 April 2011 (UTC)
@ Dmcq: try clicking this - that should point to the '06 version.
I agree that finding better sources is a better solution, but unfortunately better sources are not always immediately available, and so stopping the misuse of sources becomes much more important. We should not be putting bad information into the encyclopedia and then forcing others into an uphill battle to remove it. The burden should be on the people who want to use the source to use it correctly in the first place.
@ Enric (as well an everyone else): Your 11:24, 28 April post is a perfect example of the wrong approach to this. look at what you've done: You start out with an absolutist statement about the NSF - "The NFS is a RS." - as though the NSF were RS in some abstract, inviolable way about everything under the sun. You end up with a relativist critique - "Ludwigs2 is appealing to common sense and to our personal definitions [...] in order to reject perfectly good RS that don't agree with his personal ideas." Do I need to point out how quickly and thoroughly you would reject this exact style of reasoning if it were offered by a biblical literalist (e.g. "The Bible is the authoritative word of God; Personal interpretations of the text by people who disagree with the Word as written are of no consequence and should be rejected"). I mean, what do we call this: fundamentalist scientism? You are again arguing against a properly scholarly approach - one that uses common sense to evaluate the context of quotes and the intentions of the authors who are writing the articles we cite - as original research. But no: Trying to understand the conclusions that the authors themselves are intent on making is not OR, not unless you are asserting that I cannot use the source itself as a reliable source for what it itself is saying.
@ Hans: well said. I don't think there's anything I can add to that analysis. I think we should try resolving it here first, though, before heading to ArbCom. Would it help if I suggested some specific additions to the policy, so we could mull them? --Ludwigs2 16:34, 28 April 2011 (UTC)
I wasn't talking about problems finding the actual page, just the one about how they compiled it. Your solution to the sources is as far as I can see unworkable and would lead to even more arguments and would cut out a lot of stuff that should be in. [redacting per Ludwigs2's request on my talkpage]It's fine for you to complain about a student but the editors here are not your students and I bet you have enough trouble getting students to do things your way. Also I do not believe ArbCom would or could do anything about your concerns as stated here, what you are talking about is nowhere near their concerns yet that I can see. Dmcq (talk) 17:40, 28 April 2011 (UTC)
<shrug...> This is a matter of scholarship, and so I'm not really concerned whether other editors agree or disagree on a personal level; if we want to be a reputable encyclopedia, some change along these lines needs to be made. The only question in my mind at this point is the best way of framing it, which is why I started this thread. Since this is a change that is arguably necessary for maintaining the reputation of the encyclopedia, ArbCom may very well want to look at the problem if it turns out there is significant opposition to implementing it. However, we can leave that up to them if/when it proves necessary. --Ludwigs2 18:11, 28 April 2011 (UTC)
[redacting per Ludwigs2's request on my talkpage]Please explain why you felt my comment was personal or unnecessary. I mean exactly what I said and I stated it in a straightforward and reasonable way as far as I can see. I do not see what you are objecting to in what I have said. You said "I've given students C's and D's on papers for this kind of thing". It is simply not possible to train editors here to the same standard as students to do things the way you would like them to analyse sources. What's wrong with that? Dmcq (talk) 22:18, 28 April 2011 (UTC) Dmcq (talk) 07:36, 29 April 2011 (UTC)
I see the problem has been going on for a year. ArbCom may be right for resolving the dispute, I'm sure you must have looked at WP:DISPUTE in that amount of time. The case seems to me to be of somebody sticking in references to an old NSF document which was not clearly enough written and can be misinterpreted and some others defending the additions. Saying you need a change of policy to deal with it which would affect loads of other articles would lead to sdmins asking where's the consensus for the policy change. Unfortunately there seems to be no 'simply wrong' guidelines as far as I can see and WP:IAR after discussion and consensus on the talk page is the normal way to solve problems like this before they get to the year long dispute level. Dmcq (talk) 11:15, 29 April 2011 (UTC)
Ideally it should not be necessary to change policy at all. Unfortunately the most obvious things have to be spelled out on this project. It should go without saying that we must source what we say in our articles even if we don't say it explicitly but only implicitly by carefully building an obvious association such as: "The Hawaii attorney general's office said Obama's birth certificate cannot be released to anybody. Hawaii Revised Statute §338-13 says that birth certificates are provided upon request." But in practice we need the explicit rules against synthesis by juxtaposition in WP:SYN. Without that a case like the one in the section below would be incredibly hard to deal with. Of course this must be carefully balanced so that it doesn't impact legitimate discussions. Similarly, it should go without saying that a source is no good for what it doesn't actually intend to say, and that something said or implied in passing doesn't have the full authority of the source. But then we get those who insist this is improper second-guessing of sources, or who simply ignore the argument and only ever talk about the reliability of the source in some abstract terms. We have had another example where the abstract of an article about a technical issue around medical pseudoscience was used to support the insane claim that all pseudoscience is a health risk. Of course the author never meant to say this and nothing like that appeared in the article body. We can't force all scholars to protect all their articles against out-of-context quote-mining. An author who publishes in a medical journal has the right to expect that all readers are aware of that context. Apparently we need an explicit point in policy where we can point editors who refuse to use common sense in such situations.
I am not sure what the best approach is. Maybe something like this: "A source can only support a claim which it intends to make. Authors often use throw-away definitions or base arguments on implicit assumptions that they would not claim explicitly. At other times authors make a point unambiguously, strongly and forcefully without actually expressing it explicitly. Common sense and reading comprehension are required to distinguish these two cases. Claims that are not explicitly present in a source must not be overstated in articles." I am sure this is still problematic in some ways; I am just putting it on the table as an example of what I have in mind. Hans Adler 12:41, 29 April 2011 (UTC)
Everything is supposed to be subject to common-sense as in IAR and the policies say things more explicitly. The policy would have to say more explicitly how one is supposed to distinguish between the 10% controlling interest of the NSF an whatever the tax office or a economist or financial paper might say. None of them may make it the main topic and there may be no main topic about it, it is just something they say in passing or footnotes. Do any of them 'intend to make' the claim? Would your preferred solution mean that we couldn't use the tax office definition for instance? Dmcq (talk) 13:10, 29 April 2011 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── As I see it, there are three ways to approach this in the policy, depending on how we want to frame it:

  • As a modification of the Wikipedia:No_original_research#Using_sources section: basically expanding on the third and fourth lines, so that they say:

    Source material should be carefully summarized, quoted, or rephrased without altering the meaning or intentions of the authors. Consider the main themes the source was trying to convey within its own context, and avoid expanding beyond the scope of the source by extrapolation or quote-mining. In general, do not imply that a source makes a claim unless it is clear that it was the specific purpose of the source to make that claim. Wikipedia only attributes claims to sources when it is clear that the source explicitly and directly intended to make that claim as a conclusion of their work; trying to extrapolate from individual passages without considering the source's intentions and the statement's context is original research.

  • As a modification of the Wikipedia:No_original_research#Reliable_sources section: rewriting the last paragraqph of the section so that it reads:

    In general, a source is considered reliable within the scope of topic areas where it has particular knowledge or expertise, and not outside those areas. Passages in the source text should not be considered reliable statements when they are (i) unclear, difficult to interpret or open to multiple interpretations, (ii) inconsistent with the bulk of the given text, (iii) unrelated or only tangentially related to the topic, conclusions or main themes of the text, (iv) obvious background material, side points, or random commentary, or (v) outside the scope of reliability of a given source.

  • As a modification of the wp:SYNTH section: This would be more difficult - it's such a badly written section to begin with that I'd basically have to redefine the notion of synthesis- but the idea wold be something like the following:

    Synthesis occurs when a claim is attributed to a source - directly or through implication - which the source could be read as having meant, but which the source does not explicitly and intentionally state as a conclusion. Do not combine sources to make a claim that none of them make explicitly; find a source that does make the claim. Do not extrapolate or generalize from a claim made by a source to a different topic that the source does not cover; find another source that makes that statement about that different topic. Do not engage in quote-mining, cherry-picking, or in any way take a quote out of its context to suggest something that the source likely did not intend to say; find another source that makes that claim intentionally. Sources should be used for the claims they do make, not for the claims we would like them to make.

The other alternative is to combine the above into a separate small section on 'proper use of sources'. Really, the ideas are already there in the policy, scattered around and not very clearly put. the whole durned policy needs a major copy edit, if you ask me. --Ludwigs2 18:27, 29 April 2011 (UTC)

Well, no response. so unless anyone wants to raise some objections, I am going to try editing something along these lines into policy. I just need to mull for a while on the best way to do it from the above options. --Ludwigs2 22:24, 1 May 2011 (UTC)
Sorry, somehow I had missed that the discussion had adavanced after my last comment. I think each of your three proposed passages is too sweeping. They might be quite effective in solving the problems we have in mind, but I am afraid they would have severe side-effects, much worse than what we are trying to address. A rules that has to be ignored in every single article is not a particularly good rule. Hans Adler 22:36, 1 May 2011 (UTC)
I'd been leavving it to see if anyone else responded or if it would just die as I'd made my feelings plain enough before. But to reiterate no I don't like any of the options and don't see why anything needs to be done. Dmcq (talk) 23:33, 1 May 2011 (UTC)
@ Dmcq: well, let me just be clear - I have no intentions of letting this issue 'die' because it's something I feel needs to be addressed. I'm happy to work out the details of how it gets addressed, but this issue is not going to go away for me on the articles I work on until something is resolved in policy, and it's going to take some argument better than "I don't see why anything needs to be done" to convince me that nothing needs to be done.
@ Hans: I'm not sure I see the 'too sweeping' point. I mean, I can see how that might be true of the last two points, but the first point is really just an expansion of what's already in the policy. does it need to be trimmed back? --Ludwigs2 00:15, 2 May 2011 (UTC)
For some topics (such as my article pigeon photographer, which is currently at FAC) there are very few sources available and Wikipedia is the best source of information anywhere. Such articles can only be written through a great deal of research that is original but is not "OR" in the technical, forbidden sense. I am also active in the areas of card games and of historical units of measurement, which both have similar problems because there just isn't enough solid research. E.g. there is no reliable source that will tell us completely and accurately where Belote is played. This requires research. This is an example of an excellent source for the fact that it is played in Morocco. In fact, it's clear from the paper that it's an immensely popular game in Casablanca, among the general population. But that's not even close to the main point of the paper, which is about the Jewish community in Casablanca and discusses the social dynamics around Jews and Muslims playing this game together. (It's a fixed partnership game, and Jews make sure that they always have a Muslim partner when playing against Muslims.) When I take many such sources and forge them into a statement about the prevalence of a game, I am already in danger of getting trouble with people who misunderstand the concept of OR, and with PSTS beancounters. (The paper is a primary source for the fact that Belote is popular in Casablanca; so what? It still proves that it's true.) Your proposed text gives them even more ammunition: "unless it is clear that it was the specific purpose of the source to make that claim", "when it is clear that the source explicitly and directly intended to make that claim as a conclusion of their work". (Yes, I know: "imply that a source makes a claim" and "attributes claims to sources", but wikilawyers always ignore such precise language.) Hans Adler 05:59, 2 May 2011 (UTC)
Interestingly, this is pretty much the way I feel about topics such as acupuncture (the main difference being that there's no entrenched bias on project against avian photography to cause you problems there), so I'm not unsympathetic to your concern. But there has to be some balance here. I don't suspect that you are misquoting or making unsupportable generalizations in order to advance a position, but I frequently run across editors who do that, and it's a tiresome, article-ruining experience having to drag through the same ridiculous disputes every time. how would you suggest we write this to cover both of our concerns? --Ludwigs2 08:57, 2 May 2011 (UTC)
I just had a quick look there and I must say I immediately had problems with your edit to the lead. Not all acupuncturists believe in qi, some think it may 'just work' without any such mumbo jumbo. I do hope you have reliable sources to back up that explicit statement. I suppose in a way it shows you have some basis for your concerns, after all one would need a citation showing that was generally believed among actupuncturists rather than just have a text saying that was the belief in some acupuncture manual. Dmcq (talk) 09:29, 2 May 2011 (UTC)
Best you raise that over there. Qi is a self-evident part of the theory that lies behind acupuncture, so the statement is generally true, but I don't doubt that there might be variations is ideation that are worth mentioning. As I keep saying, the revisions are open to discussion so long as that discussion is not in the "no changes are acceptable, ever" vein. --Ludwigs2 14:57, 2 May 2011 (UTC)
Thank you for the suggestion but I read your last comment there [30] and I can think of better things to do with my life than contend with you on an article I have little interest in. Dmcq (talk) 16:37, 2 May 2011 (UTC)
lol - suit yourself. If you can't distinguish between discussion and contention, then I have better things to do with my time, too. --Ludwigs2 17:13, 2 May 2011 (UTC)

Synthesis by juxtaposition

A discussion here has introduced the concept of synthesis by juxtaposition. As I understand it, this says that juxtaposing two assertions so that they relate contextually in an article implies a need for the relationship between the assertions to be shown in a reliable secondary source. In the exampled case, an assertion supported by a Primary source was inserted closely juxtaposed with an existing assertion supported by a Secondary source, and its introduction is being opposed on SYNTH grounds — asserting that its insertion implies an unsupported argument about the relationship between the two-separate cite-supported snippets. In fairness, I'm the editor who introduced the objected-to snippet with Primary source support. I can see the point of the objection (which I've tried to describe neutrally here) as well as the point of a counter-argument made by another editor for allowing the information, thereby allowing WP Readers to judge and follow up for themselves if they wish.

I wonder whether editors here focused-in on OR policy might care to comment either here or in the exampled discussion. It does seem to me that if this opposition to the introduction of supported material has merit, it should be supported by some explicit statement of policy, and that perhaps SYNTH, here, should be extended to make an appropriate explicit statement. Wtmitchell (talk) (earlier Boracay Bill) 08:05, 27 April 2011 (UTC)

Certainly looked like synthesis to me. And even more importantly the primary source did not have anything saying it was in any way directly related to the topic of the article. One would really need a source referring to it in the context of the topic and then say what the secondary source said it said. Dmcq (talk) 09:32, 27 April 2011 (UTC)
Agreed. what's clearly happening is that that whoever added that text was trying to demonstrate a novel point (that someone in the bureaucracy was lying) by introducing evidence for it into wikipedia. It's not even really synthesis - it's actual original research on the part of a wikipedia editor in which the research conclusion is implied rather than stated explicitly. As Dmcq says, there'd need to be some published source actually laying out the steps of this argument - that the conflict between published statements and Hawaii legal code implies an active conspiracy around Obama's bc - and then we could present that argument with proper attribution. --Ludwigs2 19:10, 27 April 2011 (UTC)
Agreed. Juxtposition is an implicit statement of relation between the items. Much of Wikipedia has been written by mild uncontested synthesis, but the is neither mild nor uncontested. North8000 (talk) 10:14, 29 April 2011 (UTC)
Given that, and given that SYNTH defines Synthesis and enunciates WP policy regarding same, should not SYNTH cover Synthesis by juxtaposition? Wtmitchell (talk) (earlier Boracay Bill) 11:50, 3 May 2011 (UTC)
That's exactly what it does cover; the example about plagiarism is essentially the same as the issue discussed here about Obama's birth certificate. — Carl (CBM · talk) 12:16, 3 May 2011 (UTC)
I'm not trying to make an argument here re the Obama example, I'm trying to make the point that SYNTH could/should cover SbyJ a bit more explicitly than it presently does. See also the discussion at Talk:Natural born citizen#US Constitution - Law of Nations referenced for what appears to be a second example where more explicit information here about SbyJ could be useful. Wtmitchell (talk) (earlier Boracay Bill) 12:40, 3 May 2011 (UTC)

Is this routine calculation per WP:CALC? Second opinion please.

The Sun's winter solstice position first fully crossed into the Milky Way around the year 800 AD.

  • I suggest this was not routine WP:CALC and hence required a reference but User:Serendipodous suggests that is was per "You can determine the approximate angular width of the Milky Way. You can determine the speed at which the Sun precesses. You can determine the ecliptic longitude of the Sun at the winter solstice. You know Sun is 0.5° wide. You know the direction the Sun is travelling in. So you can calculate from that when the Sun first entered the Milky Way. Of course, the figure should be rounded, as it can only ever be approximate."
  • What do you think? Regards, SunCreator (talk) 19:38, 6 May 2011 (UTC)
It could be borderline as far as WP:CALC is concerned. I think more importantly it is original research because no secondary source talks about it, it was something they though of and calculated themselves. Dmcq (talk) 21:44, 6 May 2011 (UTC)
  • OR as it is making a notability claim. The sun crosses human created lines all the time. Arbitrarily claiming one is significant enough to note is OR. (talk) 21:58, 6 May 2011 (UTC)
It's not only not routine, the specificity of it is a house of cards. By one def, all of the stars that you can see in the sky in the milky way (galaxy) because we are embedded within it. By the other "sky feature" def, it is shaped like a fuzzy, broken up "cloud" that is in lots of places. Suggest requiring a reliable RS cite of it to sort it out. Sincerely, North8000 (talk) 00:26, 7 May 2011 (UTC)
  • I have to agree with the previous respondents. The first point of my analysis is determining whether it is even true. I looked it up using xephem and 800 seems too early for this cycle (it is not as if it would be the first ever), I get a figure more like 1450 - see [31] - the brown line is xephem's idea of the Milky Way. Of course, that is not rigidly defined, but arguing that those lines are in the wrong place is arguing a matter of interpretation anyway, which is always inherently OR, unless someone (reliable) makes the call. I wouldn't cite xephem as a source either - it has a reputation for very good accuracy and it will base its results on a far more sophisticated model than anything we would call "routine", but I still would not necessarily trust it for events that long ago without an RS to vouch for the integrity of that specific result.
    The basic issue of accuracy is probably irrelevant though - as the others have noted, unless a source can pin down the relevance of this calculation in the first place then any result, correct or not is superfluous since there is no context for its inclusion. If that is forgotten ultimately anything can be used to prove anything else with a simple assertion that it is a relevant factor. Crispmuncher (talk) 00:55, 7 May 2011 (UTC).
The calculation is very relevant to the article it's in, and deals with a very notable concept that's been getting a lot of play. Given that the concept is being bandied about mostly by internet whackos, there won't be an RS on this ever, but it needs to be mentioned. Serendipodous 04:29, 7 May 2011 (UTC)
Re whether internet whackos talking something up indicates that it needs to be mentioned, see WP:DUE. Wtmitchell (talk) (earlier Boracay Bill)
I suspect the calculation doesn't fall under WP:CALC's "routine" provision. It also, as evidence by Crispmuncher above, apparently fails "provided editors agree that the arithmetic and its application correctly reflect the sources". So I'm 'fraid not. --Icerat (talk) 05:13, 7 May 2011 (UTC)
If your point is that, to a significant degree , people are asserting this, then THAT can become the statement.....more credible and sourcable. Like "These people assert that XYZ is the case" instead of "XYZ is the case" North8000 (talk) 12:29, 7 May 2011 (UTC)
Thank you for the many replies. It seems a consensus was formed of the material being some form of Original Research. Regards, SunCreator (talk) 11:53, 9 May 2011 (UTC)

PSTS: I think you'll find it's a bit more complicated than that

The regulars at this page already know that determining whether a source is primary or secondary is not simply a matter of counting up the number of sources, so that when I say something, it's primary, and when Blueboar quotes me, it's secondary, and when SlimVirgin cites him quoting me, it's tertiary, and so forth. Or, worse, when I say something in the newspaper, it's magically secondary from the beginning, so when Blueboar quotes me from the newspaper article, his comment is tertiary, and SlimVirgin's citation is quaternary or something, and we start inventing number-names if someone else refers to SV's statement.

But most of our less-experienced editors probably don't know that, and we don't really explain it all here. We try to hit the basics, but the territory is large, the terrain is bumpy, and we don't cover it all.

I'd like to improve this situation, but after looking over the policy, I'm not sure that this is the page for it, unless we really want PSTS to be half the policy (I don't).

So here's my question: Are there any decent essays or guidelines out there that help folks figure out these details? I wrote WP:Party and person a while ago, but it mostly deals with the sloppy conflation, in which all sources are either secondary–reliable–independent–third-party–good sources or primary–self-published–non-independent–bad sources (the alternate title is WP:Secondary does not mean independent), not really with figuring out how to classify a given document (or a given part of a document, because a document can be primary for its own title, secondary for content summarized from other sources, and tertiary when it quotes a dictionary definition).

What do you think? How can we improve this? WhatamIdoing (talk) 03:25, 2 May 2011 (UTC)

  • We could always write an essay on it together? Fifelfoo (talk) 03:37, 2 May 2011 (UTC)
    If we need to, sure, but if someone else has already done that work, then I'd be equally happy to send folks to an existing page. WhatamIdoing (talk) 04:14, 2 May 2011 (UTC)
    I've occasionally done stuff on History / NOR / RS / PSTS; but, I don't have anything prepackaged for expansion. Fifelfoo (talk) 04:35, 2 May 2011 (UTC)
    I've never seems a good treatment of the issue, which may be because (like many places in policy) it's a an effort to pound a round peg into a square hole. The PST rubric is useful for people doing historical research who need some metric for the validity of different historical claims; It is not so useful for encyclopedia purposes where establishing validity is not the issue. Primary sources are highly desirable in historical research; not so much so here. --Ludwigs2 15:08, 2 May 2011 (UTC)
I think an essay (or better yet a guideline) outlining how to appropriately (and inappropriately) use primary sources (including a discussion of situations when they would be considered the best, most reliable sources... and discussion of how to use them without violating WP:NOR) is a great idea. I would love to participate. Blueboar (talk) 16:22, 2 May 2011 (UTC)
Well, there was Wikipedia:Primary_Secondary_and_Tertiary_Sources, which failed as a proposal (for unknown reasons) back in 2009. I think it was later supplanted by the wp:PRIMARY in NOR. We could co-opt that page and rewrite it as needed, but I still think we need to get away from the PST model, which doesn't fit right to our needs. What we want is (a) publishers with a reputation for impartiality and fact-checking (b) authors with a reputation for scholarship and/or a demonstrable knowledge of the topic, and (c) media that reflect established scholarship rather than momentary or idiosyncratic ideas. PST gets in the way of that as much as it helps. --Ludwigs2 17:37, 2 May 2011 (UTC)
Yes, I would start over... perhaps a good working name would be WP:Using primary sources. (?) Blueboar (talk) 20:37, 2 May 2011 (UTC)
I must admit I'm not happy with the primary/secondary/tertiary business as it currently appears. What calling a source a secondary source is needed for is to show something is worthy of inclusion. Calling something a primary source means we shouldn't include anything from it unless a secondary source shows it is relevant, we shouldn't be mining governments statistics and databases for instance. Calling something a tertiary source is simply saying it is too far away from the subject and we should do better. Things like that secondary sources discuss the information without being too involved are qualifications we put in as common attributes of sources that fulfil our purposes but don't actually describe even halfway accurately what happens in practice. What's there at the moment are guidelines to set us on the commonly accepted path but they don't really describe reality very well. As WhatamIdoing says there's a lot more to it than counting the indirections, I'd like to concentrate mainly on the purpose in the policy and leave the usual attributes to a guideline. Dmcq (talk) 20:42, 2 May 2011 (UTC)
Maybe just wp:Proper use of sources?
I think if we break down the PST thing, the concepts we're after are as follows:
  • We don't like primary sources because the authors (while usually knowledgable) are often trying to argue for new ideas which have not yet gained acceptance. that can be useful for talking about the new idea itself, but can give an unfortunate appearance that the new idea is more accepted than it actually is, which should be avoided.
  • We don't like tertiary sources because the people who write them are not usually experts and often just copy off of others, which they may do badly.
  • We do like secondary sources because we assume that secondary sources are (a) reasonably knowledgeable on the topic, (b) not trying to advocate a particular position, but rather (c) presenting a more-or-less pre-digested overview of the topic.
That implies to me that we can get rid of the PST language and talk directly about carefully considering what sources are trying to do, as we consider using them. --Ludwigs2 21:09, 2 May 2011 (UTC)
I think the other significant reason that we're down on primary sources is that they're very frequently non-independent. Folks say "primary" when they reject a corporate website (as, say, proof of the company's notability), but the real problem is "written and published by the marketing department". WhatamIdoing (talk) 05:01, 3 May 2011 (UTC)

I favor getting rid of "the PSTS thing" but I find it disturbing to read that there may be an attitude that secondary sources are somehow noble and free from agendas. It must surely be true that the better secondary sources are not trying to advocate a particular position but to extend that blessing to all secondary sources seems to be incorrect. I don't doubt that there can be many examples given of secondary sources whose authors were careful, prudent, diligent - but the virtue is tied to their being careful, prudent, and diligent, not to their being secondary sources.
The goal surely is to use reliable sources, where "reliable" surely needs to be well-characterized and well-defined. To denigrate primary sources automatically and to elevate secondary sources automatically does not contribute to good characterization nor to good definition. Of course all the PSTS rules acknowledge that sometimes exceptions are possible - but doesn't the fact that there are exceptions reveal that the entire PSTS notion is misguided for Wikipedia? If there are exceptions when primary sources are fully appropriate, for example, what really matters is whatever it is that makes the primary source usable.
Shouldn't the policy discuss that aspect of the suitability of sources and not instead get hopelessly entangled with primary/secondary/tertiary discussions? If an article is discussing, for example, the Periodic Table of the Elements isn't it close to mandatory to refer back to Mendeleev and his first publication, wherein he very much was pushing a new idea which (because it was new) couldn't possibly have "gained acceptance"? Obviously later primary and secondary sources may provide additional useful and informative material. How utterly strange it would be to discuss the Periodic Table and to refer, somewhat obliquely, to "some Russian guy" who first formulated the idea: you have to identify the original proponent and that can easily and properly involve inclusion of material from that primary source. Similarly for Darwin and evolution. Maybe for some aspects of writing about historic topics some editors like the PSTS categorization but Wikipedia policy is for all articles. If a policy fails (in part or entirely) to make sense for some class of article it should not be policy. (A policy statement on policies is clearly needed.) Minasbeede (talk) 04:48, 12 May 2011 (UTC)

I think there in particular three things that might need to be covered explicitly (here or in an essay):

  • The differences between the WP usage of primary/secondary/tertiary and the usage in the real world. This important because the first association (new) editors have when reading those terms relates to the external use they are familiar with. This can be a constant source for misunderstanding & disagreements (including the formulation of this policy).
  • Distinguish between sourcing content and sourcing notability (secondary,tertiary sources are often better to indicate notability).
  • As far as the sourcing of content is concerned the quality, accuracy and reputability of a source are almost always more important than its primary/secondary/tertiary nature. If those 3 characteristics are clearly higher for a particular primary or tertiary source than for a competing secondary source, the primary or tertiary one is to be preferred. To give a "typical" example, some (new) scientific resputable is published in some reputable academic journal and later there is an article about that result in some newspaper or general interest journal. Now in this scenario the second article is a secondary source and indeed might (more) be useful for indicating an increased notability (for general oir largers audiences), but the first article despite being the primary source is usually much better suited for sourcing the content of the WP article. Another way to kook at this might be, don't copy typos, clear mistakes or inaccuracies from secondary sources just because they are secondary, but in doubt use primary or teriary sources to correct them.

--Kmhkmh (talk) 09:15, 3 May 2011 (UTC)

Your first point (on which I fully agree) makes me think we should name it WP:Wikiprimary and wikisecondary sources, to distinguish it from real-world primary and secondary sources. ;-)
A cute name is likely to undercut the utility, however.
Would you all help out with this? We can start in a sandbox in someone's userspace while try to find the perfect name. I don't mind editors working in my userspace; User:WhatamIdoing/Sandbox 4 is at the disposal of anyone who wants to start it. We can move the page wholesale when we've got a working title for it. WhatamIdoing (talk) 18:16, 7 May 2011 (UTC)
Page since moved to Wikipedia:Identifying and using primary and secondary sources WhatamIdoing (talk) 15:54, 11 November 2011 (UTC)
I think Ludwigs2 got it just about right above. I also think that the PST thing contributes more confusion than it ameliorates. If a rewrite is drafted, one important point currently made re primary sources should probably be extended to all sources -- the project page says, "Any interpretation of primary source material requires a reliable secondary source for that interpretation." Surely that does not mean to imply that interpretation by Wikipedia editors of secondary source material does not require a reliable secondary source for that interpretation. I would restate that as something like, "Interpretation of source supported material requires a reliable source supporting that interpretation." (I would want to footnote that to clarify that WP:V requires that such a source exist, but does not require the citation of the specific source relied upon unless the interpretation is challenged or is presented in the form of a direct quote -- but I do tend to be overly wordy) Wtmitchell (talk) (earlier Boracay Bill) 20:33, 8 May 2011 (UTC)

Wp:ver/wp:nor puts too much weight on the primary/secondary/tertiary metric, and is missing two other metrics which relate to ACTUAL reliability on the item that cited them. These would be objectivity and expertise with respect to the item that cited them. The "two birds with one stone" solution would be to add those two metrics to wp:ver/wp:nor, take all 4 (primary/secondary/tertiary, current "RS" criteria, and these two new ones) cumulatively together as the "strength of the cite" and say that the strength of the cite must be commensurate with the nature of the situation (controversial, non-controversial etc.) Sincerely , North8000 (talk) 21:39, 8 May 2011 (UTC)

I agree... Our current WP:PSTS does not address reliability as it relates to P/S/T sources at all, nor does it discuss how P/S/T sources relate to notability (and even Neutrality). Whatever we call some new PST essay/guideline/policy, those should be included.
I would like to make a suggestion: Let's not think of this as a re-write of the current PSTS section... that will raise alarm bells and knee-jerk opposition from those who (mistakenly) think we are in some way out to "weaken the NOR policy"... let's approach it as creating a completely new and separate page to help editors understand all the complexities of using and citing primary, secondary and tertiary sources. Blueboar (talk) 21:57, 8 May 2011 (UTC)
Cool. Except wouldn't you say that these other metrics are something in addition to PSTS rather than redefining PSTS? North8000 (talk) 22:18, 8 May 2011 (UTC)
Here's some wording which, if added, that would take a baby step in that direction:
Controversial claims require stronger sourcing. Two additional measures of this are the objectivity and expertise of the source with respect to the material which cited it.
Sincerely, North8000 (talk) 22:57, 8 May 2011 (UTC)
Yes—but not here. The need for "heavy" claims to be supported by "strong" sources has basically nothing to do with PSTS issues.
I've been thinking that it would be nice to have an image to illustrate that concept: a tiny 'claim' (a butterfly?) being held up by a weak 'source' (a stick figure person? a fragile-looking architectural column?), next to a huge 'claim' (a globe?) being held up b a strong 'source' (Atlas (mythology)?). WhatamIdoing (talk) 23:29, 8 May 2011 (UTC)
Yes you're right, but since it's not in Wikipedia, it unfortunately isn't "related" to any policy so here's as good of a place as any. And I think that your idea is cool. North8000 (talk) 23:49, 8 May 2011 (UTC)
I've started at User:WhatamIdoing/Sandbox 4. Please feel free to improve it. It's got a long way to go before we sort out a name and move to the Wikipedia space. WhatamIdoing (talk) 23:40, 9 May 2011 (UTC)
I'd like to say thanks to Blueboar and Wtmitchell for their improvements; I haven't had time to get back to this, but I'd like to invite any interested person to comment at User talk:WhatamIdoing/Sandbox 4. And, again, although the page is currently in my userspace, you're welcome to boldly improve it. WhatamIdoing (talk) 04:27, 24 May 2011 (UTC)

Slightly expand CALC

I can't find it now, but I ran across a talk page discussion yesterday in which an editor actually asserted that our sourcing policies prohibit editors from describing quantitative information, e.g., that if the source says "X is 5 meters high, but Y is only four meters high", then editors were not permitted say "X is taller than Y".

As such a restriction is incredibly silly, could we add a sentence to CALC along the lines of, "Editors may also provide simple descriptions and comparisons of quantitative information, e.g., saying that something is bigger, taller, or heavier."

While we're at it, I wonder if we should also say that editors are permitted to round numbers, e.g., "33%" even if the source actually says 33.124623946%, or even "about one-third" (which is highly desirable when multiple sources provide slightly different numbers: 31.7%, 32.1%, and 34.2% are all "about one-third", and "about one-third" is a more accurate description of the consensus of sources on that point than any single number). WhatamIdoing (talk) 00:31, 12 May 2011 (UTC)

Why would one want to say X is taller than Y if no source says so? That would be just something an editor thought of and that does not fit into Wikipedia. It is basically original research. SO I would totally reject the first suggestion. As to the second it seems fairly reasonable in the spirit of summarizing sources, however I think it is probably better to give the range as minimum-maximum, e.g. 31.7|% - 34.2%. Dmcq (talk) 15:18, 12 May 2011 (UTC)
To me it is rather common sense than original research. And the answer to your question is, that WP authors summarize and contextualize/explain established knowledge and doing that may lead to pointing out X being greater than Y occasionally.--Kmhkmh (talk) 15:25, 12 May 2011 (UTC)
If the sources don't feel it necessary why do we? At best I can think of it might go under the same grounds as examples which are meant to illustrate the article but not provide extra information, it would have to be carefully phrased or people will drive a horse and cart through it with all sorts of stupid comparisons like how many times the size of Wales somewhere is or how heavy thunderstorms are compared to elephants. Dmcq (talk) 15:45, 12 May 2011 (UTC)
Because we combine/summarize/explain (different) sources rather than literally copying them. Also many the sources we use are written in a different context and/or for different audience, where it might not be deem necessary to to spill out various comparisons explicitly.--Kmhkmh (talk) 16:00, 12 May 2011 (UTC)
Perhaps you could point to an example where you think such a comparison would be very useful for understanding but is not provided in a source? Dmcq (talk) 17:33, 12 May 2011 (UTC)
An example could be a list of notable objects that is ordered according to a certain rule, e.g. aircraft by maximum take-off weight. You probably won't find an explicit statement in the literature on how to construct the list with the ordering rule, given the objects with the weights. Count Iblis (talk) 17:36, 12 May 2011 (UTC)
Why would we provide such a list if no source does that sort of thing? And the idea of putting in extra explanations of our own is quite worrying, I've seen that used a few times as a coatrack device to bring in stuff which wasn't directly relevant. I don't mind us just listing things in a way that lets a user sort the list by clicking a column header however they want but us doing it on our own criteria sounds wrong to me. Dmcq (talk) 17:42, 12 May 2011 (UTC)
Suppose that such lists do exist, but because old paper sources don't get updated, we will have multiple sources from which we maintain a complete list on Wikipeida. Of course, from time to time, complete lists will appear in certain aviation journals, but we would obviously not wait for that. So, e.g. when the Airbus A380 became a commercial airliner it could immediately be added to such a list, we should not have to wait until, say, a year later when some aviation magazine updates the complete list. Count Iblis (talk) 17:49, 12 May 2011 (UTC)
We already have the concept of List articles which contain notable lists which can have items added which weren't on the list which established the notability of the list. So we would need to have a notable list somewhere which did the sort like you said but was missing the Airbus figure. Primary sources can be used to add data to some extent after notability for something has been established. Dmcq (talk) 18:02, 12 May 2011 (UTC)
Ok., but then the position of that item on the list should not be considered to be an "illegal act of OR"  :) Count Iblis (talk) 18:05, 12 May 2011 (UTC)
Dmcq, you might not have noticed, but the source I gave is actually making a direct comparison of the heights of the objects. The only difference is that it is not using the phrase "taller than" to do so.
Similarly, a source might present a table or graph of data that shows the height (say, of buildings in a small town); it would not be unreasonable for editors to describe the contents of that table with sentences like "The tallest building in town is the Smallville Church, whose belltower reaches to 19 meters high." This is far more informative to the reader than the bald statement, "The Smallville Church is 19 meters high," and far more concise, appropriate, and encyclopedic than "The Smallville Bank is 9 meters high. The Smallville Market is 10 meters high. The Smallville Mansions are 12 meters high. The Smallville Church is 19 meters high."
Additionally, editors sometimes need to describe things at a different level than the source. A highly technical source might be a stellar source of information, but the editor might be writing a simple topic sentence to summarize the information at a level that a teenager can understand. WhatamIdoing (talk) 18:55, 12 May 2011 (UTC)
Having searched the discussion I still do not notice this source you say you gave. I still do not see why you would select the buildings by height if the source does not. Dmcq (talk) 22:43, 12 May 2011 (UTC)
Read the first paragraph, at the top of this section, specifically: "if the source says "X is 5 meters high, but Y is only four meters high", then editors were not permitted say "X is taller than Y"." You are arguing that given a source saying these exact words, that the editor must pretend to be either too ignorant to construct the simple arithmetic statement that 5 > 4 or too stupid to apply that mathematical statement to the subject at hand.
And the reasons for doing this are two:
  1. It is sometimes useful to the reader to get a tangible comparison to something familiar. For example, it is sometimes better for the reader to be told that a tiger can weigh four times what the typical adult human does, or that a given dinosaur is the height of a three-story building, than to be given raw, decontextualized numbers. Most people don't have a talent for picturing how long 11 meters is; most people do have a basic notion of the height of a three-story building.
  2. It is sometimes necessary and appropriate to provide a general description of things (e.g., buildings in a town), and one valid way of describing things is to note both the typical (e.g., very few buildings in Paris are more than about six stories high) and the atypical (except for the Eiffel Tower). Naming, therefore, the tallest (or oldest, or largest, or whatever obvious and verifiable facts are available) building in a town—or the professional athlete who scored the most points last season, or was paid the most, or playing the most games, or any number of similarly obvious and easily verifiable facts—is a way of describing and contextualizing data. If you have a complete list of players for a team during a season, and a complete list of how many points each scored, then "Alice Jones scored the most points" is certainly verifiable, even if the particular source you name does not use the words "most points scored" anywhere in it, and it is often preferable to the alternative of copying an entire table's worth of trivia into an article. WhatamIdoing (talk) 05:03, 16 May 2011 (UTC)
You said you gave a source, did 'source' mean the hypothetical example at the start of this discussion? Just given that information I see no point to your saying that X is taller than Y. There is no argument here for changing CALC. Have you a real world example which might have more point than the examples you gave? Personally I have little time for saying how many times China is the size of Wales. Dmcq (talk) 08:29, 16 May 2011 (UTC)
Yes, I was referring to the hypothetical source.
If the relative size is all that matters, then I see no reason to provide absolute numbers; they may be distractingly precise. "China has the largest land mass of any East Asian country" is more appropriate in an article about China's geography than a laundry list like "China has ___ km2 land; India has ___ km2 land; Japan has ___ km2 land..."
But what's really important to me is that this rule is being misunderstood as prohibiting editors from describing simple, obvious facts using words instead of numbers. WhatamIdoing (talk) 16:09, 22 May 2011 (UTC)

Linking an essay

My essay is live at WP:SYNNOT (no longer actually mine, of course, now that it's in the WP namespace, but no one else has edited it yet). I didn't get any objection in the previous section when I suggested adding links to essays. If I still don't get any objection here, I'll go ahead and add the link in the "See also" section of the project page, along with one to the other essay I mentioned. Of course, the essay has a standard supplemental-essay template at the top saying to defer to the actual policy. --Dan Wylie-Sears 2 (talk) 05:17, 25 May 2011 (UTC)

Laudible. I was immediately driven here by someone insisting that applying a definition was SYNTH (i.e., "Y fits the definition of X, therefore Y is X" is SYNTH). This seems ridiculous to me. Do you (a) agree and (b) think your essay covers this case? Graft | talk 15:44, 25 May 2011 (UTC)
I think (a) that applying a definition can be SYNTH, but usually isn't. It depends on whether it introduces a new assertion. The example in the essay of something that is SYNTH actually can be framed as the dubious application of a definition: if soldiers are missing after a battle but not among the dead or taken prisoner, they count as deserters under... . It's how the definition was applied that's the problem: it requires the exercise of some non-trivial judgment about the significance of the numbers. (That does raise the point that the ellipsis ought to be replaced with another sentence. I'll do that next.) As for (b) I think it now comes about as close as it can without making WP:BEANS. I added a section explaining that you're supposed to make arguments on talk pages that would be SYNTH if they were how something in an article is sourced. The approach to SYNTH advocated in the essay would rule out a blanket invocation of the sort the person seemed to be making, anyway. --Dan Wylie-Sears 2 (talk) 17:46, 25 May 2011 (UTC)
Sounds a good idea. There have been numerous arguments about SYNTH on this page and the OR noticeboard and it sounds a reasonable place to stick some corner cases so editors can make a better judgement. And while I'm at it could I say I think a few cases here should really have been moved to the noticeboard. People should really have some idea about exactly what they think is wrong with the policy or ideas on how to improve the page if they're on this page rather than just complaining they don't like its application in a particular case and therefore it must be wrong. Dmcq (talk) 09:41, 26 May 2011 (UTC)

Does WP:OR apply to the correct spelling of celebrity's names, or just apply to ideas and such?

I am having trouble accepting that these policies (WP:OR and WP:SYNTH) are meant for celebrity's names. I am also having trouble accepting that a verified Twitter account for a celebrity can never be used as a source for the correct spelling of that celebrity's name, especially when other reliable sources can demonstrate that the celebrity himself is actually doing the tweets, and many other reliable work credits also support that same spelling. There exists a very unique problem originating from the spelling of celebrity Norm Macdonald's surname in reliable media sources. IMO, Norm Macdonald's active verified Twitter account and his work credits (both past and present) should be considered self-evident, reliable, determinate sources for his correct surname; however, the big problem is some reliable media sources have sometimes been inconsistent when spelling Norm's surname (using MacDonald). This fact is causing all the trouble. Some say Twitter can never be used as a reliable source for surname. They also say that Norm's work credits and Twitter are somehow not self-evident and determinate, because of some inconsistent media reports of his surname. IMO, these errors in media print have no standing whatsoever on the Twitter and work credits. Those who oppose cite WP:OR and WP:SYNTH, and oppose my statement that Macdonald's surname is "sometimes presented incorrectly in the media as MacDonald." They say for me not to violate WP:OR and WP:SYNTH with that statement, I must cite other reliable media sources which actually mention the media reporting Macdonald's surname incorrectly as MacDonald before I can make that statement. I say this doesn't make any sense when we are talking about a bio article for a living celebrity with work credits(past and present) using Norm Macdonald, and they could be used as self-evident sources to support the statement that MacDonald is incorrect. --RedEyedCajun (talk) 10:26, 26 May 2011 (UTC)

Is there some reliable source saying there has been a problem? If there has been no reliable source saying there has been a problem then the article shouldn't talk about one. It is acceptable for the article to point out that some publications have written MacDonald, I can't think of anything more that should be done. There is no need to go on about incorrect or anything. It would only be if the title was Macdonald and he was born MacDonald or there was somebody had written about a problem that anything more should be said. Dmcq (talk) 11:08, 26 May 2011 (UTC)
If Macdonald spells his own name with a lowercase "d", then using a primary source in this case should be acceptable. If there is significant debate over the capitalization (i.e., it's actually controversial), then that wouldn't be appropriate, but if it's noncontroversial, then use the spelling he himself uses. /ƒETCHCOMMS/ 14:51, 26 May 2011 (UTC)
It seems to me as though using him as a primary source for the fact that his name is spelled with a lower-case "d" ought to be fine. What seems questionable is the inference that it's not also correct with a capital "D". Maybe he was born MacDonald, uses both casually, had Macdonald on his first professional deal by mistake, and just has his own professional stuff be consistent for some obscure legal reason. Who knows, stranger things have happened. I agree with Dmcq that there's no need to say "incorrect" unless some reliable source reports a problem over it. I wouldn't call it synthesis that advances a point, though: in some old versions of WP:NOR I looked at, the concern was more about making arguments, implicitly or explicitly, to advance crackpot theories. That's not going here at all. This is just an unverified piece of factual information. If the editor's opinion (or consensus on the talk page once a question is raised) is that a source presumably exists, then it can be included unsourced. Btw, this question seems as though it belongs on the notice board instead of here. --Dan Wylie-Sears 2 (talk) 03:02, 27 May 2011 (UTC)
There aren't any reliable sources (or unreliable either) that have ever reported on the problem of reliable media sources sometimes using the incorrect MacDonald version. The only controversy I can find is on Wiki 'Talk page' for Norm Macdonald bio-article. I'm told that since 2007, the title has moved back-and-forth from "Norm Macdonald" to "Norm MacDonald" about four times. Just recently, I managed to get consensus to move back to "Norm Macdonald". There are past and recent instances of the New York Times using the MacDonald version, but overwhelmingly they use Macdonald. TVGuide's website is the worst reliable source offender, using both versions on all it's webpages about Norm Macdonald. Curiously, the titles on these TVGuide webpages use the Macdonald version, but then they use the MacDonald version internally on those webpages. I surmise the overwhelming instances of using the incorrect MacDonald version are found in unreliable website sources, which I believe are propagating this surname error and causing a backwash of errors into reliable sources. Norm Macdonald's brother is Neil Macdonald, an award-winning journalist for CBC news, and the CBC website spells Neil's surname Macdonald. As far as anyone can discover, Norm has always used Macdonald in all his comedic/actor/host work credits, with the major exception being the current official SNL website which uses MacDonald. (Norm was on SNL in the mid-1990s. We don't know when this SNL website was created. A year after Norm left SNL, he made a movie using work credit of Macdonald.) The problem we are having is people keep editing using the MacDonald version on the Norm Macdonald bio-article and elsewhere across Wiki pages that refer to Norm Macdonald. I was trying to put the stmt "Macdonald's surname is sometimes presented incorrectly in the media as MacDonald" into the Norm Macdonald bio-article hoping that might cut-down on future editors using incorrect MacDonald. However, I believe we have found consensus on another possible solution of putting a banner on the 'edit page' for Norm Macdonald bio-article. Hypothetically, if only unreliable sources had ever used the MacDonald version, would the stmt need any reliable references at all? Or would that stmt still be considered a WP:OR and WP:SYNTH violation without references? --RedEyedCajun (talk) 10:16, 27 May 2011 (UTC)
If it's caused that amount of trouble on Wikipedia a note at the top of the talk page is probably a reasonable thing to do. I'd probably have something like (sometimes written as MacDonald) in the lead of the article and say no more about it. Sometimes people put in a html comment for more to stop editors changing things round like "(sometimes written as MacDonald)<!-- See the FAQ about name on the talk page -->" Dmcq (talk) 11:13, 27 May 2011 (UTC)
If you look at a broad selection of sources, check which they use, evaluate their credibility, and conclude that only unreliable sources use the capital D, you're definitely doing original research. But even if only unreliable sources use the capital letter, the conclusion still does not follow. Maybe both are correct, and the reliable sources have a practice of using a standard reference work that happens to list one rather than the other. Or he regards both versions as correct, but prefers the lower-case version. To rule out those possibilities takes some research, too. There's a question of clarity here: what exactly does "incorrect" mean in the case of a name? I normally go by "Dan", but referring to me as "Daniel" is also correct. There's one person I went to elementary school with who still calls me "Danny", and I don't mind. Is it that he has a history of correcting people when they use the incorrect version? Is it that legal documents with the incorrect version aren't binding in some jurisdiction unless an affidavit is attached to attest that the MacDonald referred to in the document is really Macdonald? You've established that the article should refer to him as Macdonald. You haven't shown that it's verifiable that the alternate is incorrect rather than merely alternate. --Dan Wylie-Sears 2 (talk) 17:56, 27 May 2011 (UTC)
Thanks. Being new to Wiki, I agree I did apparently post this question in the wrong place. "Btw, this question seems as though it belongs on the notice board instead of here. --Dan Wylie-Sears 2 (talk)" Can a Wiki Administrator move the above discussion to the notice board or what, if anything, should be done to fix my apparent posting error? --RedEyedCajun (talk) 20:21, 27 May 2011 (UTC)
It's probably OK here, IMO. /ƒETCHCOMMS/ 01:36, 28 May 2011 (UTC)

Is providing context SYNTH?

I've encountered this issue a few times and I don't think the current policy is adequate. Adding a sourced piece of information to an article without making the context clear can be POV-pushing, however adding the context could be considered SYNTH. An example. In 2009 there was a peanut butter related salmonella outbreak involving the Peanut Corporation of America. As a result hundreds of companies issued voluntary recall notices for thousands of different products. These recall notices are RS sources (albeit primary). However, in my opinion adding to an article "In 2009 ACME company recalled three of their products due to a potential salmonella outbreak" is pushing a particular POV about that company - ie there's some significant problem with the companies products. Adding the context that this was part of a nationwide voluntary recall involving hundreds of companies and thousands of products provides the context that it really wasn't a major issue for ACME company - but doing so might be considered SYNTH, since it's putting together two sources to imply a "novel" conclusion. Strictly following policy would appear to create a misleading article. --Icerat (talk) 15:09, 14 May 2011 (UTC)

To really answer this, we would have to look at the specific language in a specific article, and the sources used to support that language. Providing context is generally a good thing ... but the specific wording used to provide context may well constitute OR. So it comes down to how you provide context. Blueboar (talk) 15:25, 14 May 2011 (UTC)
But if you're using a source about the Peanut Corporate of America, that doesn't mention ACME company, and including that in the article on ACME company, wouldn't that be consider Synth no matter how it's worded? --Icerat (talk) 15:37, 14 May 2011 (UTC)
That would be a synth, but it's hard to imagine ACME would remove the product without explanation, or without mentioning the Peanut company. If they indeed didn't provide any explanation, we indeed shouldn't make a guess and provide a context that might be unrelated. That would be OR. On a side note, if it was a primary source, perhaps the fact is not notable enough to be included. Laurent (talk) 15:58, 14 May 2011 (UTC)
I guess it might be sensible in some circumstances even if it breaks OR, you'd have to get consensus on that on the talk page. However in this instance it sounds like two topics have been rolled into one and the article about the corporation is being used as a coatrack. I believe the article should be split in two, onne about the corporation and one about the outbreak and the corporation one shouldthen refer to the outbreak one for the wider view. Dmcq (talk) 16:17, 14 May 2011 (UTC)
Ok, well here is one specific example. An editor added this [32] and I attempted to add some context [33] which the other editor objected to as SYNTH and deleted. I see his point. Personally I think the info is clearly WP:NOTNEWS and trivial and doesn't deserve inclusion, but it's also so trivial I have no interest in wasting time challenging it (pretty much every edit on this page is a battle), but I'd like to get the principal of providing context and SYNTH clarified as I've been encountering it a lot lately. --Icerat (talk) 16:36, 14 May 2011 (UTC)
If a separate article had been set up for that recall, and it certainly sounds notable, then the problems in both this article and the Peanut Corporation of America article could be solved without any of the hassle. This is exactly the same problem as the last. There may be something for what you say but your examples don't provide a convincing case. Dmcq (talk) 17:25, 14 May 2011 (UTC)
I'm not sure how that would solve the problem? It would still require a reader to actually peruse through other articles/sources to get a picture. In another situation, with I think clearly SYNTH, with another article, an editor took a quote from an RS where an "expert" disparaged a company and it's products. That's fine on the face of it, because it was clearly RS. A moments research however discovers the "expert" owned and operated at least one competitor company, ie he had a clear COI. Mentioning the fact he is a competitor is something you would expect a fair source to do, but doing so in WP would be implying the "new" conclusion that this source had a COI, ie SYNTH. The COI information is RS and V. Mentioning it violates SYNTH, not mentioning it means wikipedia is actually pushing his POV, which is contrary to wikipedia's goal of producing a neutral, reliably sourced encyclopedia. --Icerat (talk) 17:41, 14 May 2011 (UTC)
Personally I'd say the contextual information would tend to indicate that the information added to the article about Amway was pretty insignificant (for someone wanting to read about Amway) and should therefore not be included in the article at all. But if it is to be included, then the context seems necessary so as to avoid misleading readers. --Kotniski (talk) 17:35, 14 May 2011 (UTC)
The original edit is definitely problematic. It seems to be an obvious coatrack addition, worded in a non-neutral tone, negative towards Amway. If the issue is to be included at all (and I think you have a case that it should not), it needs to be reworded... To avoid even potential SYNT issues, my suggestion is that you paraphrase and quote Amways own recall notice. Something like:
  • In 2009, Amway Global voluntarily issued a recall its XS Energy Bar Chocolate Nut Roll "as a result of Peanut Corporation of America’s expansion of their recall of peanut products back to January 2007." The recall notice went on to say that Amway took this step "as a precaution while the FDA continues its investigation into PCA’s activities and salmonella contamination." The company also stated that it was not aware of any reports of illness in connection with the products.
Then cite Amway's recall notice (which is already the source used for the problematic addition). Blueboar (talk) 17:37, 14 May 2011 (UTC)
Agree with all of this, but as per my comment above these feel like "workarounds" to get around a policy that, in part, can work against wikipedia's own stated goals. I've honestly no idea on an improvement, but I keep hitting this particular problem.--Icerat (talk) 17:41, 14 May 2011 (UTC)
I think there needs to be far more emphasis in these policies about editors' using sources with intelligence rather than attempting to follow rigid rules. The "rules" stated in these policies are really only dumb first approximations to what we do - in practice we aim (or should do) to reach sensible conclusions about how to select and present information so as to help readers and not mislead and confuse them.--Kotniski (talk) 17:49, 14 May 2011 (UTC)
Easier said than done when you have controversial topics and editors with very strong held POV! (particular in essentially 1:1 situations) Just to add though, this is really all related to discussions like the one on "ban foxnews". The POV and context of sources is important, but including even RS/V info on this would be SYNTH. Admittedly opening this up would open another whole can of worms, potentially making every article a soapbox for only peripherally related material, but I'm hopeful maybe some kind of "official" guidance could be provided. I just don't know what it could be. :-/--Icerat (talk) 17:53, 14 May 2011 (UTC)
No, I don't think it's possible to resolve all disagreements by means of official guidance. We can state some principles, then just hope that in any given situation there will be enough sensible people around to outweigh the POV-pushers. Now if we could find a way of ensuring that...--Kotniski (talk) 18:11, 14 May 2011 (UTC)
Well... that is in part what noticeboards like NORN, and RSN are for... to notify (usually) sensible people that there is an issue and seek guidance/help. Blueboar (talk) 22:17, 15 May 2011 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────Back to the original question though, describing some attributes of a source would be synth - yes/no/depends? (not that I can't guess the answer)--Icerat (talk) 22:35, 15 May 2011 (UTC)

It depends (was that your guess?). We have to look at the specifics of each case to say yes or no. The key factor is how you describe the attributes. Blueboar (talk) 13:21, 16 May 2011 (UTC)
I've spent a lot of time of late on business systems, developing as idiot proof a system for people to follow as possible. I hate "depends". It allows the idiots too much leeway to screwup :-). On a related note - what about attributes of key people in an organisation? For a bit of background, I'm trying to develop content guidelines for company articles. I've encountered people wanting to add information about "key people", essentially in order to push a particular POV. It might be on an executives background or an owners political donations, or a directors other corporate affiliations. This seems to me to be at least partially valuable "context" and background information but also potentially SYNTH for POV pushing. Thoughts? --Icerat (talk) 13:41, 16 May 2011 (UTC)
Same answer... with the caveat that WP:BLP requires a "better safe than sorry" attitude, demanding much higher quality RS sources and extreme caution when it comes to NPOV, NOR, etc.
We can define what an improper synthesis is (in broad terms), and tell people not to do it... but to know whether any specific sentence or paragraph fits the definition, we have to examine the specific sentence or paragraph in question and the sources used to support it. Looking through the various Amway related articles, there is definitely a lot of POV pushing (and I detect it from both sides)... and there is definitely some OR being used to push those POVs. But whether a given sentence or paragraph is a NPOV or NOR violation has to be examined on a case by case basis. Blueboar (talk) 14:04, 16 May 2011 (UTC)
I think the article has been turned into a coatrack about the salmonella outbreak and the salmonella outbreak should be treated in a separate article. It could possibly be reasonable to do a tiny amount of explanation but I wouldn't try sticking it into policy - I've had to deal with the gun lobby sticking 'explanations' of gun terms into Brady Campaign which used sources that were nowhere near mentioning anything about the topic. Dmcq (talk) 13:47, 16 May 2011 (UTC)
I agree... clearly a coatrack. Blueboar (talk) 14:04, 16 May 2011 (UTC)
I believe it absolutely essential to place an artcile into context at as early a stage as possible. This is almost always hinted at in the lede. In some articles it might be neccessary to provide a bit more in a dedicated section. For example, both the articles First Battle of El Alamein and Second Battle of El Alamein have a section called "Prelude" which puts the battles into context. Martinvl (talk) 19:37, 29 May 2011 (UTC)

"Providing context" is inevitably SYNTH, unless one uses reliable sources that provide the same context and are about the topic of the article. The only real question is if that SYNTH of a minor/NPOV/helpful enough nature to allow it, despite the fact that it is SYNTH. In some cases it's neutral in its presentation, and helpful to the reader. Very often, though, it's POV-pushing designed specifically to counter material found in the article, or to discredit sources quoted in the article. Jayjg (talk) 20:17, 29 May 2011 (UTC)

I hate SYNTH

SYNTH justifies the removal of all but a few sentences of every article in Wikipedia. Unless one paraphrases a source sentence by sentence, something will be juxtaposed in the article that's not juxtaposed in the original. If you do copy the source structure, sentence by sentence, that's not encyclopedic unless the source is an encyclopedia article. Even if the source is an encyclopedia article, copying it and altering the word order a little in each sentence is probably a copyright violation.

To be removed under SYNTH, such a juxtaposition need only be "challenged or likely to be challenged". The challenge need not have any merit whatsoever. So a zealous user of SYNTH can go through removing juxtaposed material, which yields a new juxtaposition, which can then be removed, and so on until the article is down to a set of adjacent statements from one source.

Under common sense, of course, no one would ever do such a thing. And it's not used quite exactly that way. But the policy contains no limitation on how it can be used, no principle saying what is ok. And common sense is not typical of how SYNTH is applied.

If anyone who reads the two statements in the source (in good faith, and understands them) would agree that the conclusion is supported by the source without requiring any further argument, it shouldn't be considered OR. Certainly, if you were trying to get something published in a research journal and the only new content was such a straightforward inference, it would be rejected as insufficiently original. If such a journal wouldn't consider it original research, maybe that's a clue that it isn't. At least, back in reality it isn't. Under WP policy, it is.

There are cases where SYNTH makes sense, of course. WP should have a policy something like SYNTH. There just needs to be something in the policy to limit it to such cases. --Dan Wylie-Sears 2 (talk) 12:05, 21 May 2011 (UTC)

Well... instead of just complaining about it, would you suggest how we could improve it. I think most of the regulars to this page would agree that SYNTH is probably the single most difficult policy concept to explain. We have done our best, but that does not mean we couldn't improve it. Blueboar (talk) 14:44, 21 May 2011 (UTC)
I thought I had: add a statement or two of what's not SYNTH, such as the ones I offered. The version above is an on-the-fly first draft, so I'll do a first revision. "If a typical reader who understands two sourced statements would agree that the source supports the statement without further argument, then it's probably not SYNTH." And "If a publication that presents original research would decline to publish the conclusion solely on the grounds that it lacks originality, then it's probably not OR." --Dan Wylie-Sears 2 (talk) 01:35, 22 May 2011 (UTC)
I've changed my mind about that. Don't add it to the guideline. It would quickly get too long, and change-a-policy level of consensus might be too hard to get anyway. Instead, add links to essays at the bottom of the page, and a statement in the SYNTH section noting that the essays exist. I've started one at User:Dan Wylie-Sears 2/What SYNTH is not. Feel free to contribute to that one, either directly or on its talk page -- but as long as it's still in-progress on my user subpage, I'll feel free to revert edits to the essay itself without cause: after all, besides being on a user subpage it's also just an essay, so you can always write your own. --Dan Wylie-Sears 2 (talk) 07:14, 23 May 2011 (UTC)
Ah... I see from you edit history that you recently had something you added removed from an article on the basis of SYNTH. Suggest you talk it through on the talk page of the article (to find out what aspects of your addition were considered synthetic, and reach a consensus on how best to resolve the issue) rather than trying to change the policy to allow your edit. Blueboar (talk) 14:53, 21 May 2011 (UTC)
This isn't the first time I've seen SYNTH used in a way I didn't like. It's been long enough since the others that I don't remember the details. I do remember that they contributed to me having quit editing Wikipedia for a long while. And I'm pretty sure the edits that at issue in the worst cases of indiscriminate reversion/deletion were someone else's edits to an article I was contributing to, not mine. It's not about my one recent edit. I discussed that on the talk page of the user who reverted it. --Dan Wylie-Sears 2 (talk) 01:35, 22 May 2011 (UTC)
Most policies & guidelines can become problematic in individual cases if they are taken all to literally or treated overly dogmatic. The need to be approached with some common sense and in an individual case one always has to check, whether the WP goal (writing a good, informative, correct and readable article) is furthered by them or not. If their literal or dogmatic interpretation is not serving that goal, one needs to read them in a different manner or even resort to WP:IAR (in the context of that individual case).--Kmhkmh (talk) 15:00, 21 May 2011 (UTC)
I had a quick look and what you said about 'the code is determined...' seemed very unclear and to me. It would be better to cite a source for conclusions in cases like this so people can see what on earth an editor is trying to say and fix up the language. There are enough complication in the topic without compounding them, and there's quite enough sources without needing to start complaining here about people reverting things. Dmcq (talk) 15:12, 21 May 2011 (UTC)
The problem was indeed that my edit was somewhat unclear. The article previously said that there are no introns. I stuck in a couple sentences as a minimal patch instead of rewriting the entire section to avoid saying it. The solution is for someone to rewrite the section. There are probably numerous reasons why no one has. But one is the rule that all edits have to be sourced with sentence-by-sentence in-line citations to specific page numbers or else they'll be reverted. When a rule is enforced like that, you can't just ignore it.
That's also a symptom of an over-broad policy: The problem wasn't that I was posting original research, by synthesis or otherwise. It was that the writing was awkward and unclear. But clicking revert and saying SYNTH is easier than asking for clarification on the article talk page. You don't need to understand what the person was saying. Whatever it was, it'll probably count as SYNTH.
I'm not actually convinced that citing sources even really matters for high-school-level material, where there's no controversy over the actual material. I could get a high-school biology textbook and look up in the index where it mentions that there are introns in eucaryotic genes. Would anyone ever dig up that particular edition of that particular textbook, instead of looking in the index of whatever biology textbook comes to hand? Not a chance. Even if they did, would it help? Probably not. A book will tell the facts, but there's no question about what the facts are. What's unclear is which facts I was trying to express, whether what I wrote actually means what I think it does, and how it could be stated more clearly. --Dan Wylie-Sears 2 (talk) 01:35, 22 May 2011 (UTC)
One doesn't need to have citations on every line, that's just wrong. However what you wrote 'the code is determined by' just doesn't make sense that I can see. And you have gone and stuck it in again. The code 'just is', it is interpreted to produce things, it is not determined by the proteins or enzymes it produces. Do you mean determines? At least that would be closer I believe but then there are other problems like how the DNA is folded and epigenetic information which aren't coded in the DNA sequence. Dmcq (talk) 11:29, 22 May 2011 (UTC)
One doesn't need to have citations on every line, that's just wrong. Tell that to the mass-reverters and wikilawyers. Please. As for the substantive part about the genetic code article, I responded on the article's talk page. --Dan Wylie-Sears 2 (talk) 05:33, 23 May 2011 (UTC)
A thought that goes beyond this particular case... Many wp:synth violations are unintentional... with the synth being the result of a poor choice of wording. Such issues can usually be resolved with a very simple rewording of the material. Perhaps we should mention this in the policy. Blueboar (talk) 13:24, 22 May 2011 (UTC)
The vast majority of Wikipedia (basically everything except direct quotes) violates a literal reading of wp:synth. In the 90% of Wikipedia that works, what happens is that there is a near 100% consensus that it is accurate, and citable with the sources that got it right and somewhat support it literally. But anytime you have a rule that is written so that the norm is that you have to violate it, you have a problem. In the 10% where WP is a failure, it is used as a magic bullet for wikilwayers to knock out any material that they don't like, or engage in pissing wars. The answer to this seemingly complex quandary is actually very simple. Make it the norm when tagging/removing the material that the person briefly states a concern about the material beyond just "or/uncited/synth" when tagging / removing it. They don't have to engage in or win a debate on their concern, they just have to state it. That way, questionable/questioned material continues to be subject to the same standards and toughness under wp:ver/wp:nor, but abusive uses of wp:ver/wp:nor would be vastly curtailed. Sincerely, North8000 (talk) 15:12, 22 May 2011 (UTC)
I don't know whether SYNTH is being abused in general, but there was nothing wrong with the particular invocation of SYNTH that sparked this thread; I'd have reverted Dan Wylie-Sears' edit myself if I'd been the first to come across it. There's some confusion above because Dan Wylie-Sears' edit consisted of two parts. The first part was to fix a genuine but non-obvious problem in the article. The second part was to introduce a true but counter-intuitive fact that Dan Wylie-Sears' has deduced by synthesis. Dan Wylie-Sears might have avoided some frustration if he'd explained the motivation behind the first part using the edit summary, and perhaps made the second part as a separate edit. I've commented on the technical side of the edit at Talk:Genetic_code#Quotes_for_some_of_the_challenged_statements. Adrian J. Hunter(talkcontribs) 12:33, 23 May 2011 (UTC)
I note that that you also described it there as self-evident. If something is self-evident given what's cited, it's verified by what's cited. It falls within the letter of the prohibition, yes, but there's certainly nothing involved that even remotely resembles actual original research. This is a problem with the policy. If it's really about original research, that shouldn't be included; if that's really what's to be prohibited, it shouldn't be called "no original research". --Dan Wylie-Sears 2 (talk) 19:36, 24 May 2011 (UTC)

It sounds like the specific article edits are being discussed (appropriately) at the article talk page. I remind people that this talk page is for discussing the policy itself (also, I remind people that we have the NOR noticeboard to discuss edits to articles). So, unless someone is proposing an actual change to the policy, we can probably end this discussion. Blueboar (talk) 13:07, 23 May 2011 (UTC)

I suggested above that we add some description of what's not SYNTH, then changed my mind and suggested that we add links to essays. For example there's WP:NOTOR and the one I'm working on once I get it out of my sandbox and into the project namespace. WP:PG says Links should be used only when clarification or context is needed. I think clarification is needed for OR. (I also think that the policy should be drastically weakened, but the best way toward that is clarification.) --Dan Wylie-Sears 2 (talk) 19:21, 24 May 2011 (UTC)

SYNTH is bullshit. Every piece of text in this encyclopedia is a synthesis of various other texts; that is not just the nature of an encyclopedia, it is the nature of text. Attempting to write a text that does not do this is impossible without direct plagiarism; even that, in many cases, becomes subject to transformations of meaning based on context. If the intent is to prevent novel arguments from coming up, this is not the way to do it. Writing these articles is a work of interpretation, necessarily, because writing is a work of interpretation. The idea that we can have a document that is rigorously sourced (that is, all of its "meaning" can somehow be directly attributed to outside sources) seems to me to betray an understanding of import that is completely at odds with all modern understanding of texts; probably this is why we're having the kind of arguments above, because the fundamental impossibility of producing an article that conforms to "no original research" (viz., "no original thinking") is running up against the limits of writing. Graft | talk 01:58, 24 May 2011 (UTC)

Your argument is dependent on a theory of text which is not universally held to be true. Hermeneutics, for example, is a theory of text which holds a differing opinion. Fifelfoo (talk) 02:05, 24 May 2011 (UTC)
(ec; this responds first to Graft, then more generally) Notice that SYNTH doesn't actually forbid the synthesis of texts. It specifically forbids combining texts "to reach or imply a conclusion not explicitly stated by any of the sources". And I think the examples given in the section make it very clear what we're talking about. Yes, abstractly speaking, SYNTH is ridiculous; but, then, so is NPOV (neutrality doesn't exist) and OR (nothing is "original" in the sense of having no precursor, and everything is original in the sense of not being an exact copy of other sources). But, in practice, I've found that it isn't actually all that hard to distinguish "good" synthesis from "bad" synthesis. Maybe I just haven't seen enough, but in the situations I've seen, there's usually one editor arguing that something isn't synthesis, and a half a dozen or more (often including uninvolved editors, if the issue foes to WP:NORN) saying it is. It's really no different than NPOV--one person swears up and down that their edit is neutral because it's the truth, when a bevy of other editors say, "No, you think it's the truth, but the sources disagree". I think SYNTH works pretty well, most of the time--no better or worse than our other (intentionally) vague and open-to-interpretation policies/guidelines. Qwyrxian (talk) 02:17, 24 May 2011 (UTC)
It's not that the sources disagree, at least in the cases that are relevant to possible changes in the policy. It's that the sources do agree but aren't explicit. If everyone agrees that reading and understanding the sources would force you to the conclusion (even if they think the sources are wrong about it -- this is about verifiability not truth, after all), it should be considered verifiable. But it isn't, unless the sources spell it out. --Dan Wylie-Sears 2 (talk) 18:58, 24 May 2011 (UTC)
Well, my implication is that to some extent, combining of texts will always imply something not stated by any of the sources - give or take. And as we're seeing in the argument above, with the debate turning on a single particle, SYNTH produces problems because it is a fine enough instrument that people actually do run up against its limits.
I think my primary complaint is that none of these things (NPOV, OR, SYNTH) are really useful except as vague principles (emphasis on vague), but they have been enshrined as "Wikipedia policy", which means they are mostly good as rhetorical weapons in arguments: "You are violating this wikipedia law, end of argument." SYNTH seems most easily abused in this regard, since it's the vaguest, and therefore the most broadly applicable in service of bad arguments. I'd rather see people forced to engage in an argument on its specific merits than resort to a bureaucratic line of reasoning. Ultimately I think it serves no purpose, because it doesn't actually erase the dispute - that will have to be resolved by actual argumentation anyway - it just introduces an unnecessary appeal to authority. If we really do believe none of these things, such as neutrality, are really possible, then we should understand them as limits and guidelines, not as rules, and the "policy" should be appropriately soft. Graft | talk 20:25, 24 May 2011 (UTC)
Quite the opposite; NPOV, NOR, and SYNTH are typically the last defense against the egregious POV-pushing one finds incessantly forced onto articles, and which brings Wikipedia into disrepute. One typically sees it in this kind of formulation:

According to the liberal[1][2][3] New York Times, Mr. Smith was doing X.[4] However, the New York Times also published articles by Jayson Blair, who plagiarized and fabricated parts of his stories.[5]

Feel free to substitute any other source for New York Times. While it's not typically so blatant, one certainly often sees political or similar labels pre-pended to sources as a means of [[poisoning the well]. The point is, people typically use SYNTH to invent their own arguments against a source (including claims regarding their politics, overall accuracy, and any number of other of characteristics that are typically irrelevant to the situation being described), and that's exactly what we don't want editors to do. Jayjg (talk) 20:29, 29 May 2011 (UTC)

Need to clarify image sourcing policy (WP:OI)

I think we need to clarify our policy on the types of sources required for self-published images. Currently, the policy says: "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." What is not clear here is the part that says "so long as they do not illustrate or introduce unpublished ideas or arguments" -- what is meant by this?

As an example, I am currently engaged in a dispute over File:A vietnamese Professor is pictured with a group of handicapped children.jpg. An editor is claiming that the image cannot be used because it is "self-published", and is saying that it should be removed on the grounds that WP:NOR applies to "anything challenged or likely to be challenged". However, WP:OI seems to say that reliable sources are not needed for images that are not "illustrating unpublished ideas". As far as I can tell, the only thing this image is "illustrating" is the existence of an OBGYN clinic in Vietnam that treats children who have been deformed as a result of Agent Orange exposure. This clinic can clearly be seen to exist by a Google search for its name (Tu Du Obstetrics and Gynecology Hospital), and it can also be verified that the woman in the photograph (Nguyen Thi Ngoc Phuong) works at this hospital. Furthermore, there are numerous non-free images that depict these clinics, so we have no reason to doubt that the situation depicted here is unrealistic. I personally see no reason why there is any reason to doubt the validity of the photograph, and the user challenging it has not provided any reason to doubt it, and is essentially just saying "Doesn't matter, I'm challenging it, so it has to go." The current rules are not clear enough about what to do here. What precisely are the criteria that we use to determine when an image is "illustrating or introducing unpublished ideas or arguments" or when it is not? Where do we draw the line?

Obviously, we don't want people being able to post a photo saying "This is a picture of Bigfoot, taken in Northern Canada". An image purported to be a photo of Bigfoot would be implying that Bigfoot exists, which is a WP:FRINGE view. On the other hand, in this case, we have numerous sources that say that (a) clinics such as the one depicted exist, (b) this specific clinic exists, and (c) the woman depicted works at the clinic in question, and numerous images exist that show similar situations. That is, there is no reason that I can see to doubt the validity of the image. Are photos such as this one to be considered as unreliable Bigfoot photos? If not, then how do we make it clear that even though this particular photo isn't published in a reliable source, that it is not in the same category as a Bigfoot photo? What is it about this photo that sets it apart from a Bigfoot photo?

This issue seems to pop up quite often, and I feel that we need to have a more clear policy statement on this. How could we rewrite the quoted passage above to indicate more clearly when we are to deny the use of an image, on the grounds that it is unreliable/dubious? For one, I would suggest requiring some evidence on the part of the user challenging the validity of the image that explains why they think the image is possibly not real/accurate. Also, I would clarify what is meant by "unpublished ideas or arguments". And how does this all relate to the general NOR requirement of requiring a source for "anything challenged or likely to be challenged"? -- Jrtayloriv (talk) 05:12, 25 May 2011 (UTC)

It may be a different issue. Pictures, regardless of source, may be used to create an emotional response that favors a POV. Editors should be careful that the choice of pictures does not affect the neutrality of the article. For example, BLP says, "Images of living persons should not be used out of context to present a person in a false or disparaging light." However I do not see any policy that restricts the use of user-created images. TFD (talk) 05:34, 25 May 2011 (UTC)
I understand that in some cases there might be issues with misusing images to promote a POV (i.e. putting 15 gory photos of aborted fetuses in an article about abortion). However the fact that Agent Orange and dioxins cause birth defects, and that there are clinics in Vietnam that treat children with these deformities is not a "POV", but a well-documented fact, and thus including images of these health effects or of clinics that treat them is not pushing a POV, but simply illustrating a fact that is the consensus view of the scholarly community. Furthermore, the user in question was not challenging it based on neutrality, but on the fact that it was not published in a reliable source and was "self-published". Regardless, the image in question was only used as an illustration of a more general problem (i.e. lack of clarity on when it is/isn't appropriate to use user-created images). Whatever the determination is on the particular image mentioned above, I am more interested in solving the general problem, which has come up repeatedly[34] [35]. -- Jrtayloriv (talk) 05:50, 25 May 2011 (UTC)
The reasoning behind removing this photo from Wikipeda pr WP:NOR is quite ridiculous. Please provide diffs where exactly An editor is claiming that the image cannot be used because it is "self-published", and is saying that it should be removed on the grounds that WP:NOR.--Termer (talk) 06:15, 25 May 2011 (UTC)
Sorry about not providing diffs -- it was the only discussion on the talk page, so I didn't think it was necessary. Here they are: [36] (see #3 - "are self-published photos by Alexis Duclos; no editorial oversight and not a reliable source.") ... [37] ("anything challenged has to be attributed to an RS. Self-published") -- Jrtayloriv (talk) 06:46, 25 May 2011 (UTC)
PS. the image is not even "self published" (not that it wold matter in such case) like claimed but licensed through the OTRS.--Termer (talk) 06:24, 25 May 2011 (UTC)
I don't see a problem with WP:OI. It seems unambiguous to me that your image is ok with respect to NOR. The real issue is that images of birth defects inevitably elicit a strong emotional response, making it difficult to maintain NPOV. But this isn't the place for that discussion. --Dan Wylie-Sears 2 (talk) 11:25, 25 May 2011 (UTC)
The picture seems to illustrate the section of the topic and not push anything extra forward. If the section is changed to say agent orange is alleged to cause injuries and the picture said it did cause injuries there would be a problem, but complaining when the text does not alleged is just wrong. And even if the text had said alleged which it doesn't that particular picture would probably be okay if it just said alleged victims. Dmcq (talk) 11:59, 25 May 2011 (UTC)
I think it's problematic to use that image in the Agent Orange article without a high-quality reliable source saying what it is. The caption, in particular, needs a reliable source just as any other part of the article does. See WP:V. SlimVirgin TALK|CONTRIBS 18:20, 25 May 2011 (UTC)
Given that the woman in the photo works in the clinic in the photo, and that numerous similar photos exist of this and other clinics, could you explain why you feel that it is "problematic" to use the image, and then explain how you would suggest changing WP:OI to be more clear on this type of thing? -- Jrtayloriv (talk) 18:30, 25 May 2011 (UTC)
There is no requirement for illustrations that they be actual photos, this could be a reconstruction and provides it illustrates what it is supposed to illustrate and does not advance anything new then according to Wikipedia policies it is acceptable. The same sort of thing happens about examples, there is no requirement that examples be taken from a book, only if they advance something new rather than illustrate something that's said does all the business about citation apply. All that matters is a that it is accepted as a faithful illustration. Dmcq (talk) 18:52, 25 May 2011 (UTC)
I don't know what the image was being used to show, or what the caption was. But if it's being used to illustrate something contentious, we need to see the provenance of the image, i.e. we must have reason to believe that it is what we say it is, and that it's directly relevant to the article. We also need a reliable source for any caption. Who is saying that the children in that image were affected by Agent Orange? Were some affected, all? What is the source? SlimVirgin TALK|CONTRIBS 18:55, 25 May 2011 (UTC)
I'm not familiar with the scientific background to this. Agent Orange was sprayed in the area in the 1960s. The disputed image of the children was taken in 2004. What is the scientific link between Agent Orange and the children in the photograph? SlimVirgin TALK|CONTRIBS 19:05, 25 May 2011 (UTC)

At first glance what may really be of contention (and most questioned) is the statement in the caption of the image as well as the implied statement made by putting it in the article. North8000 (talk) 03:23, 26 May 2011 (UTC)

Yes, the caption needs a source, per V: "This policy applies to all material in the mainspace—articles, lists, sections of articles, and captions ..." SlimVirgin TALK|CONTRIBS 03:32, 26 May 2011 (UTC)
I agree with you that the caption needs to change (probably to something along the lines of "Staff and patients in the Tu Du Obstetrics and Gynaecology Hospital in Ho Chi Minh City, Vietnam; Tu Du provides medical treatment to children suffering from congenital disorders that are believed by some to be the result of prenatal exposure to Agent Orange"), but I don't think that captions for original images always need a source, "so long as they do not illustrate or introduce unpublished ideas or arguments". WP:IMAGE says "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. For example, if you take a picture in your neighborhood, you do not need to produce a published, independent reliable source to prove that you took the picture in your neighborhood. However, if such sources are available, please provide them." ... what I'm asking is when we stop assuming good faith on the part of the image author, and at what point we claim that an image is illustrating or introducing unpublished ideas or arguments. Right now, there seems to be a clash between WP:V and WP:OI/WP:IMAGES. What I'm trying to determine is how we resolve this. I personally don't believe that for cases such as this one, where we don't have any reason to believe that the photo is a fake, that we should assume good faith on the part of the author. -- Jrtayloriv (talk) 15:14, 28 May 2011 (UTC)


OK, thanks for the diffs Jrtayloriv. I'm getting this now and it seems it's not about "self published". The question should be: is this photo relevant to the article about Agent Orange since its Description claims "...children, most of them victims of Agent Orange."? So new questions arise, how many children and who exactly on the image are the victims of AO? So yeah, better use images that can be verified what it shows exactly.--Termer (talk) 07:08, 26 May 2011 (UTC)

Except that in practice that's often impossible - many things can only be illustrated by self-produced images (due to copyrights etc.), and a self-produced image is never going to have received attention from a reliable source. --Kotniski (talk) 07:29, 26 May 2011 (UTC)
The question is whether it is relevant and faithful to the associated text. There is no citation needed tag on the associated section and it says things like 'The center provides medical care, rehabilitation and vocational training for children and veterans from Vietnam who have been affected by Agent Orange.' so there does not seem to be a dispute that something like this exists. It looks to me like what it says it is and I'd need some evidence before saying the person who uploaded it was wrong. In these circumstances we do not need each person in the photo identified and medical histories to be registered with the Wikipedia Foundation. Basically it is an illustration rather than a new fact and falls under WP:OI. The purpose of WP:V is 'To show that it is not original research, all material...' and OI says 'Original images created by a Wikipedian are not considered original research, so long as they do not illustrate or introduce unpublished ideas or arguments'. Dmcq (talk) 09:25, 26 May 2011 (UTC)
Taking my previous comment a step further, I think that all of the following are "claims" associated with images which are not a part of the image itself:
  1. Statements made in the file name of the image
  2. Statements made in the caption of the image used in the article
  3. Statements implied by placement of the image into the article.
I think that #2 and #3 can and should be subjected to wp:ver and wp:nor in the article. #1 is somewhat "invisible" and sort of slips through the cracks (?) but maybe one could say that nothing controversial/questioned can build upon it or use it as a basis. North8000 (talk) 13:15, 26 May 2011 (UTC)
Well you're certainly welcome to suggest a change in the policy in WP:OI. Dmcq (talk) 13:37, 26 May 2011 (UTC)
The problem remains that the statement "this image depicts...", which is made (explicitly or, usually, implicitly) in almost every image caption, is never going to be WP:verifiable for any self-created image (since no reliable source will ever have commented on that image). So in a sense, image captions are exempt from usual sourcing requirements to a certain extent. --Kotniski (talk) 13:52, 26 May 2011 (UTC)
Other places that would probably need changing for what you say you want are the guideline WP:Images#Pertinence and encyclopedic nature and the policy WP:Image use policy#Image titles and file names. Dmcq (talk) 13:54, 26 May 2011 (UTC)
Good point; in fact under the first link is the statement Generally, Wikipedia assumes in good faith that image creators are correctly identifying the contents of photographs they have taken. For example, if you take a picture in your neighborhood, you do not need to produce a published, independent reliable source to prove that you took the picture in your neighborhood. Which is the proviso I had in mind above (and which should probably be noted in this policy as well).--Kotniski (talk) 14:00, 26 May 2011 (UTC)
I think that Wp:ver and wp:nor as-is can be used for #2 and #3. Wp:ver and wp:nor need a change (which I have been promoting) that says a person should be asked to question something about the material when invoking wp:ver / wp:nor. (I.E. to actually implement / follow "challenged or likely to be challenged" ) With that huge caveat, I think that Wp:ver and wp:nor should be and are applicable to #2 and #3. North8000 (talk) 14:05, 26 May 2011 (UTC)
I'm not certain I understand what you're saying but I get the impression you want to remove everything said in the 'Pertinence and encyclopedic nature' section, especially for instance "a photograph of a trompe-l'œil sculpture of a cupcake may be an acceptable image for Cupcake", and stop any assumption that Wikipedia image creators correctly identify photographs? Dmcq (talk) 14:25, 26 May 2011 (UTC)
Basically, I'm only talking about the captions in articles, and what is implied by an instance of placement of an image in an article. I'm not talking about the images. And I'm not advocating any change to the image guideline. North8000 (talk) 15:20, 26 May 2011 (UTC)
The image description says 'English: Ho Chi Minh. Professor Nguyen Thi Ngoc Phuong, at Tu Du Obstetrics and Gynecology Hospital is pictured with a group of handicapped children, most of them victims of Agent Orange'. Have we some slight midgin of a doubt that would cause us not to assume in good faith that the image creator has correctly identified the contents of the photograph? Because the caption isn't saying anything more. And the illustration seems to me to be a faithful illustration for that section. Dmcq (talk) 17:01, 26 May 2011 (UTC)
That version of AGF could be used to argue against the existence of wp:ver and wp:nor.  :-). I wasn't weighing in on this particular situation. But now here goes. I think that the "self-published" argument for removal by the challenger is not valid. The end. Nobody has challenged the caption and so discussions about the caption are a moot point. North8000 (talk) 17:30, 26 May 2011 (UTC)
That version of AGF as you put it is exactly what WP:Images#Pertinence and encyclopedic nature says we should assume for illustrations. Dmcq (talk) 22:45, 26 May 2011 (UTC)
Thanks for that info. However, taking that categorically to include other assertions buried in the image description (and then) caption could be quite extreme. For example, take a picture of a person and say "photo of person who's life was ruined by the socialist policies of the Obama administration" could that be challenged for sourcing or it is AGF? :-) Being not a policy, wp:ver/wp:nor might take precendenc over it, or at least if the description ends up as a caption in an article. In fact, maybe that is the way to make a distinction. Sincerely, North8000 (talk) 01:20, 27 May 2011 (UTC)
By the current reading that probably would be okay if the image was truthy enough and the article was saying people lives were ruined by the Obama administration. If somebody has some sort of real doubt about its veracity though they should remove it and say why, good faith only goes so far. Dmcq (talk) 12:08, 27 May 2011 (UTC)
Yes, there's a difference between "saying where you took it" and trying to make some (possibly unsourced) point on the basis of it. Like in the case under discussion - we can take the photographer's word for it that the picture was taken at a particular centre on a particular date; or probably even that it showed Agent Orange victims if it was well sourced that many disabled Vietnamese children were Agent Orange victims; but if the last assertion is not well sourced, then we shouldn't accept any usage or captioning of the image that would imply such an assertion.--Kotniski (talk) 16:53, 27 May 2011 (UTC)
I think that that sums it up. BTW, probably the most-likely-to-be-questioned statement here is that that there are child victims in a 2004 photo. This is a statement that this is happening to people 30 years after the stuff was flying around. E.G that there is a new exposure pathway, or that it is inherited etc. North8000 (talk) 17:49, 27 May 2011 (UTC)
Again, my apologies if I'm just uninformed and everyone else knows the answer to this. Agent Orange was sprayed in the area in the 1960s. The disputed image of the children was taken in 2004. What is the scientific link (real or perceived) between Agent Orange and the children in the photograph? SlimVirgin TALK|CONTRIBS 18:08, 26 May 2011 (UTC)
SV, why don't you read the damn article instead of interjecting here? Agent Orange contains persistent pollutants - dioxins don't break down easily. This dispute probably doesn't belong on this page anymore, and should move back to Talk:Agent Orange. Graft | talk
Okay, but there's no need to be aggressive just because someone asks a very obvious question. Is there a source that says children being born in the 2000s are suffering from birth defects because of Agent Orange? I'm trying to establish whether this is widely accepted as fact, or is a significant-minority POV, or a tiny-minority POV, because that will tell us whether using this image is contentious.
There is then the separate question of whether most of the children in that image fall into that category, as the caption says. SlimVirgin TALK|CONTRIBS 18:33, 26 May 2011 (UTC)
Responding to Graft, I read the article and it didn't answer that question. Persisting in the environment does not equate to saying a significant exposure pathway still exists and is creating new cases. North8000 (talk) 18:49, 26 May 2011 (UTC)
There's a CBC article here that interviews scientists, and concludes: "But on what many consider to be the key question: is dioxin contamination the cause of the high rate of birth defects in Vietnam, the study provides no answer." SlimVirgin TALK|CONTRIBS 19:14, 26 May 2011 (UTC)

Back to WP:OI

I think that discussion related only to Agent Orange should be taken Talk:Agent Orange. I only used that image as an example, and did not wish for this discussion to center exclusively around that image. Given the above discussion, are there any suggestions for how to change WP:OI to make it more clear when it is inappropriate to use original images without reliable sources? -- Jrtayloriv (talk) 15:14, 28 May 2011 (UTC)

I would say we should include the two sentences from the guideline which I quoted above (that we assume good faith re where the photo was taken etc.), and perhaps give an example of the limits of that good faith (i.e. something(s) that would not be acceptable in an image caption).--Kotniski (talk) 15:20, 28 May 2011 (UTC)
I agree with you that the guidelines from WP:IMAGE should be included, and that an example would be helpful. If nobody else has gotten around to it by later today, I'll post up a suggested revision that incorporates each of these. -- Jrtayloriv (talk) 15:38, 28 May 2011 (UTC)

I also think that incorporating some of the concepts from Wikipedia:Images#Pertinence_and_encyclopedic_nature is a good idea (or perhaps referring the reader there). Since I'm not always convinced that people will click through, here's what it says:

I think that this strikes the right balance. In particular, I think that it is correct to focus on what the reader sees, as opposed to how the image was formed. In this article, it's more important that the reader see images of kids with disabilities that are reasonably consistent with AO exposure (something that the editors can determine by reading about the birth defects or looking at copyrighted images in reliable sources), than that the reader see images of specific, identifiable humans who are certified to have a disability that was absolutely, definitely, unquestionably caused by AO exposure (since that doesn't actually exist: we can tell what's typical for more highly exposed people and what approximately percentage of birth defects could be attributed to AO, but when it comes down to individuals, there's simply no way to differentiate between 'this person is missing a leg because of AO exposure' and 'missing a leg because of thalidomide secretly taken by the mother' or 'missing a leg for some other reason'). WhatamIdoing (talk) 23:28, 29 May 2011 (UTC)

"Controversial" in first sentence of lede, based on primary source

A number of editors have been insisting that the first sentence in the lede of an article describe an individual as a "controversial" religious leader, because one source says he described himself that way. They argue that, despite WP:LABEL specifically stating one should not use the bare word "controversial", he did a lot of controversial things, and we do have one source in which he describes himself that way, so it must go in the first sentence (or at least the lede). A larger number of more experienced editors disagree, also citing WP:PRIMARY as an issue, but so far have made no headway. Should an article use Wikipedia's voice to describe an individual based on their own self-description in one source? If possible, it would be very helpful if members of this board could express their views at Talk:Elazar Shach#Shach - "controversial and divisive"?. Jayjg (talk) 19:37, 29 May 2011 (UTC)

Translations into English of album and song titles

Should Wikipedia articles include English translations of album and song titles that were released only in foreign languages?

Some editors consider that this is WP:OR and therefore not allowable, so have renamed articles about Chinese albums from English translations to romanisation, e.g. Sky to Tiān Kōng, and removed English translations of song titles.

I think there is a consensus that it is appropriate to include the original language, with romanisation where applicable, especially where different translations are possible. The draft guideline MOS:CHINESE takes this approach.

But as this is an English language encyclopedia, (Q1) shouldn't English translations always be included? And (Q2) in cases where English-language sources do not mainly refer to an album in the original language, should article titles generally use an English translation? - Fayenatic (talk) 14:03, 12 May 2011 (UTC)

It is not OR to translate in good faith (although the exact translation may be challenged) ... but that does not answer the question of whether to translate the title of a given song/album or leave it in non-English form. That decision is a function of WP:COMMONNAME and is determined by looking at what English language sources use when discussing the song or album. If a significant majority give the name in translated form, then so should we. If a significant majority give the name in untranslated form, then we should do so. If there is no significant majority, we try to reach a consensus on which is best. (Note: if you use the non-English form for the article title, I think it very appropriate to give a translation in a parenthetical in the opening sentence) Blueboar (talk) 14:35, 12 May 2011 (UTC)
In a sense it is, depending on how its used, by adding implications and such. English translation is added when there's no emphasis such as Putting the Original name first, WP:COMMONNAME i don't trust too much and could be more refined. Lets not used proposed guidelines such as MOS:CHINESE as its only proposed, not a real one. English wikipedia doesn't meant we have to give translations to everything.Bread Ninja (talk) 01:59, 13 May 2011 (UTC)

I fully agree with Blueboar. WP:COMMONNAME is a WP policy about article names, so it must be followed until it gets changed (unless that results in such a plainly stupid result that WP:IGNORE applies, which is hard to imagine). So that settles my second question (article names).

As for my first question (English translations of song titles etc within articles), I agree with Bread Ninja that the original language should come first. However, I disagree about the need for translations. In English WP, foreign words should often be followed by romanisation if applicable, and always by English translations. If the translation is arguable or ambiguous, then multiple translations can be given. Any "implications" can be set out explicitly; see the examples of "Face", "Thee" and "Tong" on Scenic Tour. - Fayenatic (talk) 13:51, 13 May 2011 (UTC)

I would definitely agree that the lede sentence of the article should give both the original non-english "name" of the song or album and a translation. Which comes first depends on which we used for the Article Title ... there is no reason why some articles can not start: "Non-english Name (English translation) is a song by the Russian rock band X." while others start: "English translation (Non-english Name) is a song by Japanese singer Y". - Blueboar, 14:40, 13 May 2011 (UTC)
You can't use another related article that uses the very thing in question. That's like saying its ok to add original research on Pokemon the movie, just because Pokemon (manga) does the same thing (its called being bias). If a translation is ambiguous enough to put in multiple translations then yes, but for a title of media or so. We don't need to give translation for every foreign title of each track. It causes redundancy. And implications and such are definitely original research and should not be added in unless its verified. Same with the title of the albums. IF the common name is in english on reliable english sources, than yes, it can be used. but everything has to be verified.Bread Ninja (talk) 13:26, 14 May 2011 (UTC)
I concede the point about unverified implications being WP:OR, and will remove these where I find them. - Fayenatic (talk) 15:34, 29 May 2011 (UTC)
A subtle correction to Bread Ninja's last statement... everything has to be verifiable... it does not always have to actually be verified in the article. Blueboar (talk) 15:13, 14 May 2011 (UTC)
And that's the flaw i see in WP:COMMONNAME.Bread Ninja (talk) 22:32, 14 May 2011 (UTC)
Um.... I don't I follow. How is that a flaw with WP:COMMONNAME? Blueboar (talk) 22:13, 15 May 2011 (UTC)
And how are translations in track listings "redundant" to readers who cannot understand the original language? - Fayenatic (talk) 13:02, 16 May 2011 (UTC)
Its redundant to give a translation for everyone. Its a list, and the translation will take up just as much as the original name. You would have to verify the translations, because it will be added in the tracklist. Its different from merely adding them in next to a title for mere convenience of what the title means, and actually putting them onto a tracklist where it can be seen as the official translation to the tracklist. WP:COMMONNAME as long as it "can" be verified but not actually being "verified". That is highly subjective.Bread Ninja (talk) 19:05, 16 May 2011 (UTC)
In terms of space, translations looks fine using the template that you provided here. Would the column heading "Possible translations" be better?
You are getting muddled between my two questions. WP:COMMONNAME is only relevant for Q2, and there is no dispute that verification is desirable for translations in article names. However, as for Q1, it's not essential to cite references to translate song titles in track listings, as any that are wrong will sooner or later be changed/challenged by other readers. Verification could also be aided by using template:linktext to link to the original language for each word/phrase in Wiktionary, e.g. 浮躁 for Restless (Faye Wong album). - Fayenatic (talk) 20:04, 16 May 2011 (UTC)
"possible translations" seems even more off.....Its best, to just have them without translating every track. translating the album title is ok. But that doesn't mean we use the first english translation first. Especially if there is no source saying so. COMMONNAME is relevant no matter what. So we can't question it despite having a lil too much freedom for subjective things. And if they were verified, and the translation wasn't exact, then we will still use it, regardless of it wasn't exact. It happens alot in japanese. The point to Wikipedia isn't to add the truth, but to add the information we can back up. translation for every track is unnecessary. Maybe individual tracks mentioned in prose is ok, but in a track list i dont think its a good idea. It would lead some to believe that it the official translation. In prose it wouldn't as it's less confirming.Bread Ninja (talk) 19:18, 19 May 2011 (UTC)
There are a number of precedents for such article - for example, the Dutch national anthem Het Wilhelmus. Martinvl (talk) 20:15, 17 May 2011 (UTC)
Again, that's WP:OTHERCRAP. I will only trust a given featured article or good article of the same type. In fact, i believe that translation wikitable is hurting its chances.Bread Ninja (talk) 07:05, 23 May 2011 (UTC)
I think that Martinvl made that illustration in relation to Q2, as an example of WP:COMMONNAME where the article name should be foreign because that name is widely used in English media too. Nobody disputes that.
His illustration is also relevant to a third question, about having an English translation of the full lyrics. IMHO this will always be justified for an article about a national anthem: Wikipedia readers want to know what the song means. But it doesn't help with my Q1 about translating the track list in an article about an album.
Let's take this one step at a time. In the case of a non-Roman script, do we all agree that (a) romanisation is generally desirable in English wikipedia, including track lists, as English readers may otherwise be unable to relate to the text at all? And (b) is it still desirable to include romanisation even where it is merely verifiable rather than verified? - Fayenatic (talk) 17:22, 23 May 2011 (UTC)
Do we have RS that use the translation in question? If so use them. If not then it must not be all that notable. Hcobb (talk) 19:40, 23 May 2011 (UTC)
Sorry, that doesn't follow. There may be plenty of RS in the original language, but few in English. - Fayenatic (talk) 21:45, 23 May 2011 (UTC)
The motto of Greece is "Eleftheria i thanatos (Greek: Ελευθερία ή θάνατος, English: Freedom or death). IMO this format should be used for song titles.Martinvl (talk) 21:21, 23 May 2011 (UTC)
You mean romanisation + original + translation? I agree. The romanisation and the translation are both verifiable from the original. And without them, the article falls short of making sense or being fully useful to an English-only reader. Is it very different from rewriting copyrighted material in our own words; verifiable (by someone with relevant expertise) as facts from the original? - Fayenatic (talk) 21:45, 23 May 2011 (UTC)
Again, i can agree for titles, such as title of an article, or a name that will constantly be mentioned within the article. And of course there may be alot of RS in the original language, and few in the english, but that doesn't mean its an excuse to not look for it.Bread Ninja (talk) 08:49, 24 May 2011 (UTC)
OK. That leads back to a point that you raised above, on 19 May, which you said happens a lot with Japanese: reliable sources may use/repeat incorrect translations. For instance, the Chinese album title 陌生人 means Only Love Strangers (anyone who can't read chinese can verify that using the wiktionary links in that text). It was only ever sung in Chinese, but the Japanese release had a CD insert with a Japanese translation of the lyrics, and this carried an English subtitle "Lovers and Strangers"; it seems that this was used as an English title by sources such as BBC and Yahoo Music. Amazon used the incorrect translation for the album, but the correct one for the track name. The correct translation is verifiable by anyone with knowledge of the language or a dictionary. It is also backed up by Ghits: 55,000 for faye wong "only love strangers" compared to 8,800 for faye wong "lovers and strangers"; even if the 55,000 are mainly non-reliable sources, the reason why they are in the majority is that they are verifiably right!
If the incorrect version was widely and consistently used in RS then I accept that WP:COMMONNAME would apply and we would have to use it. However, in this case I do not think the usage of "Lovers and Strangers" is established so strongly that we would have to follow it.
In this situation (i.e. where COMMONNAME does not determine the answer), I do not agree that WP:V restricts Wikipedia to repeating errors from WP:RS if they are demonstrably (verifiably) wrong. Read WP:V in a nutshell: facts only need citing from WP:RS if they are likely to be challenged.
It would be useful to have a few more contributions here to reach a consensus on allowing translations that are only verifiable as opposed to verified. - Fayenatic (talk) 15:32, 29 May 2011 (UTC)
Which is why i don't trust WP:COMMONNAME a lot, and why certain wikiprojects (such as the anime and manga wikiproject) ignore it to a degree, because direct translation that sources use (regardless of reliable) versus the many that are reliable (but the number is smaller). It would be much more easier if they used common name based on the sources in the article instead of proving it outside of wikipedia.Bread Ninja (talk) 01:40, 30 May 2011 (UTC)
Sorry, I don't understand. Are you arguing in favour of using wrong translations from WP:RS? If so, would you include a statement giving the correct translation, which is verifiable even if unsourced? - Fayenatic (talk) 08:37, 31 May 2011 (UTC)
If they are verified, than yes. Wikipedia isn't about telling the truth, but telling what can be verified. So if reliable sources translate it wrong, than we go with that. Otherwise, not use translations at all. If a correct translation can be verified (not using a translator) then we're good.Bread Ninja (talk) 09:08, 31 May 2011 (UTC)
Well, you're out of step with Blueboar's advice above - verifiable is sufficient, not everything has to be verified i.e. cited from RS. WP:V says sources are required for what is likely to be challenged; and a correct translation, linked to Wiktionary, is not likely to be challenged. - Fayenatic (talk) 17:33, 31 May 2011 (UTC)

A word by word translation can be very wrong, especially for the titles of creative works. Even character by character translations can fail for things like Japanese to English, because Japanese can't handle most trailing consonants and has one sound that covers both the L and R of English. For example translate Urusei Yatsura into a short English phrase that means exactly the same thing. Hcobb (talk) 17:44, 31 May 2011 (UTC)

Good point. As for article titles, Urusei Yatsura is presumably referred to in English RS mainly by the romanisation, so that should be the title anyway by WP:COMMONNAME. Within articles, however, a translation of foreign text is never "redundant" (as alleged above) for readers who don't know the original language; so where this is difficult, it is better to give alternative translations rather than none. This should include song names in an album track listing.
I've only just realised that WP:NOR#Translations and transcriptions already says that faithful translation is not WP:OR. Sorry for not looking properly before posting here. - Fayenatic (talk) 12:49, 2 June 2011 (UTC)
Its nto about informing everything, but informing what we could verify already. A track by track transaltion shouldn't be done. If however emntioned in prose, than yes. a tracklist is merely a tracklist. its not really changing much inth e article than mentioning what the album contains.Bread Ninja (talk) 12:54, 2 June 2011 (UTC)
"Shouldn't be done" - why? You said it was original research; that was wrong. Give a policy reason. - Fayenatic (talk) 13:22, 2 June 2011 (UTC)
i have. think of it like this, when in prose, the given track or title is noted. Romanization is one thing so that people know what the given characters are. But its not like wikipedia is a translator of all foreign words mention. Please refrain from exclamation points. it could probably be noted in the note section as it just gives small details. But not a complete section dedicated to the translations. Why? because it passes off more than just a faithful translation.Bread Ninja (talk) 13:35, 2 June 2011 (UTC)
You haven't cited particular paragraphs of WP:V, WP:NOR etc to support your POV. What do you mean by "note section", "complete section"? - Fayenatic (talk) 13:42, 2 June 2011 (UTC)
the tracklist has a note section showing how the actual track is spelled or giving special notation that doesn't fit in particuliar one. However, you added an extra note dedicated to translation. I've said all i could, but really its a matter of trivia WP:TRIVIA, but thats just as subjective as keeping the translations per track that aren't mentioned other than tracklist.Bread Ninja (talk) 13:46, 2 June 2011 (UTC)
Ah, you mean like this. I was following Template:Track listing/doc which suggests using the "Notes" field, which is displayed in small print, for the original script. Is this better? I think so; it looked wrong to show Chinese characters in small print as (i) they are the original, and (ii) it is hard enough to read them at normal size. - Fayenatic (talk) 16:56, 2 June 2011 (UTC)
I think so too on that one, but i could ask about it in the given wikiproject to see what could be done about that.Bread Ninja (talk) 17:00, 2 June 2011 (UTC)
Good, please do. The outcome needs to take account of English/Roman-only titles within foreign albums, e.g. "MV" on that one, or Everything. - Fayenatic (talk) 17:18, 2 June 2011 (UTC)

Primary sources - a problem with the wording

I know this issue has come up before, though I did not find an extensive discussion of it (forgive me if that is incorrect). The policy on primary sources reads:

A primary source may only be used on Wikipedia to make straightforward, descriptive statements that any educated person, with access to the source but without specialist knowledge, will be able to verify are supported by the source.

The intent of this is fine: knowledge brought from the primary source into wikipedia must be clearly present in the source itself and must not require an expert interpretation or synthesis of the source with anything else. No problem with that. The problem is that it does not consider the skills required to read the source. It is not true, for example, that "any educated person" can read Greek or understand mathematical notation (or physics jargon, etc), so a literal application of the rule would entirely eliminate all primary sources written in a foreign language and most original scientific papers. This would be a disaster, and I daresay we are lucky that the editors of scientific articles in Wikipedia usually ignore this rule and nobody complains. I think we should try to rephrase this rule to take care of the problem without harming the real intent of the rule. Here is a first attempt:

A primary source may only be used on Wikipedia to make straightforward, descriptive statements that any educated person, familiar with the language of the source but otherwise without specialist knowledge, will be able to verify are supported by the source.

(I'm assuming that "language" includes "jargon, symbolic notation, etc".) Zerotalk 04:57, 29 May 2011 (UTC)

Sounds good to me, but maybe retain the reference to access as well. I would also prefer "to support...(statements)" rather than "to make...".--Kotniski (talk) 06:02, 29 May 2011 (UTC)
I thought that "with access to the source" was too obvious and applies to all cases of verification, not just primary sources. But I'm not opposed to it. Zerotalk 09:26, 29 May 2011 (UTC)
I disagree, you are making an unwarranted assumption about the skill set required, and "language of the source" is vague, as it seems you are suggesting that mathematical notation and scientific jargon are languages (they are not). For example, I do not speak or understand latin, but given a dictionary and other readily available resources, I could translate latin. (I say latin in this, as I can read a tiny bit of greek having had a year of it in college.) The wording as it is is much better, as the point is that we are not to engage in OR. I would argue that if the source is not readily understandable by an educated person in general, the source requires interpretation by someone with specialized knowledge, and at that point, use of that source is original research. That scientists often ignore this aspect of policy is no different than the legions of fans of various types of entertainment who ignore RS and OR requirements, but keep each other in check by that specialized interest and knowledge. Call it an application of IAR, but my take on that kind of activity is that so long as it does no harm and is done in the spirit of improving WP, it's ok, but I'd rather have the policy in place as it is and used relaxed enforcement. And if "language of the source" does include jargon, this change would open the possibility of some pretty nasty conflicts in fringe areas, and even if not, it would have the same affect in any areas concerning nationalist issues, such as any articles treating any issues in the Balkans. Think of how difficult it would make discussions of religious subjects. --Nuujinn (talk) 14:16, 29 May 2011 (UTC)
I don't see the logic in your reply. If you can learn Latin to check a source, you can learn mathematics to check a source too, so the distinction escapes me. Both Latin and mathematics are hard to learn and are not in the skills base of more than a small fraction of "educated people". You are also incorrect that the use of primary documents in scientific articles is like that in entertainment articles. There is an essential difference: science is written in a very precise manner that conveys a unique meaning to qualified readers. Genuine original research is not much of a problem in scientific articles because all of the established editors can recognise it at once and hardly ever disagree on it. That's because they understand the sources, not because they are "educated people". I'll state it stronger than before: a majority of articles in mathematics or physics beyond high-school level, and a large fraction of articles in other areas of science, would be severely damaged if this rule was enforced rigorously. What sort of rule is that? My suggested wording isn't what I'm arguing for, any solution that is better than the current one would be welcome. Problems such as the meaning of "language" can be fixed by adding explanation. Zerotalk 15:28, 29 May 2011 (UTC)

Here's an illustration. At Wikiproject mathematics#Mathematics_and_mathematical_problems you can see a list of 23 articles chosen as good articles by the Wikipedia mathematics community. At least 15 of them use citations to primary sources that cannot be understood by the great majority of "educated people". Even I as a professional mathematician would struggle to understand some of these sources. Yet these are in truth excellent articles. Zerotalk 15:44, 29 May 2011 (UTC)

Yes, and I have to ask, what is the problem? We have good behaviour leading to good articles, created and maintained by talented editors. Now, I could be a dick and go over there and argue that the articles do not conform to policy, and try to remove sections arguing that they are violations of NOR. But I don't see a problem over there that need fixing even though the articles may not conform to policy. IAR frees us from being slaves to policies, and I think this is a good application of IAR. Now, if you, Zero, for example, came to the NOR noticeboard with an example of how a passage in one of those articles is non-encyclopaedic and problematical because it relies on primary sources and the OR of editors, I might very well agree with that. --Nuujinn (talk) 16:47, 29 May 2011 (UTC)
We seem to be drifting from the original topic, which is about language - all we want to clarify (since there seems a reasonable possibility that someone might misunderstand it sometime) is that when we say "any educated person" we mean any educated person who understands the language of the source. I suppose it's obvious enough (since we couldn't possibly not mean that, since there is no language which every educated person in the world speaks, though some people might reasonably resolve the ambiguity by supposing that since this is English Wikipedia, we're talking about English-speaking educated persons, which in fact we're not).--Kotniski (talk) 17:30, 29 May 2011 (UTC)

While I agree that zero's attempt might be more precise, there is a general issue, that cannot really be fixed by rewriting. Almost all of our policies & guidelines become a problem if they are taken all to literal in individual cases and this is indeed when WP:IAR needs to be used in doubt. There are no golden formulations, that will keeping somebody from being a dick and engange in unproductive wikilawyering and pestering good authors. The only way to such deal with behaviour is a community with common sense at large. Meaning other editors need to help out to shut down the dick. So if such a case comes up on a notice board, we need to rely on having enough sane editors around recognizing that it is dick move and telling the guy to bugger off. In cases where the advanced content might make it difficult for average readers to recognize what's going on the portals and projects covering the content should be consulted (at least if they are large, active and can be considered reliable).--Kmhkmh (talk) 23:03, 29 May 2011 (UTC)

Just because problems of interpretation abound doesn't mean things shouldn't be improved when they're noticed and can be, and I think Zero's change is an improvement. The only caveat is as already noted - what is meant by "language"? --Icerat (talk) 23:09, 29 May 2011 (UTC)
I think that "access to the source" is worth mentioning, since people do occasionally try to guess the contents of a publication based on the title. I support adding something to the effect of "familiar with the language of the source", and I suggest linking it to WP:NONENG, in honor of all those editors over the years that have claimed that their personal ignorance of <insert name of any non-English language here> proves that the material in an article is unverifiable. WhatamIdoing (talk) 23:35, 29 May 2011 (UTC)

To WhatamIdoing: Access to sources is covered by WP:SOURCEACCESS, a section of WP:V, but I don't object to it being here too. To Kmhkmh: I agree that we can always fall back on IAR, but my view of IAR is that it is something for boundary cases or exceptional circumstances. It doesn't seem right if a large chunk of Wikipedia relies on it for most of its articles. To Nuujin: Yes, we have good behavior, but it is good behavior that violates the letter of the OR policy. We will keep on doing the same if the policy isn't changed, but I think it would be nice to adjust the policy so that the policy corresponds better to what we consider good behavior. Here is a revised proposal:

A primary source may only be used on Wikipedia to make support straightforward, descriptive statements that any educated person, who has access to the source and can read the language and notation of the source but otherwise has no specialist knowledge, will be able to verify are supported by the source. Zerotalk 10:25, 31 May 2011 (UTC)

Zerotalk 10:25, 31 May 2011 (UTC)

Seems good. (But I still prefer "support" rather than "make" - you use your mouse and keyboard to "make" the statement.)--Kotniski (talk) 11:21, 31 May 2011 (UTC)
I agree and changed it. Zerotalk 12:53, 31 May 2011 (UTC)

Are these primary, secondary or tertiary?

All the discussion on primary, secondary and tertiary sources assume that the source concerend is wrtten in prose. However, many sources that are cited in Wikipedia are not prose, but are in some other form - government statistics, climate tables, maps and the like. I have been invovled in a discussion regarding maps and editors cannot agree whether on their status. May I suggest that the definition extended givign the following text:

Secondary sources are second-hand accounts, at least one step removed from an event. They rely on primary sources for their material, often making analytic or evaluative claims about them[ref] or collating them in a standard manner. For example, a review article that analyzes research papers in a field is a secondary source for the research[ref] as are reliably published climate charts, tables of statistics, maps and the like, the prime criteria being whether or not the data collectors and the data evaluators/collators are the same people. Whether a source is primary or secondary depends on context. A book by a military historian about the Second World War might be a secondary source about the war, but if it includes details of the author's own war experiences, it would be a primary source about those experiences.

above contributed by Martinvl (talk)

Perhaps you'd like to read the discussion #PSTS: I think you'll find it's a bit more complicated than that above. Dmcq (talk) 11:47, 29 May 2011 (UTC)
I have read it and yes, I agree that it is more complicated than my proposed additions. However, having an overhaul of the envisaged will also be complicated, so I make my suggestions again, this time in the realisation that they might be little more than a stopgap while any new system is being worked through. Moreover, more suggestions might wewll help focus the minds of whoever is actually working on a new system. Martinvl (talk) 20:46, 29 May 2011 (UTC)
So lets get this straight, the basic purpose of this is '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'. As that says the term has two purposes instead of one which is the main problem that other discussion is about. If you look at WP:N you see this main criterion of notability '"Sources", for notability purposes, should be secondary sources, as those provide the most objective evidence of notability. What you are saying would mean that practically anything on a map could have an article written about them. The same would apply to all the words in the dictionary, to every company in the register of companies, and loads of suchlike trivia in similar reliably checked and in multiple databases. Is this what you want? Dmcq (talk) 22:42, 29 May 2011 (UTC)
Martin, while I agree with where you're coming from, it is a bit more complicated than that. Merely re-printing someone else's data in a tabular format does not turn it into a secondary source, just like quoting Alice Expert in the newspaper doesn't mean that we have a secondary source on Alice Expert (or her views). I hope you'll take a look at User:WhatamIdoing/Sandbox 4 (the newest attempt at explaining some of this more fully) and the other advice pages (like WP:Secondary does not mean independent) and let us know what you think about them, and how we could improve them. WhatamIdoing (talk) 01:03, 30 May 2011 (UTC)
I'd say the only secondary source component in that was that somebody thought the set of data was interesting, it doesn't say much about the individual entries in the actual data though. Dmcq (talk) 12:11, 30 May 2011 (UTC)
Well, they thought it interesting, and perhaps reliable as well (otherwise why would they be reproducing it?)--Kotniski (talk) 16:06, 30 May 2011 (UTC)
Statistical tables are more than just reproducing data. For example, a census consists of everybody in the country filling out forms. These forms are plainly primary data. The Census office then takes those forms and collates them, giving population totals, demographic breakdowns and the like. Are these tables still a primary source, or are they a secondary secondary source? Likewise, the primary data for maps is hundreds of survey readings. The survey office takes these readings and from them draws contours and adds other information gathered by the surveyor. Does the resultant map have the same status as a table from a census report? Martinvl (talk) 16:31, 30 May 2011 (UTC)
Noting in advance that it really is a bit more complicated than this:
Imagine that you survey five people on your street. You ask them simple questions, like how many people live in their home. You write the answers down on a piece of yellow paper.
You go home and look at your yellow pieces of paper. You type on your website, "Three people live in a home with two people, and two people live in a home with three people." You type a simple description of your data: "The average household size is 2.1 persons per home."
Both the yellow pieces of paper and your website are primary sources for the number of people living in the homes.
The scale of the project is irrelevant. It doesn't matter whether you surveyed five people or five hundred million people. This is the first, original publication. It is not based on any separate work. It is therefore a primary source.
NB that "primary source" does not mean "bad source". Sometimes a primary source is the best possible source. WhatamIdoing (talk) 19:36, 7 June 2011 (UTC)

RfC: Are maps in general secondary sources?

The question has been raised at Wikipedia:ORN#Falkland Islands about how much can be gleamed from a map and stuck in an article without using text secondary sources. It seems also that some featured articles about roads like California State Route 7 are set up with details gleamed from maps. A primary source would only be used to answer obvious questions raised by secondary sources or give some detail or accuracy to things which had already been noted, but a secondary source could be used to find out and detail new things entirely. So is Google maps for instance a primary or secondary source? Dmcq (talk) 16:14, 4 June 2011 (UTC)

Are maps secondary sources?

I believe maps should in general be treated as databases rather than collections of interesting facts unless special features are noted like 'places of interest' or the map is specially set up to illustrate some facts like a historian saying 'this shows the extent of Gaul in 60BC'. The reliability is a separate matter and they are normally reliable sources in Wikipedia terms, Google maps is a pretty reliable source anyway whereas he historical map may be fairly inaccurate just it is set up by someone who is acknowledged as a reliable source.

I believe in general Wikipedia should only treat things as secondary sources where somebody has evaluated and noted or shown some interest in the individual bits of information rather than just be paid to amass a great gob of data and stick it on a computer. In that light bus timetables, most maps, planning applications, registers of births deaths and marriages and deaths and suchlike unevaluated stuff is primary data and should not be used except in very straightforward ways indeed to answer obvious questions raised by secondary sources.

I believe for instance seeing what road it turns into at the end is an obvious query related to the topic of a road and using a map to see should be okay. However saying it goes besides the Colorado River is original research unless someone says it passes near the river. People should not be filling up articles with things they think are interesting in a map but nobody else has noted. Dmcq (talk) 16:14, 4 June 2011 (UTC)

Yes, maps are secondary sources. The original primary source of the information would be aerial photography, GIS data or field notes from surveyors/fact-checkers. The maker of a map devotes time and energy into several sorts of editorial decisions in creating a map. What will this map show? How will that be featured? What color-coding and symbology will be used? What won't be shown? What scale should be used? To mention Dmcq's example, if a road runs beside the Colorado River, and a map clearly shows the road next to the river, and both the river and the road are labeled, then I can't see how it could even be described as OR to state that fact in an article using the map as a citation. Imzadi 1979  16:37, 4 June 2011 (UTC)
I agree with the above. (Though we really need to get it clear in our collective mind why - or indeed if - it matters; apart from being on the wrong policy page to start with, the WP:PSTS section consists largely of statements about one kind of source which turn out on brief reflection to be equally true about all the other kinds of source, making the whole thing look rather redundant.)--Kotniski (talk) 16:53, 4 June 2011 (UTC)
So since secondary sources give notability you'd be quite happy for me to start up articles on anything I can see on multiple maps? Dmcq (talk) 16:58, 4 June 2011 (UTC)
Yes, provided that the map comes from a reliable source (eg Ordnance Survey) and the information that you gleaned can be verified by a reasonably educated person with no specialist knowledge of the subject, or put another way, the rules that apply to a book are no different to those that apply to a map. Martinvl (talk) 17:13, 4 June 2011 (UTC)
Also remember that the topic still has to be notable. Just because it's on more than one map doesn't automatically grant it notability. –Fredddie 17:26, 4 June 2011 (UTC)
I declare my involvement since I was the one who asked the FI question to be brought to WP:NORN. I would say that maps in general are primary sources in the same way that telephone directories are primary sources. While there may be editorial decisions in creating a map, there are also editorial decisions in creating a telephone directory (details of collation order, where the boundaries of the area covered are and suchlike).
If we can see the road runs along the Colorado river based on a map, and this has not been noted by any other source, then I can't see why we particularly care. You would assume that other sources exist, otherwise the topic would not be notable: if they have not seen fit to mention this, why should we? It may be possible to get and use information from a map in such a way as to not require us to "analyze, synthesize, interpret, or evaluate" the source, but we should not include any information that does require us to do any of these things. Pfainuk talk 17:22, 4 June 2011 (UTC)

I think that most maps are primary sources, although some, usually specialized, maps will be secondary sources. There's more to a secondary source than being one baby step removed from raw data. A peer-reviewed scientific journal article about an original experiment is a primary source. So are the lab books filled with the data that the paper reports. Similarly, both the GPS data and the map created from it are most commonly primary sources. (What make something a secondary source is being based on separate, usually previous published, almost always multiple, sources. In both these cases, the raw data and the product created from the raw are not separate [nor are they previously published or multiple sources].)
However, I don't believe that there's anything wrong with editors using primary-source maps for the typical purposes. We can look at a primary source to count up the number of exits on a highway or to identify which towns it goes through: that's exactly the sort of "straightforward, descriptive statement" that this policy authorizes. There's no "analysis, synthesis, interpretation, or evaluation" involved in saying that the road is next to a river.
It is always worth remembering that editors must WP:Use common sense. If you find yourself wanting to restrict the use of a source because it's "primary" and therefore not "good enough" to support some fact that is so completely undisputed that a non-Wikipedian would call you an idiot for caviling against it, then you should stop worrying about perfection and use the sufficiently reliable source.
I suspect that Dmcq's real issue is WP:DUE: how much weight should we give to some detail that isn't discussed in a primary source? There are two important answers:
  • You must provide a reasonably complete description of the subject. Editors must use their best judgment (informed by, but not absolutely limited to, the high-quality secondary sources that are available to them) to decide what counts as a reasonably complete description.
  • "We currently name a perfectly adequate, primary-source map as the source" is not the same thing as "no secondary source has ever published this detail before". In the example of a major river near a major highway, it is extremely likely that the local newspapers will have mentioned this fact (e.g., in the context of stories about potential flooding) at some point. WhatamIdoing (talk) 18:07, 4 June 2011 (UTC)
(edit conflict) Why is a map different than a book beyond the fact that it is visual rather than written for its primary communication method? Telephone directories are raw information, and under Feist v. Rural barely even subject to copyright protection in the United States. A map, any map for that matter, has a significant level of creativity and editorial expression/control to them, just like a book, news article, etc. Why should it matter than a map is lines and shapes instead of words and paragraphs? Imzadi 1979  17:44, 4 June 2011 (UTC)

We've had a similar discussion before: Wikipedia_talk:No_original_research/Archive_39#Regarding_maps_being_.22primary_sources.22_according_to_this_policy. --Rschen7754 18:20, 4 June 2011 (UTC)

That discussion rambled on for a long time, but did not result in a conclusion being reached. In my view, maps should just be regarded as another language. Some people just cannot read maps, other (English-speaking) people just cannot read French. Just as there is no reason to exclude sources that are in French, there is no reason to exclude maps. If somebody who is writing about widgets cites a source that is in French (because that is an appropriate source), then as long as that source can be verifed by a reasonable educated person who can read French, but does not have a specific knowledge about widgets, it is admissable. Now replace the word "French" with "maps". The same rules should apply. Martinvl (talk) 19:07, 4 June 2011 (UTC)
Actually, the discussion resulted in User:Moabdave/maps, which is the defacto practice used in the U.S. road articles. --Rschen7754 19:28, 4 June 2011 (UTC)
Well that could be turned into an essay in project space and let people comment on it properly and see what comes of it. Basically then the best we have is some private user essay at the moment?
Personally I think linking our primary, secondary, tertiary to academic rules when ours are used for a quite different purpose is silly. A major aspect of secondary sources for Wikipedia is that they analyse and discuss from a more neutral viewpoint and so confer notability on what's in them as per WP:N '"Sources", for notability purposes, should be secondary sources, as those provide the most objective evidence of notability.' The business about primary sources being missing so source that was once secondary now becomes primary is also a bit weird. I do believe some maps can be considered secondary sources as I said before, e.g. a map by a historian of where they think the extent of Gaul was in 50BC but I see from that discussion some people consider that as primary whereas they consider Google maps secondary! Dmcq (talk) 19:43, 4 June 2011 (UTC)
Is the article User:Moabdave/maps sufficiently mature that it can be moved into WP: space, maybe with a "Under construction" flag on it? I think so. Martinvl (talk) 19:49, 4 June 2011 (UTC)
Quite possibly. --Rschen7754 19:53, 4 June 2011 (UTC)
What I recall from the discussion is that maps are secondary sources (the aerial footage / GIS data is the primary source). That being said, we need to be careful what we source to a map, because maps can only say so much. That's what Moabdave's essay states. --Rschen7754 19:53, 4 June 2011 (UTC)
IMO, the real criteria is whether or not the statement made from the map is verifiable using conventional map-reading techniques. To put things in perspective, if User:Rschen7754 were to source something about a Californian highway from a map, could I, as a Brit who is map-literate, drives a car but who has never been to North America verify what he wrote just by looking at the map? Martinvl (talk) 20:17, 4 June 2011 (UTC)

Wouldn't maps be considered on a par with other reference materials... ie tertiary? Blueboar (talk) 21:07, 4 June 2011 (UTC)

I don't think the main qualification for being a tertiary source is that it is a reference book, it's in the opposite direction - reference books are normally tertiary. The main qualification for being tertiary is that they just list or summarize stuff that has already been published, they don't do an independent analysis like a secondary source does. Dmcq (talk) 21:15, 4 June 2011 (UTC)
By the way I think we should accept a tertiary source as a source of notability. Dmcq (talk) 21:20, 4 June 2011 (UTC)

Imho the whole thing is entirely misleading. Polemically speaking does it even matter whether a map is primary, secondary, or tertiary source? Asssessing the reliability and type of the map is imho much more important. Even if I I were to regard maps as primary sources only, the example from the beginning (the road is running next to the Colorado river) still wouldn't be OR, since it is essentially just "quoting" the primary source which is possible. The normal reading of a (reliable) map, i.e. simply describing, what any normal person is reading on a map sees, hardly constitutes OR.

Now if the editor were to make some more involved conclusion based on the map data (say using the map to calculate the length of the colorado river) that would be OR indeed. However that would be OR no matter whether the map is considered a primary, secondary or tertiary source.

Now whether augmenting an articles with obvious information from a (reliable) is useful or not, is imho primarily a question of editorial judgement/taste and depends on the particular case. However pulling out the big gun WP:OR to settle a potential editorial conflict in such a case is misplaced, since there is no real OR here to begin with. --Kmhkmh (talk) 01:14, 5 June 2011 (UTC)

I think too much is made of the primary/secondary classification. But anyway just like in textual sources, maps can be primary or secondary. A map made directly from raw data, such as a map made from aerial photos, surveying measurements, etc, is certainly a primary source. The claim that the aerial photos are the primary source is not convincing when the aerial photos are taken for the purpose of making the map. Consider a journalist interviewing someone — we consider the published interview to be a primary source, and don't try to argue that only the recording made by the journalist for the purpose of writing the story is primary. It isn't entirely logical but that's the usual division. A map made by distilling information from a previous map would be a secondary source, and something like a school atlas would be a tertiary source. Zerotalk 03:17, 5 June 2011 (UTC)

User talk:DGG/Archive 21 Oct. 2008#are maps primary sources? is an opinion that disagrees with you from our earlier discussion. --Rschen7754 03:35, 5 June 2011 (UTC)

For the record, we at the U.S. Roads project have 36 FAs, and most, if not all of them, use maps at some level (and the use of maps as sources has come up at least once or twice at FAC), so it's not like maps are invalid for sourcing on Wikipedia. --Rschen7754 03:35, 5 June 2011 (UTC)

One of the FACs where it came up: [38] (there may be more, I just looked through the successful FACs, not including where it failed before it passed later). --Rschen7754 04:09, 5 June 2011 (UTC)

MoabDave's articel copied into WP: space.

I have been WP:BOLD and copied User:Moabdave's article into WP:Using maps and similar sources in wikipedia articles. I have also added two paragraphs at the end about topological assessment, but have not yet attacked the primary/seconday/tertiary issue. I expect a number of changes to be made before we start linking it in to other articles, but at least it is now visible to all editors who have an interest in the use of maps as source material. Martinvl (talk)

I'd like to put this change to consensus here, if it's a standard we're supposed to be following. While I don't have a problem with the "glacial moraine" bit (given that the addition says that the words "glacial moraine" have to be explicitly stated on the map), I'm not convinced that "U-shaped valley" should be being read from contour lines. Pfainuk talk 08:50, 5 June 2011 (UTC)
Perhaps that guideline needs to be split between the three aspects of notability or due weight - what is 'pointed out' by the map, reliability - how accurate is an aspect of a map, and original research - how much is okay to read from a map.
In my view a map is no different from any other database so perhaps the essay could deal with databases in general. For instance a genealogist can read a family history in diagrams or the Salt Lake database with equal facility and there must be enormous numbers of geographic databases, I've one I set up myself for something I'm doing even.
The academic definitions of a secondary source do not correspond with Wikipedia's needs as the point of Wikipedia is to not do original research whereas doing original research is an aim of academic institutions. The academic definition corresponds more with our reliable sources requirements whereas Wikipedia has been using it for notability and due weight, and somehow it has got written into the original research page. If this essay can be brought to a fit state then maybe the whole business about primary secondary and tertiary sources can be formulated better or totally replaced as it is just unfit for purpose currently. Dmcq (talk) 09:50, 5 June 2011 (UTC)

Research vs incorrect information

The original research barrier has always confused me. It seems to imply that research = fake rubbish. If the research deals with things that can be proven though, I would have thought that would be more important to an article. The majority of published citations on Wikipedia will have had their origin in research that someone's conducted at some point in time, so why is it seen as a bad thing? For an example, I can pretty much prove that a certain minor celebrity wasn't born with the name Wikipedia says is their birth name. However as no-one's seemingly referenced it anywhere else in print or on the web, it comes under research, so the article continues to display incorrect information as fact. So I'm wondering, is it only Wikipedia's remit to collate whatever's published, even if that information is wrong? -- Analog Kid (talk) 10:20, 8 June 2011 (UTC)

  • Wikipedia is an encyclopaedic project. We provide a tertiary reference object based on secondary sources that conduct original (or unoriginal) research. If you wish to publish original research, and then use that on wikipedia, find an appropriate publication venue off-wiki that meets the reliability or high quality reliability (FAC) guidelines. Fifelfoo (talk) 10:31, 8 June 2011 (UTC)
Original research means something an editor finds out that isn't published. Very often it is true and sometimes it may actually be important. However Wikipedia is in the business of summarizing stuff that has been published in reliable sources. There is no way for editors to go around checking each others conclusions or seeing that they are not trivial. The checking of facts should be done by the reliable sources, they can sometimes be wrong but that's a different matter from having editors here pushing their ideas. And the fact that a reliable source has bothered to think about something and write down what they thought is the basic minimum criterion that something isn't trivial. So overall if an editor finds out something and it hasn't already been published then it has no place in Wikipedia. Dmcq (talk) 10:44, 8 June 2011 (UTC)
Thank you for the clarification. -- Analog Kid (talk) 10:55, 8 June 2011 (UTC)
Although it might be a good reason to exclude (or water down) other stuff that is included but is contradicted by it.--Kotniski (talk) 12:00, 8 June 2011 (UTC)

@ analog kid: Of course are correct information or correct facts the primary goal of WP. However we cannot leave the assessment or right or wrong to arbitrary WP editors, that approach would only work if the WP editors were certified non pov pushing experts editing only in their domain. Since we don't have that we use reputable (mostly secondary) literature as a replacement, as they are usually correct. However we do do not simply copy the content of any publication, but of course we select the best and incorporate corrections (but selecting a more recent correct source over an older oudated one). Another thing the sources provide us is a notability filter, as the existence of publications on some subject can be seen as a "proof" of its notability.--Kmhkmh (talk) 12:29, 8 June 2011 (UTC)

The terms "research" and "original research" seem to be getting a bit muddled here, and the concept is frequently misunderstood. Obviously, when developing an article, you have to "originate" the article by "researching" the sources. "Original research" actually means that wikipedia does not "originate" information, it merely summarizes already-published information from reliable sources. That does not mean that being from a reliable source automatically makes a bit of information notable for an article. Being reliably sourced is only a minimum standard that must be met. If information is not-notable, or for that matter if it's clearly incorrect, it shouldn't be used in an article. ←Baseball Bugs What's up, Doc? carrots→ 12:35, 8 June 2011 (UTC)

RFC notice

There is an ongoing RFC at Talk:Santorum (neologism)#Proposal to rename.2C redirect.2C and merge content that may be of interest to editors here. Dreadstar 18:06, 8 June 2011 (UTC)

Proposal to change first sentence of wp:ver

There is a proposal to change the first sentence of wp:ver at Wikipedia_talk:Verifiability#Proposal for a change in the first sentence. Your input is requested. Sincerely,North8000 (talk) 22:35, 10 June 2011 (UTC)

Secondary source clarity needed

How big a step is "one step removed"?

The example says, "An account of a traffic accident written by a witness is a primary source of information about the accident."

How much separation does there need to be for a summary or retelling of the primary source to be considered a secondary source? Using the above example, if the witness hands the account to a cop, and the cop summarizes it the next day, is that still a primary source?

What if it was a police car in the traffic accident, and the witness who wrote the original account was a cop?

Suppose then that a different office at the same police department writes the summary?

Or, what if it's a different person in a different office of that same police department, but it's one year later? How about a different person writing a summary ten years later?

I'm sure it's in the old discussions, but this should be written more clearly in the guidelines. As the policy is now written, I would assume that any second person (assuming RS) who writes a summary is then creating a secondary source.
-- Randy2063 (talk) 05:09, 12 June 2011 (UTC)

Anything written by someone involved in the event is a primary source. How close they would have to be to an event to count as "involved" would depend on your perspective, because the determination is always relational. SlimVirgin TALK|CONTRIBS 05:33, 12 June 2011 (UTC)
A secondary source should also be one that imparts some sort of notability because they notice and talk about something. That a person hit on the head writes about being hit on the head impart the event no notability. That someone else unconnected noticed the report and specially wrote about it imparts notability. Dmcq (talk) 11:22, 12 June 2011 (UTC)
────────────────────────────────────────────────────────────────────────────────────────────────────Also, if a cited secondary source merely echoes primary source, it is no more useful than the primary source which it echoes. It is important that the secondary source make analytic or evaluative claims about the content reported. Wikipedia editors are constrained from any analysis or synthesis of published material that serves to advance a position not advanced by the sources (primary or secondary), but may rely on such analysis or synthesis made by reliable secondary sources.
The distinction which WP makes between source classifications strikes me as contrived and as contributing more confusion than clarity. Any interpretation of primary source material requires a reliable secondary source for that interpretation. Likewise, since Wikipedia editors are constrained from any analysis or synthesis of published material that serves to advance a position not advanced by the sources, any interpretation of secondary source material requires a reliable secondary source for that interpretation. Said differently, any interpretation of primary or secondary source material requires a reliable secondary source for that interpretation. Why not just say, "Any interpretation of source material requires a reliable source for that interpretation."?
Is the text of a SCOTUS decision a primary or a secondary source? It is clearly an account written by people who are directly involved, offering an insider's view of an event. However, opinions contained therein invariably make analytic or evaluative claims about the event and almost invariably engage in analysis and synthesis of other sources cited therein.
WP defines Tertiary sources as "publications such as encyclopedias or other compendia that mainly summarize secondary sources", and says that WP is a tertiary source — however, the oft-cited Encyclopedia Britannica is often treated as a secondary source, and analyses and interpretations found therein are relied upon to support article assertions. WP also says that it regards many introductory undergraduate-level textbooks (one wonders: which ones?) as tertiary sources because they sum up multiple secondary sources. What is "summing up" anyhow? How does that differ from "analysis and synthesis"?
The crux seems to be distinctions made on the basis of judgements made about source reliability, and identifying reliable sources, more than on the basis of differing treatment of sources by PST categorization. Any source is reliable to support an assertion that it contains the content which can be seen by inspection (it says what it says.) — (but is that all that it says?) if a stable source engages in "analysis and synthesis" and makes an assertion resulting therefrom, is that assertion considered reliable enough to support an assertion to that effect made in a WP article? (there's the rub) Wtmitchell (talk) (earlier Boracay Bill) 23:21, 12 June 2011 (UTC)
You might want to read WP:Secondary does not mean independent. "First party", "first person", and "primary source" are all completely separate concepts. WhatamIdoing (talk) 00:48, 13 June 2011 (UTC)
Thanks everyone. That's a lot to chew on.
This is even more restrictive than I had thought. I think, at a minimum, we should amend the phrase "An account of a traffic accident written by a witness is a primary source of information about the accident" to make clear that reports put out by anyone in the entire police department would also a primary source. The witness was too obviously a primary source, but it leaves open the question about those who came along later.
-- Randy2063 (talk) 18:48, 16 June 2011 (UTC)

"Synthesis" indeed!

WP:SYN is simply the most badly worded and misunderstood bit of policy in the whole of world wide wiki land! What it should state is that it is unacceptable to link statements together as a rhetorical device, so as to imply (connote) conclusions that go beyond the premises. What it actually forbids is any kind of synthesis whatsoever! It actually forbids logic! It reduces WP to a list of direct quotations. That is all we can do according to WP:SYN. Furthermore, not only does the policy forbid more than it intends to forbid, but it fails to prevent what it intends to prevent! To illustrate this, look at the first example given on the policy page:

>A simple example of original synthesis:

>The UN's stated objective is to maintain international peace and security, but since its creation there have been 160 wars throughout the world.

>Both parts of the sentence may be reliably sourced, but here they have been combined to imply that the UN has failed to maintain world peace. If no reliable source has combined the material in this way, it is original research. It would be a simple matter to imply the opposite using the same material, illustrating how easily material can be manipulated when the sources are not adhered to...

But, you can imply the same conclusion, without combining anything, purely by sequencing, thus:

The UN's stated objective is to maintain international peace and security. Since the creation of the UN there have been 160 wars throughout the world. ... Stho002 (talk) 08:39, 25 June 2011 (UTC)

I think that your post covers several very different topics. While it was not on your main thread, I think that it is a problem that under wp:nor, as written, if followed literally and rigorously, anything but a direct quote is OR. Wikipedia makes it work most of the time by going by reasonable / common sense interpretations of the policy. But such a "conflict" between reality and policy opens it to abuse, which also happens. Gives wiki-lawyer POV warriors a magic bullet to throw out any material that they choose.
Second, you are addressing the concept that juxtaposition does imply a connection. I think that it is important to recognize this even if it is hard to pin down. Your dissection of the vulnerable example shows that it is hard to pin down.
Third, that example also has POV connotations. I.E. implying (without stating) that the UN is a failure by the mere presence of that second phrase. This points to a gaping and widely exploited hole in the wp:npov policy. This is for cases were the mere presence of material is prejudicial /POV, but where it slips under the radar of the wp:npov policy because it does not explicitly make any statement about the issue at hand. So, the example is one where wp:nor is trying to be a (weak) backstop for POV'ing that slipped through a gaping hole in the wp:npov policy.
Finally, I can't tell whether you are saying that the wp:nor policy needs tweaking (I agree) or that WP should be opened up to substantial creative thought, things created by (what the writer considers to be) logic etc. by editors. There, I would disagree.
Sincerely, North8000 (talk) 11:25, 25 June 2011 (UTC)
Your example would be considered implicit synthesis and would violate policy. TFD (talk) 12:14, 25 June 2011 (UTC)
I note that user:Stho002 is currently blocked. But I'll respond anyway. I wrote WP:SYNTHNOT, an essay that no one else has edited so far, to address problems with non-common-sensical interpretation of SYNTH. In particular, the objections here seems as though they're covered adequately by SYNTHNOT in the sections "SYNTH is not mere juxtaposition", "SYNTH is not ubiquitous", and "SYNTH is not a matter of grammar". If anyone has a better idea than SYNTHNOT for how to deal with SYNTH's potential for over-broad interpretation, I'd be glad to hear it. --Dan Wylie-Sears 2 (talk) 03:53, 3 July 2011 (UTC)
I hadn't seen SYNTHNOT previously. I like it, but I don't think it's much help here. It says that gray-area cases aren't SYNTH, just unclear writing, but, well, this is an example which came up over something I wrote, and I'm still not sure whether I was right or wrong. The discussion mentioned there has been archived, and can be seen here.
This appears to be an oft-repeated discussion; see e.g., this archive search for "synthesis by juxtaposition". I'm guessing that there are other discussions touching on this which don't include that exact search string. Wtmitchell (talk) (earlier Boracay Bill) 23:18, 3 July 2011 (UTC)
I like to imagine that SYNTHNOT would have helped, some: people alleging SYNTH would have felt a little more obliged to spell out what thesis was being implicitly advanced; people wanting to include whichever text in the various articles would have felt a little more obliged to rewrite so that an unintended thesis was not suggested (or to make it explicit and find a source for it). The argument would have focused a little more on rewriting and less on removal, more on spelling out the thesis supposedly being suggested and less on generalities about verifiability. I can't fool myself into thinking that SYNTHNOT would have helped dramatically in all cases, though. For one thing, it's only an essay. People on all sides are free to simply disregard it.
In the case of the birther article, the implicit claim is that the particular passage of law applies to the certificate in question and contradicts the statement of the AG spokesperson. That legal interpretation requires a source, given that one provision of law often is superseded by others, or has legal import at odds with its plain-English reading. The text of the law seems like something that could be included, but not in that juxtaposition. --Dan Wylie-Sears 2 (talk) 05:43, 4 July 2011 (UTC)
SYNTHNOT is a good essay. While Wales himself (I think) identifies "crackpot theories" as one of the things to be avoided by the NOR policy the policy obviously also apples to non-crackpot theories. Wikipedia chooses to not be the place for the publication of any original research. That seems to be an entirely sound policy.
I do agree with both Stho002 and Dan Wylie-Sears 2 that all synthesis should not be forbidden. Wikipedia is, after all, an encyclopedia. It is straining to pretend that for every aspect of everysubject there is a non-encyclopedic source for every useful bit of information. In articles on topics of importance (where here I'll suggest that "importance" implies that there are refereed journals that are devoted to the topic, although I'd not push this to the limit - it's more of a notion) it ought to be completely obvious that such journals are written at a high level, to be read and understood by those already competent in the field (there are guidelines for authors that say exactly this.) these are persons for which exposition and clarification is unnecessary - because, as experts, they already know. The mandated ()by extremists) reference for the results of a synthesis that is sometimes adamantly insisted on simply may not exist - and the nonexistence is most likely for a synthesis that is obvious and non-controversial. The synthesis may, however, be something that provides useful clarification or insight to the non-expert reader (as you'd anticipate most Wikipedia readers would be.)
I do not argue for nor wish to sneak in any OR by these words. The point is that synthesis may be an appropriate outcome of intelligent editing, editing that is solely concerned with making the subject matter more clear or more accessible, to highlight implications of fully accepted principles and facts in the field of inquiry being discussed in an article. If another editor perceives a taint of OR in such synthesis then let the discussion begin. It seems willfully and carelessly contrary to the nature and purpose of an encyclopedia to savagely apply an unnecessary rule. SYNTHNOT expresses it well: if the synthesis isn't an instance of OR then it is not a violation. What has happened, however, is that after the prohibition of synthesis was inserted into NOR based on examples of apparent OR done by the use of synthesis the eventual (and, given human nature, nearly inevitable) result was that of all synthesis being declared inappropriate. If SYNTH is part of NOR then, surely, the only legitimate use of SYNTH to forbid synthesis is strictly limited to cases where SYNTH introduces OR. Absent the introduction of OR there is no problem with SYNTH (insofar as violation of a basic Wikipedia principle is concerned.) It is not and can not be appropriate to find one or a few examples of any rhetorical technique that has been used in violation of the basic rules of Wikipedia and to then absolutely forbid that technique because it has been misused. If that can be done then simple declarative sentences should be forbidden, once one s found that introduces OR.
Bottom line, though: read and pay attention to SYNTHNOT. It's well thought out, well written. Then let that guide an improvement in the policy.Minasbeede (talk) 02:37, 11 July 2011 (UTC)

"Novel Synthesis"

Template:Uw-nor1 warns users not to publish "novel, unpublished syntheses". This phrase is Wikijargon and is not self-explanatory to a Wikipedia outsider. Is there a better more familiar short phrase that could replace this? Marcus Qwertyus 22:19, 28 June 2011 (UTC)

You are right. I took a swing at it. Gigs (talk) 17:21, 30 June 2011 (UTC)

Allowance of Logical Deduction in WP:NOTOR

Hi, I've been away from wikipedia for a little while, but I'm concerned about this new development. Can someone explain why it was removed. I mean deductions are like simple mathematics but with words, right? Is this not allowed for some reason? Or haha, I'm getting ahead of myself and the proper use of logical deductions is found on another policy page or something. LOL, but seriously, little help here. BETA 03:49, 16 July 2011 (UTC)

You are quite right. Logical deductions are not allowed. They are original research. WP:SYNTH and WP:CALC show the extent of the restriction. There are a number of reasons for this. Firstly we should not be writing things nobody else has noticed or thought important enough to note down. Secondly one person's logical deduction is another person's total heap of rubbish. Wikipedia aims at WP:Verifiability. Wikipedia is not a blog for writing up whatever people dreamt or 'deduced' overnight. Dmcq (talk) 10:44, 16 July 2011 (UTC)
Mmm, that's not exactly what a logical deduction is though is it. I mean it would be synthesis to skew something to advance an opinion. but logic is just that. something that can't be refuted. like "A is a Letter and B is a letter, then A and B are both letters."
Simple logical deductions and syllogisms are part of our language. If they aren't allowed then simple tasks like paraphrasing wouldn't be allowed either. because you're equating the meaning of a phrase with the meaning of a different phrase with the same meaning. We have to draw the line somewhere yes, but we have to draw it where it makes sense. BETA 16:11, 16 July 2011 (UTC)
Have you read the talk page of WP:NOTOR where the removal is discussed? In general though even in the cases they are talkiing about that more complicated calculations would be allowed in articles aimed at a higher level of competence the calculation or something similar should appear in some text anyway so we are really discussing verifiability. It is fairly obvious that if a text goes from A to B that it will go through the main intermediaries but normally they will not be explicitly cited. Nor is it necessary to have citations when illustrating a formula with an example. Saying something new or which people have not noted which you have calculated is however original research except in very restricted cases as described in WP:CALC e.g. converting feet to meters. Dmcq (talk) 18:36, 16 July 2011 (UTC)
What about opposite statements? For example, I've encountered editors who wants to continually undo my changes articles when all I write is essentially the opposite of what was written. For example, if an article says "proposed rule A is made moot by adopted rule B", why would it be considered OR to say "the subsequent adoption of rule A would then render rule the previously B moot"?? Again, I've encountered this frequently and the claim is not one of intellectual redundancy, but of OR. My edit to the article is merely an expanded explanation that would pass the verifiable test and no one would reject it, just like most reasonable people wouldn't object to some one stating Paris is the capital of France. — Preceding unsigned comment added by (talk) 03:06, 17 July 2011 (UTC)
Why would you reverse how something is written and why are you so keen to have your version included instead? It sounds to me like people are actually objecting and you are saying they are being unreasonable. I don't think your generalities about A and B are a good enough basis for a discussion about whether other editors are being unreasonable. This page is more for discussing the actual policy rather than its application in specific cases which is what WP:ORN is about but you might as well continue here if you have a specific dispute. Dmcq (talk) 16:09, 17 July 2011 (UTC)
In fact I see now you have a specific query at WP:ORN#Corwin Amendment so I guess that is the best place for asking about the dispute. Dmcq (talk) 16:28, 17 July 2011 (UTC)

Accidental original research

From time to time it happens that an editor accidently discovers a relationship that as far as he is concerned, has hitherto been unpublished. How should such a circumstance be handled?

The example that I have in mind are the articles Charles Rawden Maclean and Nathaniel Isaacs. The background is that in 1824 Maclean, then aged about 10 and Isaacs, aged 17 were shipwrecked with a group of about another 12 white men at Port Natal (South Africa) where they came into frequent contact with the Zulu people. Four years later they were rescued and went their separate ways. A few year later Isaacs published his memoirs in which he referred to Maclean as "John Ross". Thirty years later, Isaacs was in trouble about alegations of slave-trafficking in West Africa; Maclean on the other hand had risen to ships captain and had acquired a reputation for opposing slavery, particularly on his trips to what was to become the Confederate States of America. Between 1852 and 1854 he published a series of articles about his life in South Africa. The question is - Did Maclean know about Isaac's alleged activities and if so, was this the trigger for him to publish his articles?

More importantly, from a Wikipedia point of view, how should this be handled? I have no evidence one way or the other, but in my view there is enough circumstatial evidence to ask the question. I do not know if anybody else has even asked the questions. The approach that I took was to draw attention to the two events being at the same time, but to avoid posing any questions and certainly not to expound any theories. Have I taken the right approach? Martinvl (talk) 17:16, 17 July 2011 (UTC)

You would need sources that draw this conclusion. However I fail to see how a childhood acquaintance with someone who later became a slave trader would have any influence on someone who decided to oppose the trade. TFD (talk) 17:27, 17 July 2011 (UTC)
The question is "Did the rumours of Isaacs' slave trafficking spur Maclean to publish his articles about South Africa on the eve of Isaac's arrest?". [In the event Isaacs was not arrested - the relevant papers were lost in a shipwreck]. Martinvl (talk) 17:59, 17 July 2011 (UTC)
If you haven't found an answer to that in one of the sources you're unlikely to find an answer here! I suppose there are qute a few conjecture editors must come up with but can't write anything about in Wikipedia. I suppose if there was a conjecturepedia it would quickly be overrun by nutters so the only place for things like that are the talk page of the article or better one of the reference desks. Dmcq (talk) 21:56, 17 July 2011 (UTC)
I am not looking for an answer to my specific problem here - I am using my particular problem as a case study to illustrate how people doing encyclopeadic research could accidently stumble on something that could be WP:OR. My question was how far one should go in highlighting the situation without themselves doing original research. Martinvl (talk) 22:21, 17 July 2011 (UTC)
You might be looking for this: Wikipedia:These_are_not_original_research#Caveats_about_expert_material. Also check the policies/guidelines on the use of primary sources and conflicting sources.--Kmhkmh (talk) 22:35, 17 July 2011 (UTC)
It happens occasionally that WP authors come across some "novel" knowledge or perspective while doing there research or slightly less spectaluar that they discover overlooked errors and inconstencies in reputable sources. Sometimes such incidents are reported in the press as well. Personally I'm aware of 3 such cases (1 in math, 1 in geography, 1 in history). Imho they are a nice side effect of writing for WP, but to include those results in WP itself, they need to be published elsewhere first. Once that is done, i. e. they have been published & checked outside of WP, the results can incorporated into WP itself.--Kmhkmh (talk) 22:29, 17 July 2011 (UTC)
Exactly, and it has happened to me twice now. Of course, one could send the information to a reliable academician and see if they find it interesting, in which case it might be published, but until it is, it's just private knowledge and not appropriate here.

Contradicting primary and secondary sources

Every so often I've run across instances where a group or organization that is newly formed or has its own established history adopts the name or brand of another defunct organization with its own separate history. According to the new group's statements and press releases, they describe themselves and the defunct group as one in the same. I'd assume these are primary sources. There is no acknowledgement of the separate nature between old and new. There is only an insistence the present day organization is a direct continuation or a return of the old. However according to secondary sources, the situation is described as a newly established entity adopting an old name/brand. These secondary sources distinguish between what had been and what is entirely new. How does an editor here distinguish between the two without getting caught in the trap of original research? --Blackbox77 (talk) 02:36, 20 July 2011 (UTC)

The primary sources are unreliable on this issue and shouldn't be used. Use the secondary sources that describe the groups as different. They are the only reliable sources on such an issue. You can describe what the group says about itself, e.g. "According to the Group Two, they are the same organization as Group, and follow the same ideology and practices". A self published source is reliable about itself, as long as there is no obvious reason to believe otherwise. LK (talk) 03:37, 20 July 2011 (UTC)
Lawrence's answer is correct. However, I'll add that WP:Secondary does not mean independent. You should follow the independent sources in these cases, even if they are also primary sources. WhatamIdoing (talk) 03:59, 20 July 2011 (UTC)
In addition to the aspect of independent sources being generally more reliable, another criterion also points in the same direction: When two sources contradict each other, then as a rule of thumb if one of them has clearly taken the other's arguments in account (e.g. it contradicts the other by explicitly referring to it, or as in this case it's implicitly clear that the author knows about the arguments for the opposing case), then it is usually more reliable. But when there is any indication that there is a POV problem, as in your description, then that usually takes precedence. Hans Adler 05:12, 20 July 2011 (UTC)

Thanks for the feedback. Everyone's input has definitely helped me understand this dynamic further. If you're curious, the articles that prompted my question were New York Cosmos (2010) and Arena Football League. In the New York Cosmos case, the naming rights to the long defunct soccer club are being used by a recently formed organization to establish a new Cosmos club. Despite the original Cosmos not existing in any form since 1985, the modern-day Cosmos consider themselves to be one-in-the-same as the old. However, old and new are legally separate entities and reputable secondary sources - when describing these points in detail - make a distinction between separate teams. As for the AFL, the original league went bankrupt, disbanded, and auctioned off its remain assets in 2009. A newly formed group of arena football clubs with preestablished plans to start a new league bought the AFL naming rights. Now under the AFL name and starting two years removed from the original's last season, the new league claims there is no distinction between them. This again flies in the face of numerous news reports describing the separate and distinct nature of the two (especially before and after the folding and auction). In both cases, it's as if the new wants to claim the lineage of the old for itself despite the accounts of those documenting these events. --Blackbox77 (talk) 03:24, 26 July 2011 (UTC)

Secondary does not mean third-party

In the category of messes that need to be cleaned up, the text for the template {{Primary sources}} confuses "secondary" with "third-party" sources:

I'm not sure how widely this template is used (perhaps used on many pages that ought to be tagged with {{Third-party sources}}?), or what the best way to fix it is, but if anyone who understands the difference between a primary source and a non-independent source has a little while to think about it, then perhaps someone else will find an inspired solution to the problem.

Please join the discussion at the template's talk page. WhatamIdoing (talk) 05:16, 21 July 2011 (UTC)

We've made some progress, but we're running into confusion from people who don't know that WP:Secondary does not mean independent. Additionally, we have an admin who refuses to link to either WP:Third-party sources or WP:Independent sources, because there's a "clean up tag" on them (a long-ignored proposal to merge the essays into one page). In the meantime, we still have a template that seems to think that all third-party sources are, by definition, secondary sources, which you all know is nonsense.
BTW, this has resulted in a strange TFD for {{third-party sources}} (which says that even secondary sources aren't good enough, if those secondary sources are affiliated with the subject), apparently on the grounds that the nom completely agrees that even secondary sources aren't good enough, if those secondary sources are affiliated with the subject. That's at Wikipedia:Templates for discussion/Log/2011 August 4#Template:Third-party, if anyone wants to take a look. WhatamIdoing (talk) 00:20, 5 August 2011 (UTC)


Should this page point users to Wikiversity? I think so. After all, that's a large part of what Wikiversity is for. (talk) 12:39, 11 August 2011 (UTC)

Is this map a synthesis

This flag-map shows UN states and associated flags, and I'm wondering if this is a WP:SYN. The author most likely put these sources together:

- A UN map

- The flags of UN members

And put them together to create a new map. However, I cannot find any such map on UN websites or any reliable source, and my problem is that some territories, like Taiwan, are covered by flags they do not fly. I think the author reached the original (and incorrect) conclusion that both the controlled and claimed territories of a UN member fly the same flag. Obviously, this is not correct for a few territories. What do you think? Is it a synthesis or not? Laurent (talk) 13:10, 31 July 2011 (UTC)

It does not have the flgs of the countries. It has somebody's artistic interpretation of them. Flags have rules associated and if you break the rules you are either not showing the flag or you are defacing the flag. I would reject this as a persons own artistic work at best and not an informative image. Dmcq (talk) 14:47, 31 July 2011 (UTC)
That's indeed a way to look at it, but I'm sure many readers take it as a reliable map, especially since it's a featured picture (and the caption makes it sound like it's a properly researched work based on UN data). Also, if it is an artistic work, should we use it on Wikipedia? Laurent (talk) 14:56, 31 July 2011 (UTC)
There are too many maps and charts that are, in my opinion, original research, but it's hard to do anything about them except keep trying to get them removed from articles. Getting them actually deleted is difficult, in my experience. Dougweller (talk) 15:25, 31 July 2011 (UTC)
Any map contains an immense amount of information. So do other graphical presentations of information. If Wikipedia want's to fully recognize that, it doesn't need to just "crack down", it needs to treat it as it does other information. That would fully applying content policies, including citations and sourcing for them and/or their contents when challenged. North8000 (talk) 20:50, 31 July 2011 (UTC)

The image is used in only one article: National flag. As used (with a caption that reads: "A world map showing most national flags") It seems harmless, but more than a bit "gimmicky". I have long held that the purpose of an image should be to illustrate something that is said in the article. What fact or concept is being illustrated by is image? 22:31, 31 July 2011 (UTC)

That's one article definitely I think where it should not be used in as it deals explicitly with flags and talks about the various protocols, all of which is broken by that image. Images made by editors should only illustrate ideas in the article and should not introduce anything more. I think the panel of old flags further down is a far better illustration, WP:IMAGE#Pertinence and encyclopedic nature covers this. Dmcq (talk) 22:50, 31 July 2011 (UTC)
Replaced it with a panel of national flags. Dmcq (talk) 22:58, 31 July 2011 (UTC)
Good idea, I have made the same change on the French Wikipedia. Laurent (talk) 11:17, 1 August 2011 (UTC)
It's not actually a SYN violation, but that doesn't mean that we have to use it for any purpose at all. WhatamIdoing (talk) 05:16, 4 August 2011 (UTC)
Agree, it's almost certainly an aesthetics violation, probably offensive to a few countries whose flags got mangled (I'd be particularly careful of the Saudi flag, which contains a Koran verse) and really really difficult to read. HrafnTalkStalk(P) 05:26, 4 August 2011 (UTC)

Identifying and using primary and secondary sources

I have just moved the long-awaited page on Wikipedia:Identifying and using primary and secondary sources into the project namespace. We've been working on it for four months, and I'd like to thank all of the people who contributed suggestions in the many scattered discussions.

It is not a "finished product", but I think that it will help dispel some of the confusion that Wikipedia created a few years back, when a lack of education among editors resulted in our policies using "third-party" and "secondary" interchangeably. I hope, too, that it will result in primary sources being used carefully—but also in fewer people believing that "primary source" means "hopelessly unreliable source".

I'm uncertain whether we should propose it as an official guideline. On the one hand, most of us know that the WP:Difference between policies, guidelines, and essays can be quite minimal. On the other hand, we do still encounter people who believe that "essay" means "irrelevant page that can be safely ignored", even though pages like WP:BRD and WP:Use common sense are "just" essays. I'd be happy to have other people's opinions on that point.

And, finally, we need three of these pages:

One suggestion to that: another (a required?) characteristic of a secondary source is one that transforms primary and sometimes secondary and tertiary into novel thought; this to me is a really good way to measure the use of news stories (which may or may not be secondary) for consideration of event notability. A news report simply reiterating facts without analysis of their own is primary; one that attempts to make some conclusion is secondary. --MASEM (t) 18:29, 15 September 2011 (UTC)
As a general rule, I agree with you, and that is the point of the section currently titled "Not a matter of counting the number of links in the chain". But it's a fuzzy continuum, and even experts might disagree on some borderline cases, so there seemed to be some opposition to stating this as an absolute "always" fact. WhatamIdoing (talk) 23:47, 16 September 2011 (UTC)
  • There is a dispute at WT:Identifying and using primary and secondary sources over whether newspapers and newsmagazines should be generally categorized as primary sources or secondary sources. That issue could feed back to this policy and have a significant effect on which articles are deleted and what sources may be used. More input would be welcome.   Will Beback  talk  23:00, 21 September 2011 (UTC)
  • In particular, we'd benefit from attention from people who already know that WP:Secondary does not mean independent, that the act of quoting someone does not magically transform your publication into a secondary source, and that how you use the source matters as much to the classification as the source's inherent qualities. WhatamIdoing (talk) 16:08, 22 September 2011 (UTC)

Prince Michael of Albany

Discussion about specific cases belongs at WP:NORN, not on this page. Please move it there. Zerotalk 16:08, 17 September 2011 (UTC)
Thanks for the guidance, although the discussion was about a point of principle raised by a specific case, not the case itself. Discussion removed to Wikipedia:No original research/Noticeboard Radex09 (talk) 17:04, 17 September 2011 (UTC)

Interviews are not primary sources

An interview by definition is a dialog between a non-involved party and a participant in an event. This always involves a contribution in the form of the questions chosen, and often even editing control, by a non-involved party.

The Wikipedia's own interview article contains many good examples of interviews which no one questions as being tainted by being primary sources which violate the letter or spirit of WP:NOR.patsw (talk) 14:14, 25 September 2011 (UTC) In case its not clear what this edit refers to, it is to note 3 which in a long list of other examples of WP:PRIMARY includes interviews. Why interviews are primary sources is asserted not explained. patsw (talk) 14:19, 25 September 2011 (UTC)

First of all, there is nothing wrong with primary sources, but for interpretation of such primary sources we need secondary sources.
Secondly, I disagree that an interview is not a primary source, the questions are developed to get a specific response, but the actual responses are made by the interviewed person and make the interviews therefore clearly a primary source. Editorials, or analyses that are subsequently created after an interview in which actual fragment of an interview are merely illustration of the editorial analyses are the relevant secondary sources here. This is clear from e.g. The Nixon Interviews where (as far as I can see) the actual interviews are not among the sources. Arnoutf (talk) 14:35, 25 September 2011 (UTC)
First of all, there is a strong taint for primary sources expressed in this policy which trickles down to guidelines, essays, and Afd's. The subtlety of their usage by editors leading in some cases to a WP:OR problem is lost and over-simplified by means of a edit comment rm, see WP:PRIMARY. Making it clear what primary sources are and how and why usage of primary sources can leads to OR problems is the purpose of this policy. I believe that useful, good faith edits are being deleted because primary sources are defined over-broadly here, and furthermore the policy is interpreted so as to eliminate usage primary sources per se, and not to prevent the introduction of WP:OR into articles.
Secondly, an interview is always a work product of more than one person, one of whom is a non-participant. Would a panel of interviewers still constitute a primary source? The example you have given, The Nixon Interviews clearly fails the close to the event test. The Watergate break-in was conducted on June 17, 1972. The resignation was August 1974 and pardon a month later. Analysis is newspapers, magazines and books. The most famous of which All the President's Men was already movie in 1976. Clearly, we are now out of close to the event/primary source time-frame. The interviews were taped in 1977 and began to air in May of that year. There was on both sides a great deal of time and effort to prepare. Here's an example from Frost:
Now, if non-Watergate participant Frost simply wrote that down as a hypothetical question in 1977, then it would be a secondary source - but as it was an actual question, it acquires the taint of being a primary source.
I will add that an interview given in the aftermath of an incident to a police investigator or NTSB investigator is a primary source since it is close to the event and documents not interprets it. Specifically, I am addressing the most common form of interview cited in Wikipedia, the journalistic interview. patsw (talk) 17:11, 25 September 2011 (UTC)
In this case you could argue that the interview was socially close to the Watergate event, as it was Nixon himself being interviewed, although thre is indeed temporal distance.
Your quote of Frost shows the core of the issue in my opinion. While Frosts analysis is clearly a secondary analysis of the occurences; the answers of Nixon are clearly his own personal and primary view on the issue. Quoting a journalistic interview without referring to the answers of the interviewed party, and only using the prepared analysis of the journalists is a bit odd in my view. Arnoutf (talk) 18:37, 25 September 2011 (UTC)
The opening sentence mis-defines “interview”. The one offered is not even supported by the link. The applicable definition here would be “A conversation in person (or, by extension, over the telephone, Internet etc.) between a journalist and someone whose opinion or statements he or she wishes to record for publication, broadcast etc.” The one given by the OP is a fairly common type of interview, and is one that is almost always a primary source. There are other interviews, such as with an expert of reputed commentator, who was not a participant, or is not speaking as a participant. Such interviews are much more likely to be secondary sources for whatever it was that they were talking about. --SmokeyJoe (talk) 01:58, 26 September 2011 (UTC)
I agree....interviews aren't first party...secondary yes, third even. but the only way it can be primary if it was hosted up on an official site or an official page.Bread Ninja (talk) 03:06, 26 September 2011 (UTC)
No, interviews can be secondary, but usually they are primary sources. They are primary source material if the interviewee is an eyewitness giving his account. They are primary if the interviewee is the subject. They are primary if they are old interviews about a co-temporary event, and the point is to find out what the general reaction is. --SmokeyJoe (talk) 06:01, 26 September 2011 (UTC)
I don't agree...why are they primary if the subject itself puts itself int he gaze of secondary at least. If subject A is being interviewed for a popular series, and the interview itself is posted on the site that specializes on the media, how can we call it "primary"? i mean, reliable source, or reliable third party sources containing info from primary source.Bread Ninja (talk) 07:42, 26 September 2011 (UTC)
Because the primary-sourced information is not being interpreted/fact-checked by a reliable, secondary source, and because WP:CRYSTAL: upcoming series can be delayed, changed, or even cancelled. Uniplex (talk) 08:24, 26 September 2011 (UTC)
Bread Ninja, I recommend that you go read WP:Party and person. Whether or not a source is "first party" or "third party" ("second party" does not exist outside of contract law) has absolutely no connection to whether it is a primary, secondary, or tertiary source. WhatamIdoing (talk) 20:43, 26 September 2011 (UTC)

Interviews are in general pure and proper primary sources that include two (or more) involved parties - much akin to a reporter reporting first hand information (either from an accident site or from an interview room). Primary is defined with respect to involvement rather than with respect to reliability. Therefore, a primary interview may be reliable. But if the authenticity of a primary source (interview, if you may) is challenged, you will have to ensure secondary sources - and much more care has to be taken in BLPs. Wifione Message 14:55, 26 September 2011 (UTC)

Yes--this is the case in the Frost interview as well. Frost's comments are cast as his opinions, note the use of "I think", "I say", "I know", all of which suggest that his statements are best used as a primary source for his personal stance on what Nixon should do, rather than as a secondary source for Watergate or subsequent events. Yes, secondary sources can contain primary source material--a reporter, for example, writing about an event may quote someone involved, and we can treat that as a secondary source asserting that that involved person said what they said, but we need to make clear that the report stated that the involved party made that statement, and not just include the assertions contained in the statement as if it were the reporter's analysis. --Nuujinn (talk) 15:40, 26 September 2011 (UTC)
A quote in a newspaper article is a primary source within a secondary source, and many sources as a whole will not fit neatly in the box of "this is primary" or "this is secondary" since that's almost a line-by-line decision. I think it's safe to keep interviews as an example of a primary source. The "interview" format is sometimes used for secondary source information, but typically in informal settings that aren't likely to be used as sources for an encyclopedia article. Primary and secondary are fuzzy concepts that defy rigid definition and we shouldn't ask "does it meet the definition of a primary source?" we should be asking "does it robustly verify the statement in our encyclopedia article?" Interviews are rarely useful for that purpose. SDY (talk) 15:56, 26 September 2011 (UTC)
I think that is sound reasoning. --Nuujinn (talk) 16:08, 26 September 2011 (UTC)
Good summary of the issue SDY, thanks. Arnoutf (talk) 17:20, 26 September 2011 (UTC)
If there's a campaign to add some rational nuance to the statement that interviews are primary sources please sign me up. patsw (talk) 17:31, 26 September 2011 (UTC)
I think SDYs analysis is nuanced enough. The problem with naming interviews a secondary source would be that all kinds of idiots would start claiming that any kind of question-answer dialogues should be a secondary source. For that reeaon alone, I strongly suggest that we err on the safe side, which is calling all interviews primary sources. Arnoutf (talk) 18:02, 26 September 2011 (UTC)
I don't see much sympathy for my position: that the exclusion of interviews is clearly enforced in an arbitrary manner and in a manner that allows criticism the elevated status of "secondary source" and response to that criticism to be deprecated as a "primary source". This is even when the participant (our "primary source" in perpetuity) has no choice over the questions asked, or editorial control over the final published work, and regardless of the length of time elapsed between the event and publication. patsw (talk) 18:53, 26 September 2011 (UTC)
Perhaps you can point to some specific examples where the policy as worded created or failed to stop a problematic set of edits? --Nuujinn (talk) 19:13, 26 September 2011 (UTC)
Here's an absolute don't you dare argue with me here about it version from the essay Wikipedia:Identifying and using primary and secondary sources linking not to the main text of the WP:OR but to its footnote:
So are the reports of the reports of interviews primary sources as well? patsw (talk) 20:11, 26 September 2011 (UTC)
Of course you shouldn't argue with me.  ;-)
Imagine a new source that describes a previous primary source. Whether the new source is primary depends on (1) how the new source describes the previous primary source and (2) how you use the new source. It's not just a matter of counting up links in the chain. WhatamIdoing (talk) 20:39, 26 September 2011 (UTC)

Arguing "interviews are primary sources by policy" actually might make little sense depending on how you read the discussion.

If you the entry posting (and part of the subsequent discussion) is simply read as a question, then yes the simply answer is indeed yes our current policy considers them as primary.

If however it is not understood as a simple question but as a debate contribution about the policy, then it makes no sense to argue with the policy that is subject to this debate here. As the discussion here in doubt defines the content of the policy and not vice versa.--Kmhkmh (talk) 21:00, 26 September 2011 (UTC)

It looks like we've agreed here on two points:
  1. that interviews are normally primary sources (with perhaps a few exceptions), and
  2. that "primary" is not another way to spell "unreliable". WhatamIdoing (talk) 14:40, 30 September 2011 (UTC)
Because an interviewees statements are not fact checked, that has nothing to do with primary vs. secondary. It may effect reliability.
Interviewee statements are primary if they are about themselves or things related to them, secondary if they are about something they are not related to. Imagine asking a film critic their opinion on a movie in an interview. This would be a secondary source for the movie. It doesn't matter than he said the words instead of writing them down. We should remove any mention of interviews, unless it's to say they should be dealt with on a case by case basis. - Peregrine Fisher (talk) 14:52, 30 September 2011 (UTC)
Actually the interview of a critic would still be a primary source, as the specific interaction between critic and interviewer is likely to give somewhat different outcomes compared to an independent anaysis by that critic. Of course this is a somewhat grey area, as an open interview question: Please give your analysis about A, may yield a secondary analysis of A (if the interviewed party has no connection to A). But these types of interviews are fairly rare, most interviews are meant to elicit some specific opinion/piece of knowledge outside the normal vision of the interviewed person, and that is what makes most interviews a primary source. Arnoutf (talk) 19:13, 30 September 2011 (UTC)
I don't know if I agree in general, but when you say that most interviews are primary sources, you show that some are not. So, the current statement is incorrect. It's talking about all interviews, as far as I can tell. I don't think interviews are so heinous that we should throw them all out just to make sure. Is that how people feel? - Peregrine Fisher (talk) 00:53, 1 October 2011 (UTC)
Oh no. Arnoutf is mistaken in one conceptual understanding in my opinion. The reason why an interview (all interviews) are primary sources is because the moment an interviewer publishes an interview taken by him/her, by definition, the interviewer is directly involved in the interview by his primary presence/interaction with the interviewer - thus, any and every interview immediately becomes a primary source. The term primary is with respect to the interviewer's involvement (which is primary) rather than the interviewee. Thus, all interviews are primary - some may be reliable, some unreliable. Wifione Message 15:48, 1 October 2011 (UTC)
Peregrine, you seem to be confused:
Nobody is saying that "interviews are primary, and therefore they are so heinous that they should never be used". We are saying "interviews are primary, and therefore they may be used (exactly like WP:PRIMARY says that any primary sources may be used)". WhatamIdoing (talk) 21:47, 1 October 2011 (UTC)
I guess what I mean is that primary sources are treated as a second class citizen in some ways. Mainly notability. I wouldn't be super surprised if the interview example was added to preclude its use in notability discussions. Doesn't really matter now, as we seem to be coming to a consensus below on how interviews are and are not primary/secondary. - Peregrine Fisher (talk) 22:51, 1 October 2011 (UTC)

(redent) So you are saying that an interviewed critic would be a primary source for a movie? Doesn't it matter what the subject is? - Peregrine Fisher (talk) 20:06, 1 October 2011 (UTC)

(ec)I think that Wifione's main point is valid for most interviews, but there's too much variation to draw a bright line. An interviewer may, for example, provide context during the interview, giving a summation of other sources. Or an interviewee may, as has been noted, be an expert on a topic about which they are being interviewed and providing an outsider's view (such as an expert in nuclear safety commenting the history reactor failures). In the former case, I would argue the summation may be secondary, if it represents the result of investigative research, and in the latter, that the material may be secondary an expert analysis across a number of events. The key issue is that primary sources present material of a directly involved party, and as you noted above, it is not a simple matter of counting links in the chain. I would suggest that an interview is substantially no different than an article in a news paper of journal, a blog of an acknowledged expert, or any other source in that it can be primary, secondary, or tertiary, or any mixture of the three kinds of sources. --Nuujinn (talk) 20:15, 1 October 2011 (UTC)
The reason that I believe there are some exceptions to the "100% of the content of 100% of interviews are primary" idea is this:
  • If you ask a historian for a quick summary of what the Spanish were doing in the New World 500 years ago, and you get a written answer that you publish in a textbook, then that summary is a secondary source.
  • If you ask the same historian to summarize the same information, but this time in an interview, saying something like, "Just give our listeners a quick summary of what the Spanish were doing in the New World at that time", then the historian's summary is still a secondary source.
The format of the response doesn't change the classification of the response. It's not "print=secondary, spoken=primary". The information does change the classification: The answer to "Ms Historian, please tell us about your new interpretation of this particular episode" is a primary source.
But for recent events—the sort of things editors are most likely to be reaching for an interview, because textbooks don't cover them, like "Ms Filmstar, tell us about your upcoming movie" or "Mr Talkinghead, please tell us about the current mood is among voters," then that answer is a primary source. I believe that the policy defines interviews as "always" (except IAR) primary is because 99% of the interviews that editors actually want to use are primary, not because 100% of the interviews in the real world are 100% primary material.
It's not a matter of "involvement" in the sense of a conflict of interest (which is the real difference between an affiliated source and an independent one: the first has a dog in the fight, and the other doesn't). It's more a matter of separation in the sense of being outside the mental space of the primary source. WhatamIdoing (talk) 21:42, 1 October 2011 (UTC)
I think that's very well put, and I agree with your assessment of the examples you've provided. --Nuujinn (talk) 22:21, 1 October 2011 (UTC)
I have to disagree with WhatamIdoing; or perhaps we're just mentioning the same issue in different ways. Primary sources are significantly defined by the publishing/reporting party's involvement in the event being reported, not the interviewed party's involvement (There are other ways of defining primary sources too; but in interviews, the issue is involvement). When an interviewer claims to have conducted an interview (and publishes it), then the reportage of the interview by the interviewer is by definition a primary source - irrespective of the contents of the interview, irrespective of the expert profile of the interviewee, irrespective of whether there's a clear conflict of interest, irrespective of whether the source is reliable or unreliable. When an independent source publishes its analysis of this primary source interview, that is one step moved from the event and a secondary source. Where WhatamIdoing is perhaps mistaken (and please don't mind the use of the word 'mistaken'; it's just an opinion), is that WhatamIdoing is also defining the primary/secondary classification by focusing on the interviewed person's involvement. Well, that may be incorrect. The question WhatamIdoing asks ("Is 100% content of 100% interviews primary?") should be modified to be "Are 100% interviewers of 100% interviews taking the interviews themselves?" If the answer to this is yes, then all the interviews taken by the interviewer become primary. Wifione Message 06:13, 2 October 2011 (UTC)
You seem to be leaving out which article we use it for. If a critic says something about the CGI in Star Wars in an interview, and we use that as a source in Star Wars, it's pretty obvious to me that this is not a primary source for Star Wars. No one is involved with Star Wars in this case. How is it primary in this situation, if you feel that it is. - Peregrine Fisher (talk) 00:50, 3 October 2011 (UTC)
Peregrine, with respect to an interview of a critic about CGI in Star Wars, the fact that an interviewer is claiming to have first hand interviewed the critic, makes this a primary source - as the interviewer has become personally and physically involved in the interview. The primary nature of this source is to do with the interviewer's personal presence at the interview table; and not necessarily with which article this interview is being used. The article and the manner in which this primary interview source is used would determine whether the source can be used reliably or not. For example, if the interview is being used solely to claim that CGI was (or was not) used in Star Wars, and if this claim is challenged, then you would need reliable secondary sources to support this claim. That is why it is policy that one should not base articles or even material purely on primary sources like interviews. Wifione Message 04:50, 4 October 2011 (UTC)
Quibble Wifione's word choice: Not to "support", but to show that the fact is interesting enough to be worth mentioning. To establish that others have written about it. --SmokeyJoe (talk) 05:49, 4 October 2011 (UTC)
:) Wifione Message 07:00, 4 October 2011 (UTC)

(Here we set questions of reliability aside.) If a critic says there is CGI used in a movie, this is primary source material. It is a statement of fact without interpretation. If he said that there is a lot of CGI material used [by implication, compared to other movies], then it becomes secondary source material, being basically a statement of opinion by the critic. If he said “The CGI used was impressive” it would be secondary source. If he said (on the basis of counting pixels and frames) that 20% of the visual component of the movie was CGI, this would be primary source information. If he says that a high proportion of the CGI involved saturated colours, this would be primary source information. If he says that the CGI was kaleidoscopic, it would be secondary source information. If he said that he felt the movie was overly kaleidoscopic, this would be primary source information (a first account of a viewer’s impression).
Primary source material is material about the facts and nothing but the facts. What happened. What was seen. Who said what. The reactions of the people there. Primary source material is reliable or not. No amount of mere repetition converts primary source material to secondary source material. Embellishment does not create a secondary source (instead, it converts a reliable information to unreliable information). Simple processing, such as counting, averaging, cataloguing numerically or alphabetically does not create a secondary source.
Secondary source material is a commentator’s creative addition of meaning to the facts. It is commentary, interpretation, analysis, comparison or similar. The secondary source component of the information is the product of the author and is distinct from the facts. Secondary source material is not so much subject to questions of “reliability” (perhaps the author didn’t have his facts right and so his commentary is confused, or his analysis was faulty, but set that aside). The value of a secondary source to us is “reputation” of the source.
“Interveiw” broadly defined is so wide as to be useless to us. We need to know the purpose. Is it to find or discover the facts (ie it is a primary source)? Or is it hear the interpretation of a reputable expert (ie it is a secondary source)? --SmokeyJoe (talk) 00:49, 4 October 2011 (UTC)

RFC Notice - proposed changes to WP:Verifiability

I draw your attention to the RFC at WT:Verifiability#RFC - Compromise proposal re first sentence proposing changes to WP:Verifiability. As this RfC relates to a core policy (one that is deeply inter-connected with WP:NOR) it is hoped that we can receive comments from a wide spectrum of the community (and especially those who regularly work on this page). Please swing by, read the proposal and accompanying rational, and leave a comment. Blueboar (talk) 12:48, 6 October 2011 (UTC)

Incorporating WP:NOTOR

Conflict between sources

There will be times when sources have conflicting facts and opinion. While generally comparing and contrasting these conflicts in of itself is not original research (see WP:NOTOR) great care that synthesis or unsupported conclusions based on those conflicts do not appear in an article. These source conflicts fall into two broad categories: factual and summation.

A factual conflict is where a direct statement (such one source stating as a certain phrase first appearing in 1920) can be proven via reliable sources to be facially wrong (other sources demonstrate that the phrase in fact appeared before 1920).

A summation conflict is when sources disagree regarding a detail regarding the subject.

For example 2002 Ingle's Endodontics 5th edition stated "(i)n the 1930s, editorials and research refuted the theory of focal infection". while 2006 Carranza's clinical periodontology stated "(t)he focal infection theory fell into disrepute in the 1940s and 1950s", and the 2009 Textbook of Endodontology stated that while it had lost its influence "dental focal infection theory never died"

The resolution of such conflicts is a WP:NPOV issue as long as unsupported synthesis, conclusions, or violations of WP:Weight are not made. Noting in the article there is a conflict is NOT OR but presenting an unsupported reason for the conflict is.

For example the Jesus myth theory article had to address the problem of differing definitions and classifications of what Jesus myth theory even was. The solution there was to note that there was no exact and agreed-upon meaning and then listed some of reliable sourced examples to illustrate the conflict. However an actual reason for the conflict was not provided as no reliable source even suggesting why there was a conflict was found.

Another such example from the Jesus myth theory article was how to handle the obvious conflict between Sir James George Frazer stating "My theory assumes the historical reality of Jesus of Nazareth" in volume 9 of The golden bough in 1913 and Schweitzer's grouping of Frazer with John M. Robertson, William Benjamin Smith, and Arthur Drews in his 1913 and later editions The Quest of the Historical Jesus as well as stating in his 1931 autobiography Out of My Life and Thought "I especially wanted to explain late Jewish eschatology more thoroughly and to discuss the works of John M. Robertson, William Benjamin Smith, James George Frazer, Arthur Drews, and others, who contested the historical existence of Jesus." Here as above the nature of the conflict was noted with an additional reliable source added that expressly stated there was a conflict but no reasons for the conflict were provided as no reliable sources for that could be found.

As both of these examples show describing the nature of a conflict is NOT OR though presenting reasons not even suggested by the material definitively is OR.

Editors should resist the temptation to synthesize conflicting sources or in an article use one source to prove another wrong. Editors should also remember that a source only meets Verifiability regarding what it actually says not what the editor thinks it says. The same with statements made in the article itself.

I think some of the points of WP:NOTOR should be incorporated into this article especially the "Comparing and contrasting conflicting facts and opinion is not original research" part because as Wikipedia_talk:Verifiability#Contradictory_sources shows some editors have some really bizarre ideas as to what counts as OR.

I would even go as far as to suggest a WP:SOURCECONFLICT section giving examples of what is and what is NOT OR is needed.--BruceGrubb (talk) 06:02, 20 September 2011 (UTC)

Here is a rough draft of the proposed section:

I moved the draft up top to facilitate discussion. That draft does not cover the problem of sources that have different or contradictory facts. In some cases, there is a well-known dispute over the facts, like the date of birth of Alexander Hamilton. More frequently, there are simply multiple sources with different facts and no explanation for the difference, and no way of determining which one is right. Could we add a sentence or line about those disputes?   Will Beback  talk  22:55, 21 September 2011 (UTC)
All this section does is establish that noting such conflicts in of itself is NOT OR; much of what you want above is in reality beyond the scope of WP:NOR and belongs in WP:NPOV. I've added two passages addressing this and directing the reader to WP:NOPV which is where these types of conflicts should be handled. --BruceGrubb (talk) 04:04, 22 September 2011 (UTC)
So Bruce, the last I heard is that you were in a dispute at Conspiracy theory (or some such article), and a reliable source claimed that the term was first found in 1909, but you found a publication from 1906 that used the term. As I understand it—and I've not bothered to look at the talk page—you want to present your discovery of an earlier use in the article not only as proof that the term was used in 1906, but also as proof that the first source is wrong.
Do you think this change to NOR would allow you to make such a claim? WhatamIdoing (talk) 04:46, 22 September 2011 (UTC)
WP:NOTOR already states "Comparing and contrasting conflicting facts and opinion is not original research, as long as any characterization of the conflict is sourced to reliable sources." and is in the see also list at the bottom of WP:NOR. Putting this into WP:NOR itself with a more detailed explanation would not "allow" anything but rather would point out to editors that such issues are a matter for WP:NPOV and NOT OR.
Take a look at Rklawton's statement of "Looking for sources using the phrase "conspiracy theory" is indisputably original research, and we simply don't publish our own work here" presented in Talk:Conspiracy_theory and you will see why we really need this.
"The term "original research" (OR) is used on Wikipedia to refer to material—such as facts, allegations, and ideas—for which no reliable, published source exists" so looking for a source is NOT OR because we are trying to find an existing, reliable, published source.
"The first recorded use of the phrase "conspiracy theory" dates back to a history article from 1909." (Knight, Peter. "Plots, paranoia and blame". BBC News 7 December 2006) met Verifiability by any reasonable standard--a direct quote by a Senior lecturer in American Studies from the University of Manchester in a well respected paper.
"The fact, however, which makes the conspiracy theory completely illogical is that the political leaders in the slave states were not united in support of the southwest-ward movement, nor those in the free states against it." (Garrison, George Pierce (1906) Westward extension, 1841-1850 Edited by Albert Bushnell Hart LLD Professor in history in Harvard University
"The authoritative action of this caucus, taken in connection with the array of Northern contemporary and later writers that support the conspiracy theory," (History of the United States from the compromise of 1850 copyright 1895 New York, Harper)
"I must content myself with saying that the class-conspiracy theory of economic development may generally be considered false,..." (The Economic review: Volume 1 Christian Social Union (Great Britain) Oxford University Branch 1891 Page 540.)
and several other also met Verifiability by any reasonable standard. Instead of a reasonable WP:NPOV talk of how to address this we had a lot of nonsensical OR claims by editors who clearly did not understand what OR actually meant.
This was again hammered home in Wikipedia_talk:Verifiability#Contradictory_sources where an editor claimed OR even through it had been demonstrated that the conflict between Frazer and Schweitzer being used as an example itself had been mentioned by a reliable source.--BruceGrubb (talk) 06:29, 22 September 2011 (UTC)
Sure, but where's the properly published, reliable source that directly says "BBC 7 was wrong when they said that the first use was in 1909"? Note that I'm not asking, "Where's the real-world proof that they're wrong?" but "Where's the source that said they're wrong?" WhatamIdoing (talk) 19:52, 23 September 2011 (UTC)
The above proposed section already states "Editors should resist the temptation to synthesize conflicting sources or in an article use one source to prove another wrong" so why are you wasting our time with a non issue?
Furthermore exactly WHAT is OR about the following?
"Peter Knight states "The first recorded use of the phrase "conspiracy theory" dates from 1909" (ref); however, the phrase "conspiracy theory" also appears in Garrison, George Pierce (1906) Westward extension, 1841-1850 Edited by Albert Bushnell Hart LLD Professor in history in Harvard University pg 31(ref) and The American: a national journal: Volumes 19-20 May 10, 1890 Page 67(ref)"
The answer is NOTHING. It does NOT "directly state" Knight is wrong but rather what it actually states totally meets Verifiability. Deal with what is actually being presented rather than wasting our time with nonsensical side treks into fantasy land.--BruceGrubb (talk) 20:23, 23 September 2011 (UTC)
What's wrong with saying "Properly Published Independent Reliable Secondary Source X says ___; however, Primary Source Y (which was found by a Wikipedia editor) proves Source X wrong" is that it violates this policy.
You could say, on the basis of the 1890 source itself, "The phrase has been used at least as early as 1890." You may not say anything that implies that any other source is wrong unless you have a source that say the other source is wrong. It's not good enough for BBC 7 to be wrong; to add your proposed sentence, you need a source that says BBC 7 is wrong. WhatamIdoing (talk) 22:07, 23 September 2011 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── Let me make this a little plainer. You want to juxtapose Source X (BBC 7) with Source Y (old documents you found) to produce the conclusion that Source X is wrong.

The policy says: "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."


  • Source "A" (BBC 7) says 1909.
  • Source "B" (the old publications you found) says 1890.
  • The implied conclusion "C"—what the reader will take away from the sentence—is "so BBC 7 is wrong".

The rule is:  :do not join A and B together to imply a conclusion C that is not mentioned by either of the sources". The implied conclusion C (that BBC 7 is wrong) is not found in any source. You therefore are prohibited from saying anything that could be reasonably interpreted as implying that BBC 7 is wrong. WhatamIdoing (talk) 22:14, 23 September 2011 (UTC)

Except WP:NOTOR expressly states "Comparing and contrasting conflicting facts and opinion is not original research, as long as any characterization of the conflict is sourced to reliable sources. If reliable references cannot be found to explain the apparent discrepancy, editors should resist the temptation to add their own explanation."
Each individual part of the compound sentence I provided meets the Verifiability requirement: Peter David did indeed make that statement in a RS (a fact) and the phrase "conspiracy theory" is indeed found in the specific reliable sources provided that predate 1909 (again a fact) and NO explanation for the discrepancy was provided per WP:NOTOR. Per WP:NOTOR THERE IS NO OR period end of sentence go not pass go, do not collect $200.
User John Shandy correctly pointed out we can NOT as you suggest "say, on the basis of the 1890 source itself, "The phrase has been used at least as early as 1890"" because "it doesn't assert any claim about when the phrase was first used." In other words you yourself presented a OR solution to the problem.
In fact sources before 1890 were also found (two primary court documents, the official record of the Parnell Commission (1889), Congressional edition 1888 Senate of the United States United States, and a 1871 issue of The Journal of Mental Science)
Things get even messier as in his book Conspiracy theories in American history ABC-Clio Knight said it was implied that the phase first appeared in 1909.--BruceGrubb (talk) 05:05, 24 September 2011 (UTC)
As I have said elsewhere, you are "characterizing" this fact as being wrong through juxtaposition. You may not do this.
How many more people, on how many more pages, are going to have to tell you that this is impermissible before you'll quit beating this dead horse? It's not going to get up and run, no matter how hard you flog it. WhatamIdoing (talk) 14:37, 30 September 2011 (UTC)
To put it as blunt as possible--is it verifiable that the phrase "conspiracy theory" appeared before 1909--yes or no? We don't want to hear anything else but if the statement "the phrase "conspiracy theory" appeared before 1909" is verifiable. Yes or no.--BruceGrubb (talk) 06:05, 2 October 2011 (UTC)

Oh, are you arguing about this here as well? At least it seems a bit clearer (to me) from some of the above exchanges what the argument is about. If you want my opinion, I don't see anything necessarily wrong with saying "source X says P, but source Y implies Q" (where Q contradicts P). This might have the side-effect of implying something about the worthiness of source X, but too bad - our mission is to provide source-based information, including information about contradictions between sources. --Kotniski (talk) 12:01, 4 October 2011 (UTC)

I find it interesting that once I throw down the "is it verifiable that the phrase "conspiracy theory" appeared before 1909--yes or no" gauntlet the discussion stops dead.--BruceGrubb (talk) 07:12, 8 October 2011 (UTC)
My I suggest that the reason the discussion has stopped dead is because you have raised the same issue on half a dozen other policy pages - and have already received replies to it on those pages... WP:Forum shopping is noticed eventually, and people are going to stop replying to you if you refuse to listen to the replies you have already received. Blueboar (talk) 14:20, 8 October 2011 (UTC)

Dissenting View

Preamble: I use Wikipedia at least 15-20 times a day. Mr. Wales' epistemological theories about compiling knowledge do not bear on a vast majority of articles as they are usually curt and often quite well written summaries of well settled and self-evident fact. To make myself clear, I often use Wikipedia to ensure myself that I am on the same planet as the body of well-settled understandings about either a person, a country, an institution, a scientific theory, or sometimes even just the spelling of a word. I am astonished how often I use Wikipedia to help me remember an actor whose name I have forgotten but whom I associate with a particular film. I don't read the articles critically: and most often I use what can readily be found in Wikipedia as a jumping off point for further and deeper investigation. I think that this is the way Mr. Wales wants us to use Wikipedia. I am reminded that in academe, it is borderline plagiarism to use an encyclopedia as a source. So I suppose that the inverse of the proscription against using an encyclopedia for research is that an encylopedia should not contain native citable material. Since one can not cite an encyclopedia in academe, if one wanted to cite material from an encyclopedia without being charged with plagiarism, it would be necessary to revert to the encyclopedia's stated sources. While permitted, using an encyclopedia to organize academic research is looked down upon and is not the way to get ahead in academe. I respect Mr. Wales' views on the matter. It is his ball, and if we don't want to play by his rules, then I suppose he has the prerogative to go home taking his ball with him.

But: there are instances where Mr. Wales has painted himself into a corner.

I know of one particular incident where there is an article about a "person" [in the Mitt Romneyesque sense] which is largely a self-laudatory auto-biographical paean to the subject of the article. The subject of the article uses a fully-staffed PR department which then hires ostensible independent and regular Wikipedia writers to correct the article in question: especially when cited facts that are offensive to the person appear in the Wikipedia article. I think it is also safe to say that the subject of the article has made donations to Mr. Wales: and that these donations have influenced Mr. Wales as to what un-cited but high favorable content is allowed to remain in the article and as to what highly citable but offensive content must be expunged. It is also clear that this person has a direct tie-line to the Wikipedia legal department: and makes is editorial requests "lawyer to lawyer".Puckslider (talk) 03:27, 8 October 2011 (UTC)

Reporting an event

If a reporter writes an article based on what happened on a press event, would the reporter then be considered a primary source or not? // Liftarn (talk)

It depends on what he says. Is he just repeating? Is he just rewording for yet another newspaper or magazine? If so, it is still primary. Newspaper articles reqorded taken from the news wires are still primary. Language translations are still primary. Such mere-repetition content is typically unsigned.
Does he add his own creative interpretation to the story? Does he commentate or analyse? Does he present an alternative context or perspective to the original press event? IF so, it is a secondary source. --SmokeyJoe (talk) 12:01, 11 October 2011 (UTC)
The issue is at Talk:Energy Catalyzer#Ny Teknik and primary sourcing. The reporter wrote about how long the device ran. Is he then a primary source since he was there? // Liftarn (talk)
This is a FRINGE subject, isn't it? --SmokeyJoe (talk) 10:26, 12 October 2011 (UTC)
Yes... in this case the report is a Primary source for that information. It is an eyewitness account. Blueboar (talk) 12:51, 11 October 2011 (UTC)
So if a reporter writes an article about something he saw we can't use that unless another newspaper article is written about the first article? That sounds strange. // Liftarn (talk)
It depends what he wrote. "writes ... about" suggests a secondary source. Then there are other criteria. The existance of the secondary source suggests that the something seen is wikipedia-notable. Generally, more than one such secondary source is required. The secondary source might be discounted if the author is somehow involved. --SmokeyJoe (talk) 10:26, 12 October 2011 (UTC)
Is "eyewitness account" really the right criterion? I've attempted to point out events in movies I've watched and had them deleted on the ground that there is no review that noted such events. Such events are even more verifiable than the sort of eyewitness accounts you're referring to, since anyone can immediately watch the movie to verify the event. The editors doing these deletions evidently have some clear notion of why such an eyewitness account is inadmissible, suggesting either some problem with this eyewitness-account criterion or some misunderstanding on the part of the deleting editors. --Vaughan Pratt (talk) 07:27, 12 October 2011 (UTC)
Right criterion for what? We can use primary sources, but we have to be careful how we do it... --Nuujinn (talk) 09:02, 12 October 2011 (UTC)
According to Blueboar a review would be considered a primary source and we would not be able to use it. // Liftarn (talk)
A "review" is almost certainly a secondary source. --SmokeyJoe (talk) 10:26, 12 October 2011 (UTC)
I think it rather depends on the nature of the review. Blueboar is absolutely correct that a review written by a reporter who witnessed something is a primary source. For example, a movie reviewer is a primary source for the plot of a movie. But it is also true that reviews often contain secondary source material. If a reporter includes an overview of related technology in a review of a product, that material would be secondary, just as a movie reviewer might be a secondary source for statements about an actress's career. But if a reporter participates in a technology demonstration and takes measurements, they are a primary source for those data--the issue is degree of involvement.
We can use primary sources, but we have to be careful. Primary sources are fine for non-controversial material, but not as sources for experimental data about a controversial topic. Even results from independent testing firms would be primary material. As an example, white papers produced by same are often funded by stakeholders in the outcome, and while the funding entity does not conduct the tests or the results of the analysis, it is not uncommon for them to define the parameters of the tests, what exactly is to be tested, and how the tests should be conducted, and therefore have considerable influence over the results. --Nuujinn (talk) 10:35, 12 October 2011 (UTC)
The terminology can be difficult, yes. To me, "reviewer" in "movie reviewer" is often a misnomer. --SmokeyJoe (talk) 10:48, 12 October 2011 (UTC)
I want to emphasize a common misconception above: "So if a reporter writes an article about something he saw we can't use that"
Yes, you can use primary sources!
You just need to be careful about how you use them. If a reporter writes that the traffic light turned blue with orange and lavender spots, then you can use that source to say that the reporter (said he) saw this—assuming that saying this is both WP:DUE and consistent with your very best editorial judgment.
What you want to avoid is:
  • overemphasizing what the primary source says (violating WP:DUE),
  • including trivia and irrelevancies (offending your best judgment), and
  • exceeding the very narrow limits of the primary source (say, by presenting this one-time, possibly mistaken eyewitness report as the absolute, widely agreed upon, universal truth, which would violate WP:NOR).
This particular policy only deals with the third item.
You might find it helpful to read Wikipedia:Identifying and using primary and secondary sources. WhatamIdoing (talk) 16:29, 18 October 2011 (UTC)

"Primary, secondary and tertiary sources" — Hard Facts

From my reading on this, Wikipedia would frown on going with a primary source, such as a published birth certificate, over a secondary source which comments on it. For certain types of factual information (that is, information that isn't subject to interpretation/opinion), a primary source would be preferred, not a secondary source. The guidelines in this page seems to be focused on facts that are susceptible to subjective interpretation (such as an account of an event) rather than taking into account the existence of hard-facts. Even (US) copyright law takes hard-facts into consideration and exempts them from copyright (thus, the tired-example of being able to legally copy phone numbers from one phone book to make your own phone book). — al-Shimoni (talk) 11:21, 12 October 2011 (UTC)

Outside the natural sciences there's hardly such a thing as a "hard fact". Document and artefacts can be forged or falsely dated or categorized. That's why we prefer quality/expert secondary sources that have surveyed, assessed and analyzed various primary sources with their domain knowledge. This minimizes the chances of falsely using primary sources. Another problem is that many primary sources are not easily accessible which poses issues with the confirmation/verification by the community. Having said that however we do allow the the use of primary sources within in reason and where they are appropriate anyway.--Kmhkmh (talk) 14:14, 12 October 2011 (UTC)
Would that include copies of records such as birth certificates and marriage license that are now in the public domain or social security records made public by the US government (only after the person has become deceased) and easily accessed via internet? Much of this information is heavily utilized by genealogists where primary is preferred over secondary sources. I've seen people kill citations for birth information here on WP because it was a primary source and not a secondary. — al-Shimoni (talk) 10:08, 18 October 2011 (UTC)
Yes, that includes birth certificates. The primary reason for this is that there are a lot of people named "Robert Smith", and the birth certificate doesn't say something unambiguous like "Robert Smith, who is going to change his name and become a celebrity when he grows up".
You are correct that there are times when primary sources are best. It just happens that birth certificates and marriage licenses aren't (usually) one of those times. WhatamIdoing (talk) 16:20, 18 October 2011 (UTC)

Let me give you an example of why primary sources are a problem. Here is a quote from a book published in 1821 (Samuel Seyer (1821). Memoirs historical and topographical of Bristol and it's neighbourhood: from the earliest period down to the present time 1.  Unknown parameter |Publisher= ignored (|publisher= suggested) (help))

See Old Style and New Style dates for the calendar problem. To see a talk page example where this is currently a problem check out Talk:List of Sheriffs of London.

The above was to emphasise a point WhatamIdoing was making. The other problem is one of original research. The use of primary sources without restriction would allow original research to be carried out on Wikipedia, which most editors, not being specialists, are not qualified to asses. For example I was involved in a talk page dispute over whether an eccentric World War II British commander of special forces was under the command of the theatre general or reported directly to Whitehall. The most authoritative secondary source said that he was under the theatre commander. The editor who said he was not, was producing primary sources from an an archive to prove his point. But because he was producing only some primary sources, it was impossible to say if they were representative or significant etc and there was no reason to think that either of us was qualified (expert enough) to decide on the merits of those primary sources. If Wikipedia was to allow such research it would need to do peer review, something we are not set up to do. It would be far too easy for us to publish rubbish (the social science and arts equivalent of perpetual motion machines). So the advise of someone who wants to put a new research into Wikipedia, which will usually be based on primary sources, is go and get it published in a peer review journal first and then cite that in Wikipedia. So two major reasons we discourage the use of primary sources: It is easy to misinterpret them, and if they have not been published in a reliable secondary source it is easy to slip into OR by using them. -- PBS (talk) 20:39, 18 October 2011 (UTC)

which most editors, not being specialists, are not qualified to asses I think that's a typo, but it's a beautiful one. But I agree with your point regarding use of primary sources being perilous in regard to crossing the line into OR. --Nuujinn (talk) 20:54, 18 October 2011 (UTC)

Interviews count as primary, but it says for GNG you need secondary sources.

Discussing this over at Wikipedia_talk:Notability#Interviews_count_as_primary.2C_but_it_says_for_GNG_you_need_secondary_sources. Please join the discussion there on which article to edit and how. Dream Focus 20:31, 18 October 2011 (UTC)

As far as I can tell, we decided at Wikipedia_talk:No_original_research/Archive_56#Interviews_are_not_primary_sources that this policy has incorrect language. I tried to fix it, and was reverted. Changing these policy pages is such a biatch, that I didn't try again. But since almost everyone agreed it is currently wrong, we should change it. - Peregrine Fisher (talk) 05:17, 19 October 2011 (UTC)

The discussion at notability seems to be tending to the idea that, if there's a problem, it's a problem here. Should there be a serious effort to fix this page? SamBC(talk) 14:21, 28 October 2011 (UTC)

It's not an easy problem to fix. While IMO some interviews contain secondary material, nearly all of the interviews that Wikipedians actually want to use are primary sources. We don't (to use an example current at WT:N) usually bother citing media interviews with scholars for basic facts about imperial Rome (an example of secondary material in an interview). We find a better, more obviously secondary source (like a history textbook) for that information.
The sort of interviews that editors actually use tend to be broadcast on Entertainment Tonight, or clips from press conferences with politicians, or chats with daytime talk show hosts, and their focus is on what "I" say about "me": here's my opinion on the book I wrote, my explanation for the bill I'm proposing in the legislature, what I thought about the character I portrayed in my new film, why I left the band to start my solo career, my new plan for achieving a lasting peace in the Middle East, etc. Those are all primary sources.
As a result, if an editor actually needs to know how to classify a source, and he's looking at an interview, then this very simple advice is probably right about 99% of the time. If I were going to introduce the complexity of identifying that 1% of secondary material, I think I'd do it in a supplementary page like Wikipedia:Identifying and using primary and secondary sources. WhatamIdoing (talk) 19:29, 28 October 2011 (UTC)
The discussion there seemed to be more about suitability for establishing notability than about classification. Including folks essentially saying that some primary sources might be good for establishing notability, which is contrary to current wp:notabiity wording. Hate to play hot potato, but seems like more of a wp:notability guideline question. North8000 19:43, 28 October 2011 (UTC)
We would completely agree at WP:N that an interview done with a person that is primarily about that person is a primary source regardless of who published it. We'd also consider eyewitness or bystander interviews to be of the same realm - they're primary sources about an event they are not an expert on. The problem we face is when person A is interviewed about topic B which they are an expert on or of, such as the example you state of a professor talking about ancient Roman history, an actor describing what his role in an upcoming movie is, or a video game developer discussing concepts they used going into the game. Taking secondary sources as those that analyze, review, critique, and otherwise transform basic facts into new knowledge, interviews like this are nearly always secondary for topic B, while primary for the person who is giving them. That is, from a notability standpoint, we regularly accept these types of interviews, as long as they have been conducted and published in an independent reliable source, as sufficient for notability. The problem, of course, is that footnote #3 here classifies interviews without prerogative into a primary source which simply isn't true. Looking at footnote #3, I'd say the same is true of "opinion pieces"; they are primary for the author but secondary for the topic being written about. In otherwords, #3 is making several over-zealous assumptions on what are primary sources simply based on the format of the source and not actually what the source is about. --MASEM (t) 19:53, 28 October 2011 (UTC)
To some extent, I agree with the comment at 19:43: The problem is that the GNG suggests that primary source coverage is absolutely worthless. Unfortunately, the requirement for secondary sources was added in the dark ages, when several senior editors didn't realize that WP:Secondary does not mean independent. Independent primary sources (e.g., Masem's example of eyewitness reports from bystanders) may not be entirely sufficient, but they do suggest a certain amount of notice given to the subject by the world at large.
And this is how we use these: People wrongly claim in AFDs that breaking news reports are "secondary", because they're trying to finesse the GNG's rejection of all primary sources. WhatamIdoing (talk) 17:34, 29 October 2011 (UTC)
    • ^ source 1
    • ^ source 2
    • ^ source 3
    • ^ New York Times.
    • ^ source 5