Wikipedia talk:Reliable sources

Initial comments

Todo:

• Fix internal links (done)
• Write "Common knowledge" section (done)
• Write "Science" section (done)

(I have specific ideas for the latter two; the first, I or someone else will just need to RTFM. -- Beland 08:43, 28 Feb 2005 (UTC))

Yay, drafting is finished. -- Beland 03:30, 5 Mar 2005 (UTC)

Good wark, Beland! JFW | T@lk 01:21, 6 Mar 2005 (UTC)

Semi-common knowledge

Beland, this is extensive, and I've only glanced at part, but ... Can you clarify whether this allows for material that is not known to be published but is common knowledge and undisputed among people with at least moderate knowledge of the subject? Maurreen 09:02, 5 Mar 2005 (UTC)

Maurreen, you have identified the grey area. Some knowledge is so obvious that it requires no proof in sources (e.g. the Sun rises in the East). Yet, for me it is obvious that high levels of parathyroid hormone cause hypercalcemia (elevated blood calcium levels), while others would like to see proof that I'm not making this up. While this is probably undisputed, it is falsifiable due to the (relative or potential) ignorance of the readership. JFW | T@lk 01:21, 6 Mar 2005 (UTC)

Does this mean we have a consensus to keep it gray? :)

Maybe we should just have a sentence somewhere along the lines of "Material which is undisputed should not be removed solely because a source is not identified"? Maurreen 01:29, 6 Mar 2005 (UTC)

I didn't intend for this page to encourage anyone to remove anything; what to actually do when you find an undocumented fact or unreliable source is the question that's being hashed out at Wikipedia:Confirm queried sources right now. -- Beland 01:56, 6 Mar 2005 (UTC)

Everything is allowed, of course, but as for what's recommended as the most reliable...

There's a distinction between facts that one can personally verify and facts that everyone who is moderately familiar with a particular area "knows". For the second kind of "facts", I would recommend seeking reliable published sources or people with personal experience who can serve as primary sources. If no such sources exist, you could either report the statements as widely believed but unverified, or be more specific and reference "less reliable" sources and characterize their reports as unverified or speculative. Even in scholarly communities, it's easy for something to become "widely known" without actually being true. But on the scale of reliability, there are worse things.

If this isn't clear from the front page, maybe it needs to be improved. -- Beland 01:51, 6 Mar 2005 (UTC)

I was thinking of facts that one can personally verify. But no biggie either way. Maurreen 01:53, 6 Mar 2005 (UTC)

Beland, I hadn't looked at your most recent changes. I think that clarifies the matter well. Thanks. Maurreen 01:55, 6 Mar 2005 (UTC)

Beland, this is really good work. Thank you for doing it. I've slightly changed the definition of fact, and I hope this is okay with you. It was: "A fact is a piece of true information about the universe, whether that be a historical event, or an ongoing social or natural phenomenon. "I changed it to: "A fact is an actual state of affairs, which can be an historical event, or a social or natural phenomenon. To say of a sentence that it is true is to say that it refers to a fact." It says essentially the same thing as before, but the latter is more consistent with how the word "fact" is used by philosophers. Do say if you prefer your own version. SlimVirgin 10:45, Mar 6, 2005 (UTC)

I have improved on your improvements. Yay! -- Beland 02:12, 11 Mar 2005 (UTC)

;-) You have, indeed. "Claims of fact" is a lot better. So what is the status of this page? I'd like to start linking to it but don't know how to describe it or how it relates to Cite sources. SlimVirgin 02:29, Mar 11, 2005 (UTC)

Common knowledge

I find the "Common knowledge" section troubling. For example, it says

Wikipedians have should have direct personal knowledge of the facts reported. This does not include hearsay - hearing or reading about something or even learning about something in school. (See the above section "Hearsay".) For those things, you (or someone else) should be able to cite a reliable source.

If someone says "I was there and I saw it", how can we evaluate this claim? Perhaps the person is telling the truth, perhaps lying, perhaps his memory is terrible, perhaps he was hallucinating.

You should evaluate them like you would any other primary source. The author of a famous book could easily also have any of these faults. For some things, having a real live person with an eyewitness account you can communicate with and ask to clarify their statements and whatnot, is certainly better than reading an eyewitness account. And getting several eyewitnesses to different instances of the same phenomena to agree is often an even better verification. But it depends on the subject matter. -- Beland 03:10, 12 Mar 2005 (UTC)

I also find this sentence difficult:

No particular technical expertise should be required to understand or verify "common knowledge" facts.

Many proponents of crackpot theories insist that no particular expertise is required to understand or verify their theories. As examples, how would you deal with people who made the following assertions as "common knowledge"

"The Jewish lobby controls American foreign policy, especially as regards Israel".

"The Israeli settlements are illegal under international law".

Are these statements acceptable in Wikipedia articles as "common knowledge"? Jayjg ^(talk) 03:27, 11 Mar 2005 (UTC)

I would say that these are both statements on a subject about which there is a significant amount of known controversy, so whether or not one has or needs legal expertise to understand these statements is not a necessary question to decide. Though in general, I would not recommend applying "common knowledge" verification to any legal subject. If something is illegal, there should be plenty of written evidence which is more authoritative than the testimony of a random person off the street. This is another reason not to use "common knowledge" verification for the second statement. People are also perfectly free to say, "I don't agree. I think the opposite is true." If there's no rough consensus, then a claim based on "common knowledge" grounds should be removed. -- Beland 03:10, 12 Mar 2005 (UTC)

Beland, sources must be provided; issues of fact, legal or otherwise, cannot be decided by consensus. SlimVirgin 03:40, Mar 12, 2005 (UTC)

I tried to copy edit the Common knowledge section in line with Jay's concerns, but found that it all seemed to contradict Wikipedia:No original research, so I deleted it. It's "common knowledge" that we're trying to keep out of Wikipedia. Everything must be verifiable with reference to credible, published sources. That doesn't mean that a reference has to be provided for every single claim, but there should nevertheless be references available if an editor challenges an edit, and if none are available, any editor may delete the claim. Also, there are a few references to incorrect facts, or "facts, true or false." There's no such thing as an untrue, false, or incorrect fact. A fact is an actual state of affairs, true by definition. SlimVirgin 02:15, Mar 12, 2005 (UTC)

Also, I've just discovered that another page has been created called Wikipedia:Common knowledge. SlimVirgin 03:16, Mar 12, 2005 (UTC)

Have just looked at it and it's the common knowledge section created as a new page, half an hour after I deleted the section from this page. It contradicts Wikipedia:No original research, Wikipedia:Cite sources, and Wikipedia:Reliable sources. I don't know what's going on. SlimVirgin 03:21, Mar 12, 2005 (UTC)

Yes, "Common knowledge" generally contradicts "No original research", which is why it is troubling. I'm not sure why User:Beland recreated it as an article. Jayjg ^(talk) 05:46, 14 Mar 2005 (UTC)

I posted an explanation to Wikipedia talk:Common knowledge. -- Beland 05:50, 14 Mar 2005 (UTC)

Independent primary sources

Continuing the copy edit: I like the discussion about different types of sources, but found this one problematic:

What is an independent primary source?

Independent primary sources:

• Each had direct personal experiences which they are recounting
• Have not discussed their experiences with each other, which could contaminate their memories of events
• Do not have a common influence which could taint their stories in the same way.

That sounds as though we are saying editors may do their own original research and then use it. If we're not saying that, what do others feel the purpose of this section is? SlimVirgin 03:00, Mar 12, 2005 (UTC)

There are plenty of primary sources who aren't Wikipedia editors. Like say we're researching an article on the dot-com explosion. Say the two founders of a company have both written books on the subject. We'd want to compare what they have to say. But what if they used the same ghostwriter? Did they compare notes after the fact but before publishing so they'd have a consistent story? A more independent source might be the book written by the CEO of a competing company, or maybe the company's janitor, or a journalist covering the company during the period of interest. Maybe the page needs some more diverse concrete examples to get the point across. -- Beland 04:57, 12 Mar 2005 (UTC)

I think the point that needs to be stressed throughout is that all material in Wikipedia must have been published in a credible publication already. The way you have written the primary-source section makes it sound as though editors are allowed to go off and do their own research (as though they're journalists), then come back here and report the results. They are not allowed to do this, and this cannot be stressed often enough, because many of them think if they have personal knowledge of something, they're allowed to chuck it in the encyclopedia. Here's an example: there's a rumor that the Apple logo (apple with a bite out of it with gay-pride colors) was designed around the suicide of Alan Turing, who invented one of the earliest computers, and who was homosexual in the UK when it was illegal there; and who was therefore persecuted, and killed himself (it appears) by biting into a poisoned apple. This was in the Alan Turing article without a reference, and I couldn't find one. So I recently e-mailed the designer of the Apple logo, and he told me the truth behind the design. However, so far as I know from him, this has not been published anywhere. Therefore, I can't use that knowledge in Wikipedia. Even if you see something with your own two eyes, it can't go into Wikipedia until it has been published elsewhere first. This is why I'm unclear what you're trying to achieve with Wikipedia:Common knowledge, and I'd really appreciate it if you could explain it to me, because it seems to be going against everything that several of us are trying to achieve at Wikipedia, which is accurate, well-referenced, encyclopedic entries. SlimVirgin 05:09, Mar 12, 2005 (UTC)

Well, given the discussion on Wikipedia talk:Confirm queried sources, not everyone would seem to share the same view. Quoting User:AlanBarrett:

I do not believe that it is a productive use of editors' time to track down sources for claims like "rain is composed chiefly of water", or "animals can die of starvation", or "the nose is an organ associated with the sense of smell", or "the round orange thing visible in the sky is called the sun", or any other item of undisputed general knowledge.

I can see his point. On the other hand, some of these facts do deserve some explanation with references to outside sources. How do we know the liquid in rain is the same as the liquid in the ocean? It certainly tastes different. On the other hand, something like defining the word "sun" in reference to a commonly visible object is not something one really needs to turn to an outside source for. (Certainly you would if you wanted to explain what the sun actually was.) In fact, sometimes Wikipedia can give you a better sense of the meaning of a word than a professionally prepared dictionary, both because the entry is longer, and because we are able to be more up to date. On the other hand, I would trust the etymology in the dictionary more than what Wikipedia says, at least for our current state of accuracy on those types of facts. Unless there's a referenced source, of course. And there should be.

In your particular case, the designer of the logo is the ultimate primary source. Their comments should be cited directly. If those comments have not been published, I would be happy with a simple Usenet posting, as long as there is no significant doubt of the authenticity of the post. It's also worth noting that authors of reputable books cite personal correspondence all the time. I'm not sure it's worthwhile to exclude this from Wikipedia, especially if it can be excerpted directly. I like that "No original research" prevents Wikipedia from becoming like an academic journal, where people publish their experiments and novel theories. But I'm not sure I like excluding "scoops" all together. Especially when there's really no question that they are accurate.

I'm sure there are plenty of published works about the meaning of the word sun, and of course the science of the sun. But there are very few, if any, published sources about how people in wheelchairs ride AC Transit buses. And yet there are many Wikipedians who could confirm the accuracy or non-accuracy of something I choose to write on it. In practice, these types of facts are tolerated in Wikipedia; in fact, they represent a significant chunk of its content so far. I think these facts should stay, and I'm not too worried that no sources are cited. I don't think it's "original research" as Wikipedia:No original research defines it, because the ideas presented are not particularly novel; they are mundane and easily verified. Or maybe I'm wrong, and there is a contradiction with the policy as written. If so, then maybe that policy should be changed to reflect current practice. Unless we decide that our current practice should move toward the existing policy instead. Personally, I see a lot of benefits to "consensus-based fact finding", but I also see a lot of potential pitfalls. I've re-written Wikipedia:Common knowledge as a statement of when this technique is not appropriate, and left the remaining situations in a gray area, because it seems not everyone so far agrees on how they should be treated. Hopefully we can at least agree where the black should be, as a starting point for determining if there is any white left.

I've also clearly labelled these two pages as proposals, not policy, lest anyone be confused. I'm interested to see what other perspectives we'll be getting on them in the coming weeks. -- Beland 05:50, 12 Mar 2005 (UTC)

Beland, if you're saying material can be published without references, you're up against the entire culture of Wikipedia, notwithstanding what any individual editor says. Tell me: you say you'd accept a single Usenet post from the designer of the Apple logo. How would you verify it was him? Please take me through the steps. SlimVirgin 05:58, Mar 12, 2005 (UTC)

Slim, I don't get this statement: "If you're saying material can be published without references, you're up against the entire culture of Wikipedia."

To me, that is at least inconsistent with the number of articles without sources. Maurreen 07:15, 12 Mar 2005 (UTC)

This common failing points to flaws in the articles themselves, not in the policy or theory. Jayjg ^(talk) 05:51, 14 Mar 2005 (UTC)

Well, if something like that is published in a public place (actually, Wikiquote might be even better than UseNet, since that's what it's for), then other editors can e-mail or otherwise contact the source and verify the authenticity of the statement. It's not unlike how newspaper editors check their reporters' quotes - they call up the source, and ask if the reporting is accurate. I would say that if two independent reporters agree, and there's no particular reason to doubt them (they're not known hoaxsters or habitually sloppy; even better if they're regular Wikipedia fact checkers), then the quote has been satisfactorally verified. -- Beland 21:41, 12 Mar 2005 (UTC)

It's very unlike how reporters verify material. Reporters don't just make a couple of phone calls then chuck stuff in the newspaper; or at least they're not supposed to. They usually go to college or university for a number of years to learn how to do it, or serve an apprenticeship; they carefully verify the identities of people they speak to, by telephone and by meeting them (it's considered bad form to interview by e-mail); they verify the information with other, independent sources; they submit the story to their editor; it gets checked by a managing editor, proofreader, copy editor, fact-checker (just one or all four, depending on the set-up); and if it's a sensitive story, with the lawyers, editor-in-chief, and publisher; and if it's very sensitive, with the owners. Wikipedia doesn't have these checks and balances and that's why we rely on information from publications that do have them. SlimVirgin 09:11, Mar 14, 2005 (UTC)

Also, if a book author cites a personal letter from someone as a source, and that is cited in Wikipedia, how is the verification process really any different? Try to contact the source, check for other secondary sources that might confirm or deny the quote (or more often, a slightly skewed summary of a quote), check against what else is known about the source, etc. Maybe we all trust the book author more than a random Wikipedian, but it really depends on the individuals involved. We can also ask questions of the Wikipedia editor, but likely not the book author. Quotes are relatively easy to verify. If a Wikipedian did a science experiment, it'd be hard for other Wikipedians to try to replicate the results, especially since most of us wouldn't have the right expertise. The appropriate thing is for them to publish in a regular academic journal, that people that do have the expertise and ability to attempt replication actually read and write for. Likewise, a novel scientific, medical, legal, or historic theory (or the like) should be put through the grinder of expert peer review before it hits Wikipedia as fact. Otherwise, a lot of things would end up here that we can't appropriately evaluate and accuracy would suffer. I think it's also poor form for people to cite their own research, because this essentially becomes self-promotion. That's what "No original research" is for. But many Wikipedia articles do cite single studies a lot when they have interesting relevant results, and that's informative.

I'm not sure I follow your points here. If a book quotes a letter, we quote the book. We don't need to do any further verification. SlimVirgin 09:11, Mar 14, 2005 (UTC)

I don't see what the point would be of having a policy that says, "no you can't say how people in wheelchairs get on buses until some schmoe publishes a description on their blog". -- Beland 22:11, 12 Mar 2005 (UTC)

If you were really to put your mind to it, you could find a published source regarding how wheelchair-bound people get on buses, because buses have been designed in a certain way to facilitate wheelchair boarding, and therefore the designers of buses will have written this down somewhere. But no one is going to challenge you over edits related to such issues. It's a question of commonsense. SlimVirgin 09:11, Mar 14, 2005 (UTC)

Here's an actual example where I think that insisting on a published source is too restrictive. I made this edit to the article about the Woolworth Building in lower Manhattan. My edit added a paragraph about the effects of the 9/11 attack. The basis for my edit was that, as of 9/11, I was employed in that building. Before the attack, I often saw tourist groups being guided through the ornate lobby. After the attack, we couldn't get into our office for two weeks, and when we did it still took several more days to get back phone service and electricity. I no longer work there, but I go there on business, so I know that there's now a sign near the entrance saying that tourists can't go past it. There's now a guard at a desk in the lobby who verifies that you have business there. I have no idea whether of any of this information has been published anywhere, let alone in a "reputable" source. This isn't "original research", just something I've observed. Should this paragraph be removed unless someone can find the information reported elsewhere? Maybe such a source could be found "[i]f you were really to put your mind to it," but maybe it couldn't; in any event, I don't think that scouring the Web or the stacks at the New York Public Library to confirm this point is the highest priority for someone who's willing to spend time improving Wikipedia. Of course, a different standard will apply if an editor claims to have bribed Giuliani and wants to assert that in an article, based on alleged personal observation, without further reference. JamesMLane 12:07, 19 Mar 2005 (UTC)

Importance

---

From Wikipedia:Cite sources#Merge.

What are the standards of importance for sources? (prompted by recent dispute on Talk:Myth) Sure, we have to Wikipedia:Cite sources, but who and what kind? Does the source need to be a professor of the article's topic? Do they have to be an "expert"? A professional? A friend or relative? Should creationist sources be allowed? Hyacinth 03:19, 1 Mar 2005 (UTC)

For example: mythology.

May no discipline other than mythology comment on myths, or just not on Wikipedia? Why? What about when other fields studies include myths, such as music with texts or tone painting depicting myths. Would a source simple need to be respected in both fields? Actually all three: mythology, poetry, and musicology? (four including history?) Freud may not be an expert in the history of myth, but that does not mean he may not have insight. Also, he was an example, and possibly a poor one, so don't waste your words tearing him apart. Hyacinth 06:13, 1 Mar 2005 (UTC)

I don't know if there is an easily defined standard, but if the purpose is quoting someone on the definition of the term, that individual ought to be highly respected and sourced by other scholarly published references in the field. To use your example, Freud, although he may be entertaining, certainly would not count as notable for defining a field other than psychoanalysis and its offshoots (say, dream analysis or something along those lines). DreamGuy 03:32, Mar 1, 2005 (UTC) (from Talk:Myth#substance)

Shouldn't this be part of Wikipedia:Cite sources or a related policy? It's really a different issue to the importance of different subjects. ··gracefool |☺ 01:39, 26 Mar 2005 (UTC)

For some topics/encyclopedia articles, there are dozens of books written by people who have each spent years researching the topic. For other articles, we're lucky to find a couple of magazine articles written by journalists who spent a couple of days, total, interviewing a couple of people.

* For less-documented articles, I'd rather have those 2 "less reliable" references than no references at all.

* But for well-documented topics, I'd prefer the well-researched books over the magazine articles.

What do you think about setting a guideline of "10 to 20 references are preferred. In articles with more than 20 references, the less-reliable references (and any quotes that depend on them) may be deleted." ? --DavidCary 07:01, 30 December 2005 (UTC)

primary vs. secondary sources

Continuing in the same track as the "Independent Primary Sources" discussion item above, I think that our discussion of primary vs. secondary sources is naive and incomplete. The subheading "Get close to the source" doesn't describe Wikipedia's practices, actual or ideal. For example, if we had a subject with considerable scholarship--say, the childhood of a famous US President--then we should absolutely use primary sources, but only in order to check that the secondary sources we are using are actually based on the evidence they claim to be based on. Just reading the primary sources, then synthesizing a new analysis of the subject, in this case, would be original research (or "novel conclusions" as I prefer). DanKeshet 23:58, Jun 2, 2005 (UTC)

Availability/stability of sources

Another aspect of reliability of sources, that is their availability, started in a new section on the page;

the section What to do when a reference link "goes dead" imported from wikipedia:cite sources, seems more on its place here than in that style guideline.

The corresponding section on the talk page of that guideline read:

Dead link details

Perhaps much of the details at Wikipedia:Cite_sources#What_to_do_when_a_reference_link_.22goes_dead.22 should be in Wikipedia:Manual of Style (links), Wikipedia:External links or Wikipedia:Footnotes. The summary in this article should kindly remind DON'T DELETE DEAD LINKS and point at the details of restoration methods. (SEWilco 04:18, 17 October 2005 (UTC))

• Wikipedia:External links deals, basically, with the "external links" section, and advises that you should delete dead links: clearly the wrong advice for links used as citations. Wikipedia:Manual of Style (links) also really doesn't discuss links as references, except for a passing remark on the discouraged use of inline links rather than fuller references. Wikipedia:Footnotes is about footnoting mechanisms: this would be entirely off-topic there. So this would appear to be the correct place to take up the topic. We might want to refactor material between Wikipedia talk:Cite sources and Wikipedia:External links, but it wouldn't be a matter of moving that one section. Wikipedia:Manual of Style (links) and Wikipedia:Footnotes would certainly be wrong: this isn't a style issue, and has nothing to do with footnoting mechanisms. -- Jmabel | Talk 04:52, 17 October 2005 (UTC)

All in all I think the section should maybe better be moved to wikipedia:reliable sources, because "being available" is of course one of the many aspects of "being reliable" - for instance a web source that is unavailable 90% of the time, is IMHO not the most reliable source, if there's another website with similar content being available 95% of the time. --Francis Schonken 06:58, 18 October 2005 (UTC)

That material is in WP:CITE. There's no need to repeat it here, and the English was hard to understand in places. WP:CITE is about when to use links and where to put them. This is about quality of sources. SlimVirgin ^(talk) 07:38, 18 October 2005 (UTC)

What is a reliable source?

In theory, the policy of quoting reliable sources makes sense. As the policy so eloquently states, you don't rely on the advice of a plumber for dentistry. However, the definition of a reliable source is often in the hands of the rich, the powerful and the well-educated. We take it for granted that scholars are more objective than used car sellers, that teachers know more than students, that the young know more than the old (or is it the other way round?)

You see the problem?

Are we taking it for granted that the 'experts' have the authority? Would that the real world was so simple and straightforward.

Take this example. Once, a scholar made a translation of the New Testament that was roundly condemned by some of the most learned scholars of the day. In fact, he was accused of wilfully mistranslating the holy text. He was finally captured by church authorities, strangled and burnt at the stake. His translations were bought up by the clergy and destroyed.

Yet that man was William Tyndale. His translation of the Bible is acknowledged to be pivotal in setting the standard for later translations. Some of his translation choices that were so contentious at that time are now accepted by all.

Or, let's take a contemporary example. The policy as it stands says that Hamas should never be relied on. Really? Is this plain commonsense or is it evidence of a particular world view?

Michael Glass 10:43, 19 October 2005 (UTC)

I don't agree with much of the above, but I do wonder about the mention of Stormfront, Hamas, etc. as essentially uncitable on anything but their own opinions. Wouldn't it be OK to cite them, for example, on the death date of one of their own members? -- Jmabel | Talk 03:11, 1 November 2005 (UTC)

Jmabel, you didn't make clear what you disagreed with. The example of Tyndale ws to illustrate the point that the conventional experts aren't always right. Michael Glass 13:14, 2 November 2005 (UTC)

Groups like Stormfront may be used as primary sources about themselves, just not as secondary sources about anyone else, and even when using them as sources about themselves, it has to be done with caution. We don't repeat every wonderful thing they might choose to describe themselves as, for instance, even if we were to attribute it. The date one of their members died could be said to be about them, so I wouldn't see a problem with using them as a source for that. SlimVirgin ^(talk) 03:58, 1 November 2005 (UTC)

Rather than a binary "reliable" vs. "unreliable", perhaps it would be better to think of sources in a spectrum, where each individual source has other sources that are "more reliable" or "less reliable. For example, this "a general ranking of sources, starting with the most desirable". (Perhaps there's a better illustration of the idea that "conventional experts aren't always right" -- such as those Guglielmo Marconi proved wrong, or "There is little doubt that the most significant event affecting energy is the advent of nuclear power ... a few decades hence, energy may be free -- just like the unmetered air...." -- John von Neumann, scientist and member of the Atomic Energy Commission, 1955.) --DavidCary 22:13, 27 December 2005 (UTC)

Tertiary sources: using other contemporary encyclopedias as sources for our own articles?

I am quite surprised to learn that WP:RS apparently recommends making use of tertiary sources such as other encyclopedias (here). Do we really want to promote the use of articles from other encyclopedias as sources for our own articles? It doesn't sound good to me; in fact, I think it's quite embarassing to have articles citing 'Encyclopedia X' among their references (or worse, as the only source). I think we should instead stress the absolute importance and primacy of primary sources (which is what I would think belongs in section called 'get close to the sources'). See also this and its talk. — mark ✎ 12:32, 3 November 2005 (UTC)

Mark, what do you mean by primary sources? SlimVirgin ^(talk) 12:37, 3 November 2005 (UTC)

I expected that question — while I was penning down my concern I thought I should try to find a better or more inclusive wording, but the real issue I wanted to draw attention to was the apparent recommendation to use other encyclopedias as sources.

What constitutes a primary source differs in every field of study, I think. I'm working mainly on languages and linguistics, and I would consider specific grammars or descriptive articles in academic journals primary sources for articles on languages and the like; for articles on linguistic topics, I think primary sources are broader works by academic linguists as well as theoretical articles in academic journals.

On second thought, I should note that secondary sources that transparently cite their (primary) sources are OK to me, too — the real problem of most mainstream encyclopedias is probably that they usually don't cite their sources. To give an example: I would find it very embarassing if Niger-Congo languages was based on EB2005's 'Niger-Congo languages', even if their article is quite good and complete and as such reliable. Luckily, our article is based on a lot of primary and secondary sources instead. Now why is the latter scenario better than the first, even though both might (possibly) have resulted in similar articles? Or am I wrong and isn't it better to tell our editors: "Hey, rather than grazing other encyclopedias, you could go read some real books and articles?" — mark ✎ 12:59, 3 November 2005 (UTC)

Anyone any thoughts about this? — mark ✎ 10:35, 5 November 2005 (UTC)

I generally wouldn't use an encyclopedia for most source purposes, but I would use it to fact-check what we've already got. -- Jmabel | Talk 04:27, 6 November 2005 (UTC)

One tertiary source I do often use is book reviews of secondary sources. -- Jmabel | Talk 04:27, 6 November 2005 (UTC)

Just something to think about: academic papers and other primary/secondary sources are meant to serve as sources for research. Encyclopaedias, on the other hand, are meant for fact-checking and as portals for further research, so the former set is probably the better set to use as a source. Ingoolemo ^talk 02:13, 7 November 2005 (UTC)

I don't see a reason why other encyclopedias can't be used as sources, as long as we are assured that they are reasonably reliable. One advantage of using an encylopedia over research papers is that encylopedias only contain information that is agreed upon by a majority of mainstream scholars, while "primary sources" may contain new theories and viewpoints of researchers that do not enjoy extensive support. I was myself struck in a seemingly endless discussion with an editor who wanted to introduce a new viewpoint published in a journal as The True History of the Hindu-Arabic numerals. deeptrivia (talk) 00:23, 1 January 2006 (UTC)

the problem is that general encyclopedias like EB no longer rely on experts, except for "prestige" signed articles. (They kept Einstein's 1911 article on relativity for over 50 years, and put his name in all their PR.) Eb relies mostly on freelancers who do batches of articles at $x per-word. (I worked at a library nearby in Chicago where the freelancers worked.) The freelancers check a few (unnamed) reference books and write up their article, which goes unsigned. They also update the signed articles usually without consulting the original author. (those are signed "X.") The headquarters staff does NOT fact-check the article, it only copy-edits for EB style. When people write in to point out an error, unless it is a howler the note is usually filed away until the next scheduled update, which may be 5 years from now. The result is that EB is ok for high school kids but not a very reliable source for Wiki. Much better are the many new topical encyclopedias that are indeed written by experts and each article is reviewed by scholars. Rjensen 13:10, 25 February 2006 (UTC)

Blogs

I noticed blogs are not viewed as a reliable source for anything other than information about the blogger. Should we not clarify that this can vary, however? For example, in Malaysia, the government keeps the mainstream media on a tight leash, and non-mainstream media is very limited and cannot cover everything. Prominent politicians (such as Lim Kit Siang, Fong Po Kuan and Shahrir Abdul Samad) have thus started blogging to keep the Malaysian internet populace informed. Are these blogs citeable? Some of them have already been cited in Malaysian articles. (For instance, Fong, an alumnus of the International Islamic University Malaysia, has had her blog cited in the article on IIUM.) Should this practice be continued? I think maybe we should be permitted to cite blogs where the bloggers give their full names and make it clear they are not using a pseudonym. Just an idea. Johnleemk | Talk 10:54, 5 November 2005 (UTC)

Good point, John. I think the page does say something about blogs being acceptable where the blogger has some relevant professional interest, but we could add something like "and where it's clear that the blogger is using his or her real name," and perhaps use your example to illustrate the point. SlimVirgin ^(talk) 18:21, 5 November 2005 (UTC)

So I'll add it, then? Johnleemk | Talk 12:03, 6 November 2005 (UTC)

Sounds good to me. SlimVirgin ^(talk) 21:28, 6 November 2005 (UTC)

It might be interesting, in Wikipedia space to start a page, maybe even a project, related to blogs as sources: which ones qualify as citable and to what extent (citable for facts, citable as significant sites of opinion). Or perhaps an equivalent of AfD for deciding a source was citable? Here are two examples I would want to be able to cite for tech-related articles:

• Sorting it All Out, Michael Kaplan's blog on MSDN, which I can say first-hand is often the "rough draft" of official Microsoft documentation in its area, and sometimes out years earlier.
• Lawrence Lessig's blog, why should this be any less citable than his books?

-- Jmabel | Talk 05:21, 6 November 2005 (UTC)

That's a good idea. Andjam was wondering today whether there was a page where he could make a note of a poor source, because of some errors he found in a book. [1] It got me wondering whether we could set up a page for editors to note books that might seem like good sources but that are flawed in some way. I wonder if that could tie in with your blog-page idea. SlimVirgin ^(talk) 05:36, 6 November 2005 (UTC)

That is an interesting idea. Books and webpages, IMO, aren't very much different, except that books tend to be more accurate because their publishers pour a lot of money into them. Dedicated webmasters can create material just as good as or even better than some books on the same topic. It would be good to have a way of finding out "Is this reliable?" Johnleemk | Talk 12:03, 6 November 2005 (UTC)

This could make for an incredibly cool Wikipedia-space project, probably also with potential to spawn encyclopedia-space articles about many of the more useful sources. I don't have time to work on it right now, but I think it would be a great idea. -- Jmabel | Talk 21:01, 6 November 2005 (UTC)

Does anyone have the time to get it going? I'll drop Andjam a note about this discussion, and Slrubenstein. SlimVirgin ^(talk) 21:27, 6 November 2005 (UTC)

With regards to books, I've discovered that wikibooks has the option of describing perceived errors. For example, I've added to wikibooks:Errata/0949853054 some of the errors I suspect ISBN 0949853054 of having. Andjam 12:53, 7 November 2005 (UTC)

Blogs, as are most other webpages, are fairly poor sources for anything, and if a print source is extant that is highly preferable. Blogs are invariably about primary material elsewhere, and bloggers very frequently regurgitate opinion formed elsewhere, even if not cited by name.

The exceptions to the rule, as stated above, need to be inspected closely to come up with some sort of a working definition of what exactly constitutes a "citable blog". A blog by a prominent personality may in itself be citable. But are these statements actually "on the record" as far as journalists etc are concerned? If they are, then these may be valuable sources of information.

Another conundrum is the matter of affairs conducted exclusively on blogs (i.e. formation of public opinion solely on blogosphere with little meatspace antecedent). This merits further discussion. JFW | T@lk 22:08, 6 November 2005 (UTC)

The current disctinction that we have between primary sources and secondary sources allows a blog to be used as a primary, but not a secondary, source. Wikipedia:Reliable sources#Evaluating secondary sources explains what is required to be viewed as a reliable secondary source. Very few blogs qualify, and if they did they might not be called "blogs". I think that if online newssources fill the existing qualification they could be used. I don't see a need to change the wording of the guideline. If there were a blog by US VP Cheney for example we could use it as a primary source for his opinions, but not as a secondary source for reliable information. On the other hand, an official "office of the Vice President" website would probably be viewed as a usable secondary source. -Willmcw 23:32, 6 November 2005 (UTC)

I like the idea of the Errata Wikibook. Should these guidelines suggest that, when spot-checking a book to see if it really does say what some wikipedia article claims it says, we recommend also checking the Wikibooks:Errata for that book? --DavidCary 07:01, 30 December 2005 (UTC)

Another problem with outright banning blogs as sources is over at List of political epithets. Since political epithets ofter originates and are used in the blogs and rarley gets mentioned in acedemic papers published in peer-reviewed journals it means the no blogs rule can be (and has been) used for selective censorship. For instance it may be impossible to find any non-bog source to words like "freeper", "idiotarian" and "moonbat". // Liftarn

A quick search turns up the following : "Those of us who believe that our president has to be accountable are not all 'deranged moonbats.'" (George W. Bush is no FDR, Los Angeles Times, November 18, 2005). "Entire Web sites spring up to monitor what journalists do, to call us 'Idiotarians,' to condemn us as anti-Semitic or anti-Muslim ..." (The truth is out there somewhere, The Toronto Star, January 8, 2004) "The war supporters, many of them known as 'Freepers' because of their devotion to FreeRepublic.com, insist it is unpatriotic and demoralizing to protest the war ..."(A Weekly Battle Over War in Iraq; Two Camps Stand Ground by Walter Reed, The Washington Post, October 30, 2005) -- Dragonfiend 15:34, 9 January 2006 (UTC)

But none of those are primary sources, only secondary sources and some thus argue that they are not usable. // Liftarn

When the term has no currency outside of blogs, then blogs are a primary source for the fact that is is used in blogs. But if it hasn't precipitated out of the blogosphere, it has to have quite a presence in the blogosphere to reach encyclopedic notability. - Jmabel | Talk 02:15, 10 January 2006 (UTC)

It may be used outside of blogs, but then perhaps only when people talk about the term and that is thus a secondary source and not usable. See Talk:Self-hating_Jew#Critical_of_Israel_etc. and you understand the problem. // Liftarn

Why is a secondary source "not usable"? We use secondary sources all the time. History books, for example. In a word usage case like this, a secondary source would generally be preferred: trying to gather evidence from primary sources for notability of a word use borders on original research, really more of a lexicographer's job than an encyclopedist's. -- Jmabel | Talk 00:34, 11 January 2006 (UTC)

It is User:Urthogie and User:Jayjg who demands a primary source that is not a blog, personal website or of a biased organisation or simmilar. The problem is that those who use it as a political ephitet (i.e. the primary source) are... well, not very unbiased so it's unusable for that reason. The usage may ofcourse be covered in newspaper articles, books, academic papers and so on, but those are secondary sources and they are claimed to be invalid. // Liftarn

Can you point me to the basis on which you are saying that secondary sources are not usable? As I said above, this is a case where I would consider a secondary source not just acceptable, but preferable. -- Jmabel | Talk 08:07, 12 January 2006 (UTC)

"Unfortunately those are secondary sources, so they may be true, or may not be.", "so you could you provide a primary source, of him saying that directly, like not an editorial about him", "What other primary sources do you posess", "Although this may be a scholarly, encyclopedic source, it is not a primary source", "What you'll need to prove me and Jayjg wrong is primary source proof.", "your sources you just provided are either secondary sources", "it has to have been in one primary source used that way" (all quotes from User:Urthogie). He and User:Jayjg demands a primary source that is not a blog, personal website or of a biased organisation or simmilar. // Liftarn

Speaking of sources: you are giving no indication where any of this is written, so they are devoid of context. I see nothing from User:Urthogie at Talk:List of political epithets, the talk page of the article where you said you were being prevented from using these citations. All I see from Jayjg there are some objections to including "Anti-Semitism" and "Judenhass" as political epithets. I'm not sure if I agree with his rationale, but that is another matter. -- Jmabel | Talk 06:37, 13 January 2006 (UTC)

As I wrote earlier it's from Talk:Self-hating_Jew#Critical_of_Israel_etc.. The comments from Talk:List of political epithets are archived by now. // Liftarn

The ban on blogs is absurd. For one, the term is entirely too vague to be in any way meaningful, as it covers everything from Washington Post columnists to Livejournal posts. --The Cunctator 05:34, 7 March 2006 (UTC)

Bulletin boards

Forgive me if this point has already come up.

"Posts to bulletin boards and Usenet, or messages left on blogs, are never acceptable as primary or secondary sources. This is because we have no way of knowing who has written or posted them."

A reasonable policy in principle, but what happens if the author or poster can be reliably identified as a public figure? Shouldn't exceptions be allowed in such instances -- ie. shouldn't Wikipedians be permitted to cite these sources to reference the opinions of such figures? CJCurrie 23:52, 7 November 2005 (UTC)

This is just a guideline. There may sometimes be a good reason for not following it in particular cases. Friday (talk) 23:55, 7 November 2005 (UTC)

An example of where not to follow it would be an article about a Usenet group (i.e. alt.romath). Without some use of Usenet postings as sources, it'd be virtually impossible to write articles about them. It be ideal if we could draft language to make explicit some common-sense exceptions. -Willmcw 00:00, 8 November 2005 (UTC)

I question the blanket provision against citing, only with regard to determining if something is a neologism or not. In AfD discussion one often sees "ran Google search, did not find this term, likely to be a neologism, delete as non notable" or similar. A recent AfD I was a party to seemed to turn on whether a term even existed or not. Copious evidence was presented that large numbers of blogs/forums/discussion groups and so forth used the term, suggesting that the term does exist and is used by significant numbers of people. If you can demonstrate that hundreds or thousands of sites use a term, is that not sufficient to demonstrate that the term is widely used, even if you cannot produce a verifiable source like the New York Times actually using it? I do not suggest that a single board is sufficient, or that you should take claimed meaning verbatim, but it does seem a valid argument against neologism. (if it is not, then the converse given in the lead sentence of this paragraph, is, in my view, not valid either) Thoughts? (note that this does not mean that an article might not nevertheless be deletable on other grounds) ++Lar 22:30, 12 December 2005 (UTC)

This is currently an issue for Paleoliberalism. But it cuts two ways, depending whether the blogs are being used as a primary or a secondary source.

Right now, the bulk of the article is about a usage that can only be found (at least so far) on blogs and personal web sites. It is found on enough blogs and personal web sites that it clearly has some currency, and should certainly be mentioned in the article (although, I don't really see why the article should largely consist of a near-manifesto for what is essentially libertarianism by another name). On the other hand, there is an uncited claim that the usage derives from Ludwig von Mises. Should be verifiable, you say? What's the citation? Well, a Google search reveals that one of those blogs claims that von Mises said this in private correspondence. Right.

Which is to say, that the blogs, taken collectively, are a perfectly good citation as a primary source for this word being used, with a more or less consistent meaning, in the libertarian blogosphere, but they are not an appropriate secondary or tertiary source for the claim that von Mises coined this usage in private correspondence. (And I guess I should copy most of this to the talk page of that article.) -- Jmabel | Talk 06:34, 13 December 2005 (UTC)

References/external links name-change proposal

There's a proposal at Wikipedia_talk:Verifiability#References_title_misread_as_non-web_External_links to change the References header to "Sources", and External links to "Further reading". So far, the proposal has been accepted by all the editors on the page, but because Wikipedia:Verifiability is a policy page, I'm putting it out for further discussion before changing it.

The reason for the proposal is that using "References" and "External links" is confusing. Sources are supposed to be listed under References, and any further reading is listed under Further reading or External links. But many editors think that any external links, whether used as sources or not, should go under External links, so then they list any material that isn't online, like books, under References, even if not used as a source. To cut through all this confusion, the proposal is to change the headers to Sources and Further reading, which are self-explanatory, and don't make the online/offline distinction. Comments would be welcomed. SlimVirgin ^(talk) 22:06, 25 November 2005 (UTC)

Experts

I think that for sources on minor points of an article, a personal web page of a known expert in his field is better than nothing (I would prefer a peer-reviewed paper by a known expert, but these are not always available).

For example: William Kahan is a known expert on floating-point arithmetics, and one of the architects of IEEE 754. He has a web page with tons of papers, some published, some unpublished. When he says something on the issue, I think he is a more reliable source than many non-specialists who publish on related issues in peer-reviewed papers. Yet, according to our rules, we cannot cite his papers! This is absurd.

Another example: if on the article of Objective Caml I wanted a source for some statement I make (say, about the efficiency of garbage collection), I would quote what Xavier Leroy, designer of the system, said on the issue. He is probably the best expert on what he did himself. Yet, by our own rule, we should prefer to quote some non-specialist writing a textbook on programming. Again, this is absurd.

So I propose a change of policy: for want of better sources, we can quote reputed experts writing on their own topic on home pages, Usenet postings etc. Of course, this should not be the only source, and should only be done for uncontroversial specific specialized aspects of a topic. David.Monniaux 11:14, 1 December 2005 (UTC)

What are minor points of articles, what is a reputed expert, and what is for want of better sources? To give an example, I and a few other editors have been trying to keep the original research and fringe theories of Roylee (talk · contribs) out of Wikipedia (much is documented here). What policy captures the problem better than Reliable Sources as it stands now? — mark ✎ 12:14, 1 December 2005 (UTC)

1. Minor point: something interesting and worthy of mentioning, but which is not one of the defining characteristics of the topic being discussed. Considering IEEE-754, for instance, the fact that floating-point operations on x87 can incur "double rounding" is an interesting but minor point, while the division into mantissa and exponent is a definining characteristic.
2. A reputed expert is a person that is consistently esteemed by his peers as a specialist of a particular domain in an established field (as in "peer review"). This is how peer-reviewed publications work: at some point, you have reputed experts making up "editorial boards" and "program committees". In the above case, William Kahan is a reputed expert on floating-point because he was one of the main designers of IEEE-754, is a Turing Award laureate and an ACM Fellow.
3. For want of better sources: when we do not have a textbook or peer-reviewed paper discussing the same issues.

So there is clearly a difference between people like Kahan or Leroy, on the one hand, and unknown Wikipedians like Roylee. David.Monniaux 14:49, 1 December 2005 (UTC)

I think the page does already say we can use blogs and personal websites if a well-known professional in the field has set one up under his or her real name, though we can't use Usenet postings because there's no way of knowing who has posted them. However, even though personal webpages of this kind are sometimes allowed, they should be used with caution, because there's no form of third-party fact-checking when someone self-publishes, and we have to ask ourselves why someone well-known and good at what they do has to rely on self-publishing information that appears to be worthy of publication. Other publications may have good reason not to want to publish it, and if they don't want to, we probably shouldn't want to either.

By the way, published articles don't have to be in peer-reviewed journals, David, just published somewhere reputable. SlimVirgin ^(talk) 14:29, 1 December 2005 (UTC)

Important RfC: please weigh in

An RfC has been opened recently concerning an editor who has spread fringe theories and original research over a wide array of articles (>100) making clever use of cross- and self-referencing, thus making his contributions looking reliably sourced and verifiable to editors who assume good faith. Finding a solution to this problem is of imminent importance to Wikipedia's future reliability and verifiability. Please weigh in at Wikipedia:Requests for comment/Roylee. — mark ✎ 14:09, 3 December 2005 (UTC)

Don't be lazy

One of the very last headlines is don't be lazy. This is an inappropriate exhortation in Wikipedia. It implies that editors have a duty to expend a certain effort on Wikipedia and that some may not fulfill this duty. This assumption is not correct. Any person chooses completely feerly how much time and effort she donates to Wikipedia. The instructions and rules can only concern how this work is done. --Etxrge 08:04, 11 December 2005 (UTC)

I see this as only one step removed from "first, do no harm". -- Jmabel | Talk 20:27, 11 December 2005 (UTC)

What about a book now rejected by the authors on their homepages?

What about the case of two authors who wrote published books on a certain subject that they now vehemently reject on their respective homepages. Can I cite the book, but not their homepages on which they give reasons for the rejections of their books? See Talk:Sathya_Sai_Baba#Abuse_of_referenced_links Andries 22:05, 17 December 2005 (UTC)

Wikibooks:errata

Someone recently pointed me to Wikibooks:errata. Sounds like something that may be of interest to many people who watch this page. -- Jmabel | Talk 18:24, 30 December 2005 (UTC)

Orphan journal articles

One thing that has come up a couple of times is people citing "orphan" scientific papers (that is, ones that have never been cited anywhere else) on various topics, for instance on Nuclear power (see the edit history), or perhaps the canonical example of Traumatic masturbatory syndrome. Would it be worth specifically noting that orphan papers don't generally count as reliable sources? --Robert Merkel 06:40, 4 January 2006 (UTC)

Certainly orphan papers should, at best, be used with caution. I'd say this is much more of a rule in the sciences than elsewhere, and (to state the obvious) should be a pretty serious warning flag in a much written-about subject matter area like nuclear power. I'd be less worried about citing, for example, an "orphan" dissertation about (for example) a commune in rural France; it might simply be the only academic writing on the (obviously legitimate) topic. -- Jmabel | Talk 07:55, 5 January 2006 (UTC)

Newspapers preferred over blogs

I'm trying to get a good sampling of third-party feedback on the use of newspapers as a source instead of blogs. In the case of the Robert Clark Young article, I have found newspaper sources that cover much of the same material as the blogs that have been used in the article. When I replaced the blog sources with the newspaper sources, Alabamaboy reverted every single one of my edits. Also, very strangely, he accused me on the discussion page of being Mr. Young himself!

This is the Wikipedia Guideline I am trying to follow with my edits:

"Publications with teams of fact-checkers, reporters, editors, lawyers, and managers — like the New York Times or The Times of London — are likely to be reliable, and are regarded as reputable sources for the purposes of Wikipedia. At the other end of the reliability scale lie personal websites, weblogs (blogs), bulletin boards, and Usenet posts, which are not acceptable as sources."

Thus, I have replaced the blog sources with newspaper sources. Again, let me stress that this has not led to much change in the text of the article itself--what I'm trying to do here is change the nature of the sources so that they themselves comply with Wikipedia Guidelines.

Could Alabamaboy and I get some feedback on this? I'd like a few editors to go over to the Robert Clark Young history and compare both versions of the sourcing--the one using newspapers, and the one using blogs. Thank you. Berenise 01:11, 6 January 2006 (UTC)

Editors should be aware that there are three reasons the article was reverted: 1) Berenise made the changes despite a lack of consensus and my objections on the Talk:Robert Clark Young. In short, the online references are refered to in the newspaper and print articles, making the online sources primary sources. The article also has many print sources which complement and add to the online sources. 2) The edits made the article less NPOV b/c they removed opposing viewpoints. While these references may be online, they are from credible named sources who are considered experts in their respected areas. 3) There is a strong possibility that Berenise is Robert Clark Young. Young previously edited the article about himself and most of Berenise's edits since coming to Wikipedia have been to the Young article. I'm trying to clear this up with Berenise; once she proves she is not Young I'd love to get opinions from other editors about this situation. For full details, see Talk:Robert Clark Young. Best,--Alabamaboy 01:17, 6 January 2006 (UTC)

Berenise seems to have started this same conversation several places, so I guess I will paste my comments here as well.

Before seeing Alabamaboy's comments I had written:

1. I've crossed paths with Alabamaboy quite a few times, and he usually knows what he's doing. Yes, in the case of an article where the subject of the article has been known to meddle with the article and where a new user appears and seems focused almost entirely on that article (I didn't verify this independently, if it is a mischaracterization it is his not mine), it is reasonable to ask that person if they are the same person as the subject of the article. Conversely, it would also be reasonable to assume good faith.
2. In general, if particular material can be cited from newspapers rather than blogs, we should prefer the newspapers. The only exceptions I can think of offhand are (1) if the newspaper was quoting or paraphrasing the blog (which makes the blog the primary source and the newspaper secondary) or (2) if the blog is that of someone directly involved in the story. For an example of the second case, (and I have given only a brief glance at the article), it would be entirely appropriate to quote Robert Clark Young's own blog, if he has one, or that of Melany Neilson, etc.

Once I saw Alabamaboy's comments I added that I have no opinion on the claim that Berenise might be Robert Clark Young, but it sounds like the blogs are the primary sources, in which case (as I wrote above, not knowing this) they should be included. -- Jmabel | Talk 06:58, 7 January 2006 (UTC)

Wikipedia:Notability (websites)

I've recently rewritten the above page, and there is currently discussion on the talk page regarding the viability of certain websites as reliable sources. Comment would be appreciated. Steve block talk 16:47, 6 January 2006 (UTC)

Foreign Media

How does this apply to foreign media used as primary sources ? Secondary sources ?

The article "Flying humanoids" was sourced from Jeff Rense's website, which has cited Mexican media sources for sightings of these creatures in Mexico, and used Texas media as sources for material reported in sightings of these things in the US, mainly in Texas. Martial Law 21:26, 15 January 2006 (UTC)

I'm not sure I understand your question. The criteria for reliable sources are pretty much the same regardless of what country the media are from ("foreign" is all a matter of where you are standing). The only clear exception is that one would be wary of media from a country with a highly censored press. -- Jmabel | Talk 06:35, 26 January 2006 (UTC)

This guideline in relation to No Original Research

This page says in bold, "We may not use primary sources whose information has not been made available by a credible publication. See: Wikipedia:No original research." Of course, if you then go to NOR, it says: "This is not 'original research,' it is 'source-based research,' and it is fundamental to writing an encyclopedia."

This part of this guideline is fundamentally wrong and in contradiction to what NOR actually is. Unless the editor's writing is defining new terms, introducing a theory, making or refuting an argument, etc, then examination of a primary source, even without publication is entirely acceptable. The primary source is' the source for verifiability.

This also extends to things personally witnessed. If an editor wrote "The airplane fell out of the sky", because they witnessed it, that is not original research. If an editor wrote "The airplane fell out of the sky because terrorists blew it up." then you have a NOR problem, the editor is introducing a theory on why the airplane fell out of the sky. SchmuckyTheCat 08:39, 16 January 2006 (UTC)

I think when the nit-pickers say "No original research", they sometimes mean "Verifiability". And someone simply witnessing the aircrash is not (in Wikipedia terms) verifiable unless it is reputably reported. Although I understand the people who place verifiability above truth, I have little sympathy if this is applied strictly to ordinary, work-a-day, articles. Thincat 16:22, 26 January 2006 (UTC)

Why is it not verifiable? The event is a primary source, and presumably, there is a crash site. Primary sources are absolutely critical in encyclopedia writing. SchmuckyTheCat 23:53, 29 January 2006 (UTC)

Well, I think it is verifiable but the Wikipedia consenseus is that it is not.

• Wikipedia:Reliable sources: "Wikipedia articles should use reliable published sources".
• Wikipedia:Verifiability: "Facts, viewpoints, theories and claims in articles must only be included if they have already been published by reliable and reputable sources".
• For example, Wikipedia talk:No original research#Eyewitness testimony

I think this is a great shame and that people should be able to put in "what they know to be true". However, I agree that a reference can properly be demanded by other editors and that if other editors dispute the "fact" it can properly be removed. Clearly cranks can put in all sorts of facts that they "know" (and others may think that I am one of the cranks) so there needs to be protection against this. Thincat 12:40, 30 January 2006 (UTC)

I don't consider other users opinions on talk pages as authorative. I'll continue to reject the opinions of people who think primary sources aren't allowed. There is a fundamental contradiction here, and I'll stick to the KISS principle and the original sense of NPOV. If it doesn't formulate NEW opinion, conjecture, defintions, etc, then it's not original research. SchmuckyTheCat 17:46, 30 January 2006 (UTC)

Company web sites

Is a company's official web site a "reliable source", especially for basic details of the company such as when it was founded and their primary product lines? I realize that sources independent from the company may be better, but often the best source of information is the company itself. -- Kaszeta 16:01, 22 January 2006 (UTC)

Generally, yes. It's certainly not a reliable source for how their product compares to other products, of course, but the kind of things you are listing, yes. It's important to cite clearly, though, on anything where there might be a matter of controversy. For example, a claimed founding date is sometimes the founding date of the oldest of several merged companies. -- Jmabel | Talk 06:39, 26 January 2006 (UTC)

Okay, as a more specific case, I've had a disagreement about Flying Buffalo with another editor who insists that the "No Original Research" tenet should keep me from using Flying Buffalo's own web site as references and removed the citations. Since it's a somewhat obscure (but notable nonetheless) company, their own site is probably the best resource. They do make claims on their site that could use verification (like being the first computer-moderated play-by-mail company), but getting basic facts like the company's history and product line should be reasonable, no? -- Kaszeta 14:30, 26 January 2006 (UTC)

It seems to me that Wikipedia:Reliable sources#Personal websites as primary sources is a pretty close analogy. You can use their web site for information about the company but probably not about anything else. It looks to me that is just what you want to do. However, things are more subtle, as you say. What if they say they were the first compamy to do so-and-so? Is this a statement about themselves or about someone else? I think the "great caution" bit is far too cautious and there shouldn't be a problem unless the actual content quoted is genuinely in some doubt. Anyway, if you say "the company's web site says "...", surely you can't go far wrong. But I'm not an expert! Thincat 15:55, 26 January 2006 (UTC)

In terms of things like a listing of their products, this should be uncontroversial, to the point where it is hard to imagine a good-faith objection. In terms of anything else, if there is any doubt the text can make it clear that this is the company's own claim. - Jmabel | Talk 22:48, 29 January 2006 (UTC)

Wikinews

I'm having some trouble with unreliable sources around... and while I know they're true, I admit they aren't all that reliable, so I can't back it up. It's quite frustrating. After thinking a lot about it, I came up with a question that is all about its implications... Is Wikinews reliable? If not, will it ever be? • Ekevu • ♥ • ★ 16:28, 26 January 2006 (UTC)

Court cases?

A number of times I've noticed that someone cited a court case to support a particular fact: for example, "Lynchburg Circuit - Criminal - CR91003195-00." I've been unable to find anything on WP:RS stating that this is acceptable or that this is unacceptable. Thoughts?

Seems to me that this should not be acceptable. This material is not available online; in most cases cannot be ordered online for the corresponding courthouse; in some cases, it can be ordered for money. However, in most cases, one needs to actually go down to the court - in this case, Lynchburg, VA - and physically file a request to see the court records. That makes verifying the material virtually impossible for any editor not living in the locality. --Pierremenard 19:50, 30 January 2006 (UTC)

• Sucks, don't it. But that's the nature of primary source material. If it's a court case, and you have an actual legal citation (the court, the docket, which document in the docket, which page on that document, which line on that page) then in most American jurisdictions you can probably get the clerk of the court to verify what it says or order a copy of just that page. SchmuckyTheCat 07:26, 9 March 2006 (UTC)

I am a crowd

It seems to me that some people add information to a page without knowing if their information is correct or common knowledge; they don't even know if any experts or a majority of the population believe it to be true. They might even acknowledge it's a myth that a lot of people are mistaken about. But they take a "one-person survey" -- themselves -- and rationalize that if one person believes it is true, they cannot be the only one. Considering the millions of people there are in the country, or in the world, surely, there must be many people with the same opinion. So, they add their statement and qualify it with the introduction, "Many people believe that...", thinking that whether it is true or not, it is an important statement about society that many people believe it. GUllman 19:07, 31 January 2006 (UTC)

Could someone please comment on this? Is this a valid way to contribute to an article -- adding your own ideas and then waiting to see if anyone deletes or adds support to it. In a way, it could be considered original research -- surveying the Wikipedia community to see how widespread an idea is, without having to cite an authority or a reliably performed survey. (Sometimes, there is none for those topics that no other encyclopedia covers.) This page should mention this type of edit, and whether it is a reliable source or not. GUllman 20:28, 2 February 2006 (UTC)

Bulletin boards

The current version states

"Posts to bulletin boards and Usenet, or messages left on blogs, are never acceptable as primary or secondary sources.

This is because we have no way of knowing who has written or posted them."

This sounds in some specific cases as excessively skeptical: some people have a prominent internet presence under their real name and there can be no reasonable doubt about their identity. What about people who list e.g. their own yahoo groups on their homepage? In those cases there can be no reasonable doubt about the identity of the posters. Andries 13:14, 4 February 2006 (UTC)

What if we are talking about websites or software? What the admin of the site posts should be pretty reliable. WP 10:15, 2 March 2006 (UTC)

Sources in languages other than English

I disagree strongly with much of the section "Sources in languages other than English". I started some of this discussion a few weeks back on WP:V; my issues were generally not addressed there; instead, the material seems now to have been moved here, so I will start over on this page. I will try to confine myself to my major issues with this; the following list could easily be tripled in length.

Much of this section seems to me to be part of a process of dumbing down. Much as when people object to particular sources on the grounds that are not universally available either online or at the average public library, ease of verifiability is deemed to trump scholarship and accuracy. To me, the section reads like it was put forward by someone who does not do much research with non-English-language sources. It is weighted entirely to the convenience of a monoglot English-speaker who feels that language should never be an obstacle to checking all sources for him-/herself.

• "… English-language sources… should always be used in preference to foreign-language sources." Should people who read Arabic prefer to use a translation of the Qu'ran to the original? Should I prefer the (often hideous) quasi-official translation of a Romanian government document to the original? I think not.
• "For example, do not use a foreign-language newspaper as a source unless there is no equivalent article in an English-language newspaper." Two questions (both rhetorical, I'm afraid):
  1. Many of our contributors, especially on matters outside the English-speaking world, do not have native English. Are we actually telling them that not only do we expect them to write in English but to track down their research sources in English?
  2. Does this mean that if I translate an article from Catalan or German I am expected to re-do all of the research rather than assume good faith on the part of the original researcher?
• "Readers may not be able to read source materials in other languages…" and readers may not have access to a particular book you cite, and readers may not be familiar enough with the subject matter to judge correctly handle a generally but opinionated source, and so on. Such is life. Not all sources that can be successfully used by a capable researcher reasonably expert in the topic can be successfully used by everyone who walks in off the street. Learn to live with it.

As I said, I could go on. If enforced, these rules will discourage a lot of good work that has been going on. - Jmabel | Talk 06:38, 5 February 2006 (UTC)

I agree completely. No serious academic community would ever have a rule like this - I think it makes the English Wikipedia community look incurious and ill-read. As a practical matter, Wikipedia's extraordinary linguistic diversity means that an editor who has questions about a source written in a language they don't understand can probably find someone who can help them here or on another Wikipedia. This guideline should go. CDC (talk) 06:21, 8 February 2006 (UTC)

I agree with Jmabel. I'd recommend following standard academic practice as it applies to the topic of the article. If there is an important primary source in a foreign language, it should be cited, along with pointers to English translations if they exist. (The idea behind the recommendation to use "English sources" also misses the point that English itself has changed considerably during its recorded history, and the reader cannot be assumed to understand even early modern English well enough either.) --MarkSweep (call me collect) 08:49, 8 February 2006 (UTC)

I am editing accordingly. - Jmabel | Talk 01:11, 13 February 2006 (UTC)

I support the new version. Good job. --MarkSweep (call me collect) 01:25, 13 February 2006 (UTC)

But I see that someone has now restored "English-language sources… should always be used in preference to foreign-language sources", which tends to undercut the rest of this. - Jmabel | Talk 05:27, 27 February 2006 (UTC)

Well, that makes no sense. Sources should be consulted primarily on the basis of their reliability and relevance. It would make little sense to write, say, an article about 17th century European philosophy without mentioning primary and secondary sources in French, German, etc. Of course, whenever English translations are available, those should be cited as well. For recent events, there may not be any high-quality English sources available. In that case it would make no sense to insist on using a low-quality English source instead of a high-quality non-English source. In other words, the proposed requirement to prefer English sources should at most hold ceteris paribus, when all sources can be considered to be of comparable quality. --MarkSweep (call me collect) 07:05, 27 February 2006 (UTC)

Blogs == personal blogs only?

I notice in the section of "Personal sites as secondary sources", the sentence is "Personal websites and blogs should not be used as secondary sources".

Does that mean personal blogs, or does it mean all blogs? The context, to me, implies the former.

Is there a valid distinction between personal blogs (e.g. "hey, here's my blog") or a more focused topical blog, ranging from Slashdot to BoingBoing to Google Sightseeing and even to Daniel Drezner's personal but extremely economics-focused and heavily-traveled blog [2] ?

- Keith D. Tyler ¶ 20:55, 17 February 2006 (UTC)

Citing a FAQ - FAQ's not credible sources unless written by credible authors

Someone wants to cite a FAQ that was written by people unqualified to comment on the subject --just written by various Joe Blows on the internet. I strongly believe it's not a credible source. What do I do? Is it arbitration? What's the procedure to contest the credibility of a source? (The policy should say something about FAQ's, actually). RJII 21:43, 18 February 2006 (UTC)

Citing a FAQ - Credibility of a source is based on the source

Someone wants to prevent me citing a well-written and researched FAQ "An Anarchist FAQ" that was written by many people involved in a huge collaborative effort (just like wikipedia) over 10 years - the website where it's hosted is the largest anarchist website on the internet and the FAQ is updated constantly. I strongly believe it's a reliable source, and contains very useful information which is good for wikipedia. What do I do? Is it arbitration? He's already on probation, so I doubt it'll make much difference. What's the procedure to contest the credibility of a source? (The policy should say something about FAQ's, actually). Infinity0 ^talk 21:52, 18 February 2006 (UTC)

It's a partisan source: Partisan political and religious sources should be treated with caution, although political bias is not in itself a reason not to use a source. Widely acknowledged extremist political or religious websites — for example, those belonging to Stormfront, Hamas, or the Socialist Workers Party — should never be used as sources for Wikipedia, except as primary sources i.e. in articles discussing the opinions of that organization or the opinions of a larger like-minded group, but even then should be used with great caution, and should not be relied upon as a sole source.

What's being sourced, and in what article? Steve block talk 22:01, 18 February 2006 (UTC)

Anarchism#Anarcho-capitalism: Opponents argue that such relationships are not fully consensual, but coercive in nature (for example wage slavery) [3]

The source is being used as a primary source to give evidence for the above claim, that "opponents of a-capitalism think... etc", since Infoshop.org are "opponents of a-capitalism". Infinity0 ^talk 22:04, 18 February 2006 (UTC)

Somewhat OT, but pertaining to the passage from the project page quoted here: is there any reason why "Stormfront" & "the Socialist Workers Party" are mentioned in all 3 examples of an unreliable source? This repetition suggests the outlines of a now-forgotten flamefest, & rewriting these sections would improve the article. (BTW, I'm not a fan of either, so using them in one example is not out of line.) -- llywrch 17:07, 22 February 2006 (UTC)

Citing a fact-checked blog

We have an interesting situation in Hurricane Katrina. Robert Lindsay has independently been keeping the tally of deaths caused by the storm, and now has a more recent total than the now outdated Tropical Storm Report. His blog is properly sourced with independent primary sources, and those sources are publicly available, and I've fact-checked them without coming to any negative results. Since he has been following the story for quite a while, I think that his blog does check out as a reliable source for the death toll, as long as it continues being properly referenced. What is everyone else's view on this? Titoxd^{(?!? - help us)} 20:44, 27 February 2006 (UTC)

Support use. Perhaps in the cite footnote inside the article you are working on mention that the next two sources on the list are primary sources the blog used and that spot checks validate his accurate citation, then add the next two on the list as pass through primaries? (where 2 is some arbitrary small number). If nervous, caveat that the user is advised to revalidate the blog's use of primaries for themselves? ++Lar: t/c 22:29, 27 February 2006 (UTC)

Is Wikipedia a reliable source?

I don't think that Wikipedia meets its own guidelines for reliability of sources; yet editors are encouraged to wikify their articles. The topic deserves mention in the main page as a notable exception to the guideline. -ikkyu2 (talk) 05:03, 6 March 2006 (UTC)

I don't see how it could be a reliable source, since the editors have no apparent qualifications to editorialize. Whereas normally, papers are written by named persons with academic qualifications, there is no reason to trust the editorial judgement of Wikipedia writers. I would think that citing the editorial content from Wikipedia in a scholarly article would be laughable. RJII 19:53, 6 March 2006 (UTC)

Is there any difference between wikification of a word in a Wikipedia article and citation of an off-site link, for the purposes of sourcing? -ikkyu2 (talk) 04:54, 7 March 2006 (UTC)

Blogs

It says "Personal websites and blogs may never be used as secondary sources." but I want to use a blog as a source. The situation is that the blog is used by an individual to publish his criticism of politics. In that case, I think it is perfectly acceptable to cite the blog as a source for the views of the owner. Would anyone object to my updating this article to cover this?Eiler7 19:34, 7 March 2006 (UTC)

I think it already covers that. Blogs may be used as primary sources for the opinions of their writers. -Will Beback 23:05, 7 March 2006 (UTC)

The wording could be improved. It says "A personal website or blog may be used only as a primary source, i.e., when we are writing about the subject or owner of the website." but it would be better to say that "when we are writing about the opinions of the owner". If that change were made, then the next sentence should be removed. Eiler7 10:10, 8 March 2006 (UTC)

Self reference

We've got two places where it explicitly states that things can be used as sources about themselves:

• At Evaluating_sources it says "Groups like these may be used as primary sources only i.e. as sources about themselves..."
• At Personal_websites_as_primary_sources it says "A personal website or blog may be used only as a primary source, i.e., when we are writing about the subject or owner of the website."

In both of these instances it seems to me that the last person we should be taking information from, and also appears to explicitly contradict the verbage higher up that says "Wikipedia articles may rely on primary sources so long as what they say has been published by a credible publication." Taking bias into account even in the cases of reputable entities, I'd suggest that we should never use anything as a source about itself. If there do not exist third party sources, we should be evaluating if the item belongs in Wikipedia at all.
brenneman^{T}{L} 03:33, 8 March 2006 (UTC)

I think you want to use somebody as a source of what *they* say, with attribution. If the NY Times prints something negative about somebody, and doesn't say their response; it seems fair to quote the person's web site (putting it in quotes, and attributing it properly). Also, we have to consider trivial items. If NY Times says somebody is 30 years old do you want to write "(born circa 1976)" or do you want to use the exact birthdate given on their web site that matches the age? Also, occasionaly (very occasionally), the subject of the story is a reliable source. For some prominent awards, the award granter is the source information of who they gave the award to. They may be a better source than some newspapers. Danielle House was widely reported as being Miss Canada. Except she wasn't. The official site of the pageant organization sets the record straight, even though it's not a neutral party. The pageant is even a legitmate source to use in the Miss Canada International article itself. In this example, the topic was written about very heavily (warranting inclusion), but on a point a fact, very inaccurately. But, in general, of course, one shouldn't use some as a source on themselve. On a more serious example, the government is frequently a source on itself. If we did what you suggest literally, we couldn't use a lot of census data. --Rob 11:41, 8 March 2006 (UTC)

This is quite inconsistent with the rest of our policy. We have no problems using novels as sources for their own plot summaries, or television shows as sources for their own credits. Put another way, if this rule were applied to other types of sources, then, combined with NPOV, if someone were to publish a book of untrue information about someone, their own refutation of that information would not be considered reliable, and we would have to report the untrue information as published. Lemme get back to editing the Siegenthaler article...

Unless the intention is a blanket ban or deprication of online sources - which seems rather strange for an online encyclopedia with a particular strength in newer topics that aren't covered by more traditional encyclopedias - not only should we accept that any source is reliable when talking about itself, but frankly some of the other bans on Internet sources are absurd. The Bogdanov Affair, for example, would need to be heavily revised in a way that makes a complete fraud look accurate and important if we throw out USENet sources.

Our judgments of reliability cannot be used as blind hammers. Judgments of reliability and accuracy take skill, nuance, and attention - they are the absolute worst things to condense into 30 KB guidelines. Entire college courses are taught on the subject of identifying reliable sources. Discussions of reliability routinely echo through the upper echelons of the academy. Of all the places where unsubtle letter-of-the-law based process is destructive to Wikipedia, this is the biggest. Phil Sandifer 14:02, 8 March 2006 (UTC)

Reply to Aaron Brenneman: I agree with your view, and I think it entirely respects the spirit of the policy. Consider the following excerpt from the original verifiability policy before this edit of SlimVirgin.

Self-published sources and other sources of dubious reliability may be used as primary-source material in articles about that source. For example, the Stormfront website may be used as a source about itself in an article about Stormfront, so long the information is appropriate, not unduly self-aggrandizing, and not contradicted by third-party sources.

The key ingredient here is the concept of "primary source", a jargon to say "raw material". The main point is that the content (includiong any opinion, analysis, interpretation, evaluative claim, etc.) that is found in a primary source is only used as raw material. When you use a document as a primary source, you do not have to and usually should not restrict yourself to report the interpretation, analysis, viewpoints, etc. that are expressed in that source. For example, in accordance with WP policy, an official trial transcript can only be used as a primary source. Normally, this transcript contains the views of witnesses and arguments of lawyers. Some secondary source could refer to this transcript to attack the witnesses and the trial itself. This is the normal way to use a primary source: you analyse it, interpret it, etc. Usually, you must collect many primary sources before you do that. On the contrary, if you were to use the transcript as a secondary source, you would have to directly report the viewpoints of the witnesses and the arguments of the lawyers. The policy says that we cannot report the content of a personal website in this way. Instead, you must have a critical attitude. Étincelle (formerly Lumiere)

You may think that this does not explicitly say that we must use separate reputable secondary publications to source our critical view of the website. However, elsewhere in the policy it is clear that we cannot report primary sources without secondary sources that provide the interpretation, evaluative, etc. aspect. The first paragraph that explains this is the following.

In some cases, where an article (1) makes descriptive claims the accuracy of which is easily verifiable by any reasonable adult without specialist knowledge, and (2) makes no analytic, synthetic, interpretive, or evaluative claims, a Wikipedia article may be based entirely on primary sources (examples would include apple pie or current events), but these are exceptions.

Clearly, this says that we need secondary sources to support any interpretation, etc. of primary sources. The exceptions are articles such as Apple pie, in which the interpretation, evaluative, etc. aspects are considered to be negligible. Étincelle (formerly Lumiere)

The fact that a WP article must report on secondary sources was reinforced in the next sentence (before SlimVirgin made this edit):

Wikipedia articles include material on the basis of verifiability, not truth. That is, we report what other reliable secondary sources have published, whether or not we regard the material as accurate.

The emphasis is mine. Étincelle (formerly Lumiere)

The fact that a WP article must report on secondary sources is further supported in the following paragraph of the current guideline:

A secondary source summarizes one or more primary or secondary sources. A tertiary source usually summarizes secondary sources. Wikipedia is a tertiary source.

BTW, it is clear that a tertiary source can also provide links to primary sources. For example, it makes no sense to request that a review article, say on molecular surface, cannot give a reference to some original pictures or any other form of primary sources that are important ingredients in the secondary sources that are reviewed. Certainly, primary-source material remain interesting and can be cited in a tertiary source. This is not at all in contradiction with the fact that a tertiary source reports on secondary sources. The logic is simple and easily understood. Étincelle (formerly Lumiere)

In my opinion, SlimVirgin is either confused about primary/secondary sources or simply he changed his position recently because he made this edit to put the emphasis on the fact that a WP article must report on secondary sources (in a way that is consistent with common sense and the above) and then later it is again SlimVirgin that removed the term "secondary" from the sentence in this edit. It does not matter. My argument does not rely on one specific sentence, and certainly not on the authority of one editor. It relies on the overall policy and common sense. -Étincelle (formerly Lumiere) 19:29, 8 March 2006 (UTC)

Discussion of some examples

Just briefly, with regards to a few of the examples raised above of "valid" self-reference:

• NY Times says somebody is 30 years old - Here we'd have to balance the possible gain from using the person's purported actual birthdate. The cases where a person's exact age would be of encyclopedic merit would be ones such as eligibility for a sporting event or accusations of statutory assault. In these cases, we're better of taking the Time's version. The question to be asked here is why would the times not have an exact date, and why would the source's version be any more accurate? It's not like they remember being born.
• Miss Canada - This is probably a non-issue, as the correction would be reported by credible sources once promulgated.
• Census data/Government - This is only "tail swallowing" if we view "the government" as a monolithic entitity. I can say with some authority that the Australian Bureau of Statistics wouldn't release survey results on itself. We could accept an oversight committee's publications on the Iran-Contra Affair, but not a National Security Council memo on thier own spending.
• Unpublished refutations - Someone writes a column in the NY Times saying I eat babies. I publish on my web site that I only eat muslim babies. If no other source repeats my refutations, it's entirely sound for us to report only that "brenneman eats babies". It's not our job to "redress wrongs" in this manner, and the editorial decision to reproduce this material is not ours to make. This would in fact be a violation of WP:AUTO second-hand. The far-fetched Siegenthaler example ignores the fact that the refutations were widely published in reputable sources.
• Novels and television shows - Here, unless the item is of such fringe nature that we must question its inclusion, there should be a plethora of sources. If we're taking plot summaries or casting credits from the novel or film itself, its due to laziness not because we are compelled to do so.
• The Bogdanov Affair - I'm not familair with, but the blurb above makes me deeply concerned if the only opposing voice comes from USENet sources. Homework for me.

Barring the last, none of the examples adress the question of information that is only available from the source. There is a dynamic tension between the desire to have wide coverage and the essential need for accuracy, and if we must choose between excluding an item and reporting dubious material, there really should be no choice.

brenneman^{T}{L} 23:50, 8 March 2006 (UTC)

I don't think it should be seen as "can a website be a primary source about itself?" Since, by definition it pretty much is. I think the pertinant question is, "can a website be a reliable source?" Note I've removed the "about itself" because, when it comes down to it, whether a website is being used as a source on itself or something else, we must evaluate the reliability of it's information.

How about some qualifying language to that effect then? Something like, "Before relying on the information provided by a website, consider the purpose of that site. While the purpose of a website might not affect the accuracy of its content, it might indicate that the material has been altered or manipulated to change or influence its meaning. Some sites try to persuade the reader to a particular point of view, distorting the contents in obvious or subtle ways. In general, look for websites with a non-biased, balanced approach to presenting information." This is adapted from an article here, written by a subcomittee of the American Library Association on working with websites presenting Primary Source material in a second-hand fashion (on-line libraries, and such).

I don't think a policy prohibiting websites as the sole primary source in any article is the way to go, since, as others have stated, each instance should be carefully evaluated on its own merits. InkSplotch^(talk) 00:22, 9 March 2006 (UTC)

Dust Breeding and other audio Doctor Who stories have to be scrapped then - no published books on them that I know of. All either primary sources or websites for those plot summaries. And you're really suggesting that for movie credits we should go to Maltin or IMDB instead of consulting the film? That's absurd. Phil Sandifer 00:35, 9 March 2006 (UTC)

• InkSplotch - I'm actually referring to a more general case of self reference, but the most common occurance of it would probably be websites. And there is some confusion still about what "primary source" means. If I surf to a baby-eating webpage and read an account of some carnovire eating babies and then I type a summary of that account into a wikipedia article, the website is not the primary source there: I am the primary source in that example. This is explicitly stated on the page in the form of report from [an] eye witness If, however, babyeaters.org has a section that says "we've got 10,000 members worlwide" then they are the primary source, but that's not what the vast amount of "direct reporting" on wikipedia is. Finally, I'm not suggesting that we prohibit websites from being the sole primary source, but that we include strong language regarding anything that can only be sourced from itself.
• Snowspinner - It's actually not absurd at all. Your quote that "Entire college courses are taught on the subject of identifying reliable sources." is actually the heart of the issue: Wikipedia editors will not have taken those courses, and are thus not equipped to evaluate primary sources. If you'd like a glimpse into the horror that allowing "direct observation" of this nature yields, have a look at Final Destination 3#Deaths. And yes, if information on the Doctor Who audio story you've mentioned cannot be found anywhere except for itself, not even in a science fiction magazine or an industry guide, than why are we including it in Wikipedia?

brenneman^{T}{L} 01:31, 9 March 2006 (UTC)

So basically, it requires specialized knowledge, and so we're not allowed to do it? That takes anti-expertise to a whole new place. Phil Sandifer 01:34, 9 March 2006 (UTC)

I fail to understand how your reply relates to what I just said. I agreed with your earlier comment that evaluation of primary sources is specialised knowledge. How is this anti-expertice? - brenneman^{T}{L} 01:37, 9 March 2006 (UTC)

Because it requires specialized knowledge, we can't trust Wikipedia editors to do it, therefore we must not engage in it. Mind you, my point wasn't that evaluation of primary sources is difficult - it's that evaluation of ANY source is a complex, nuanced, and not easily reduced to simple rules. You are arguing that because of this, we can't do it. Phil Sandifer 01:39, 9 March 2006 (UTC)

I can't read all of the above, so I'm sorry if this has been said already, but perhaps some of the confusion over primary sources is that what counts as a primary source is relative to the position of the observer. If I drive my car into a wall, I'm a primary source for the police regarding what happened, but the police and I are primary sources for a reporter. The reporter writes a story based on what we say and that story becomes a secondary source. But in 500 years, when that day's paper is dug up from underneath the rubble of the newspaper's offices, it will be a primary source for historians, as an artifact of its time.

As for reliability and personal websites, these may be used as sources of information about the author or the group that maintains them (with caution, and so long as reliable third-party sources don't contradict them), but may not be used as third-party sources i.e. as sources of information about anyone else. This is because anyone can add anything to their own website. Exceptions to this rule-of-thumb would be when a well-known person sets up a personal website, and there is no doubt as to his or her identity. So if Bill Clinton wrote up on his blog the real story behind the Lewinsky affair, we wouldn't ignore it.

The key question to ask when evaluating a source is: "how much fact-checking or editorial oversight is this likely to have been subjected to?" The answer with most good newspapers and publishing houses is that there's a lot of fact-checking by the reporter, copy editor, page editor, managing editor, editor-in-chief, possibly the lawyers, possibly the publisher. Whether that process always works is another matter, but at least it's in place, and we have to trust that it works most of the time. With the Stormfront website, on the other hand, there is almost certainly no process in place whereby what they publish is checked for accuracy or legal problems, and that's going to be the case with most self-published sources (whether political websites run by small groups, or blogs, or vanity-press books. That's why we allow those sources to be used as sources of information about themselves only. SlimVirgin ^(talk) 02:27, 9 March 2006 (UTC)

Aaron, you might be right about the confusion. If you surf to a website claiming first hand accounts of baby eating, and draw upon it to write an article here, that website is a primary source. You and Wikipedia are secondary. This is the heart of "No Original Research"...we must remain at least a secondary source. Again, I'd like to stress, "Primary Source" and similar terms are simple definitions. The issue is reliability, and I think any language which addresses information that "can only be sourced from itself" is missing the point. Reading the label on the back of a CD, or in your example, reading babyeaters.org's mission statement is not original research. Extrapolating from that into the habits of baby eaters is. InkSplotch^(talk) 02:56, 9 March 2006 (UTC)

• Yeah, we're getting down to some very fine points but I think that we agree. If we're writing an article about baby-eating, then the account on the website is indeed a primary source, but if we're writing an article about the www.baby-eaters.org it's not. But again, I agree with you that that's not actually the point.
  brenneman^{T}{L} 04:14, 9 March 2006 (UTC)

(carriage return, edit conflict) Basically, it all comes down to how that source is being used, and the available alternatives. A source that is being used for self-publicity purposes is in my view, not acceptable, while a source that is being used for descriptive (e.g. getting the cast of characters from a movie) is perfectly fine. Also, if we can spare not having to use the primary source, but rather a secondary source, everything is all right. Titoxd^{(?!? - help us)} 02:31, 9 March 2006 (UTC)

For the record, in response to Snowspinner's example, yes, I absolutely think that for the purposes of an encyclopedia that does not allow original research, IMDB is a better source for the credits of a movie than "the movie itself", since relying on "the movie itself" really means relying on an editor saying "I saw that movie, and it totally fucking said that Jenna Jameson played Uhura in Star Trek VI." Nandesuka 02:39, 9 March 2006 (UTC)

Caught this on an edit conflict. IMDB, unfortunatly, is also subscriber-edited. It's information is no more reliable that an editor saying, "I saw that movie, and..." What you're calling into question here is the reliability of the editor. If you can rent the DVD yourself, look at the credits, and see Jenna's name there, it's verified. Now, if IMDB and several other sites report the credits as a typo, then you have reliable sourcing for an entirely different kind of entry. InkSplotch^(talk) 02:56, 9 March 2006 (UTC)

For credited roles, IMDB would be less reliable than the film itself, if it's a reputable film-maker (e.g. not a minor film maker, claiming to have a star in their film). Since the maker of the film will provide the credits of the film, in printed form, in publicly available places, its verifiable, and doesn't have the problems of original research implied above. For *uncredited* appearances, imdb (which does list those) would be a better source than watching the film (e.g. somebody watches a film, and says a big star appeared before they were famous, but as uncredited extra). There's a risk of taking a good rule, stretching it to its literal extreme, and defeating the purpose of the rule. --Rob 03:23, 9 March 2006 (UTC)

I'd like to point out that I didn't suggest that we use IMDB as a source, simply that "looking at the CD case" shouldn't count as a source. That's an eyewitness acount, and we don't use those. One point that we're also missing by focusing on these sorts of examples is that the vast majority of itmes like who starred in what film would be widely available elsewhere. A more appropiate example would be something like List of YTMND fads. The page doesn't have any sources, and the only likely sources are from direct observation of these "fads" in action. We've established pretty firmly that just because User:Schmoo thinks leader Ecks is a dictator isn't enough to include Ecks in List of dictators, so why do we loosen that requirement elsewhere? - brenneman^{T}{L} 04:14, 9 March 2006 (UTC)

• Woah. Looking at the cd case isn't making me primary source. Being in the studio is making me primary source. The CD is primary source in this instance, I'm secondary source. I'm not an eyewitness, and the information is easily verifiable and also citable to a reliable source, unless you are suggesting that a CD case is not a reliable source for documenting its contents. That's an argument I'd like to hear. I think Jimbo's comments on this sort of thing way back when are instructive, information which is verifiable and which can be easily presented in an NPOV fashion...is valid material for an encyclopedia and also qualifying verification thus: When I say 'verifiable' I don't mean 'in some abstract fantasy theory' I mean actually practically verifiable by Wikipedians. I would say that citing a CD case isn't failing any of these. Of course, if one article was built on the back of a CD citation, it'd probably fail an afd, but that's what afd is for, not this page. Steve block talk 17:39, 9 March 2006 (UTC)

I agree with SlimVirgin that the difficulty maybe in the understanding the concept of primary/secondary sources. One key ingredient, which is very close to what SlimVirgin explained, is that most documents or publications, if we ignore WP policy, can be used both as a primary source or as a secondary sources. Almost nothing is a primary source in itself. If we say that it is a primary source, we mean that it is normally used as a primary source. For example, when we say that a trial transcript is a primary source, we mean that the normal use of a trial transcript is as a primary source. However, trial transcripts can in some cases be used as secondary sources. (I am not talking about the WP policy here, but in general.) Similarly, when we say that a scientific publication is a secondary source, we only mean that a scientific publication is normally used as a secondary source. However, a scientific publication can sometimes be used as a primary source. Something is used as a primary source, if we add our interpretation, analysis, evaluation, etc. to it. Something is used as a secondary source, if we directly report the view (interpretation, analysis, evaluation, etc.) that it contains, and do not add or remove anything that would significantly change this view. For example, if one writes an analysis of a car accident, an eye witness report of this accident is normally used as a primary source, which means that there is no intrinsic interpretation or evaluation, etc. attached to this report. Such an interpretation or evaluation, etc. depends on the secondary source that reports on it. For example, the author of this secondary source might decide that he will not trust the eye witness ability to evaluate distances and completely ignore this aspect of the report. However, we can imagine a situation where an eye witness report is used as a secondary source. This would be the case if the eye witness added his own viewpoint (evaluation, interpretation, etc.) in the report and that we decided to directly report this viewpoint in an article without presenting an external view on it. -Étincelle (formerly Lumiere) 04:32, 9 March 2006 (UTC)

Therefore, it is almost by definition that, if we use something as a primary source, we need a secondary source to support our interpretation of this primary source. Otherwise, we do some form of original research. The exception is when there is only one natural interpretation and there is no controversy, such as is supposedly the situation in the apple pie article. In this case, we can present the primary sources (pictures of pies, the different recipes, etc.) with their natural interpretation without providing a secondary source. (However, I am not sure that the apple pie article would remain a good example if my mother was an editor...) -Étincelle (formerly Lumiere) 04:32, 9 March 2006 (UTC)

• Late weigh in: Let's make sure we understand our terms, eh? A primary source of information is information that is unmediated: it is from the participant. A secondary source is a mediated source: it is anyone or anything that has consulted the primary sources and compared them, analyzed them, considered them, and constructed an account. A tertiary source is one that is an analysis, compilation, and narrative of secondary sources. When we say that Wikipedia is a tertiary source of information, we mean that it is not only autonarrative but also that it is not a simple review. Phil's mention of novels as sources about themselves is both non-germane and germane. An article about a novel that considers only the novel itself is a secondary: it is reviewing and weighing the primary. That is a thing we are not supposed to do. After all, what, qualitatively, is the line between one person saying, "The Beggar's Opera was produced in 1728, and it's main characters are Macheath, a highwayman, Polly Peachum, Lucky Lockitt, Peachum, a fence, and Lockitt, a jailer" and someone saying, "In the third act of The Beggar's Opera, the inversion of class becomes potentially dangerous as not only gender, but also the basis of the marriage contract are called into question by Macheath's failure to announce who it is that he "loves," and love itself becomes yet another counterfeit good in the mercantile exchange of theft and fencing?" We call one of them an article and the other "original research," but why, if we get to use the play as a source itself? In fact, what is at stake is interpretation and analysis. Can you get facts from a website about the website? No. No one writes on The Beggar's Opera after just reading the play. He or she reads the introductions, a few headnotes in anthologies, an article or two, goes to a class or three where the play is discussed, and therefore produces a tertiary account (a synthesis of multiple secondary sources). If all you do is read the poem and write the article, you'll do a crap job of it, at best, and you'll defy WP:NOR in the process. The same should be the rule for any website. If there is no secondary research then we can't talk about it. There should be whole worlds of things that we can't have an article about. It doesn't mean they're trivial or bad or not as great as Ulysses or Iliad. It means that we're following our rules and being an encyclopedia rather than a LiveJournal blog or a fan site. Geogre 04:34, 9 March 2006 (UTC)

Unless NOR was rewritten by someone since I last looked at it, there was an important modifier regarding the use of primary sources, whereby one was forbidden to make "novel" interpretations based on primary sources. Phil Sandifer 04:42, 9 March 2006 (UTC)

I can assure you, the criticism I wrote above is pretty much trite. If I heard it from a student, I'd have to stifle a yawn. It would, however, be without references, and therefore it would be rejected, because it is coming solely from reading the play. Further, you imply improperly that original research is groovy, so long as your observation isn't new? That's absurd and untrue. Geogre 06:32, 9 March 2006 (UTC)

No, it's not absurd and untrue. From your example above, the synopsis of the characters, their basic role and publication date are verifiable by someone else who picks up the play. The analysis and interpretation you also provide as an example is not verifiable, it's analysis and interpration. The first is allowed as examination of a primary source, and the second is not, as it's NOR. It may be absolutely agreed analysis and interpretation but it's still NOR. If it's a notable work, then common interpretative themes should be easily referenced to secondary sources. SchmuckyTheCat 07:45, 9 March 2006 (UTC)

Unless I misunderstood Geogre, I think I agree I disagree with him. The only thing One thing is that Geogre seems to assume that something is always either a primary source or a secondary source. This is not consistent with the usual notion of primary/secondary sources. SlimVirgim gave a good example. The analysis of the accident by a reporter, say of the NY times, is normally used as a secondary source. However, if someone decided to study the NY times and to collect different articles published in that journal, the analysis of the car accident would become a primary source in that study. The difference is that, as a primary source, the interpretation of a document depends on the secondary source that reports on it. Whereas, if a document is used as a secondary source, we must directly report on it and respect any interpretation that it may contain. More about my disagreement in the section #More on primary/secondary sources Vs No original research -Étincelle (formerly Lumiere) 05:13, 9 March 2006 (UTC)

But that's what NOR says. Your interpretation of Beggar's Opera seems to me a border case. But on the other hand if someone doesn't cite a source when claiming that The Pearl is about greed, I'm going to be really hard pressed to pull it out. It's also worth noting that in the humanities, obvious results tend not to be published. If everyone who looks at a novel identifies common themes, nobody publishes "the theme of this novel is X." Unless there's a Cliff Notes of the novel, it's really, really tough to get secondary sources for basic claims about philosophy, literature, and art. Which is why we have the "novel interpretation" clause in NOR. Because otherwise we can't have articles on those subjects. Phil Sandifer 20:24, 9 March 2006 (UTC)

We're about to get into semantics and philosophy. Yes, to a degree, things are secondary to something. However, the distinction I wish to draw is those things that are secondary to the event and those that are secondary to accounts of it. What is important is the fact that we are talking about being secondary to a narrative of the primary experience, rather than whether we are in relation to a previous iteration. The website itself is the lived life or the driver of the car. When that is written about, the account is secondary. Any analysis of other things written about is tertiary to the act or event (the verb of literature, the totality of the accident as event, the report of the accident). This distinction is important, as, when we are in relation to narrative rather than event, we are capable of re-examining and because the narrative is a selection and therefore an analysis and judgment. Geogre 06:27, 9 March 2006 (UTC)

More on primary/secondary sources Vs No original research

I realize now that I did not understood what Geogre was saying. This is where I disagree:

A tertiary source is one that is an analysis, compilation, and narrative of secondary sources. When we say that Wikipedia is a tertiary source of information, we mean that it is not only autonarrative but also that it is not a simple review. -Geogre

On the contrary, if you do your own analysis of the secondary sources (an analysis that goes beyond understanding and organizing them) instead of reporting (and organizing) the analysis that is already provided in the secondary sources, then these secondary sources are not really used as secondary sources. By definition, if you use something as a secondary source, you must respect the interpretation or analysis that is provided in that source. You cannot take a viewpoint on it as if it was raw material to analyse and interpret. Indeed, you must restrict yourself to a summary or review and an organization of these secondary sources. Otherwise, if you take a novel viewpoint on the content of these secondary sources (including their analytical or interpretative content), you are actually using these secondary sources as if they were primary sources. Also, a tertiary source cannot ignore the interpretation or analysis that is provided in the secondary sources and build its own novel interpretation or analysis of the primary source material that is contained in these secondary sources. A tertiary source such as Wikipedia can only report on secondary sources. Of course, it can cite the primary sources that are used by these secondary sources, as long as it does not do some novel work that analyses or interprets these primary sources, which is what a secondary source does. -Étincelle (formerly Lumiere) 16:41, 9 March 2006 (UTC)

I would also like to comment on the following:

However, the distinction I wish to draw is those things that are secondary to the event and those that are secondary to accounts of it. What is important is the fact that we are talking about being secondary to a narrative of the primary experience, rather than whether we are in relation to a previous iteration. [...] Any analysis of other things written about is tertiary to the act or event (the verb of literature, the totality of the accident as event, the report of the accident). -Geogre

My understanding here is that Geogre considers that the value of a tertiary source is that it uses material that already contain some evaluation and judgement. He must have in mind a student that would directlty write about a movie without consulting secondary source materials, which often result in a less thoughtful analysis of the movie. The analysis is likely to be more complete, etc. if the students use secondary sources. This is a very valuable viewpoint, but it is not at all the purpose of the requirement that we only report on secondary sources. On the contrary, the idea is that we cannot add any novel interpretation of primary source material in Wikipedia. If a student writes a tertiary source about the movie, it would have to restrict himself to only organize and report secondary sources. This might not be what the professor expects, but it is what Wikipedia policy requests. However, to add some flexibility, the WP policy allows for exceptions, and an article about a non controversial movie might be one of these exceptions. -Étincelle (formerly Lumiere) 16:41, 9 March 2006 (UTC)

Globalized/localized interpretations

Might I suggest a more atomistic approach? Some websites can quite adequately describe the organization they talk about, most particularly if the organization itself would be considered a reliable source. An engineering institution (my background) would be a good example. They are definitively a reliable source (they publish academic journals, professionally accredit people etc) but their inner workings are not reported anywhere since they are uninteresting to the media at large, and even to the technical media. Some of the detail is useful to Wikipedia though. I think it would be ok to cite the webpage they offer that says "here is how we are structured" and be quite sure that the information was verified, and reliable.

If the people/organization running the website is/are not a reliable source, then it seems only natural that the website they produce on themselves is not reliable either. So I'm inclined to think this is a case-by-case issue, rather than one of over-arching definitive philosophy. Judge the reliability of the webpage on the reliability of the source publishing it. And temper this with a need to retain a neutral pov and no advertising. Thus a blog site that says "we hvae trhee thousdan membres" and they started last week would not be a reliable source anyway, and their claim to 3000 members is not reliable either. -Splash^talk 18:24, 9 March 2006 (UTC)

Which part of the policies or guidelines says that we should never use websites of organizations as reliable secondary sources? As far as I know, the general rule is only for personal websites of individuals. The policy says that dubious sources should also be used as primary sources only, but to my knowledge it does not say that all websites must be used as primary sources. IMO, it should be mentioned that partisan organizations that are reprentative of a given viewpoint can be used as secondary sources about this view point, but not about another viewpoint. For example, a very religious organization that is not reputable as being representative of the catholic viewpoint cannot be used as a secondary source to present the catholic viewpoint. It can be used as a secondary source for its own viewpoint as long as this viewpoint is notable. This is in accord with the Neutral Point of View policy which says that we should attribute views to there prominent adherents. We do not exclude significant views, we attribute them to their adherents. -Étincelle (formerly Lumiere) 18:50, 9 March 2006 (UTC)

Uh, I was answering Brenneman's orignal questions. I didn't invoke any guidelines, I said what I personally would regard as a good way to do things. I agree that it is reasonable to say "Source X says Y", but not to rely on an unreliable Source X to verify itself. I am getting at relying on RandomForum's home pages attesting to their massive notability and cultural influence when plainly they are an unreliable source in the first place. Which I think is what Aaron had in mind. It's not really a Point of View that "my forum has 3000 members", it's a claim of actual fact that we should not rely on an inherently unreliable source to provide about itself. (Disclaimer: Not all forums/blogs are inherently unreliable. I'm just using an oft-occuring example.) -Splash^talk 21:25, 9 March 2006 (UTC)

Perhaps that you vaguely had in mind that when editors evaluate a source (in the discussions in the talk page), they should try to have third party opinions. As important as this issue may be, it is a completely different issue. The purpose of the main page is different. In the main page, assuming that the given forum is a notable topic, it is perfectly fine to have an article about this topic and write in it "The <name of forum> website reports that it has 3000 members.", even if we cannot find a third-party source to verify this information. We do not exclude a viewpoint or some information because we are not sure if it is true. Instead, we attribute it to the correct ideological groups, etc. BTW, you are right that we should distinguish viewpoints and controversial facts. However, often the policy do not distinguish these two cases. In both cases, the solution is to attribute it to the correct group.

Don't misinterpret me. I am not saying that we never exclude some material. For example, it is clear that if some material is presented as a scientific contribution in the article, then the source should be a reputable scientific publication. In my viewpoint, it is not the information per se that we try to avoid or exclude. What we want to avoid is the fundamental mismatch between the information (including the way it is presented in the article) and the group to which it is attributed. Again, what is important is that the attribution is not misleading. Also, in accordance with NPOV, the viewpoint of a tiny minority has no place in Wikipedia except in a separate article that is specifically about this viewpoint. So, we can exclude some material from articles. Finally, none of this is against your idea that the criteria should be adapted to the topic. On the contrary, the policy is clear that what is a reputable source should depend on the material that is sourced, which is in accord with your idea. -Étincelle (formerly Lumiere) 22:30, 9 March 2006 (UTC)

Books of an apparent limited circulation

I am engaged in a discussion with User:Waya sahoni on Talk:Joe_Byrd_(Cherokee_Chief)#Disputed about the source of the rather pejorative statements made in the article itself. Inline in the article are links to an online newspaper (http://www.yvwiiusdinvnohii.net/Cherokee/News/FebMarch97/MP970329marshals.htm as one example) which, by their context, seem to be offered as corroboration for the statements made. The links which I have reviewed do not seem to support the phrasing used in the article.

At any rate, User:Waya sahoni is now stating that his source is in fact a book: "Martial Law in America, The Near Death Experience of the Modern Cherokee Nation. Wm. R. Wayland 2002-2003" (see: Joe_Byrd_(Cherokee_Chief)#References), for which he will not post an ISBN number, and for which Google returns no hits on either its title [4] or on its author's name [5]

He wants my mailing address to mail me a copy of the book. I have declined, not wanting to give him my address, and feeling that even if I were to possess one copy of the book that would not make me a reliable source on the issues expressed, and certainly not validate the book itself.

Question (finally): could a book of unverified provenance and limited circulation be accepted as WP:RS? If so, what would be the criteria to determine its value? -- talks_to_birds 00:12, 12 March 2006 (UTC)

The following is an excerpt of the WP:NOR policy:

In order to avoid doing original research, and in order to help improve the quality of Wikipedia articles, it is essential that any primary-source material, as well as any generalization, analysis, synthesis, interpretation, or evaluation of information or data, has been published by a reputable third-party publication (that is, not self-published) that is available to readers either from a website (other than Wikipedia) or through a public library.

The emphasis is mine. -Étincelle (formerly Lumiere) 00:25, 12 March 2006 (UTC)