Wikipedia talk:No original research/Archive 33

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Archive 30 Archive 31 Archive 32 Archive 33 Archive 34 Archive 35 Archive 40

PSTS proposals

Proposals from talk page. Examples of proof of concept that No Original Research can and should be explained without using the words "primary" and "secondary". Currently it appears there is a consensus that that the primary/secondary/tertiary concept should exist in wikipedia policy space somewhere even if not at NOR, so removal of that from wikipedia policy space is not the proposal here.

Some talk and Proposal 1 from Moving the discussion forward

Are we agreed that what the policy should be talking about is material... and not entire sources. Such material might be the entire source (such as a court transcript or a historical document) or it might be part of a source (such as the data section of a scientific research paper) ... but what is OR is the misuse of the material? Blueboar 21:03, 15 September 2007 (UTC)

It's a step in the right direction. Go for it. Be bold. --Minasbeede 21:24, 15 September 2007 (UTC)
If others agree as well... can we please stop spending time arguing on and on about the use of "Primary Source" vs. "Secondary Source" and what they mean? And can we stop beating a dead horse over the point that some sources contain both primary and secondary materials. Let's agree to change the policy to talk about "material" instead. Now... Are we agreed that some types of materials are "primary" and some "secondary" (or "tertiary")? Blueboar 22:33, 15 September 2007 (UTC)
Suits me. .. dave souza, talk 22:51, 15 September 2007 (UTC)
It would be a positive step in the right direction. Vassyana 23:12, 15 September 2007 (UTC)
I think so as well. ≈ jossi ≈ (talk) 23:49, 15 September 2007 (UTC)
I'm not sure I fully agree. I do agree that sources may contain material that is either primary or secondary, or both. And I do agree that original research may occur when an editor makes conclusions based on material in a source, and those conclusions are not also found in that source or another source. However, the OR does not arise because of the material. The material itself is what it is, and if you cite it for what it is, you cannot commit original research. Where original research occurs is when an editor cites a source, or primary or secondary material within a source, and implies that the cited source supports their new, unpublished conclusions. COGDEN 17:18, 18 September 2007 (UTC)

How about just saying what we mean? Say "uninterpreted raw data" versus "interpretation of that raw data" instead of "primary" versus "secondary"? As in: Uninterpreted raw data is to be used sparingly and always in connection with properly sourced interpretation of that raw data. Any interpretation of the raw data that is not actually in any of the sources cited is called "original research" at Wikipedia and is not acceptable content for a Wikipedia article. WAS 4.250 00:43, 16 September 2007 (UTC)

I like it but suggest some changes/concerns:
  • "used carefully" instead of "used sparingly"
  • "always in connection..." poses certain problems
    • it may be appropriate to cite the data without reference to interpretations, particularly in lists or infoboxes
    • it may be that two sources provide the same basic data, one being more detailed, and another being more widely-available. I've read scholarly articles (e.g. Gilham and Marx) which comment on the opposing sides in the Battle of Seattle (N30) using less-complete sources while one of the editors has more-complete sources (e.g. police rosters from the city archives).
      • in this case, it seems perfectly appropriate to cite both sources, and then the interpretation of one source.
      • I'm not sure about citing details from one source and getting the interpretation from works using the other source. I suspect some limited use is NOR but it can easily cross into OR.
  • tightening "not actually in" to "not in" Jacob Haller 02:21, 16 September 2007 (UTC)
  • "raw data" might not be the best word for narrative sections of histories, but the same principles apply. Jacob Haller 02:23, 16 September 2007 (UTC)

Hmm. A birthdate is raw data. A quotation is raw data. All photos are raw data. So are you really sure any of this is what you mean to say? - Jmabel | Talk 02:43, 16 September 2007 (UTC)

Excellent points. How about:

  • Uninterpreted raw data is to be used carefully. Any interpretation of the raw data that is not in any of the sources cited is called "original research" at Wikipedia and is not acceptable content for a Wikipedia article. For example:
    • Numbers are easy to misuse and should be presented in a context similar to the source, such as income and population data for a city given in an article on that city or a birthdate in an article about that person; rather than in an article or section comparing populations or birthdays to make some point not made in the sources (richest city; oldest Jewish poet)
    • Quotations are easy to misuse and must not be used to make a point that in the original context was not explicitly being made.
    • Images and their caption should not be used to make a point that is not made in the text of a proper source.
    • Any use of old text or text written in a context other than for general current popular consumption can easily be misunderstood as to its significance or meaning and must be dealt with carefully by knowledgeable editors who have the prerequisites to evaluate. For example, math articles, science articles, and sources written in a language other than English require editors with specific prerequisites to properly evaluate the sources. Also words change in meaning over time and have different meanings in different contexts. "Jon is gay" means something different depending on context and era. "Suffer the children" in the King James Bible shows a famous change in the use of the word "suffer". The prerequisites must the common enough for editors to be able to check each other. Wikipedia can not allow raw data that requires requires the editor to be an expert in order to properly context that raw data, as we require all our content to be evaluatable by many (but not all) editors. WAS 4.250 03:11, 16 September 2007 (UTC)

Proposal 2

Within source material we will find a mixture of facts and interpretations. Some sources will be basic observations that offer little in the way of analysis; other sources will offer analysis of information and draw conclusions. When adapting material for use in Wikipedia, it is important to only introduce facts and observations that already exist.

A Wikipedia article or section of an article can employ statements of facts only if the material (1) comes from a reliable, verifiable source, (2) is used only to make descriptive claims and (3) never to make analytic, synthetic, interpretive, explanatory, or evaluative claims. Contributors drawing on statements of facts should be careful to comply with all three conditions.

Typically, statements of fact will tend to be found in materials such as historical documents; personal diaries, census results, video or transcripts of surveillance, public hearings, trials, or interviews; raw tabulations of surveys or questionnaires; written or recorded notes of laboratory and field experiments or observations; and artistic and fictional works (such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs).

Wikipedia articles can include analytic, synthetic, interpretive, explanatory, or evaluative claims (1) only if such claims come from a reliable, verifiable source, (2) the point of view is clearly identified and accurately represented, and (3) the analysis, syntethis, interpretation, explanation or evaluation matches that in the source. In general, it is this type of source that is of most use to Wikipedia as an encyclopaedic article is a summary analysis of a topic and using raw facts to build an article will often require too much original work that could be challenged.

Further, it is most likely that works which are a comprehensive analysis of the topic will be most suited to providing a comprehensive demonstration of a particular viewpoint. Cherry picking individual concepts from different sources may suggest editing to support a point of view, and this can be especially so in the case of using passing references.

Useful analysis is typically found in academic journals, Government Inquiries, and media sources such as serious documentary programs and a few respected newspapers.

Wikipedia strives to be a superb encyclopedia in its own right. Since Wikipedia does not want to be derivative, materials found in encyclopedias or similar reviews are often of limited value for Wikipedia research. Annual Reviews and Encyclopedia Brittanica articles often provide extensive bibliographies that are valuable tools for identifying important materials, and therefore of great use to Wikipedia editors. Nevertheless, these materials do not necessarily have the same content policies as Wikipedia and for this reason should not be viewed as authoritative. However, some encyclopedias and other material, such as Annual Reviews, have signed articles, and often articles that explicitly promote the author's own views which can be viewed as source material in their own right.

Proposal 3

Sticking to the Sources

Within Wikipedia articles we will find statements of fact and statements of interpretation, analysis or conclusion. It is important to cite appropriate sources to back those different types of statements. Statements of fact should be cited to reliable sources that contain that fact. Statements of interpretation, analysis or conclusion should be cited to reliable sources that contain the same interpretation, analysis or conclusion, often called secondary sources (see WP:PSTS). Collections of facts which tend to lead the reader to a certain interpretation should be supported not only by a reliable source for each fact, but also by a reliable source that contains the same interpretation.

Editors cannot include their own interpretations of previously-published facts, unless that interpretation is either 1) an obvious and non-controversial consequence of the facts or 2) can be attributed to a reliable source. Nor can editors expand on an author's interpretations of fact, unless that expanded interpretation is also found in a reliable source.

When there are a number of reliable sources that interpret a particular piece of material, we need to be especially careful not to insert or imply our own interpretation of the original material.

For example, that Franklin D. Roosevelt was born in Hyde Park, New York and that he became president of the United States are both verifiable facts that may be cited in a Wikipedia article. The dates of each event are easily verified, and it would be perfectly admissible to say that he "was born in Hyde Park, NY, and later became president of the United States" without finding a source that specifically uses the word "later", or otherwise explicitly compares the dates. However, any statement about the effect his birth place had on him or on his career would require separate citation, since such a statement would not an obvious or non-controversial consequence of these easily verified facts.

Proposal 4 Example without PSTS

The definitions of PSTS are not critical to NOR because they are not used outside of the PSTS section. I have rewritten that section without using the defined terms: primary, secondary, and tertiary:

Use of sources

Research that consists of collecting and organizing material from existing sources within the provisions of this and other content policies is encouraged: this is "source-based research," and it is fundamental to writing an encyclopedia. However, care should be taken not to "go beyond" the sources or use them in novel ways.

  • Factual sources - Facts that have been published by a reliable source may be used in Wikipedia, but only with care, because it is easy to misuse factual sources. For that reason, anyone—without specialist knowledge—who reads the factual source should be able to verify that the related Wikipedia passage is true to the factual source. Any interpretation of factual source material requires a reliable interpretive source. Examples of factual sources include archaeological artifacts; photographs; historical documents such as diaries, census results, video or transcripts of surveillance, public hearings, trials, or interviews; tabulated results of surveys or questionnaires; written or recorded notes of laboratory and field experiments or observations; and artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs.
  • Interpretive sources draw on factual sources to make generalizations or interpretive, analytical, or synthetic claims. A journalist's story about a traffic accident or a Security Council resolution is a interpretive source, assuming the journalist was not personally involved in either. An historian's interpretation of the decline of the Roman Empire, or analysis of the historical Jesus, is a interpretive source. Wikipedia articles should rely on reliable, published factual and interpretive sources used appropriately.
  • Some publications such as encyclopedias that sum up other sources are more reliable than others; within those sources, some articles may be more reliable than others. For example, articles signed by experts in Encyclopaedia Britannica, and encyclopedias of similar quality, can be regarded as reliable interpretive sources. Unsigned articles may be less reliable, but they may be used as factual sources so long as the encyclopedia is a high quality one.

An article or section of an article that relies on a factual source should (1) only make descriptive claims, the accuracy of which is easily verifiable by any reasonable, educated person without specialist knowledge, and (2) make no analytic, synthetic, interpretive, explanatory, or evaluative claims. Contributors drawing on factual sources should be careful to comply with both conditions.


I tried not to change too much, to remain as true as possible to the original. But once the sacred cows are banished, I think it shows this is not a good example of clarity. It needs lots more work, but we can't get that done until we get over the religious arguments over PSTS, and actually get down to the real work of editing. Dhaluza 01:30, 23 September 2007 (UTC)

What is a factual source? Are we now reinventing distinctions? ≈ jossi ≈ (talk) 01:37, 23 September 2007 (UTC)
Yes, in fact, I was trying to show an alternative to the confusing primary/secondary distinctions using plain English. I think if you ask a friend or relative who has never seen Wikipedia "what is a factual source?" and "what is a primary source?", you will probably get a more reasonable answer to the former than the latter. Dhaluza 11:21, 23 September 2007 (UTC)

Proposal 5

Use of sources

Research that consists of collecting and organizing material from existing sources within the provisions of this and other content policies is encouraged: this is "source-based research," and it is fundamental to writing an encyclopedia. However, care should be taken not to "go beyond" the sources or use them in novel ways. This policy places certain restrictions on the use of sources to preclude editors from including original research in Wikipedia articles.


Wikipedia does not bear witness to any fact. All statements of fact in articles must be supported by a previously published reliable source. Factual statements must be made with caution, because misstatement of facts is a common problem in Wikipedia articles. The facts should be presented so anyone—without specialist knowledge—who reads the cited source can verify that the related Wikipedia passage agrees with the source. Any interpretation of facts or evidence requires a reliable source for that interpretation. Wikipedia editors must be careful to not make statements of fact based on their own observations or evidence, unless they are also supported by reliable published sources.

An article or section of an article that makes a statement of fact should (1) only make descriptive claims, the accuracy of which is easily verifiable by any reasonable, educated person without specialist knowledge, and (2) make no analytic, synthetic, interpretive, explanatory, or evaluative claims. Contributors making statements of fact should be careful to comply with both conditions.


Wikipedia does not publish original thought. Using raw evidence or statements of fact to make generalizations or interpretive, analytical, or synthetic claims is original thought. Wikipedia editors must be careful not to include their original thoughts in Wikipedia articles, but must instead depend on reliable published sources to make these interpretive claims.

Some publications such as encyclopedias that sum up other sources are more reliable than others, and within those sources, some articles may be more reliable than others. For example, articles signed by experts in Encyclopaedia Britannica and encyclopedias of similar quality can be used for interpretation of facts and evidence. Unsigned articles may be less reliable, but they may be also be used for interpretation so long as the encyclopedia is a high quality one.

Archive 30 Archive 31 Archive 32 Archive 33 Archive 34 Archive 35 Archive 40


Would a translation of a primary source be a primary source or secondary source? So, for example, would an English translation of the Bible or the Qur'an be a primary source, or secondary source?Bless sins (talk) 23:30, 2 January 2008 (UTC)

A translation is in essence the same work... just in a different language. So a translation of a primary source is still a primary source. Blueboar (talk) 01:48, 3 January 2008 (UTC)
That is, in general, not true. While it is possible in some cases for a translation to be "in essence the same work", it often is not. In particular, translations from one culture, especially one from a different era, to another must translate concepts and lack of concepts from one language to another. Consider the difficulty in translating a religious text from a culture lacking the concept of a monolithic God. Further, human words always have multiple meanings, and any translation will translate from one set of meanings to a different set of meanings; not to mention an entirely different set of connotations. Poetry is especially difficult ... WAS 4.250 (talk) 11:13, 3 January 2008 (UTC)
The King James Bible is a primary source for a direct quote from it, but is a secondary source for the meaning, expressed in English, of the texts it was written from. And neither one is the best source for an issue like "Do the ten commandments forbid 'murder' or 'killing'?" The whole "primary" versus "secondary" issue is a distraction from the real issue which is the issue of proper use of sources for claims expressed in Wikipedia. Both primary and secondary sources can be misused and neither can be blindly accepted merely due to their being primary or secondary. WAS 4.250 (talk) 11:31, 3 January 2008 (UTC)
So are you saying that a translation is primary source? Secondary source? No general guideline - matter must be decided on a case by case basis?
BTW, the sources I had in mind were English translations of the Qur'an, and English translations of sayings of Muhammad (collected from hadith and Sira).Bless sins (talk) 10:09, 4 January 2008 (UTC)

Recent edits

SlimVirgin made a series of undiscussed edits to the policy.[1] Also, I linked to the relevant section of WP:V.[2] I am just posting this so people are aware and the changes may be reviewed. Personally, I think changes are mostly a positive revision. My only nit is that "material" should be clearly used as either a word for references or a word for article claims. I think the usage both ways muddles things a bit. At the least, I would recommend restoring the "claim" wording (or "article statement" as a replacement) in the first paragraph of "Reliable sources". Vassyana (talk) 18:05, 3 January 2008 (UTC)

  • I have made a suggested revision of wording in a sandbox.[3] Vassyana (talk) 18:32, 3 January 2008 (UTC)
I'm still going over the new revision, but I'd agree with SV's apparent belief that using "claim" is poor word choice. Imagine the WikiLawyering possible over what constitutes a claim...Someguy1221 (talk) 18:08, 3 January 2008 (UTC)
This was a light copy edit, not a change, Vassyana. Please stop this language of "SlimVirgin makes undiscussed edits!!!" whenever I fix some punctuation or tighten wording. There was no substantive content change. SlimVirgin (talk)(contribs) 18:42, 3 January 2008 (UTC)
It didn't change the substance (in my opinion, someone else could disagree), but it was a significant rewording. They were undiscussed edits. I didn't revert them. I didn't impugn them or you. On the contrary, I agreed explicitly with the changes, barring a nit regarding a single wording choice. I don't understand this antagonism when I'm essentially agreeing with you. Vassyana (talk) 09:05, 4 January 2008 (UTC)

Following on from SlimVirgin's rewrite of part of the Sources section earlier today, in which she addresses the editor directly[4], can I suggest something like this for the first para under the Sources header:

In general I prefer to see policy written in terms of "you must do this" rather than the passive "this must be done" as it engages more directly with the reader. I think further policy rewriting in this style would be beneficial too. Comments?  —SMALLJIM  23:59, 3 January 2008 (UTC)

Without expressing a preference about any particular wording, I want to agree with Smalljim's preference for active rather than passive voice. At times, this policy has been confusing because it was unclear whether it was discussing something written in a non-Wikipedia published source, something unpublished but not written by the Wikipedia editor, or something written by a Wikipedia editor. When we write this policy in the active voice, we tend to make it clear who it is that that wrote the material we are interested in. --Gerry Ashton (talk) 00:50, 4 January 2008 (UTC)
Active voice is of course preferred over passive, but second person pronouns are not necessarily appropriate. It is probably better to say "editors should..." and not "you must do this". We want policies to be descriptive rather than prescriptive. While policies should describe good editing behavior from the editor's perspective, it should be readable by anyone. For example consider the perspective of a mediator or arbitrator reading the policy. Dhaluza (talk) 00:59, 4 January 2008 (UTC)
Policies are descriptive and prescriptive, obviously, or we wouldn't be bothering to write the descriptions down. SlimVirgin (talk)(contribs) 01:22, 4 January 2008 (UTC)
We write it down to describe how we do things, not to prescribe how we do things. That is a fundamental difference. Dhaluza (talk) 02:14, 4 January 2008 (UTC)
Why would we want to describe if not to prescribe? SlimVirgin (talk)(contribs) 02:16, 4 January 2008 (UTC)
Because that is not how things work in a consensus based system. A consensus means that people agree to follow the consensus opinion. So we try to describe the consensus opinion as best we can, and then we expect people to follow it by consensus. We do not prescribe that because of a "past consensus" you must do things in some rigid precise way. The distinction here is that when differences arise, we invite further discussion to interpret the consensus, and allow that consensus can be difficult to describe and subject to change. In the prescriptive case, the rules can take on a life of their own, and become part of a self-sustaining bureaucracy, which Wikipedia is not. Dhaluza (talk) 10:04, 4 January 2008 (UTC)
So you're saying that "you should do this" is shorthand for "current consensus indicates that you should do this, else you will probably be reverted". Sounds sensible. Is there then any place in policy wording for must?  —SMALLJIM  12:06, 4 January 2008 (UTC)
No place; Wikipedia policies should be phrased in terms of "should" instead of "must". — Carl (CBM · talk) 13:10, 4 January 2008 (UTC)
Is this your opinion or is it an existing consensus, Carl? The exhortation appears in all three content policies.  —SMALLJIM  13:42, 4 January 2008 (UTC)
It's very established in practice. In WP:V, the only "must"s are in the nutshell and in a single sentence referring to NPOV, which is a foundation issue not a local policy. Moreover, there are many more policies than the content policies (protection, blocking, deletion, etc.). The biggest violator of the "should" convention is this particular document. I encourage you to look through the policies and search for the word "must" to see how infrequently it is used. — Carl (CBM · talk) 13:54, 4 January 2008 (UTC)
Point taken. I'd expect it to appear more often in the "big three" policies. I think it's really being used as shorthand for "current consensus strongly indicates that you should do this, else you will almost certainly be reverted and warned about it too" (cf above). You seem to be recommending the downgrading of some of the uses of must in this policy.  —SMALLJIM  14:31, 4 January 2008 (UTC)
I agree with "descriptive, not prescriptive". Wikipedia is driven by consensus. Policies and guidelines are written to document what consensus has established. They describe consensus. They don't prescribe behavior. Anything else would go against the spirit behind "Wikipedia does not have firm rules" and "Wikipedia is not a bureaucracy". (And thus "should" is preferred over "must".) —DragonHawk (talk|hist) 13:42, 4 January 2008 (UTC)
I'm shooting off at a bit of a tangent here, but if it's true that policies and guidelines document current consensus—and I don't see why that shouldn't be so, it seems a very "wiki" way of doing it—then all the recent discussions that have taken place here (and at WP:EVAL) about the definitions of primary and secondary sources with reference to what external organizations say about them are not only irrelevant but seriously erroneous. What we should therefore be doing here these days is looking around article space, AfDs etc, noting which types of sources are being accepted as primary and which as secondary, and then documenting the results at WP:PSTS. Maybe if we did this it would become evident that much of the P/S distinction is irrelevant…  —SMALLJIM  14:47, 4 January 2008 (UTC)
I completely agree that WP:NOR should include discussion of how the articles that we consider to be the best use their sources, and which uses of sources are commonly rejected. I think that would be more useful than an attempt to choose an arbitrary definition of primary/secondary, which many people will find hard to use because it confliucts with the definition they have already learned. — Carl (CBM · talk) 15:51, 4 January 2008 (UTC)
Yes, I do now seem to be arguing for two incompatible points of view.[5] I think that's the point you're making. Never mind, it's said to be good exercise :-) And both do express my dissatisfaction with the existing PSTS.  —SMALLJIM  16:18, 4 January 2008 (UTC)

"consensus changing" tag

I see we are back to people wanting to tag the policy with a "consensus may be changing" tag. I am of mixed feeling about this, but if we are going to have a tag in the article then it should only apply to the PSTS section (as that is what all the debate is about). Consensus is definitely not changing as to the bulk of the policy. Blueboar (talk) 15:52, 4 January 2008 (UTC)

Sorry folks...

.... but these discussions are becoming nothing more than a pissing contest. I am taking this page off my watchlist for one month. See you around then. ≈ jossi ≈ (talk) 16:49, 4 January 2008 (UTC)

When two parties to an argument that should be resolved by compromise, which is to say an argument in which both parties have flaws and valid points to their argument, attempt to win the argument by offending or damaging each other. There's no winner, both parties wind up damaged.

Fullstop's description of a primary source

I'm copying this from above, in case it gets lost. It's Fullstop's description of a primary source in Wikipedia:Evaluating sources, and I'm trying to see here how people's minds are working about this, so it would be good to clarify it. She wrote:

In the humanities, materials that are potential objects of academic research but do not themselves constitute academic research are considered 'primary sources'."

When I asked what she meant, she wrote:

  1. A primary source is a novel* intimation(abstraction, concept, conclusion, conviction, determination, design, doctrine, experiment, estimate, exposition, explanation, hypothesis, idea, impression, interpretation, judgment, notion, perception, rationale, scheme, solution, story, suggestion, supposition, theory, thought, view) that has (to anyone's knowledge) never been made/scrutinized before. Additionally, any statement that does not reflect scientific method is a primary source.
  2. A secondary source is a methodical exploration(acknowledgment, analysis, examination) of a primary source.

In both cases, "source" is relative to the target statement that draws on that source. Of course, someone is going to argue about those definitions, but thats not what you asked for. :) This still isn't what I was referring to when I said "I once wrote a complete PSTS in 4 sentences". I can't find it, its buried somewhere in the archives.

BTW: As far as WP need care about contestable determinations, those that have never been scrutinized/reiterated by any other source "do not exist." This is an implication of the due-weight clause in WP:NPOV since editors may not assign weight themselves (which would again be OR) but must have a source to do it for them. -- Fullstop (talk) 23:53, 3 January 2008 (UTC)

I have no idea what any of the above means. A primary source is not "a novel intimation (abstraction, concept, etc) ... that has (to anyone's knowledge) never been made/scrutinized before." Is the problem here that people still don't understand what a primary source is?

Fullstop, if I'm sitting next to you in a restaurant when you drop your fork, I'm a primary source for anyone wanting to investigate whether you dropped your fork. Even once the story is written, published, scrutinized, I am poly-graphed, the fork is DNA tested, people dig us up in 1,000 years and do more tests on the fork-dropping incident, my story is still a primary source. Is there something about this that is still unclear? SlimVirgin (talk)(contribs) 01:22, 4 January 2008 (UTC)

You're a primary source, unless you go on to assert how or why the fork was dropped--then you would become a secondary source. Dhaluza (talk) 02:10, 4 January 2008 (UTC)
Well, first, that's not what Fullstop says, and I'd like to hear what his understanding is, given that he's a key objector to the policy.
But to address your point, it would depend what I said. If I said "I think she dropped her fork out of fatigue, because her eyes looked pretty red, and I think she might be the same woman I spotted drinking a bottle of vodka in a bar last night," I'd still be a primary source. But if I said, "I think she dropped it because I read an article somewhere that that particular brand of fork is really slippy to hold," then, yes, secondary source. But how often do we need to get into these details on Wikipedia? Dhulaza, Fullstop, or Vassyana, please show me an actual example of when you last had to wrestle with this minutiae in an actual article? SlimVirgin (talk)(contribs) 02:22, 4 January 2008 (UTC)
I don't think we need to go into that kind of minutiae. I also disagree with Fullstop's framing of primary/secondary. Vassyana (talk) 09:02, 4 January 2008 (UTC)
The point was that the difference between a primary and secondary source is not so black and white, even if we agree on a precise definition (and I do agree with your clarification). As to the point of how often we get into this minutia, this source typing invites us to get into this unnecessary minutia, because it treats different types of sources differently based on their nature. What we should be talking about is whether a source is used appropriately, regardless of type. So in your examples, both your primary source opinion and secondary source analysis would be irrelevant, and the source typing is completely unnecessary. Dhaluza (talk) 09:51, 4 January 2008 (UTC)
Dhaluza, you wrote that this kind of source classification invites us to get into this minutiae, but it really doesn't. For the vast majority of articles, a source will either be a primary or secondary source. I've edited a couple of articles where there was a source who was both primary and secondary -- primary because involved in the issue, secondary because he started pontificating about issues he had no knowledge of -- and in fact, it has become a big issue in Holocaust historiography, in particular. You have sources who were there, and who can attest directly to some of the facts on the ground, but then they start speculating about how, if the Allies had done this or that sooner, then such-and-such could have been avoided, and of course they're not in a position to say that. Yet they insist that, as primary sources, they know more about the Holocaust than the historians who are secondary sources, which has given rise to a debate about so-called "survivor versus expert discourse," which is basically a debate about the value of primary versus secondary sources. But the areas in which this ever becomes an issue on Wikipedia are few and far between, and when the issue does arise, it makes for an interesting article, so long as it's explained properly. But that's no reason to throw the classification out. You seem to be saying that any classification we use should be perfectly and invariably illuminating, but life isn't like that, and nor is research into life. :-) SlimVirgin (talk)(contribs) 01:36, 5 January 2008 (UTC)
We have seen on this talk page (and others) how a question over what is original research morphs into an argument over what is a primary/secondary source. I think this is counterproductive, and ultimately the PSTS construct in the context of NOR does more harm than good. It may be useful in other contexts, but that is not for us to discuss here. In your example, if the survivors are including their own opinions in articles as WP editors, that is OR plain and simple--no need to decide if it's primary or secondary. If editors are including survivor's accounts from published RS alongside historical analysis, then it is an issue of NPOV and relative weight, not OR. Whether the PSS distinction is relevant in the NPOV context is not an issue we sould be discussing here. Dhaluza (talk) 20:40, 5 January 2008 (UTC)
First, you asked for *in the humanities*. Empirical methodology is not "in the humanities."
Second, iff I were to transport your fork-dropping "example" (nothing real-life on hand?) to the humanities,...
  • for example in exegesis of a fictional encounter with slippery silverware, then your observations and suppositions are both -- from the point of view of the person who is analyzing the novel -- a primary source. The analysis thereof is then a secondary source for whoever cites that source.
  • for example in a history of fork-dropping, where yours is the original narrative, then your observations and suppositions are both -- from the point of view of the person who is analyzing the narrative -- a primary source. The analysis thereof is then a secondary source for whoever cites that source.
  • for example as in the religion of fork-dropperism, where yours is divine revelation, then your revelation and doctrine are both -- from the point of view of the person who is analyzing the beliefs -- a primary source. The analysis thereof is then a secondary source for whoever cites that source.
  • for example in the writing of an encyclopedia -- which is a purely philological excercise -- a summary of the hypothesis on fork-droppage is an example of the use of a secondary source because the conclusions are -- as with anything experimentally/empirically determined -- always dependent on a preceding dataset. That dataset, the observation, is the primary source material.
The fundamental problems of the one-size-fits-all approach of primary/secondary differentiation have been rehashed a gazillion times (see almost any request for interpretation assistance, including the current [#Translation]]).
Ipso facto, sources are sources are sources are sources are sources are sources. Sources only become "primary" or "secondary" when the purpose and context for which they are being cited is known.
It will not make any difference whether you rewrite the policy to have it your way once more or not. The talk/edits/reverts are bound to continue ad-nauseum as long as there is any differentiation artifice. What you are trying to do has already been done seven ways to Sunday, though usually without misrepresenting anyone.
-- Fullstop (talk) 11:56, 4 January 2008 (UTC)

SV said "I'm trying to see here how people's minds are working about this". So here's my take - it seems to me that Fullstop is expressing a concept somewhat similar to the one I set out last month[6] which can be summarized as:

  • A primary source is (a) an "observed thing" that is relevant to the topic; or (b) someone describing their own thoughts about the topic or their experiences of it.
  • A secondary source is someone discussing other people's thoughts about the topic or their experiences of it.

The benefit of these definitions (see above link for the caveats) is that they are truly exclusive: in relation to a particular topic there can be no doubt(*) as to the P/S classification of a source. This is in contrast to the existing definitions with their easily-misinterpreted and non-exclusive terms very close and one step removed. And this formulation gives the "right answers" too(*).
(* as far as I can tell)  —SMALLJIM  12:57, 4 January 2008 (UTC)

An issue with this proposal is that it doesn't seem to distinguish between a newspaper article reporting eyewitness testimony and a peer-reviewed scientific paper. If the goal of the source typing language is to point editors to the most reliable sources, we want to encourage the use of peer reviewed papers(which will hedge claims that are merely speculative), and point out that eyewitness testimony must be taken with a dose of salt even if it has appeared in a newspaper article. — Carl (CBM · talk) 13:59, 4 January 2008 (UTC)
If you mean that the above definitions don't say which is preferable to use, I agree - that's the main job of WP:RS, not PSTS. The full proposal does include a "using sources" section though (I've copied the model to my user space here for further work). As far as I can see, source typing has just two purposes: 1. to define "secondary source" for the purposes of WP:Notability and 2. to warn WP editors that (the type 1b) primary sources may not be reliable or neutral (yes I know, secondary might be just as bad).  —SMALLJIM  19:14, 4 January 2008 (UTC)
Smalljim, can you tell us how these two purposes relate to concept of NO ORIGINAL RESEARCH... if not, we have yet another reason to move this section to some other policy or guideline. Blueboar (talk) 19:31, 4 January 2008 (UTC)
Here's what NOR looks like without the PSTS section. The "Using sources" section should probably be expanded a bit to compensate, but it doesn't appear to be a big loss to me. PSTS as a separate guideline would be fine, I think.  —SMALLJIM  20:05, 4 January 2008 (UTC)
I agree... and this is something that multiple editors have been saying for several months now. Unfortunately, there is a small but vocal group who seem to think PSTS is in some way vital to the concept of NOR and resist any attempt to shift it to its own guideline (or to move it to some other policy or guideline where it more directly relates). I just wanted to know your thoughts on the subject. Thanks. Blueboar (talk) 20:16, 4 January 2008 (UTC)
I also agree. While PSTS is no doubt useful in some contexts, it is not essential to understanding NOR. On balance, it actually detracts from that understanding because it is unnecessarily complex, and somewhat ambiguous in practice. In a conflict situation, this ambiguity has been seen repeatedly to start new side arguments over what is primary vs. secondary, rather than focus on what is or is not original research. So, taken all together, PSTS has an unfavorable cost/benefit ratio. Dhaluza (talk) 01:14, 5 January 2008 (UTC)
But I think the complexity is being invented by people who themselves don't understand the issue (and these people tend to congregate here because they want to change it -- but it's being used all over Wikipedia without problems).
As I wrote above to Blueboar, the distinction is: "A primary source is one close to, and with direct knowledge, of an issue. In contrast, a secondary source is at least one step removed, and provides a secondhand account or a general overview."
Seriously, where is the complexity in that? SlimVirgin (talk)(contribs) 01:42, 5 January 2008 (UTC)

Tags on PSTS

Recent attempts to tag the PSTS section to highlight the discussions here have been removed with edit summaries to the effect of "Please do not..." and some form of tag, deface, etc, without discussion, even though the use of the tags have been discussed on this talk page. What is the objection to placing a tag to show the topic is being discussed and consensus may be changing? The tags exist by consensus, so they do have appropriate uses. And there is active discussion, as the links in the last tag show. Yes, that particular tag is rather large, but every other tag that has been tried has been removed as well, including one designed as a compromise. If the tags are unsightly, should we just remove the policy tag at the top instead? Dhaluza (talk) 00:15, 6 January 2008 (UTC)

I would very much object to removing the Policy tag at the top. NOR is policy... and I think there is solid consensus for that designation. There is also solid consensus for the bulk of the sub-sections within NOR. The only section that is being discussed and debated is the PSTS section. If anything does get tagged, it should be that secton and that section only. Blueboar (talk) 04:11, 6 January 2008 (UTC)

The two versions

SlimVirgin's version Alternate version[1]
For the purposes of Wikipedia policies and guidelines, primary, secondary, and tertiary sources are defined as follows:
  • Primary sources are sources very close to the origin of a particular topic. An eyewitness account of a traffic accident is an example of a primary source. Primary sources that have been published by a reliable source may be used in Wikipedia, but only with care, because it is easy to misuse them. For that reason, anyone without specialist knowledge who reads the primary source should be able to verify that the Wikipedia passage agrees with the source. Any interpretation of primary-source material requires a reliable secondary source for that interpretation. To the extent that part of an article relies on a primary source, it should:
  • only make descriptive claims about the information found in the primary source, the accuracy and applicability of which is easily verifiable by any reasonable, educated person without specialist knowledge, and
  • make no analytic, synthetic, interpretive, explanatory, or evaluative claims about the information found in the primary source.
Examples of primary sources include archeological artifacts; photographs; historical documents such as diaries, census results, video or transcripts of surveillance, public hearings, trials, or interviews; tabulated results of surveys or questionnaires; written or recorded notes of laboratory and field research, experiments or observations, published experimental results by the person(s) actually involved in the research; original philosophical works, religious scripture, administrative documents, and artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs.
For the purposes of Wikipedia policies and guidelines, primary, secondary, and tertiary sources are defined as follows:[2]
  • Primary sources are sources very close to the origin of a particular topic. An eyewitness account of a traffic accident is an example of a primary source. Primary sources that have been published by a reliable source may be used in Wikipedia, but only with care, because it is easy to misuse them. For that reason, anyone—without specialist knowledge—who reads the primary source should be able to verify that the Wikipedia passage agrees with the primary source. Any interpretation of primary source material requires a reliable secondary source for that interpretation. To the extent that part of an article relies on a primary source, it should:
  • only make descriptive claims about the information found in the primary source, the accuracy and applicability of which is easily verifiable by any reasonable, educated person without specialist knowledge, and
  • make no analytic, synthetic, interpretive, explanatory, or evaluative claims about the information found in the primary source.
Examples of primary sources include archeological artifacts; photographs; historical documents such as diaries, census results, video or transcripts of surveillance, public hearings, trials, or interviews; tabulated results of surveys or questionnaires; written or recorded notes of laboratory and field research, experiments or observations, published experimental results by the person(s) actually involved in the research; original philosophical works, religious scripture, administrative documents, and artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs.[3]
Unsourced eyewitness accounts or other unsourced information obtained from personal experience should not be added to articles, as this would cause Wikipedia to become a primary source for the added information (see Wikipedia:Verifiability).
  • Secondary sources are secondhand accounts at least one step removed from an event. Secondary sources may draw on primary sources and other secondary sources to create a general overview; or to make analytic or synthetic claims. Wikipedia articles should rely on reliable, published secondary sources. All interpretive claims, analyses, or synthetic claims about primary sources must be referenced to a secondary source, rather than original analysis of the primary-source material by Wikipedia editors.
  • Secondary sources are references that draw on research and other references to make interpretive, analytical, or synthesized claims, or create a general overview.[4][5] Secondary sources are valued for analysis, broader context, and a more independent view. However, they can be biased by the views of their author(s) and/or publisher, so caution is required to preserve the neutral point of view. Care should also be taken to avoid undue weight and ensure the information cited is used in context.
  • Tertiary sources are publications such as encyclopedias or other compendia that sum up secondary and primary sources. For example, Wikipedia itself is a tertiary source. Many introductory textbooks may be considered tertiary to the extent they sum up multiple primary and secondary sources. Tertiary sources can be helpful in providing broad summaries of topics that involve many primary and secondary sources. Some tertiary sources may be more reliable than others, and within any given tertiary source, some articles may be more reliable than others. WP:Verifiability#Reliable sources describes some criteria for assessing reliability of sources.

Appropriate sourcing is a complicated issue, and these are general rules. The decision as to whether primary or secondary sources are more suitable on any given occasion is a matter of common sense and good editorial judgment, and should be discussed on individual article talk pages.

  • Tertiary sources are publications such as encyclopedias or other compendia that sum up many secondary and primary sources. For example, Wikipedia itself is a tertiary source. Many introductory textbooks may also be considered tertiary to the extent they sum up widely accepted results of large amounts of primary and secondary sources. Tertiary sources can be useful in providing context and avoiding original research in topics where there exist very large amounts of primary and/or secondary sources.

All sources should be used in a way that does not give rise to new analyses, syntheses or original conclusions that are not verifiable. Where interpretive claims, analysis, or synthetic claims are included in Wikipedia articles, use appropriate sources rather than original analysis by Wikipedia editors.

Appropriate sourcing is a complicated issue, and these are general rules. The decision whether primary or secondary sources are more suitable on any given occasion is a matter of common sense and good editorial judgment, and should be discussed on individual article talk pages.


  1. ^ This is the previously standing version, barring the secondary sources section which is/was a proposed replacement.
  2. ^ Various professional fields treat the distinction between primary and secondary sources in differing fashions. Some fields and references also further distinguish between secondary and tertiary sources. Primary, secondary and tertiary sources are broadly defined here for the purposes of Wikipedia.
  3. ^ Definitions of primary sources:
    • The University of Nevada, Reno Libraries define primary sources as providing "an inside view of a particular event." They offer as examples: original documents, such as autobiographies, diaries, e-mail, interviews, letters, minutes, news film footage, official records, photographs, raw research data, and speeches; creative works, such as art, drama, films, music, novels, poetry; and relics or artifacts, such as buildings, clothing, DNA, furniture, jewelry, pottery.
    • The University of California, Berkeley library offers this definition: "Primary sources enable the researcher to get as close as possible to what actually happened during an historical event or time period. Primary sources were either created during the time period being studied, or were created at a later date by a participant in the events being studied (as in the case of memoirs) and they reflect the individual viewpoint of a participant or observer."
  4. ^ Borough of Manhattan Commmunity College, A. Philip Randolph Memorial Library, "Research Help:Primary vs. Secondary Sources" notes that a secondary source "analyzes and interprets primary sources", is a "second-hand account of an historical event" or "interprets creative work". It also states that a secondary source "analyzes and interprets research results" or "analyzes and interprets scientific discoveries".
  5. ^ The National History Day website states simply that: "Secondary sources are works of synthesis and interpretation based upon primary sources and the work of other authors."


These are the two versions. I've copy-edited mine slightly to make it tighter than the version that was on the page. Vassyana's is copied directly. By all means, explain how hers is an improvement. SlimVirgin (talk)(contribs) 19:27, 2 January 2008 (UTC)

I corrected the inaccurate labeling of that as "my" version and included a brief footnote for explanation. I've also created a notes section so the footnotes are visible. Vassyana (talk) 02:21, 3 January 2008 (UTC)
It was most certainly the standing version before your changes. My change was purely a replacement of the secondary sources section in this instance. Vassyana (talk) 17:31, 3 January 2008 (UTC)
I have a few questions about the changes. Why was the last paragraph of the primary sources section removed? Why was the second to last paragraph removed? Why is there an additional exhortation about tertiary sources' reliability? Isn't the vast bulk of tertiary sources (encyclopedias, textbooks, review articles) more reliable than the average source? A few questions that immediately come to mind. Vassyana (talk) 02:30, 3 January 2008 (UTC)
Yes, should the objection to addressing the bias of secondary sources be equally applicable to addressing the reliability of tertiary sources? Dhaluza (talk) 10:08, 4 January 2008 (UTC)

What seems a bit disproportionate is that WP:NOR should contain a "warning" regarding the use of primary sources, while this is not balanced by a "warning" regarding the use of secondary or tertiary sources (which have their own WP:NOR issues). For example: Source A writes an opinion piece in a newspaper. That would make source A a primary source. Source B is a book that pretends to cover the domain on which Source A had written his opinion piece. In this example Source B is published by a respectable publishing house, which would make Source B a secondary source. But suppose Source B misquotes Source A. This happens all the time, and scarcely less in high-standard publications, so in this example I go from the assumption that Source B can not be excluded for WP:V/WP:RS reasons, so this is not a WP:V/WP:RS issue. Source B has also some valid points in its own right, so the source couldn't be excluded for WP:NPOV reasons either. In this case it would nonetheless be incorrect to give the secondary source precedence over the primary source, when it comes down to presenting the view of Source A in Wikipedia.

Here's an actual example of how "easy" it is to misuse a secondary source: [7]

Examples regarding tertiary sources can also be given: since tertiary sources condense condensations of primary sources, something might have gone lost (without such tertiary source becoming an unreliable source for that reason). Such tertiary source might give hints which Wikipedians try to expand a bit more (for example for WP:MOS reasons, trying to make fluent sentences - or to doctor a succession line of kings where the appropriate designation would be "warlord" - see for example Haakon the Red article based on tertiary sources like Nationalencyklopedin), easily leading to misrepresentations in Wikipedia.

WP:NOR should indicate such easily made abuse of secondary and tertiary sources, or alternatively, no provisions singling out primary sources for this danger should be included at policy level (i.e. move the "easy abuse" aspect to a guideline like WP:RS or an essay like WP:EVAL). I think Vassyana and those who support the version above on the right would like to follow the first of these alternatives. For myself, I rather tried to follow the second, for which I wrote WP:WITS, which does not differentiate between p/s/t sources on the point of "ease of abuse" for a proposed policy-level formulation. I really believe that in Wikipedia practice all types of sources are abused as easily, maybe for different reasons, but definitely "as easily" and as often. Anyway, if "ease of abuse" would be the only reason for explaining the p/s/t distinction at policy level (and then targetting suspicion on only one of the three types), that would be counterproductive. I think there are other (!) valid reasons for introducing the p/s/t distinction at policy level, and that's another reason why I wrote WP:WITS: some things that imho are valid at policy level are included there, without currently being clearly covered by WP:NOR (I mean, maybe somehow they are obliquely covered in WP:NOR, but in a way too confusing language). --Francis Schonken (talk) 12:20, 5 January 2008 (UTC)

While I agree with your observations, I disagree with your suggestion that this could be addressed in WP:NOR. These misuses of sources are not related to original research. OR is going beyond the source, and using it to support a statement not directly contained in the source. Made up facts, or facts synthesized to advance a point, are OR. Using a published source to say what it says, but in an unencyclopedic way is not an OR issue. For example, using a self-published source to include self-congratulatory material or advertising from that source, (without characterizing it as such) is not OR, because the material is not original. It's not appropriate, but that's a different issue. Dealing with conflicting sources could become an OR issue if the editors try to resolve the conflict outside of the sources, but that is again going beyond the sources. We need to keep WP:NOR focused on OR, and not expand it to cover all appropriate use of source problems. Dhaluza (talk) 13:20, 5 January 2008 (UTC) - updated 12:10, 6 January 2008 (UTC)
To whom were you replying? My suggestion remains NOT to address this in WP:NOR ("at policy level" as I called it). Really, I think we have more consensus than the words indicating disagreement suggest. --Francis Schonken (talk) 13:24, 5 January 2008 (UTC)
Here's my more elaborate answer: Wikipedia:Evaluating sources#SWOT analysis on primary, secondary and tertiary sources Wikipedia:Sources - SWOT analysis --Francis Schonken (talk) 19:02, 5 January 2008 (UTC) updated 17:07, 6 January 2008 (UTC)

All of the examples given of Primary Sources can in principle be verified by tracing back to the original author, photographer, archaeological digger, etc.
"Religious Scriptures" is a glaring exception, their provenance is usually unknown. Yet they are cited as a source in some purported "history" articles (including God's gift of the "Promised Land" to the "Chosen People"). You won't find this in any peer-reviewed academic journals, which we are supposed to prefer. Fourtildas (talk) 06:59, 6 January 2008 (UTC)
Au contraire! Scriptures a subject for academic study, both within secular institutions of higher learning and under longstanding scholarly traditions of the major faiths of the book. The Bible is not a reliable source for what God did. Assuming God himself wrote it, it is a self-published promotional work that contains disputed claims and cannot therefore be cited to verify these things. However, the bible is obviously a reliable source on simple questions as to the content of the bible. Barring translation concerns and versions, it may be a source (but it is still better to find a secondary source, even for the bible). Wikidemo (talk) 08:07, 6 January 2008 (UTC)
Fourtildas, Wikidemo do your comments above have anything to do with the comment I wrote above it? If they do: I don't see how (e.g. I didn't give any example of a primary source), please explain. If not, please use indents appropriately --Francis Schonken (talk) 12:10, 6 January 2008 (UTC)

Archive naming

Would anyone object if I went through and renamed/moved all the archive pages so their names were of the form "Archive 1" rather than "archive1"? Doing so will mean User:HBC Archive Indexerbot will be able to index them, and {{archives}} will be able to auto-list them. —DragonHawk (talk|hist) 04:09, 4 January 2008 (UTC)

Seems reasonable--we should follow standard convention. Dhaluza (talk) 10:28, 4 January 2008 (UTC)
Okay, that's done. There are a freaking huge pile of subpages under this one. Prefix Index. A few of the archives even have their own archives. Unfortunately, the archive indexer bot only knows how to archive one level of pages (per index), so to find a discussion that's in one of the sub-archives, one will still need to go digging. But it should be better than it was, at least. —DragonHawk (talk|hist) 05:12, 5 January 2008 (UTC)
Thanks for doing that. It looks like a big improvement over how it was previously. wbfergus Talk 20:08, 5 January 2008 (UTC)
Yes, indeed. Thanks for the good work. Its much better now. -- Fullstop (talk) 21:47, 5 January 2008 (UTC)

Thanks, everyone, for all the thanks! :-D —DragonHawk (talk|hist) 23:18, 6 January 2008 (UTC)

A new look at OR in general

I have been thinking about a question that SV posed above... "what is the distinction between something's being OR and being the product of OR?" This isn't really an answer to her question... but the result of my thinking about it. I share it as a different approach to OR that does not involve source typing. This isn't proposed language... or even a suggested change... just my thoughts on a different way of thinking about the issue of OR.

As I see it, there are two forms of Original Research that we want to address in this policy.... for the moment I will call them: "Direct OR" and "Indirect OR."

"Direct OR" is fairly obvious... it is when an editor adds previously unpublished facts or observations to an article. It might be the discussion of something that the editor adding the information did or observed himself/herself. If you conduct a scientific experiment and discover something new, anouncing your discovery in Wikipedia without its being published elsewhere is Direct OR. If you find a previously unpublished letter written by George Washington in your attic, and write about it in Wikipedia, without publishing it elsewhere, you are engaging in Direct OR. In otherwords, Direct OR results when an editor uses Wikipedia as the place of first publication for something he/she actually did or discovered. Direct OR also results from an editor discussing his/her thoughts about this unpublished material ... his/her analysis or conclusions about it.

"Indirect OR" is not as obvious... Essentially it is when an editor adds an original analysis of previously published material, or states an original conclusion based upon previously published material. The original synthesis of previously published materials falls into this category. The facts themselves may come from any source type... primary sources, secondary sources or even tirtiary sources. What makes it OR is that the analysis, synthesis or conclusion is novel and unpublished. In other words Indirect OR results when the editor is uses Wikipedia as the place of first publication for his/her analysis, or conclusions... his/her thoughts about something that has been published.

Any thoughts? Blueboar (talk) 17:03, 6 January 2008 (UTC)

Good points. It might be useful to make that distinction in the policy -- direct and indirect OR. SlimVirgin (talk)(contribs) 17:19, 6 January 2008 (UTC)
Re. "Direct OR": aka we don't make primary sources, as in the first point of Dhalusa's #Proposed replacement for PSTS
Re. "Indirect OR": aka we don't make secondary sources, as in the second point of Dhalusa's #Proposed replacement for PSTS
I think here is an advantage to stick to terminology that has currency (primary/secondary sources) over inventing new terminology (direct/indirect OR). --Francis Schonken (talk) 17:48, 6 January 2008 (UTC)
These distinctions remind me of what is typically required in papers in humanities classes written by university undergraduates. Their own fact-gathering or untethered opinions will be graded poorly, but analysis of published sources will get good grades. Simple summaries of published sources will get mediocre grades, at best. --Gerry Ashton (talk) 18:23, 6 January 2008 (UTC)
Not sure what you're trying to tell us? Are you getting nostalgic about the time you got good grades for analyses of published sources? WP:NOT#OTHOUGHT has a suggestion for you in that case: Wikinfo.
Maybe making summaries of published sources has something "mediocre", not if one does it on Wikipedia scale. --Francis Schonken (talk) 19:10, 6 January 2008 (UTC)
My point is that some new editors will have a mindset that analyzing and drawing conclusions from published sources is the normal approach to writing papers, and will be reading this policy from that point of view. --Gerry Ashton (talk) 19:19, 6 January 2008 (UTC)
So, are you in favour of the approach of Dhalusa's #Proposed replacement for PSTS, making clear in a few sentences that Wikipedia is different on this point?
Or do I misunderstand and are you in favour of letting new editors in the illusion that Wikipedia is what it is not, and leave the cleanup to more experienced users?
Or do you not have an opinion on this point? --Francis Schonken (talk) 19:33, 6 January 2008 (UTC)
I would prefer an aproach that avoids using the words primary, secondary, and tertiary. I think a big problem with the proposed replacement is that it promotes the word tertiary into a decisive role, where in the past, it was just mentioned as a point of information. Since some people only divide sources into the categories of primary and secondary, and don't acknowledge the existence of a tertiary category, I don't think that word should be given such a decisive role.
Further, I think it is possible to describe an entire publication as tertiary, but distinguishing between secondary or tertiary on an article-by-article, or paragraph-by-paragraph basis is not feasible. --Gerry Ashton (talk) 19:42, 6 January 2008 (UTC)
The source types are generalizations. The actual source type could be further broken down sentence by sentence, or even phrase by phrase as well. And I certainly have problems with the specific definitions of PSTS as well. But, I think using the primary, secondary and tertiary terms in context, without attempting to making precise definitions here in this policy, is a reasonable compromise between what we have now (which probably causes most editors' eyes to glaze over) and something that avoids the terms altogether. Dhaluza (talk) 00:46, 7 January 2008 (UTC)
That would be fine, if we can make people refrain from objecting to individual edits on the basis of these generalizations. But I think people will revert small edits, and use these generalizations as justification. --Gerry Ashton (talk) 01:05, 7 January 2008 (UTC)
Old habits die hard. And bad habits die harder. Dhaluza (talk) 01:19, 7 January 2008 (UTC)

primary over secondary sources

I am fully committed to secondary sources for any interpretive, explanatory, analytical, synthetic, claim or generalization. However I do think there the matter is simply whether or not something happened, I think we need stronger language in favor of aprimary source. Perhaps this is best discussed under the WP:V policy, but I bring it up here first because I sure do not want to erode NOR.

The case: some people claim Marx was an anti-Semite. This is based on his having written certain things. People (including me) have been incorporating material into the Karl Marx article. This material so far falls into three categories. I have no problem with two, but a serious problem with the third:

  1. quotes from essays, articles, books, or letters written by Marx - primary sources, but as long as they are not used to make synthetic claims, okay by me.
  2. quotes from secondary sources that interpret quotes by Marx. This is fine by me as people have produced secondary sources for a variety of views i.e. interpretations of what Marx "meant."
  3. quotes form secondary sources that report what Marx said

It is this last material I take issue with. I am not challenging the interpretation of what Marx said, I am challenging whether Marx actually said it or not. Here is the thing: virtually everything marx has written has been published. Aside from various collections of essays and articles, and books, there is the Complete Works, which any major library has. I believe that if we cllaim marx wrote something, we need to trace it to the document where Marx actually wrote it. I hope it is very clear that I am not suggesting to use a primary source to forward any argument. All I am concerned with is the veracity of whether Marx did or did not say something. Given how contentious this topic is, I believe we have an obliation to check facts and make sure that if someone claims he said it, there is a source that proves it. Again, I am not talking about claims about what he meant. Slrubenstein | Talk 23:50, 6 January 2008 (UTC)

This is exactly the Source A / Source B hypothetical example I gave above in #Discussion - so how do we handle this? --Francis Schonken (talk) 00:53, 7 January 2008 (UTC)
Obviously if you have both a primary and a secondary source for the quote, you should cite both. And certainly if there is a conflict, you should assume the secondary source is in error and use the primary. But as a practical concern, the original source of the quote may not be readily accessible. In this case, as long as there is no reason to doubt the reliability of the secondary source, we should allow the quote to be cited from that source. Yes, this does allow the unlikely possibility of error to occur--but I look for guidance to the principle in WP:V, "the threshold for inclusion in Wikipedia is verifiability, not truth." So the quote is verifiable from the secondary source, even if there is a slight chance it may not be true. I think we can take that chance, because in the worst case we would only be furthering an already published misconception, not starting a new one. This happens all the time in other published works, so the problem is not exclusive to WP. Our advantage is that if the primary source is found to correct the error, we can correct it immediately, rather than having it continue in its previous form forever. Dhaluza (talk) 00:58, 7 January 2008 (UTC)

Based on Dhaluza's comment, I see two relevant criteria. I wonder if someone would use these as the basis for proposing something to add to this or the V policy:

  • Citations from primary sources should be provided when (1) the quote in question is controversialor part of a controversy and (2) works by the quoted author are published.

The key thing here is, the more controversial the interpretation, the more questionable (I propose) any secondary source is in its account of the primary source material. IF the primary material is published, we really ought to encourage providing a citation for the primary source. The point here is NOT to use the original source for a quote to trump someone's interpretation of the quote, but rather because it is the most reliable source for the quote itself. I think something like this needs to be in policy and I invite others to propose and help craft an actual addition/clarification. Slrubenstein | Talk 01:22, 7 January 2008 (UTC)

I basically agree with a couple of caveats. First, the controversy must be an external one where reliable sources disagree, not just an internal Wikipedia controversy where editors disagree. If the subject is controversial, and the secondary source is likely biased in the controversy, then there is reason to doubt the accuracy of the quote. Second, we should not use the term primary source, when what we mean is the original source. The primary source designation is ambiguous and subject to interpretation (which may only escalate the dispute). Original source has a plain language meaning which is much clearer. Also, you should still cite both sources: the original source for accuracy, and the secondary source to show its relevance to the subject. Dhaluza (talk) 10:37, 7 January 2008 (UTC)

Okay, I proposed something - why don't you improve it, and we can see what othes thing? Slrubenstein | Talk 11:22, 7 January 2008 (UTC)

I think your proposal is misplaced here. This is not an OR issue at all. It deals with the nuances of handling reliability of sources under special circumstances. If the editor was concluding that Marx was antisemetic based on direct quotes of Marx from a primary or secondary or even tertiary source, that would be OR. Source type is irrelevant in this case. So I don't think it is related to PSTS either, regardless of whether PSTS is even related to NOR. So this probably belongs at WT:V as you said. Dhaluza (talk) 14:02, 7 January 2008 (UTC)
  • It is not merely that "secondary sources [should be used for] any interpretive, explanatory, analytical, synthetic, claim or generalization", but rather that any statement (evaluative or otherwise) should accurately reflect what someone else said.
    It just so happens that 'analysis' requires something else to analyze, thus making the analysis a secondary source for the material being analyzed.
    As such, there is nothing but "secondary sources for interpretive, explanatory, analytical, synthetic, claim or generalization."
    All the -ives and -orys exist only in secondary sources, and are what make a source a "secondary" one.
  • There is no relationship between "original research" and source-type:
    P. Misrepresenting an original Marx statement to assert that Marx was anti-Semitic == OR
    S. Misrepresenting a commentary on Marx to assert that Marx was anti-Semitic == OR
    T. Misrepresenting a summary of commentaries on anti-Semitism in Marx to assert that Marx was anti-Semitic == OR
    a) All three (primary, secondary and tertiary) source types can be misused to assert that Marx was anti-Semitic.
    b) While "original research" creates a primary source, the "research" itself can be with any kind of source.
    c) Verification that a 'pedia statement is not a misrepresentation of a source is possible with any kind of source.
Thus, what we need is stronger (& on topic!) language to prevent OR.
-- Fullstop (talk) 22:30, 7 January 2008 (UTC)

Synthesis vs. synthesis advancing a position

I take exception to this [8] edit by Slrubenstein because it no longer makes it clear that it is synthesis advancing an unpublished position that is prohibited, and not all synthesis. Synthesis that serves to summarize and organize published material is acceptable. --Gerry Ashton (talk) 00:05, 7 January 2008 (UTC)

I agree. And this is a case of changing long-standing policy without discussion as well. Dhaluza (talk) 01:11, 7 January 2008 (UTC)
Thee is no change of policy. And there was considerable discussion concerning revising the nutshell, ove a period of weeks. As for the specifics: the version I restored makes it clear that syntheses need to be backed up by sources, which is what our policy requires. The prior version makes it sound like no synthesis, even one from a reliable secondary source, is permitted. That violates our policy. Slrubenstein | Talk 01:59, 7 January 2008 (UTC)
Your definition of synthesis is too narrow. Straightforward organizing of published material, without introducing new ideas, can be considered synthesis too. But many would read the changed policy as prohibiting this. The reaction might be either to obey the overly broad policy, or more likely, regard it as absurd and ignore the entire policy. --Gerry Ashton (talk) 04:10, 7 January 2008 (UTC)
Your use of synthesis is wrong. From the Wiktionary: Synthesis, Noun: the formation of something complex or coherent by combining simpler things. That is all synthesis means for the purposes of this policy. That is all it will ever mean. The word has a clearly established definition which should not and will not be changed just for Wikipedia. Wikipedia does not have it's own newspeak. What your edit did in effect was suggest that the combination of any two sources anywhere within an article, even in separate sections, was prohibited. That is why the qualifier "that serves to advance a position" is added to the end of any mention of synthesis, because every article on Wikipedia is a synthesis. -- Y|yukichigai (ramble argue check) 09:05, 7 January 2008 (UTC)
Okay, I made a change to reflect your point. Be that as it may, the prior version, "Content should not be synthesized to advance a position" was unaceptable, as it prohibitted any synthesis advancing a position. A synthesis advancing a position is of course perfectly acceptable as long as it is the position of a verifiable source! Slrubenstein | Talk 11:25, 7 January 2008 (UTC)
Not exactly. A synthesis in a Wikipedia article to advance a point is not acceptable, regardless of whether that point is verifiable. We can of course report on a published synthesis. For example, a mathematical proof should be included if the proof is published, but not if it is the original work of the editor. Dhaluza (talk) 12:16, 7 January 2008 (UTC)
Insofar as I can tell the current revision of the nutshell says just that. -- Y|yukichigai (ramble argue check) 13:22, 7 January 2008 (UTC)
To put this in easy to understand mathematical forms...
  • A<citation> + B<citation> = C ... where C is an editor's conclusion
Is a clearly form of OR and is not acceptable.
  • A<citation> + B<citation> = C<citation> ... where the citations are to different sources
Might be a form of OR and might not be... it depends on whether C draws it's conclusion directly from A and B.
  • (A + B = C)<citation>
is clearly not a form of OR and is acceptable. Blueboar (talk) 14:28, 7 January 2008 (UTC)
I would add one caveat to the above. It might be possible for someone to misrepresenta given author as indicating conclusion "C" is accurate or indicated, when in fact no such direct statement was made by the author, who simply put forward evidence "A" and "B", then I would think that such a statement should be ruled out as a synthesis. The difficulty in this case is knowing whether there might not be additional evidence "D" put forward by the author which would make conclusion "C" contrary to his evidence. I am the first to acknowledge that it is difficult to imagine a situation in which such might arise, but in the some fields it might arise. John Carter (talk) 14:59, 7 January 2008 (UTC)
Yes. I think the intention of the "equations" was that the source explicitly concludes C from the facts A and B. Otherwise, what we'd have was covered previously, i.e., A<citation> + B<citation> = C, with no citation for C, and the citation for A and the citation for B are the same source. --Abd (talk) 15:35, 7 January 2008 (UTC)
This looks a bit like a backdoor policy change, and needs more discussion outside the talk page for the policy. --John Nagle (talk) 16:57, 7 January 2008 (UTC)

Is PSTS too complex?

SlimVirgin wrote "A primary source is one close to, and with direct knowledge, of an issue. In contrast, a secondary source is at least one step removed, and provides a secondhand account or a general overview." That contains a few confusing points. Is a source a person or a thing? This says sources are people, because only people have knowledge. Wikipedia editors can't use their own knowledge, nor knowledge from interviews, we can only use inanimate, verifiable sources. So the reader of this definition has to make a mental translation between this definition, which is about animate sources, to the book and journals they are reading, and mentally reword the definition so it applies to things, and not people.
Another confusing point is secondhand account; I think what is intended is that an author, who has no firsthand knowledge of a subject, viewed or listened to media, or conducted interviews, became knowledgeable about a subject, wrote something, and had it published someplace other than Wikipedia. There's a lot of places where a reader's interpretation of this could go astray; who is it that has no firsthand knowlege, witnesses who were interviewed? the author? the publisher? the Wikipedia editor? --Gerry Ashton (talk) 05:33, 5 January 2008 (UTC)
Books contain knowledge. There is no need to get excited about the difference between the author and the products of the author. --SmokeyJoe (talk) 08:44, 5 January 2008 (UTC)
Sources are either people or their products. For Wikipedia, everything has to be published, so a source for us is mostly the text. But we do think of sources as people too, because sometimes the authority of a text is bound to the authority of the person; for example, when we use a blog. The "mental translations" will therefore have to continue, Gerry, but at least we can rest assured that we're helping our readers stave off Alzheimer's. :-) SlimVirgin (talk)(contribs) 12:26, 5 January 2008 (UTC)
SlimVirgin, you ask where is the complexity. Well the problem with the current method of source-typing arises from the fact that the definitions of primary and secondary sources are not exclusive and they do not occupy the same plane. I think we'd all agree that ideally there would be a sharp dividing line between the two definitions (piano keys are either black or white). Failing that, a continuum from one extreme to the other would be acceptable, because when a source is placed on that plane it's then a matter of reaching consensus as to which category it falls in (is that shade of grey "dark" or "light"?). However, what we have is a state where there is no continuum between the primary and secondary definitions to negotiate (that cloud is grey, but what shade of orange is it?).
Look at the current PSTS definitions (as they are as I write this):
  • Primary sources are sources very close to the origin of a particular topic.
  • Secondary sources are accounts at least one step removed from an event.
See what I mean? The definition of "primary" talks about closeness, origins and topics. But the definition of "secondary" is concerned with accounts, steps and events. There's no sharp dividing line, nor is there a continuum to negotiate.
To just take one element, how far away from the origin does a primary source have to be before it stops being primary - and, most importantly, what does it become then? Another: is the "event" the same thing as the "origin"? What is the "event" in a scholarly book about carrots? I suggest that you only think the definitions are not complex because you have had long experience of using them as tools. But new WP editors won't have this and there is considerable potential for confusing them.
Finally here's a test. Pretend you are a new WP editor, and using just the above definitions and none of your experience of using them, source-type these two cases (the article is about the accident):
1. A newspaper item about an accident that happened outside the newspaper office just an hour or so before going to press.
2. An eyewitness account of the accident written down ten years after the event.
 —SMALLJIM  13:36, 5 January 2008 (UTC). Revised 14:49, 5 January 2008 (UTC)
The newbie editor unfamiliar with the penchant for instruction creep is probably going to run away screaming.
But the newbies who stay - very likely because they has an agenda to publish - are not going to care until someone hits them over the head. In which case they will argue that "SV said" 1. is a primary source, and 2. is a secondary source. And they'd be formally correct.
But even per the ideas behind SV's notion, if 2. cites 1., then 2. becomes a secondary source. Thats normative horsepuckey. (continued in next outdented comment)


There is no such thing as a pre-definition of "primary" or "secondary" as a whole. To return to SV's construct of the fork-dropping: If her publication is being cited for the fork-drop observation, then the citing statement is using her publication as a primary source. If her publication is being cited for her debunking of the previous study by ForkedUpInc, the citing statement is using her publication as a secondary source. The only time pre-definition works is when a publication is pure fiction/historiographic account that refers to no preceding material.
That Plato cites Aristotle doesn't cause Plato to be defined as secondary source. Only a statement that cites Plato on Aristotle is using Plato as a secondary source. Use ... As.
As the {{Notable Citation}} banner at the top of this very page demonstrates, even the assumption that "Wikipedia is not a primary source" is patently false.
-- Fullstop (talk) 15:59, 5 January 2008 (UTC)
I think we can safely assume that in practice whenever we need to decide which type a source is, we already have a particular topic/context in mind. So… are you agreeing with my assertion that the current definitions of primary and secondary are severely flawed?  —SMALLJIM  17:14, 5 January 2008 (UTC)
Actually, the majority of the editors I have to deal with don't have their sources before hand. They know what they want to say, and look for sources to support it.
But ideally the editor will already have a particular context in mind. But then, can anyone read minds?
And yes, the definitions are severely flawed. They will also forever remain flawed because the premise is flawed.
Thanks to gbooks, OR is predominantly from so-called "secondary" sources.
-- Fullstop (talk) 19:13, 5 January 2008 (UTC)
To reply to your first point, if the editor already knows what they want to say, they must already be aware of the topic/context, so when they do find a source, 'typing' it will not involve those complex considerations you listed. But that point does raise an interesting idea that could usefully be included in some best practice guideline: "Don't write anything until you have at least seen a reliable source for it". Thanks for the rest of your reply.  —SMALLJIM  20:31, 5 January 2008 (UTC)
Agree that the concept of source typing for identifying OR is fatally flawed. Regardless of whether the definitions can ever be improved to a point where they are easily understandable without "specialist knowledge" (an ironic self-contraindication from this policy) the question remains as to it's relevance to NOR. Can we fully describe what NOR is without using defined terms like this? I submit that we can, and we should, and I have made a proposal to that effect for a streamlined replacement for PSTS below. Dhaluza (talk) 20:13, 5 January 2008 (UTC)
Ah, I think you've replied to a question I didn't ask! As you probably know from yesterday's comments here, I agree with you that source-typing is of little use in identifying OR - I've had a quick look at your proposal, and will comment. In case you haven't seen it, I've already made a copy of WP:NOR with the PSTS section excised, and it doesn't appear to be fatally damaged by the loss. However we do know that source-typing is of use in some circumstances, so can I ask your opinion on whether the current definitions of primary and secondary sources are flawed (per my long reply to SV above)?  —SMALLJIM  20:57, 5 January 2008 (UTC)
Actually I was responding to Fullstop, but to answer your question, the problem with the PSTS definitions here is they are trying to be all-encompassing, and the fallacy of that is reflected in their complexity and persistent ambiguity. For example, the internal definition of primary source conflates factual, original (i.e. first-hand) and close (i.e. self-published) sources, which are different concepts that may not fit together in different contexts. We have also seen that the external definitions are different in the hard sciences, information science, and the humanities. So trying to reconcile all these differences shows the PSTS concept fails in the context of NOR (whether it is useful in other contexts is another matter). The bottom line, as you point out, is that PSTS is not essential to NOR. Dhaluza (talk) 21:10, 5 January 2008 (UTC)
Apologies, I misread the indentation - thank you for replying. I think I can take it that your reply indicates your general agreement with me - I'm sure you'll say if this is not the case. However, to take up your point about the futility of attempting to create all-encompassing PSTS definitions, I don't think there is any great difficulty in creating all-Wikipedia-encompassing primary and secondary definitions. This can be done if we abandon our reliance on those external definitions, which were, I believe, created for specific, limited, academic purposes and are thus ill-equipped to deal with our need to cover the whole of human knowledge.  —SMALLJIM  00:09, 6 January 2008 (UTC)
As I've said before, and other have too, the Law of primacy is a problem with redefining terms in existing use. So, even if it were possible to come up with an all-encompassing definition for WP purposes (and this is a very tall order), people will have difficulty separating that definition from their already familiar one, and this will only cause confusion and misunderstanding. There also is a widely expressed objection to creating neologisms for a new definition as well. So, the best thing is to just describe what it is in plain language, and avoid jargon altogether. Dhaluza (talk) 00:21, 6 January 2008 (UTC)

(< outdent) I've answered the objection about primacy before; basically, only those with an academic background will be aware of the existing external defs and such people (1) are in the minority on WP, and (2) should have the mental capacity to easily assimilate the change. For the "very tall order", please see this. Finally I agree that creating neologisms is generally to be avoided, but sometimes the cost/benefit ratio tilts towards doing so; spending hundreds of hours discussing issues to which there can be no simple solution is a good example - look at Wikipedia talk:Evaluating sources for instance.  —SMALLJIM  13:03, 6 January 2008 (UTC)

I'm not sure that the primacy is limited to academics--many educated editors will have been exposed to this in some form. Notwithstanding this, looking at your definitions, Type 1a calls the "thing" the source, but we discourage using the editor's observations of the thing--we generally want to see a published observation as a primary source. Type 1b includes first-hand opinions, but we generally treat these as secondary sources when they are expert opinions based on other people's primary source observations. I liked your type designations because they are a reasonable compromise between redefining terms and creating neologisms. But at the end of the day, what is the purpose of source typing in relationship to NOR? Isn't NOR primarily related to how the source is used on WP, not how it was originally written? Dhaluza (talk) 13:26, 6 January 2008 (UTC)
To take your last point first: yes, we've already agreed that this has little to do with NOR - I need to clarify in the lede that this is intended as a separate essay or guideline: "PS2".
Regarding the rest, you and I have actually been here before (search Archive 32 for "type 2"), and in this edit[9] you apparently approved of the type-1a and 1b distinction (factual sources vs close sources). Type-1a is a necessary addition to cover things like archaeological artefacts, which although being mentioned as an example of a primary source in the existing PS definition, are not actually covered by the definition itself (unless "very close to" includes "that are" - see the WP:PSTS definition).
It's important to note that there are no statements in my PS2 definitions about the reliability or preference of use of any of the source types. As a guideline, PS2 would refer to WP:RS to provide such guidance, though I have included brief "Using sources" notes (these are subject to adjustment).
Lastly, you say this: "Type 1b includes first-hand opinions, but we generally treat these as secondary sources when they are expert opinions based on other people's primary source observations." Well, in my model, any opinions that are based on other people's PS observations are secondary sources because they (paraphrasing) "discuss at some length other people's thoughts and statements about the topic." Note that the first rule says "always take the last option that fits..." Hope this is clear.  —SMALLJIM  20:50, 6 January 2008 (UTC)
My point is that we also value expert opinion that is based on evaluation of the evidence itself, not simply other people's testimony. So for example, a historian who makes a judgment based on evaluation of the historical evidence should be considered a secondary source for that opinion, but would not under this construct, while a non-expert who gathers expert testimony would be. Dhaluza (talk) 10:49, 7 January 2008 (UTC)
I don't think you've really given me enough information there to comment without making some assumptions, Dhaluza. So I'll make them, and you can tell me if they're wrong - hope that's OK.
I'll assume that your historian has written an article about the start of WWII in a reliable journal, and has cited all her sources; some of which were written at the time of the event, and some are later analyses by historians (both these are "thoughts and statements about the topic"). Based on that evidence she has come up with a new opinion on how the war started. If that's a fair summary of your example, then my model would state that her article is clearly a secondary (type 2) source for the start of WWII. It fits the definition perfectly: "discusses at some length other people's thoughts about the topic and the statements they made about its properties". I'm not sure what the non-expert is doing with the expert testimony that he gathers, perhaps you could explain what you intended there.  —SMALLJIM  18:55, 7 January 2008 (UTC)
I was thinking of the much more usual case we run into all the time in the many, more pedestrian, articles. For example, say we want to explain why the sky appears to be blue. We might use a source written by a qualified physicist, and paraphrase his/her expert opinion. Different physicists will give slightly different explanations, so each is unique and original to each author, even if they all say basically the same thing. So under a strict reading of your criteria, each expert's own opinion would be a primary source. But because these are interpretive statements, we need to use a secondary source to satisfy NOR. Now obviously for the sky is blue case, we may have such a large body of sources that we can find sources that talk specifically about other sources, or even tertiary sources that already summarize the topic. But for many of the more obscure subjects, we may not have such an embarrassment of riches. Dhaluza (talk) 02:56, 8 January 2008 (UTC)
Hmm. I think you're arguing that my model gives the "wrong" result in this case, but doesn't it give the same result as the existing PSTS? Isn't this similar to the issue regarding journal articles that was discussed last month? Both models probably need an exception clause to deal properly with that type of source.  —SMALLJIM  16:25, 8 January 2008 (UTC)

"The sky is blue" -- do I need a source?

An important point seems to be missing from this policy article : where is the line between self-evident facts and original research? If I write "George Washington was a male", do I need a source? Please clarify in the article. Emmanuelm (talk) 18:59, 8 January 2008 (UTC)

You can ignore this rule if its application would lead to a rediculous result, as with all rules. Someguy1221 (talk) 19:07, 8 January 2008 (UTC)
If the facts are truly self-evident, it's unlikely they'll be challenged, so a source may not be necessary. If a statement is challenged, the facts may not be self-evident to everyone. In this case it should be easy to find a source for your statement. As for the "sky is blue" example, this is particularly relevant and interesting reading. Chaz Beckett 19:49, 8 January 2008 (UTC)
Thanks, you answered my question. I still think there should be a paragraph in the article explaining this. Emmanuelm (talk) 16:02, 9 January 2008 (UTC)
See WP:V. .. dave souza, talk 18:05, 9 January 2008 (UTC)
I looked at WP:V and still cannot find a clear statement about obvious or self-evident facts. Could someone please write a paragraph to clarify this? I suggest this title : "The sky is blue" -- sourcing common knowledge. Emmanuelm (talk) 16:17, 10 January 2008 (UTC)
Keep in mind that WP:V requires that facts be verifiable and not verified. Anything that is truly self-evident should have a wealth of sources to support it, and I doubt that anyone would, in good faith, argue that no source exists to support that the sky is blue. Someguy1221 (talk) 19:25, 10 January 2008 (UTC)
Which part of "Material challenged or likely to be challenged, and all quotations, must be attributed to a reliable, published source." didn't you read? If it's obvious or self-evident it self-evidently is unlikely to be challenged. Unlike "the sky is blue", which it clearly wasn't here today. A dirty grey earlier, black now. Dunno where you live. .. dave souza, talk 16:59, 10 January 2008 (UTC)
In any event, the mere fact that Wikipedia is not a bureaocracy absolves us of the need to riddle policy pages with disclaimers to eliminate every possible technicality or loophole. While I am at times disappointed to see that some people take the rules too literally, it's generally sufficient in such a dispute to point out that a rule should be ignored when its application fails to accomplish its goal. Someguy1221 (talk) 19:20, 10 January 2008 (UTC)

Facts must be backed by citations

The nutshell currently contains “This page in a nutshell:…Facts must be backed by citations to reliable sources that contain those facts.” Could someone point to where we say this on the page? Sorry, but I don’t see it anyplace. Nor do I see “nor a forum for promoting one's own point of view.” except for under “Related policies”. I do not see where we articulate that as part of WP:NOR; it is clearly WP:NPOV and should not be listed as part of “This page in a nutshell:…” as it is not part of THIS page. Brimba (talk) 15:23, 9 January 2008 (UTC)

The citations thing is inherent in the concept of NOR, so I don't really see a problem with that -- the only way to show you're not engaged in OR is to provide a source. I think the POV part is simply to emphasize that the analyses and syntheses of Wikpedians trying to advance their own POV isn't welcome. SlimVirgin (talk)(contribs) 15:45, 9 January 2008 (UTC)
I think 'backed by citations' goes back to the hoary days when the WP:V page didn't yet exist. It is a determinant of 'original research' in the real world.
'nor a forum for promoting one's own point of view' is not at all related to 'original research'. "not a forum" and soapboxing are covered at WP:NOT.
-- Fullstop (talk) 15:56, 9 January 2008 (UTC)
We are saying ALL facts must be cited then, and not just those that have been challenged? This would be a major change in policy. “I think the POV part is simply to emphasize that the analyses and syntheses of Wikpedians trying to advance their own POV isn't welcome.” While I don’t disagree with the logic, that should be in the lead, not the nutshell. “Nutshell entries must be brief overviews, with a very high signal to noise ratio.” -the nutshell should remain a quick-reference; and not act as a scaled down lead. The lead as written is taking on the form of Wikispeak; it should be in plain simple English, not something where the new user has to vett every word as he/she reads it. It’s written by active editors FOR active editors. It should be written by active editors for new users in language they will understand at first glance. Active editors will take the time to read the lead, and maybe the policy in detail, new user will just be turned off and ignore it all. The emphasis becomes noise. If I take off my glasses the nutshell just becomes a blur, if I can take off my glasses and it still breaks down into words then we are on the right track (actually I have read it so many times now that this test no longer works). Brimba (talk) 16:18, 9 January 2008 (UTC)
I guess in short, the question is, are we willing to give up some usability for some exactness, or should the nutshell emphasize usability over absolute technical correctness, not to mention bells and whistles. Brimba (talk) 16:25, 9 January 2008 (UTC)
Hows, whys and wherefores don't belong in a nutshell. They don't even belong in a lead.
The nutshell may just as well (should) simply read "Thou shalt not misrepresent sources" or "Statements must represent sources with fidelity." (or some such).
Something that reads "Statements must represent sources with fidelity" already has 'sourcing' as a prerequisite. It does not need to explain that statements must be sourced (which is a WP:V issue anyway). It then does not have to bend over backwards to explain just what-kind-of-source-under-what-circumstance may or may not be subject to this core policy.
And nobody thinks they are inappropriately "advancing a position." The way editors (here) are nitpicking over superfluous clauses, one would think that they themselves want a built-in escape hatch. I can almost hear the conceit: "Gee, I don't have biases. It couldn't possibly apply to me!" -- Fullstop (talk) 17:32, 9 January 2008 (UTC)
This seems to be an artefact of trying to merge RS into V, and would make more sense by changing "backed by" to "attributable" –
Facts must be attributable to reliable sources that contain those facts.
Given there is the merger proposal and V is policy while RS is a guideline, there's a good case for simplifying it as –
Facts must be attributable to reliable sources that contain those facts.
That then makes it consistent with "Any interpretation, and any synthesis which advances a position, must be attributable to reliable sources...." Fullstop's right that we must encourage editors to avoid involuntary OR, and this phrasing meets that point. By the way, the sky sure isn't blue today. . . dave souza, talk 18:03, 9 January 2008 (UTC)
It may as well read "Any interpretation, and any synthesis which advances a position, must be attributable to reliable sources...."
An interpretation/synthesis is always advancing some position or the other. Why otherwise would the author have made it? -- Fullstop (talk) 20:03, 9 January 2008 (UTC)

Slim's current version looks really good to me. Note that "reliable, published source" is piped – it's not just RS, the whole requirement of V comes in. .. dave souza, talk 16:54, 10 January 2008 (UTC)

Thats fine. A bit convoluted perhaps, but sane. -- Fullstop (talk) 20:24, 10 January 2008 (UTC)
Two bullets are better than three. I further reduced the word count a bit as well. Dhaluza (talk) 19:58, 12 January 2008 (UTC)

Google Earth

Does looking at something on Google Earth and using a description of what was found constitute original research?

I've found a few articles, mostly stubs about various minor airfields, where the author looks on Google Earth and reports on what state it appears to be in... i.e. whether or not it's active based on finding or not finding aircraft on the tarmac, whether or not it's in a state of disrepair, etc.

I think using Google Earth in such a manner does constitute original research; Google Earth only provides an image and it's up to whoever's looking at it to decide what exactly it can mean. One could argue that it's a source and therefore could be used in such a way.

Jdkkp (talk) 18:26, 11 January 2008 (UTC)

I personally believe this to be original research, given the age of some of the photos on google earth and the ambiguity that can exist in interpreting them. LinaMishima (talk) 18:32, 11 January 2008 (UTC)
That's what I was thinking. There's plenty of other examples but the one that spurred the question was Burevestnik Airport. You can look at the airfield in Google Earth and draw conclusions, but that's exactly right, the photo could be several years old, particularly areas like that that are not particularly prominent. And it would be original research to go there and look at it and then say in the article, "I didn't see any airplanes there." Guess I'll remove that from the article. Jdkkp (talk) 18:38, 11 January 2008 (UTC)
Using Google Earth would have to be done very carefully to steer clear of OR. It would probably be OK to say that a satellite photo accessed through Google Earth showed X number of aircraft on the field, as that would be a descriptive statement easily verified without specialist knowledge. But drawing a conclusion that the airfield was active or inactive from that observation would be both questionable, and OR. For example, the visible aircraft could be junked and abandoned there, or invisible aircraft could be inside hangars. If the standard 'X' symbol on the runways were visible in the photo, it would be OK to say that the airfield was closed at the time of the photo, as that is easily verified as well. But that does not mean that the airfield is permanently abandoned, as it may have been closed only temporarily, or reactivad.
As for the Burevestnik Airport article, I dropped a fact tag on one item, and replaced the prior description of the imagery with one that should pass muster. I agree that identifying specific make/model of aircraft, whether aircraft apparently in one piece are airworthy or not, or speculating on the age of features shown is well outside the bounds of OR. But a simple description of the obvious features is OK, and the fact that a former interceptor base apparently has no interceptors any more has relevance. Dhaluza (talk) 18:48, 12 January 2008 (UTC)

Here's a fun example

Does reading the Producer credit on a TV show and then reporting that credit on the producer's article constitute OR? I don't think it does since there's no 'interpretation' of the primary-sourced data being done, but I'm running into an editor (here) who insists that it is. The second half of the questions is: if it's acceptable to note that the person produced the episode, and you provide the credit for every episode of the show made, is it OR to count the number of episodes the person produced and say "He produced ## episodes of Show"? Torc2 (talk) 00:58, 14 January 2008 (UTC)

The show is a published source, and reading a source is not OR. Dhaluza (talk) 01:08, 14 January 2008 (UTC)
Dhaluza is entirely correct, there is no OR in stating the contents of a primary source, and TV show credits are highly reliable. As for counting the episodes someone is credited in, it is technically weak synthesis, I believe, however it might well be allowable. LinaMishima (talk) 01:14, 14 January 2008 (UTC)
If the series is incomplete, I would rephrase as "he produced ## of the first ## episodes of Show" for the sake of durability, but I agree that no OR is occurring here. Limited synthesis is always okay, as long as you're not trying to imply some specious argument with it (like "he liked Show A more than Show B because he produced twice as many episodes of it"). Dcoetzee 09:02, 14 January 2008 (UTC)
Regarding synthesis, it is worth noting that most human languages rely upon weak synthesis in order to follow what it being said correctly. More importantly and controversially, researchers rarely publish material that is the product of weak synthesis, such as the trivial counting of something from sources, or observations that those trained in a field would find trivial to make but are deemed complex by those not similarly trained. LinaMishima (talk) 09:38, 14 January 2008 (UTC)

Thanks everybody. I think this issue is closed. Torc2 (talk) 09:16, 14 January 2008 (UTC)

Infinite verb forms

The best way to make a bad impression, in writing, is to replace a command ("stand! speak!") with an infinite-verb clause ("standing and speaking will be a good idea.") They are on opposite sides of the spectrum; one is forceful, the other apologetic. --VKokielov (talk) 00:21, 15 January 2008 (UTC)

I don't know how to fix it so it will be right. I don't know what tone you're aiming at here. As to me, I think the whole affair can be rewritten, because it is a rule, an imperative. But that is only my opinion. --VKokielov (talk) 00:32, 15 January 2008 (UTC)
Here is a way, starting from my attempt:
To prove that you are not presenting original research, please cite reliable sources which contain relevant information, supported as it is presented and pertinent to the subject matter at hand. --VKokielov (talk) 00:34, 15 January 2008 (UTC)
Yes, commands are stronger, but there's no reason to assume editors will disobey them. We can be persuasive without being authoritarian. Torc2 (talk) 00:41, 15 January 2008 (UTC)
The tone we're aiming for is not one of command, VK. SlimVirgin (talk)(contribs) 00:55, 15 January 2008 (UTC)
If you are issuing a command, show it for what it is. The rules of honest discourse demand it. It is a minor point, but it's important, because to leave it the way it is gives the whole thing a legalistic feel. "Authoritarian", I think, is a bad word to relate to writing, because all good writing is authoritarian; everything else wiggles and squirms. (That used to be an aesthetic principle, by the way, and it is still a principle of human perception). --VKokielov (talk) 02:39, 15 January 2008 (UTC)
Policy pages on Wikipedia often wiggle and squirm, unfortunately, because so many people edit them. Everything gets tweaked back and forth a thousand times until any semblance of good writing has disappeared. :-) SlimVirgin (talk)(contribs) 02:47, 15 January 2008 (UTC)
Understood. Like I said, I'm only a disciple. Editing pages under one hundred thousand eyes takes gusto and a certain subtlety that you get only through experience. --VKokielov (talk) 02:52, 15 January 2008 (UTC)

Is this OR? "A number of people have done this"

I'm having problems with an editor who says this line of mine here is OR:

A number of former Muslims have supported the publishing of the cartoons.

I'm telling him that I'm backing up this line and there are lines that come after it where former Muslims have supported something. Before you say its OR, look at what today's featured article also says:

Trembling Before G-d has had a wide impact especially within the Orthodox Jewish world, where the reception has been roundly negative.

This is also not sourced but it isnt OR, is it? How do I argue that what I'm doing is not OR? The editor wants that line to be referenced from a source. I cant find any source that says that, yet, it is obvious from the following 3 lines that indeed, "a number of former Muslims have supported the cartoons". How do I deal with this? --Matt57 (talkcontribs) 17:25, 12 January 2008 (UTC)

The problem is how Matt57 is not saying "A number of former Muslims have supported the publishing of the cartoons", but the problem is how Matt57 is backing up this line. There are sources that person X has supported the cartoons. There are also different sources that say person X is a former Muslim. But there are no sources that say both of those things. For example:
Matt57 is using two sources to say "former Muslims have supported the publishing of the cartoons". Is this not a vio of WP:SYNTH?Bless sins (talk) 17:42, 12 January 2008 (UTC)
No, that is a logical deduction. But saying "a number" if the refs only support a number = 1 is at best WP:WEASEL. You could say "at least one", but the relevance of a statement like that would be questionable at best. At least one person can probably be found to support almost anything. One is the loneliest number.... 19:02, 12 January 2008 (UTC) —Preceding unsigned comment added by Dhaluza (talkcontribs)
The question is not of one. Other references have been/can be provided. Secondly, I find the Smith and Jones example here to be logically deducible.Bless sins (talk) 19:20, 12 January 2008 (UTC)
More specifically, it is a one-step logical deduction that requires no assumptions. The problem with characterizing that as a logical deduction is that it involves assumptions and intermediate steps, which is what makes it a synthesis to advance a position. Since there are several former Muslims commenting on the subject, it seems to be reasonable editorial discretion to group their comments together under that heading as was done. This is precisely the type of synthesis we need in articles to make them readable. This is not synthesis advancing the editor's point, it is reporting the references' points of view in juxtaposition with related information. Dhaluza (talk) 20:11, 12 January 2008 (UTC)
It might be better to just be specific and mention that Hirsi Ali supports the cartoons. One of the links is a basic fact about Hirsi Ali, so I don't think it's a SYNTH issue. The notion of Hirsi Ali being ex-Muslim cannot be disputed really, so a cite pointing that out may not even be needed in that article (WP:CITE). It's like saying "Watt, a professor of Islamic studies, said X" - everyone knows Watt was a professor of Islamic studies, so there's no need for a supporting ref in that case. ITAQALLAH 19:08, 12 January 2008 (UTC)
Ok, got it. Everyone knows that Hirsi Ali is a woman, so I can say "some women, like Hirsi Ali, have supported the cartoons." Everyone knows that Hirsi is a politician, so I can say "some politicians, like Hirsi Ali, have supported the cartoons." Everyone knows that Hirsi ali is Dutch, so...This opens up a lot of possibilities.
Secondly, is the argument that "everyone knows Watt was a professor of Islamic studies" appropriate? Where do we draw the line on "everyone knows"?
Finally isn't it against the spirit of WP:NOR that wikipedia is the only source on earth that draws a connection between former Muslims and opinions on the cartoons? If there were reliable sources to support the claim, then I'd agree to its inclusion.Bless sins (talk) 19:20, 12 January 2008 (UTC)
Re: your first point, I don't agree. It depends upon what is most relevant in the context of the discussion. It might be relevant to introduce her as a dutch politician, given the political side of the cartoons debate, or it might be relevant to just say that she's a critic.
Re: your second point, I'm just talking about how we introduce people in a factual and neutral manner. Other cases are dealt with on their own merits.
Re: your last point... as I said above it might be better to simply specify Hirsi Ali. It can also always be appropriately reworded if there are multiple sources so as to avoid the issue of potential OR or weasel wording (i.e. Critics such as X and Y support the cartoons because... ). ITAQALLAH 19:33, 12 January 2008 (UTC)
It would not be WP:OR to collect and organize existing well sourced and undisputed information that directly pertains to the subject of an article. If we know that X and Y say A (that pertains to the article), and both X and Y belong to group G, we can say that some people from group G say A, and give X and Y as examples. Crum375 (talk) 20:33, 12 January 2008 (UTC)
Please see my comment on "19:20, 12 January 2008". Former Muslims isn't the only category, 'women', 'dutch', 'politician', 'immigrant' are all categories that Hirsi belongs to.Bless sins (talk) 22:36, 12 January 2008 (UTC)
Re:Itaqallah. How do we determine "what is most relevant in the context of the discussion"? The controversy was not just religious, it got political, it is alleged that Muslim immigrants to Denmark had a bit to do with inciting it etc. There are several relevant categories here. None of which are supported by reliable sources.
Once again, the last point: isn't it against the spirit of WP:NOR that wikipedia is the only source on earth that draws a connection between former Muslims and opinions on the cartoons? No source in the entire world does this - if there was another source, I'd gladly accept it, and there would be no dispute.Bless sins (talk) 22:36, 12 January 2008 (UTC)
I have decided to take out the sentence because it violates WP:WEASEL as [pointed out by Dhaluza. WP:WEASEL discourages usage of phrases like "some people have said". This is equivalent to "a number of former muslims have said/supported...". Fine, so I took it out. However I still dont know why a featured article would have the same kind of WEASEL wording or, how this is not weasel:
Trembling Before G-d has had a wide impact especially within the Orthodox Jewish world, where the reception has been roundly negative.
I found another way of saying it, which is to name the people in the first sentence. This is not weasel however its not the best:
Former Muslims like Ayaan Hirsi Ali, Ibn Warraq, Wafa Sultan have supported the publication of these cartoons.
So whats the best way of summing up the whole paragraph without it being a WEASEL issue? This summing up is done everywhere (e.g. the featured article) and as Dhaluza said, we need this kind of synthesis to make the articles readable. --Matt57 (talkcontribs) 22:37, 12 January 2008 (UTC)
Yes, it's better to name several examples, but the question remains as to what the relevance of "former" Muslims is. I think that's the point that's being challenged. Have any of them explained -- "I'm not a Muslim anymore, and I support publishing the cartoons because ..." or has any secondary sources said that lots of former Muslims support it? I'm not saying it's necessarily OR, just that I don't get the connection myself. SlimVirgin (talk)(contribs) 01:02, 13 January 2008 (UTC)
It doesn't need any special significance other than being a common factor that can help organize the information. Since several people who self-identify as former Muslims have commented, we can group their comments on this basis. This may or may not be the optimal organization of the material, but it is a reasonable one. We do not need an outside source that organizes the information in this way, and we don't need the sources to specifically link their comments to their self-identification. Also the fact that the sources are former Muslims gives them a different viewpoint from people who are or who never have been Muslim, so there is a logical, as well as an editorial basis for this. Dhaluza (talk) 19:01, 13 January 2008 (UTC)
What would the editorial basis for it be? That's what I don't get, especially as they were given their own section and were therefore being actively categorized as ex-Muslims, rather than it being mentioned in passing. People who have never been Muslims aren't identified as such (at least I hope they're not), even though that would be just as "logical," given that their lack of involvement with the religion would play a role in how they see things. SlimVirgin (talk)(contribs) 19:05, 13 January 2008 (UTC)
Simple organization. Imagine you are preparing a Powerpoint presentation. You need to divide the material into slides, and the slides into bullets. You need to organize the material so it logically fits together and flows. Each slide gets a title. Sometimes the titles don't have any special significance, other than simple organization. Dhaluza (talk) 21:03, 13 January 2008 (UTC)
Thanks Dhaluza for saying it well. SV, there's a number of different things: these people were asked specially what they thought of the cartoons so much so that they spent quite some time on talking about the thing (Wafa Sultan's video here e.g.) and these single individuals were being interviewed about it. A group of 13 went and signed a manifesto in response to the violence. Given that the affair is religious in nature, people want to know what different religious groups, Muslims and former Muslims want to think of the affair. As to why we should group them together and not just disperse them like anyone else, well we have sections explaining what Muslims thought of the affair. The voice of former Muslims also deserves its own place. Its an important voice of dissent. Thats why news sources went after them to get their opinions. --Matt57 (talkcontribs) 01:08, 14 January 2008 (UTC)
Grouping according to undisputed categories is not WP:OR per se. However, selecting specific groups could violate WP:NPOV or WP:UNDUE, in principle. If there is already a source that tells us that group X said Y, then we can report it that way. But if, for example, we only have individual statements from people about a controversial subject, and we categorize the statements ourselves, we could be introducing subtle bias and violating WP:UNDUE by selecting only those specific groups and not others. Crum375 (talk) 01:19, 14 January 2008 (UTC)
Yes, but there the problem is in being selective, not in organization. I agree that using suggestive, biased, or derogatory headings would be inappropriate. But in the present case, the sources self-identify with the group, so that does not seem to be an issue. Converting from a religion is not all that unusual. Dhaluza (talk) 12:03, 14 January 2008 (UTC)
The reporters who interviewed the former Muslims did not do so arbitrarily. They felt that the perspective provided by former Muslims was notable, we can source it, and it certainly makes good sense in reporting to include that, as well. The perspective of those who would have once been highly likely to agree, but having left the group now may have other views certainly seems like a good 'insider's view' tack to take on the subject. The only debate I see is, would former Muslims have any axe to grind against the religion, making their comments more sour grapes than honest analysis, and that's up to the editorial discretion of the publishing body, not us. I think the inclusion offers a unique perspective, one which might not be considered otherwise by the casual reader, and which does offer an expanded view of the topic. Since it can be sourced, the only question is, as Matt57 presents, how best to phrase it to fit NPOV? ThuranX (talk) 00:38, 16 January 2008 (UTC)
As I noted above, organization of undisputed information (e.g. self admitted "ex Muslims" said X about the article's subject) is acceptable, in principle. The problem is selection — if there were "ex Hindus" who said something, should we include them as well? Aren't there near infinite such possibilities? And how do we decide relevance and notability without introducing POV or violating UNDUE? IMO we need a source that actually says "A number of ex Muslims said X" — we can't just group them on our own, as their inclusion, while excluding other possible groups, can be viewed as POV. If the reporters interviewed these people because they specifically wanted to get opinions from ex Muslims, then show that the reporters said that, and we can then include it. Crum375 (talk) 01:01, 16 January 2008 (UTC)

(Undent)No, I understand, and I thought I was making it fairly clear that the reporters whose articles we use as secondary sources already found that line, and we can work form what they've put out there. Do you really think the perspective of ex-hindi is particularly relevant to the Muslim reaction to the cartoons? I don't, and neither did any reporter, as I've never seen such an article. But to find people who once thought like the objectors, and now do not, and to go to them and ask 'As former insiders to this thinking, what say you?" is, as I said before, a logical angle on the story. And finally, as I said before, we're not choosing an arbitrary threshold of notability of perspectives on the thing, we're taking our cues from the members of the Fourth Estate. Most of the editors of wikipedia aren't professional writers of any ilk, be it comedy, fiction, technical, or journalistic. But the producers of most of our secondary sources are. They're far better at finding interesting, yet relevant approaches to stories. We don't have to find new angles, just cover what's been covered. Clearly, since there's more than one 'ex muslims' article, and zero 'ex-hindi' articles, the notability of the minority viewpoint as relevant has been established. Since there aren't any ex-hindi articles, i think that alone establishes the notability of such view - there is no notability. ThuranX (talk) 05:01, 16 January 2008 (UTC)

Proposal: Make explicit when sourced lists NOT WP:OR

WP:NPOV_tutorial#Moral_and_political_points_of_view says:

On certain topics, there is naturally less "expertise" and scientific thinking, and more "opinion". This is especially the case of topics such as morals or religion, based on faith, as well as politics.
We should then list all points of views, according to their importance, and, if possible, be precise as to who holds them...

I have noticed that some articles will make a statement like "these are the views of this group" or "there are several views" and present a relevant and sourced list and no problem. Other pages you try to do the same thing and people delete the comments, saying WP:OR or WP:SYN, thought it seems to be it's really usually because of WP:Idontlikeit. On the other hand I ran into an example today that LOOKED ok to me, except it needed more sources, but I wasn't sure and just brought up the topic for discussion.

Therefore I think an explicit statement about this lists issue and explanation of and link to WP:NPOV_tutorial#Moral_and_political_points_of_view is necessary somewhere on WP:OR, probably at Wikipedia:OR#Synthesis_of_published_material_serving_to_advance_a_position. It should say something like:

However, it is not synthesis to state that some people believe "A" and then list various sourced opinions "A". Also, per Moral and political points of view on certain topics, especially the case of topics such as morals or religion, based on faith, as well as politics, listing all points of views is often necessary and not original research or synthesis.

Comments? Carol Moore 05:16, 15 January 2008 (UTC)CarolMooreDC talk

Do you have any examples of what you think are good and bad articles in this regard, it will make understanding your point simpler, and so help people to know exactly what they are commenting on. Thanks. LinaMishima (talk) 17:00, 15 January 2008 (UTC)

I see a potential problem here with undue weight. Let me explain. On the Earth article do we note that some people believe the Earth is flat? Similarly on the Buddha article do we list all 36 forms of belief? I'm an inclusionist and so initially you might think I'd say yes. However, we should acknowledge that enormous articles become unwieldy. Therefore I'd suggest that if there really are a *large* number of alternative views, or alternatively a number of minority views, that a new article detailing those views be created and linked into the main article. That would then satisfy proponents of Undue who believe all the extra text just clutters, and proponents of inclusion who believe all voices should be heard. Wjhonson (talk) 21:44, 16 January 2008 (UTC)

Carol is trying to change the WP:NOR policy to get around WP:NEO; she is trying to do original research, finding a large series of primary sources that use a phrase, and then stringing them together to draw her own conclusions about them. Jayjg (talk) 03:40, 17 January 2008 (UTC)

Specialist knowledge

This policy does not prohibit editors with specialist knowledge from adding their knowledge to Wikipedia, but it does prohibit them from drawing on their personal knowledge without citing their sources

I think this is going to stop wikipedia from getting access to a lot of specialist knowledge, particularly in the physical sciences. There is a lot of specialist topics that live in the heads of only a few people on this globe and that then often in a rather scattered fashion. To write a good introductory piece for wikipedia that gives easy access to the topic for outsiders is inevitably somewhat of a creative act. If an expert is at all willing to do that they will often not have (or take) the time to find a suitable citation for every sentence they write. I think the most you can hope for is that they come up with a relevant text book or so. I think this needs to be relaxed and it probably can be relaxed as long as there is no opposition on the topic.

For the record, my changes to paramagnetism are almost exclusively drawn from my personal knowledge and trying to find quotes for everything I remember is really beyond my means. So: if anyone wants to revert it to its previous (rather abysmal) state: be my guest.

Jcwf (talk) 20:37, 9 January 2008 (UTC)

It could perhaps use a rewording. But to put your concerns at ease, that section is meant to discourage specialists from adding their personal interpretations of prior research, and findings they've made in their own unpublished and/or unreviewed research. And while ideally every claim made in an article is cited, the fundamental policy here, verifiability, requires that content be verifiable and not verified. Someguy1221 (talk) 19:35, 10 January 2008 (UTC)
I agree. The real policy governing this is WP:V. Things need to be verifiable, which means that you can and should include your specialized knowledge in Wikipedia pages, so long as nobody would disagree with you. For anything likely to be challenged, however, you need a citation. This is how things have worked for years on math and science pages, where there are a lot of conclusions that are simply not disputed by anybody. You can say E=MC2, for example, without actually having to cite Einstein's article. COGDEN 03:55, 11 January 2008 (UTC)
As long as what you are adding is verifiable in the literature, you will be fine. The spirit of NOR is not to prevent specialists from editing articles, it is to have a page to point to when we remove crank theories from articles. A sense of purpose and perspective is necessary. The best editors (specialist or not) of math and science articles do ensure that they don't add their personal interpretation, but stick to what is commonly accepted in their profession. The scientific citation guidelines give useful advice for sourcing science articles. — Carl (CBM · talk) 17:54, 17 January 2008 (UTC)

Nutshell grammar

SlimVirgin's revert[10] of Yukichigai's grammar correction in the nutshell is surely incorrect. When two singular nouns ("analysis" and "synthesis") are separated by "or", a singular verb should be used. See, for instance, OWL Online Writing Lab.  —SMALLJIM  23:31, 13 January 2008 (UTC)

Aren't you forgetting ideas, which is plural? Crum375 (talk) 01:04, 16 January 2008 (UTC)
"Ideas" is one of many subjective compliments of the sentence modifying the singular subject "analysis or synthesis". The plurality of "serve" is determined by the subject, not the subjective compliment(s). Yes, English is f%$#ed up like that, but that's how it is. -- Y|yukichigai (ramble argue check) 01:12, 16 January 2008 (UTC)
P.S. After reading the OWL link, I'd like to point out that this situation is specifically addressed by point number 5 on that page. -- Y|yukichigai (ramble argue check) 01:20, 16 January 2008 (UTC)
The simple solution is to replace "that serve(s) to advance" with "that advances".
But that sentence is perhaps altogether too wordy...
1. The "of published facts, arguments, or ideas" is superfluous because it doesn't matter what the analysis/synthesis is of.
=> Articles may not contain any new analysis or synthesis that advances a position not clearly advanced by the sources.
2. Analysis is a prerequisite of synthesis, so "analysis or synthesis" is also just "analysis":
=> Articles may not contain any new analysis that advances a position not clearly advanced by the sources.
3. Fix the temporal (new -> not already):
=> Articles may not contain analysis that advances a position not already clearly advanced by the sources.
Someone else please take it from here, i.e. resolve the double negative (not .. not => only) etc.
-- Fullstop (talk) 03:43, 16 January 2008 (UTC)
Just noting the first point, "that serve(s) to advance" should not be replaced with "that advances." OR enthusiasts are sneaky sometimes, and arrange information in a way that will lead to a conclusion, without explicitly advancing it (synthesis by juxtaposition of irrelevant facts). It's best if such a thing doesn't seem "OK" by the nutshell. Someguy1221 (talk) 03:48, 16 January 2008 (UTC)
I wonder if the phrase we want is "seeks to advance"?  —SMALLJIM  11:38, 16 January 2008 (UTC)
Same plural issue. Re Someguy1221's comment: Irrespective of whether its "serve(s) to advance," "seek(s) to advance" or "that advance(s)", a "sneaky" editor is going to deny allegations of misuse anyway. -- Fullstop (talk) 13:36, 16 January 2008 (UTC)
My point was not about plurality - that's settled: the verb should be singular. I was responding to Someguy1221's more subtle point which suggested that there is a difference between the meanings of "advances" and "serves to advance". I don't see much of a difference in this context, but there is a difference when "seeks to advance" is used - it emphasises the intention of the writer.  —SMALLJIM  19:36, 16 January 2008 (UTC)
Of course, but I think it's easier for the rest of us if he doesn't have a leg to stand on. One word means the distinction between "hoping readers reach an unsupported conclusion" and "handing readers an unsupported conclusion," and I just think it's better that the nutshell itself not leave room for doubt. Someguy1221 (talk) 19:24, 16 January 2008 (UTC)

I'm a little ambivalent, but I can certainly see a problem with "seeks to advance". Using good faith it might be hard to make an argument that we can sense the motivation of another editor. It seems to me like "seeks to advance" is calling into question, in a negative light, why the edit was made. Also it seems a little more slippery, i.e. lightweight. "Serves to advance" seems more firm and definite, while "seeks" seems a little more like a request to engage in long wars over what exactly what is being "sought". Just my opinion, sorry if that's sounds a bit confusing. Wjhonson (talk) 22:20, 16 January 2008 (UTC)

Good points. "Seeks to advance" would also cover the case where an inept editor tried but failed to advance a position - even though the failure would mean that there was no evidence of the attempt! We can't be sure what another was thinking, so you're right, "seeks to" doesn't belong here.  —SMALLJIM  09:54, 17 January 2008 (UTC)
People need to stop picking away at this. It was fine as it was -- singular or plural both fine -- plural sounded better. "Seeks to advance" speaks to intention; "serves to advance" doesn't. SlimVirgin (talk)(contribs) 10:03, 17 January 2008 (UTC)
Umm, which "plural" do you speak of? -- Y|yukichigai (ramble argue check) 11:41, 17 January 2008 (UTC)
The one you changed. SlimVirgin (talk)(contribs) 12:08, 17 January 2008 (UTC)
Well it was grammatically incorrect. Regardless of how it "sounds" it would be more than a little embarrassing to have the nutshell of one of our more important policy documents sporting a rather blatant grammar error. -- Y|yukichigai (ramble argue check) 12:55, 17 January 2008 (UTC)
Which sounds more natural to you (ignoring online grammar tutorials for the time being)?
"You shouldn't drive your car or ride your bike when their tires are flat."
"You shouldn't drive your car or ride your bike when its tires are flat."
SlimVirgin (talk)(contribs) 13:06, 17 January 2008 (UTC)
That's not quite an appropriate example. More appropriate would be...
"You shouldn't use your car or bike when their tires are flat."
"You shouldn't use your car or bike when its tires are flat."
Incidentally, the online grammar tutorial is from the English department at Purdue, not just some random website. -- Y|yukichigai (ramble argue check) 13:11, 17 January 2008 (UTC)
But you didn't answer the question. My point is that what sounds more natural depends on the sentence, and you don't learn that nuance from reading websites about grammar. More importantly, there's a lot of poor writing in this policy, because people won't allow any changes without picking at them a thousand times, so the least of our worries is the nutshell. It was fine then, and it's fine now. SlimVirgin (talk)(contribs) 13:15, 17 January 2008 (UTC)
What sounds natural to me is grammatically correct English. I'll admit for the more obscure bits of grammar it can pass unnoticed, but a plurality mixup isn't an obscure bit of grammar. It is, as I said, quite blatant, and obviously I wasn't the only person to notice. -- Y|yukichigai (ramble argue check) 13:22, 17 January 2008 (UTC)
Okay, well, this isn't the place to discuss it, but if you're a native English speaker, and if you have an ear for what constitutes good writing, you should ignore "the rules" as explained on websites (even those run by Purdue), and listen to what sounds good to your ear. :-) SlimVirgin (talk)(contribs)

I just don't get it...

(I must admit that I've not read this whole discussion page so I apologise if this has already been covered).

I'm new to Wikipedia but not to academia. So far as the latter is concerned, quite within the meaning of the phrase 'original research' are items that, from existing (perhaps well-established) sources, are collected and presented under a single cumulative article.

In agreement (I guess) with most Wikipedia editors, it seems to me that wikipedia articles aren't for the purpose of editors drawing original conclusions from a body of research (indeed this would be a common goal amongst encyclopaedia editors in general); yet surely this is different from ruling that original research should be avoided. By the academic measure (or in my opinion any regular way of regarding it), all well-sourced Wikipedia articles in fact represent original research. I mean to say, how exactly does one come up with citations without doing any research? Why the need to make a distinct (and what seems to me unusual) definition for Wikipedia?

I suppose it's too late to bring this up though. Brambinger (talk) 06:54, 16 January 2008 (UTC)

In Wikispeak you are confusing "research" with "original research". In this land, original research involves the creation of *new* information, not the repetition of old information, even merged, and paraphrase. If I write an article on geese and because of my own experiments conclude that geese love cheese, that would be original research. If however I read 12 books on geese and merge various bits and pieces of geese-lore together into a goose article, that is "research" but its not "original". All I'm doing to splicing together old information. Wjhonson (talk) 07:54, 16 January 2008 (UTC)
Oh yes, I understand the Wikipedia concept. I just think the term 'original research' is misleading, as original research, in some sense, is being conducted in the creation of any legitimate article. I know a man who contributed to Britannica, and indeed he would say he needed to conduct research to complete his articles. Whether or not the term 'original' changes the emphasis or is just redundant is probably a matter of taste. Anyway, I'm aware this discussion isn't going to result in any changes. —Preceding unsigned comment added by Brambinger (talkcontribs) 08:03, 16 January 2008 (UTC)
The original problem being that while you understand it, it was found that many people did not understand it. And so the perhaps redundant emphasis on the distinction that we're trying to draw. We all conduct research in order to write the articles here, as you said. We're just trying to seperate "research in published sources" from "research of things which aren't published or aren't sources", such as geese themselves. Wjhonson (talk) 09:42, 16 January 2008 (UTC)
Although buried under all the indirect speech/pitiful prose, what the policy is trying to say is "hey you, editor! Don't include your own original research."
The distinction lies with whether the product is an editor's own, or someone else's. To that end, in wikispeak, the former is "original research", while reference to the latter is "source-based research." -- Fullstop (talk) 13:51, 16 January 2008 (UTC)
Yes, don't include your own brand new thoughts. However with the addition, that you also cannot include anyone else's brand new unpublished thoughts. Just because it's your neighbor telling you that geese like cheese, doesn't mean you can now include it. The policy really is saying that you need to use sources that come from fixed media. If you listen to a speech by your town mayor, you also cannot include it, unless you can point to some place where the speech is preserved in fixed media. Of course that gets more to verifiability as well. Wjhonson (talk) 18:42, 16 January 2008 (UTC)

As you can see directly above, I too remain confused about the difference between research and original research, especially on this issue of lists. Examples:

  • a phrase that is commonly used in a neutral way has many well sourced uses. However, because it is rarely if ever defined as that common usage, some editors insist that ONLY a negative interpretation of the phrase which is very commonly commented upon by "reliable sources" can be used. Therefore a group of editors insist that NO mention of the the neutral uses of the phrase can be listed because that would be WP:OR.
  • an article has a statement "the tenets of ___ism are" and then there's a list, with some items sourced and some not. If they were all sourced, would that make it research or WP:OR? Or only if the reliable source clearly said - this a tenet of ___ism.
  • a description of what is in a book. There are two listings of contents by reliably sourced reviewers. And then the phrase "other revelations include: lists about 10 more with references to page numbers when the info is contained in the book.

Are these research or original research? I think this section has to make it very very clear. That's why I want to put something about lists in per the above. If no one comments in a week or so I'll just put up my version as I understand it at that time and then I'm sure someone will comment if it's wrong! Carol Moore 03:19, 17 January 2008 (UTC)CarolMooreDC talk

Whatever you call it, it is not a violation of the original research policy to state that certain information is contained in a published book. It could be undue weight. If the book were not published, it could be a violation of both the original research policy and the verifiability policy. --Gerry Ashton (talk) 03:31, 17 January 2008 (UTC)
Actually, Carol, you are dealing with a neologism, one which is not found in any dictionary, encyclopedia, or standard reference work. You have been told time and again that WP:NEO states explicitly:

An editor's personal observations and research (e.g. finding blogs and books that use the term) are insufficient to support use of (or articles on) neologisms because this is analysis and synthesis of primary source material (which is explicitly prohibited by the original research policy).

No editors "insist that ONLY a negative interpretation of the phrase which is very commonly commented upon by "reliable sources" can be used." Rather, they insist that, per WP:NEO, "we must cite reliable secondary sources such as books and papers about the term—not books and papers that use the term." Jayjg (talk) 03:41, 17 January 2008 (UTC)
I stated this in general terms as problems arising in several articles which illustrate confusion on WP:OR per this thread's title. In the specific case you reference, there is legitimate debate over whether a phrase used for at least 30 years is a "new" phrase. I don't think it is.Carol Moore 05:44, 17 January 2008 (UTC)CarolMooreDC
I agree with Jayjg, if I understand correctly. As I see it, we can do what the sources do. If the sources use a certain word to mean a certain thing, then we can use the word to mean the same thing. If the sources make an assertion, then we can make the same assertion in different words, with or without prose attribution, or quote it. If the sources state something about a word, we can state the same thing about that word. But if the sources merely use the word, we can't state something in mainspace about the word based on that. To include the word in a list would be, in effect, to state something about the word, therefore OR and not allowed in that case. Of course, the inclusion of any particular piece of information is also subject to other policies, to whether it fits neatly into a well-organized article, and to consensus among the editors. --Coppertwig (talk) 13:41, 17 January 2008 (UTC)
the sources is sometimes the problem. For controversial topics, one can find sources, respected secondary or trtiary sources, to support almost any position. We are then left with the job of deciding which is the better source, or synthesizing a position from them. There is no substitute for intelligent judgment in editing. DGG (talk) 16:08, 17 January 2008 (UTC)
  • "we can do what the sources do, etc." sounds good and maybe that phrase and/or whole description needs to be inserted in relevant part of article. And then use both paragraphed text and lists and even boxes as examples.
  • thanks for mentioning list which I did not know exists!
  • controversial and conflicting sources a problem - like if one was trying to "List" the tenets of communism or capitalism!
  • and of course we assume we are not dealing with roving bands of 5 or 6 or 10 editors who will challenge any source on defacto POV grounds like WP:Idontlikeit in order to push a viewpoints they don't like out of articles about their favorite topic.
  • I just deleted something as WP:OR because someone inferred that an expert was talking about a specific organization because he quoted specific points made by that organization, even though he did NOT name the group specifically in his text. I told him to find someone who actually made those claims against that specifically named organization. I hope I got that one right! :::Carol Moore 19:01, 17 January 2008 (UTC)CarolMooreDC


We keep getting sidetracked by arguments over details... what makes something a primary source, under what circumstances could a primary source become a secondary source, etc. etc. etc.

Let's refocus. I am still having difficutly understanding how the current PSTS section directly relates to the concept of NO ORIGINAL RESEARCH. Maybe I am being dense... but if so, then I am not alone. Could someone please explain it to me (us) in blunt simple language? Blueboar (talk) 18:35, 5 January 2008 (UTC)

Actually making a primary source into a secondary source is one form of OR; the other is making Wikipedia a primary source. Really that is all there is to it, and if we just said this in the PSTS section we would be done. OK, maybe we would need to add that Wikipedia is a tertiary source, but we don't need all the excess source typing baggage that PSTS has accumulated over time.
One important use of source issue is the opposite of your example, making a secondary source into a primary one. The PSTS section does not address that important point at all. Dhaluza (talk) 19:09, 5 January 2008 (UTC)
If you are dense, then you are not the only one. I asked this numerous times over the last 4 months, and as far as I can remember, on the rare occasions it was answered, the answer was along the lines of either "It's been here a long time and there is consensus to keep it" or "This is a very importatnt part of determing OR. We use it all the time". When asked to clarify the first type of answer, the reply was that since nobody objected, there was obvious consensus for it (being snarky, but sound familar?). The second type of answer, when asked for some examples of how it's been used in the past to 'solve problems' was just ignored. If have a feeling that this will also suffer the same fate (being ignored). wbfergus Talk 19:17, 5 January 2008 (UTC)
This isn't to say that Wikipedia doesn't encourage original research for your own purposes (a science experiment you do for yourself) so long as you don't publish it as fact on Wikipedia. If you publish it in a book, it will be eligible for a source in an article. —Preceding unsigned comment added by (talk) 20:53, 20 January 2008 (UTC)

Proposed replacement for PSTS

I line with my comments above, could we replace the current PSTS section with a streamlined version that focuses on NOR only, similar to this:

Wikipedia is not a publisher of new facts or thoughts, but is a compendium of knowledge drawn from existing material published by reliable sources.

  • Wikipedia does not publish newly discovered previously unpublished facts—that would be original research making Wikipedia a primary source for those facts.
  • Wikipedia does not publish new commentary or analysis—that would be original research making Wikipedia a secondary source for those thoughts.
  • Wikipedia does publish collected and summarized facts and analysis—that is source based research making Wikipedia the tertiary source it is.

I think this stays true to the original intent of the policy as reflected in the history, and also uses the familiar terms, but without redefining them. It makes the important points easy to understand, dropping the unnecessary complexity accumulated over time that has expanded the definition of NOR beyond its original purpose. That material can be absorbed by other policies or guidelines, existing or proposed, if it is useful to their specific purposes. But anything that is not germane to NOR should not be included here. Dhaluza (talk) 19:57, 5 January 2008 (UTC)

Precisely the approach I proposed at Wikipedia:Wikipedia is a tertiary source#Wikipedia is an encyclopedia. Except that I didn't use the term "source based research" which seems a good idea. --Francis Schonken (talk) 20:07, 5 January 2008 (UTC)
I like it for the most part. However, I'm wondering if something can be done with "newly discovered facts". That seems a bit vague in how it relates to this (OR). It kind of sounds like primary source couldn't be used if they contained something new, and I think it's supposed to mean that Wikipedia doesn't publish those unless they come from something already published. Is that correct? wbfergus Talk 20:36, 5 January 2008 (UTC)
Good point, I've corrected it with a strikeout. Dhaluza (talk) 20:46, 5 January 2008 (UTC)
Dhaluza: I doff my hat. Your suggestion may be just the ticket out of the quagmire.
A few suggests with regard to readability/grammar,...
  • Wikipedia does not publish previously unpublished facts. Such material is original research and would make Wikipedia a primary source for those facts.
  • Wikipedia does not publish new commentary or analysis. Such material is the product of original research and and would make Wikipedia a secondary source for those thoughts.
  • Wikipedia publishes only existing facts and analysis. This is source based research and makes Wikipedia the tertiary source that it is.
Your suggestion is fundamentally good, and these are just minor readability issues.
Perhaps it might be a good idea to put 'tertiary' first (i.e. the lines in "descending" order). That might compensate for the lack of distinction between "source based" (presently at the end) and what is effectively 'published but unacknowledged' (in the middle). Just a feeling though.
-- Fullstop (talk) 00:53, 6 January 2008 (UTC)
I like this line of thinking a lot and I think it has promise. I particularly like the idea of using the PSTS terminology, but only as actually needed, and without making Wikipedia policy a slave to all the typology's unintended and sometimes undesired implications. I think this is a promising way out. If we are going to continue to use the terms "primary", "secondary", and "tertiary" at all, I think it would be best to keep the order as the one Fullstop proposed. This wording is limited to statements describing Wikipedia and its requirements only. It avoids making any claims about the outside world. I believe this approach is a very sound one. Best, --Shirahadasha (talk) 01:33, 6 January 2008 (UTC)
  • I like this general approach. It's the first time I've seen something that's tied the PSTS material into NOR in a way that has actually helped me make sense out of it, while not making me worry that it is going to cut too broadly. It's also clean and simple and clearly and explicitly states what is prohibited. --Lquilter (talk) 01:46, 6 January 2008 (UTC)
  • I think Fullstop's wording is attractive. Compared to the existing policy wording, it explains much more concisely what is and what is not allowed by the policy, and, more significantly, it does so with far greater clarity, indeed with truly exceptional clarity in my opinion. - Neparis (talk) 02:14, 6 January 2008 (UTC)
Yes, I think Fullstop's version is an improvement. I would like to tweak it a bit further, and propose the following refinement:

Wikipedia is not a first publisher of new material—it is is a compendium of knowledge drawn from existing material published by reliable sources.

  • Wikipedia does not publish previously unpublished facts or observations. That is original research, which would make Wikipedia a primary source for those statements.
  • Wikipedia does not publish new commentary or analysis. That is the product of original research, which would make Wikipedia a secondary source for those thoughts.
  • Wikipedia only republishes existing facts and analysis. This is source based research, and makes Wikipedia the tertiary source that it is intended to be.
-- Dhaluza (talk) 02:59, 6 January 2008 (UTC)
I agree! This is a step in the right direction. Although avoiding analysis is nigh impossible. Some might argue that making a description out of a set of events is constructing an analysis. Also, this policy shouldn't be used to challenge the claims given in sources or the sources themselves. --Benjaminbruheim (talk) 05:09, 6 January 2008 (UTC)
To be more precise; when text is challenged by this policy, it should focus on the additional claims resulting from OR; not entitle the removal of facts --Benjaminbruheim (talk) 05:16, 6 January 2008 (UTC)
  • I also agree that this is an improvement for the purposes of NOR, but it is only half the work done. Don't forget that definitions of primary and secondary will still have to exist somewhere, because they are referred to by guidelines, WP:Notability in particular, but also WP:WAF (others too, I'm sure). There are 300-odd links to the shortcut WP:PSTS, too.  —SMALLJIM  10:47, 6 January 2008 (UTC)
The other half of the work has already been preformed performed, it is at WP:WITS. --Francis Schonken (talk) 11:02, 6 January 2008 (UTC) - typo correction 19:20, 7 January 2008 (UTC)
Um. I can't see where WP:WITS being the complement to this has been discussed - please direct me. But anyway wouldn't Dhaluza's proposal have a greater chance of gaining consensus if WP:PSTS was initially spun off as it stands?  —SMALLJIM  13:37, 6 January 2008 (UTC)
Re. "I can't see where WP:WITS being the complement to this has been discussed" - that's a different question. I answered: "[...] it is only half the work done. Don't forget that definitions of primary and secondary will still have to exist somewhere, because they are referred to by guidelines, WP:Notability in particular, but also WP:WAF (others too, I'm sure).":
Answering your new question: WP:WITS has been discussed in several places: for example Wikipedia talk:No original research/Archive 32#WP:WITS and Wikipedia talk:Wikipedia is a tertiary source. The relation between WP:WITS and Dhaluza's proposal is discussed above on this page in #Discussion, which will show you that the question would rather need to be "where is it discussed that Dhaluza's proposal is *derived* from WP:WITS?"
Re. "wouldn't Dhaluza's proposal have a greater chance of gaining consensus if WP:PSTS was initially spun off as it stands?" - It was already spun off, the spinoff currently is named Wikipedia:Evaluating sources. But whether that spinoff is a success, is a matter of debate, e.g. (Wikipedia:Evaluating sources) started as a subpage to NOR to explain the primary/secondary thing, but then it spread out and started giving general advice about sourcing, often doing nothing but state the obvious (...) (my bolding) --Francis Schonken (talk) 14:42, 6 January 2008 (UTC)
It may be that I misunderstood your "The other half of the work has already been preformed...". I assumed that "preformed" was a typo for "performed", which led me to think that you were saying that WP:WITS was already accepted as part of Dhaluza's proposal. You're saying the opposite is true: that Dhaluza's proposal is derived from WP:WITS, but I'm not at all sure that those commenting favourably here on that proposal are aware that they're including WP:WITS in their approval. I certainly wasn't (obviously), and I would reiterate that I think Dhaluza's latest refined proposal[11] stands the best chance of being accepted if it is accompanied by a simple removal of the existing PSTS text to a new WP:PSTS guideline (or even being initially added to NOR). I don't think that consensus for the WP:PSTS shortcut to be moved to WP:EVAL or WP:WITS would be easy to achieve, but maybe I'm misreading the sentiment here.  —SMALLJIM  17:45, 6 January 2008 (UTC)
I like this approach a lot. It is short, easy to understand, and most importantly directly ties the entire PSTS concept into the broader concept of NOR. Well done. Blueboar (talk) 14:13, 6 January 2008 (UTC)
Hmm .. I'm not sure it's accurate, and it isn't comprehensive. First, what is the distinction between something's being OR and being the product of OR? Also, the republication of existing facts and analysis -- it misses out the crucial point that, in some cases of primary-source material, we don't republish it unless it's been published by a secondary source e.g. in BLPs. SlimVirgin (talk)(contribs) 14:17, 6 January 2008 (UTC)
I agree it's not all inclusive, but then neither is the current policy section on PSTS, and that is the reason there has been so much discussion about PSTS over the last 4 months. I don't see how the current PSTS wording clarifies those points either. Regarding the "some cases of primary-source material", wouldn't that be better covered in the other policies, BLP and others? After all, Kenosis and numerous others have stated many times that these policies have an inherent interaction amongst them, so explaining the rare exceptions in regards to OR in those policies makes more sense than adding a list of exceptions in this policy. At least, that's my opinion. wbfergus Talk 14:47, 6 January 2008 (UTC)
I should also say for the record though, and also to alleviate Slim's concerns and others, that while I think this proposal is far better, I would also be opposed to it being implemented until such time as there is an "official" replacement/home for the PSTS "stuff", whether it is the WITS or the Evakutaing Sources or something else. wbfergus Talk 14:51, 6 January 2008 (UTC)

(outdent) As I see it, Dhaluza's proposition is a generic solution that addresses the thousands of comments of the last six months.
The 'cons' of the proposition first:

  • it is only inviting to people who want to collaborate and make the 'pedia a better place.
  • it is not a magical miracle cure.
  • it is not a weapon to beat people with.
  • it is not convoluted enough to make people who "won't accept a clearer explanation" happy.
    For those who "make the language as tortured as possible," simple language is "not comprehensive."
  • it is not vague, and thus impossible to invoke when moving goalposts.
  • it is not inflexible, and thus particularly suitable to the dynamics of a wired community.
  • it is "not confusing, not tortured, not ill-considered, not implying things we didn't mean to imply, not language gone on holiday."
  • it no longer subverts NOR and other policies.
  • it would be the end of endless talk, and might cause editors to have to resume contributing.

The 'pros' of the proposition are:

  • the proposition is not a fundamental change to either NOR policy or to PSTS.
  • it uses simple English, thus addressing the "tortured language" issue.
  • it is short and to the point, consistently applying Occam's razor, and thus addressing "Which version is more succinct, and which do you think most editors will understand more easily?"
  • the proposition actually (finally!) gives PSTS some relevance to NOR.
  • it obviates the need to define "primary" and "secondary", thus...
    • avoids the issue of arbitrariness.
    • avoids any need to compensate for borderline cases
    • avoids redefining words that "every educated native English speaker understands."
    • avoids contradicting the terms "used by universities and professional researchers and publishers."
    • prevents sight of purpose to be lost.
  • it no longer subverts NOR and other policies.
  • it is in policy language, and clearly obligatory. Recommendations and options - ala "should" and "may" - are words for guidelines and weasels.
  • it leaves the vagaries of definition(s) to mainspace articles. If there is such a thing as a non-tortured definition of ps/ss, then this should be in the appropriate articles(s) anyway. Those articles are linked to.

Now if people "would only collaborate, ... instead of constantly resisting, ... it would be a win-win situation."
-- Fullstop (talk) 21:05, 6 January 2008 (UTC)

Smalljim and SlimVirgin, you said that the proposed WP:NOR language is incomplete because it does not discuss the needs of e.g. WP:BLP. I'd lile to follow up on this. Why not discuss WP:BLP issues on WP:BLP itself, and so forth with the other policies? Suggest discussing only the requirements actually needed for each policy on that policy page. A more comprehensive discussion of sourcing could still be done elsewhere, but as a guideline. I'd agree the current version of WP:EVAL could be improved. Why not improve it? Best, --Shirahadasha (talk) 02:01, 7 January 2008 (UTC)

Oops, I've just seen this. My point wasn't that the language of the replacement is incomplete; I was just trying to forestall future problems by observing that replacing the existing PSTS section with it would break the 300-odd links that point to WP:PSTS. I don't think we can leave WP:PSTS pointing to Dhaluza's replacement because it says something completely different (that's the whole point), and I don't think it's a decision that can be deferred because of the large number of links involved. So what is the proposal regarding that shortcut - is it to add it to WP:WITS?  —SMALLJIM  09:02, 8 January 2008 (UTC)

A problem

I've been re-reading Dhaluza's proposal and I think the second sentence of the second bullet is actually wrong. The bullet says:

  • Wikipedia does not publish new commentary or analysis. That is the product of original research, which would make Wikipedia a secondary source for those thoughts.

Adding OR in the form of commentary or analysis to WP makes WP a secondary source for the facts or earlier research that the commentary or analysis is about, but it makes it the primary source for that commentary or analysis itself. It's that question of context and what is the topic again. It's no different to publication in a book: If I write a book that contains my new commentary or analysis of WWII, it's a new secondary source for WWII, but the primary source for my new commentary.  —SMALLJIM  00:06, 7 January 2008 (UTC)

I have a different problem. When one analyzes published primary or secondary sources and forms a new conclusion, one's writing is a secondary source; everything in one's writing is a secondary source, including the new conclusion. So the second bullet is correct.
The problem is that while Wikipedia should be a tertiary source when viewed as a whole, it is OK for Wikipedia to be a secondary source if you're going to judge it on a paragraph-by-paragraph basis. For example, a Wikipedia editor creates a four paragraph summary of a work of fiction, relying only on the work of fiction for those paragraphs. Those paragraphs, judged alone, would be a secondary source. But they're OK; Wikipedia allows the use of primary sources in this way. We would hope the rest of the article uses secondary sources, so the article as a whole would be a tertiary source. --Gerry Ashton (talk) 00:21, 7 January 2008 (UTC)
Sorry, Gerry. I disagree with your first paragraph - I think you're forgetting about the context. See below for a fuller explanation.  —SMALLJIM  13:23, 7 January 2008 (UTC)
SmallJim, I think you are using a somewhat impractical definition of secondary sources, at least for WP purposes. Under this model, expert opinion would always be a primary source, but we treat this as a secondary source. Dhaluza (talk) 00:32, 7 January 2008 (UTC)
Actually, there is no conflict at all. :)
Assuming that the "That is" of the second bullet refers to "new commentary or analysis" and not the publication of it, the bullet is actually saying precisely the same thing SmallJim is saying in the WWII-book example. Which is: publication of his "new commentary or analysis" ... "would turn Wikipedia into a secondary source."
And the rest of the material in the book is ok, and covered by the third bullet. -- Fullstop (talk) 00:51, 7 January 2008 (UTC)
GA, I think we should consider the summary of the work of fiction as part of the tertiary source work as a whole. Yes, this means this part passes from primary source to tertiary, skipping secondary, but that just shows the problem with the whole primary/secondary/tertiary continuum. For our purposes, we are concerned with secondary sources containing interpretation, analysis, etc. Simply describing the primary source does not make the description secondary from that perspective. Dhaluza (talk) 00:32, 7 January 2008 (UTC)
GA, Upon further reflection, there is no conflict with the plot summary example, because "A tertiary source is a selection, distillation, summary or compilation of primary sources, secondary sources, or both." So the summary is not a secondary source, it is a tertiary source. —Preceding unsigned comment added by Dhaluza (talkcontribs) 10:54, 7 January 2008 (UTC)
A Wikipedia editor who faithfully reproduces "a four paragraph summary of a work of fiction" is not engaging in "original research" because there is nothing new in a faithful reproduction. ::(When the conditions of WP:NOTE are met, someone else will have already reviewed the work of fiction anyway, and the WP editor's summary is an unembellished equivalent of that review)
-- Fullstop (talk) 01:19, 7 January 2008 (UTC)
I think it would helpful in addressing these types of problems to simply stop using the words "primary", "secondary", and tertiary, and simply ask the question: Is it WP:OR? One could come up with reasons why Wikipedia should or should not consider plot summaries WP:OR. These reasons may be unique to Wikipedia's special needs and situation. It simply doesn't matter whether the outside world classes them primary, secondary, or tertiary for some other purpose. Plot summaries may be one of those situations where the PSTS model simply isn't useful for Wikipedia's needs and using these labels here only confuses - asking what type a source simply doesn't tell us anything useful. We need a model and language that enable us to discuss Wikipedia's actual needs and give reasons that make sense in WIkipedia's actual context. Best, --Shirahadasha (talk) 02:22, 7 January 2008 (UTC)
perfectly true, but SmallJim's and Gerry's points were with respect to Dhaluza's proposal. That proposal (see section above this one) attempts to find a sensible replacement for PSTS, i.e. that among other things gets rid of the need to define ps/ss/ts in policy.
So, even though source-typing has no relevance to NOR, what SmallJim and Gerry are pointing out still needs to be examined, because proper objections (not just the banal "insufficient" or whatever) are *good*. They show that people are thinking, and they can help point out where the weaknesses are.
But I am grateful for your comment too: looking into the abyss takes its toll and I for one occasionally need to be told to blink. Thanks. -- Fullstop (talk) 03:28, 7 January 2008 (UTC)

I know it's presumptuous of me as a relative newbie here, but I'm going to be bold and assert that you are all wrong! You are all forgetting the Thanks for the responses, but I'm sorry to have to admit that I still don't "get it". To me, you all seem to be forgetting the well-established principle that sources can often be either primary or secondary depending on the context, and the context here is the new commentary/analysis, not the topic to which it's being added. Looking at the wording of that second bullet again:

  • Wikipedia does not publish new commentary or analysis. That is the product of original research, which would make Wikipedia a secondary source for those thoughts.

I have assumed that "those thoughts" refers to the "new commentary or analysis" (in the same way that in the first bullet "those statements" clearly refers to the "previously unpublished facts or observations"). So let's replace it:

  • Wikipedia does not publish new commentary or analysis. That is the product of original research, which would make Wikipedia a secondary source for that new commentary or analysis.

Does that make what I'm trying to say any clearer? I can't see how the first publication of anything can be a secondary source for itself! That just doesn't make sense to me. Where's the "one step removed" (in authorship) that's required by the WP:PSTS definition of secondary source? Yes, that new commentary or analysis is one step removed from, and is therefore a secondary source for, the topic of the article to which it's added - but that's not what the bullet says.

A long-winded, but correct formulation of that bullet might be this:

  • Wikipedia does not publish new commentary or analysis on any topic if that commentary or analysis is created by its own editors. Doing so would be publishing original research, which would make Wikipedia a secondary source for that topic (and a primary source for the new commentary or analysis).

If any of the above reasoning is incorrect, I'd be grateful if someone would explain where I'm going wrong, because, based on what I've learned here over the past couple of months, it seems quite straightforward to me.  —SMALLJIM  13:12, 7 January 2008 (UTC). Refactored 22:49, 7 January 2008 (UTC).

It is more than presumptuous, Smalljim, it's forgetting who reads these comments. We are all wrong? Specifying the reference more tightly, all the arguments presented here are wrong? How likely is that? It's one thing to assert that some important point has apparently been overlooked in the discussion so far, but quite another to personalize this. I have another concern, myself. I consider [WP:NOR]] to be problematic in itself. The core principle is verifiability, [WP:V]]. As is noted in the present article, original research *can* be presented on Wikipedia, *if it is notable and is attributed,* provided that the fact asserted (that so-and-so commented, reported, concluded, claimed) can be verified. The problem addressed above is new analysis, done by an editor of Wikipedia. Wikipedia editors are presumed to be anonymous, and there is strong pressure here to avoid Wikipedia becoming a place to publish new analysis; it happens on Talk pages all the time, to be sure, but that analysis may be essential to the formation of editorial consensus. However, what if the analysis is obvious from the sources, and it is "new" merely because it has never before been stated -- as far as the editors know? I'd say that, routinely, such analysis is done here; editors look at all the available sources and "summarize" them. That summarization is a new analysis. Only exact quotation of secondary or tertiary sources would be otherwise. So we must back up a little bit. What is a problem is a new *controversial* analysis, that is *not* obvious from the sourced facts, that can reasonably be challenged.
Sometimes a new analysis ("summary") is properly controversial even when it would appear verifiable by examination of the primary source, so "summaries" should be handled with care. I've been dealing with an example: for a long time, Instant-runoff voting claimed, in the introduction, that "Robert's Rules of Order recommends IRV." I'm somewhat of an expert in the field, and I read Robert's Rules of Order on this, and the claim seemed substantially true to me. However, careful re-reading eventually led me to quite a contrary conclusion. That statement was a summarization, taking a complex discussion and reducing it to a single "sound bite," one quite valuable for promoting a particular POV, and its placement in the introduction, where the detail necessary to make an NPOV description of the facts may be inappropriate. The report was moved deeper into the article, and altered to actual quotation of the source, so that it could not be challenged as POV. (And something like an edit war began, more recently, over attempts to put, once again, a summary into the introduction. Naturally, the summary omitted the critical material.)
My point here is that summarization is essential to encyclopedic work, yet summarization is original research *even if NPOV*, unless the summarization is an exact quote, in which case it is not summarization at all, it's quoting a source. This is where peer review comes in. Summarization may exist in peer-reviewed publications, but peer-review does not guarantee that summarization existing in peer-reviewed articles is not biased. Peer-reviewed journals publish unvalidated opinion, particularly from experts, on the basis that the opinion is, to the community of review, notable. Thus peer-review establishes notability, not reliability, in fact. To determine reliability, one would have to review all subsequent comment on the original publication, it makes the research task daunting. There is a simpler standard, which is consensus of the editors, particularly when the editors include editors with various points of view, each set likely to come up with contrary material. A stable article with many editors has effectively been peer-reviewed, quite likely, just as print encyclopedias may submit preliminary drafts of articles to experts in the field. --Abd (talk) 15:20, 7 January 2008 (UTC)
Well obviously I didn't include you in that statement, Abd, seeing as you have not commented here before. I would hope that those who are following this discussion understand that I was only referring to the remarks posted in this subtopic that comment on mine which started it. But thank you for giving me the opportunity to assure everyone that there was no intention to impugn anyone. And, of course it's not very likely that everyone else is wrong, which is why I said I was being presumptuous and bold. If I am wrong, though, I still don't "get it" and hope for enlightenment.  —SMALLJIM  16:31, 7 January 2008 (UTC)
In retrospect I've struck the offending remark and I apologise for getting overexcited. I hope I haven't offended anyone - it certainly wasn't my intention to do that.  —SMALLJIM  22:49, 7 January 2008 (UTC)

Would this be an improvement?

How about changing the bullet involved as follows:

*Wikipedia does not publish new commentary or analysis. That is the product of original research synthesis, which would make Wikipedia a secondary source for those thoughts.

Would this be an improvement?

If there is basic agreement on this language and the issue is only tweaking a few words to address the problem discussed above, perhaps we could move towards closure. Even if we can't agree to exact wording on the proposal, if there is consensus that the proposal's general approach would be an improvement over what's currently there, that would seem to be a basis for moving forward. Best, --Shirahadasha (talk) 03:25, 8 January 2008 (UTC)

I definitely don't want to stand in the way of progress here. Regarding the issue I raised, I'd hoped (expected!) that everyone would immediately say, "Oh yes - of course!", and agree to change the text appropriately. But since that hasn't happened, I see that I must be failing to appreciate something important: if I started a new section to ask for clarification, would someone spare the time to help me?  —SMALLJIM  11:30, 8 January 2008 (UTC)
Looks OK to me (after the minor corrections). Dhaluza (talk) 13:28, 8 January 2008 (UTC)
The current nutshell is better. There is more to OR than synthesis, and we don't want WP to be a primary source either. SlimVirgin (talk)(contribs) 13:40, 8 January 2008 (UTC)
This section is on one bullet point of a proposed replacement for the current PSTS language (see two sections above for the proposal and one section above for the problem with the bullet point.) Not the nutshell. Sorry for the confusion. --Shirahadasha (talk) 14:01, 8 January 2008 (UTC)
Sorry, my mistake. :-) SlimVirgin (talk)(contribs) 14:29, 8 January 2008 (UTC)

What about this:

  • Wikipedia does not publish new commentary or analysis by its editors on any topic. That is original research synthesis, which would make Wikipedia a secondary source for that topic.

Is it correct?  —SMALLJIM  23:29, 9 January 2008 (UTC)

I very much dislike "That is original research synthesis, which would make Wikipedia a secondary source for that topic." Some of my reasons are:
  • "Original research synthesis" is a made-up phrase that no one will understand unless they have been following this talk page for a long time.
  • There is nothing wrong with Wikipedia being a secondary source, when judged on a small piece of text. It is only when viewed as a whole that Wikipedia should be a tertiary source. Secondary sources sometimes form novel conclusions based on reviewing other publications, and sometimes just summarize the other publications. It is fine for Wikipedia to be a secondary source that summarizes other publications. —Preceding unsigned comment added by Gerry Ashton (talkcontribs) 23:39, 9 January 2008 (UTC)
Rephrasing the original then, is it correct to say that WP becomes a secondary source for any new commentary or analysis that is added by its editors to a topic?  —SMALLJIM  00:40, 10 January 2008 (UTC)
Yes, but Wikipedia is also a secondary source for any quotation or summary added by its editors. "Quotation" is significant in that it shows that someone gained access to and took notice of the quoted source. Of course, most summaries are new. --Gerry Ashton (talk) 00:51, 10 January 2008 (UTC)
Thanks Gerry. Setting aside quotations and summaries for a moment, may I just confirm one point, please: are you saying that Wikipedia is a secondary source for any new (i.e. previously unpublished) commentary or analysis that is added by its editors to a topic?  —SMALLJIM  14:19, 10 January 2008 (UTC)

New revision

Based on the comments to date, I think the "product of original research" and "original research synthesis" are just roundabout ways of saying "also original research". So based on this, can we consider the following alternative:

Wikipedia is not a first publisher of new material—it is is a compendium of knowledge drawn from existing material published by reliable sources.

  1. Wikipedia does not publish previously unpublished facts or observations. This is original research, which would make Wikipedia a primary source for those statements.
  2. Wikipedia does not publish new commentary or analysis by its editors. This is also original research, which would make Wikipedia a secondary source on that subject.
  3. Wikipedia only republishes existing facts and analysis. This is source based research, and makes Wikipedia the tertiary source that it is intended to be.

-- Dhaluza (talk) 19:40, 12 January 2008 (UTC)

Well I think it's much improved, but can you remind me what you intend to do with it?  —SMALLJIM  19:53, 16 January 2008 (UTC)
I don't see the need to extend the issue of primary/secondary sources, which I believe will never have complete agreement, into the nutshell. — Carl (CBM · talk) 20:26, 16 January 2008 (UTC)
This is not a proposed nutshell, it's simply a proposed revision to the existing PSTS section. The box simply distinguishes the proposal from discussion. Best, --Shirahadasha (talk) 20:35, 16 January 2008 (UTC)
That's my mistake, sorry. In the context of PSTS, this is fine with me, although it is somewhat more brief than the current text. — Carl (CBM · talk) 21:04, 16 January 2008 (UTC)
Also support the revision. --Shirahadasha (talk) 22:04, 16 January 2008 (UTC)
  • I like this revision, although I still think it's not really necessary to have the PSTS language in there at all. However, given that it's in there, this could be a useful compromise. --Lquilter (talk) 20:32, 21 January 2008 (UTC)

Conflation of concepts in PSS

I think one of the problems we have in trying to sort through the issues with the definition of Primary and Secondary sources is that there are multiple dimensions of observation compressed into a one dimensional measurement. This results in the conflation of separate and not always compatible concepts. The discussions here have revealed several different criteria for separating primary from secondary:

  1. Closeness: The involvement of the subject in the publication process; e.g., an autobiography is primary while an independent biography is secondary.
  2. Succession: How many hands the idea has passed through; e.g., a first-hand account is primary, while a second- or third-hand account is secondary.
  3. Publication: Whether something is an original publication or a republication; e.g., The second-hand account is still primary upon first publication
  4. Opinion: Whether something is fact or opinion; e.g. census statistics are primary, while assigning meaning to them is secondary.
  5. History: The time period of the publication relative to the age of the subject; e.g. Whether something was written at the time or later, and/or how long ago.

There may be other nuances as well. We run into problems when these different criteria produce different results. For example, an eyewitness is a primary source, and a newspaper publishing the eyewitness account is secondary under all but the publication criteria. But if the reporter is also publishing their observations and analysis, we have a more ambiguous situation. And if another newspaper simply re-publishes this newspapers primary source material, is it still primary or does it then become secondary? There have been numerous other examples given here of other problematic taxonomies as well. The conflation of these ideas in the current PSTS definitions is a problem, and I don't see how to untangle this Gordian Knot. Dhaluza (talk) 11:32, 7 January 2008 (UTC)

No. Mere reporting (repeating without commentary), of something doesn’t make a secondary source. Most news is primary source material, even if it is second hand. A secondary source makes some (any) kind of transformation of the information. And you’ve been at this for some time. Maybe it is complicated, like measuring the surface area of a cloud? --SmokeyJoe (talk) 11:53, 7 January 2008 (UTC)
I agree that a primary source should stay primary through republication, but I have seen other comments that are not completely consistent with that. Also, we use news reports as secondary sources for establishing Notability, so is your comment that "most news is primary source" in conflict with that interpretation? BTW, your analogy of measuring the volume of a cloud is not complicated, per se, because we can come up with an unambiguous definition for it; but actually making the measurement would be difficult. This is the opposite of the problem we face here. Dhaluza (talk) 12:10, 7 January 2008 (UTC)
Using news (usually newspapers) to demonstrate notability would seem to produce a complicated problem with source typing, but it doesn’t seem to. Perhaps it's because the existence of the policy WP:NOT#NEWS cuts off attempts to primarily use news. Newspapers are full of primary source material (the news) interspersed with secondary source material (editorials, reviews, feature stories). Making the news doesn’t demonstrate notability. A burst of news reports is explicitly excluded by WP:N. Having a publication say something about you, something that is not news, demonstrates notability. I initially said volume, but meant surface area, which is complicated. --SmokeyJoe (talk) 12:26, 7 January 2008 (UTC)
A short burst of news coverage is excluded by an explicit exception in WP:N, saying it is not sufficient, not that the coverage is primary source. Continuing news coverage over a period of time would be evidence of Notability. You might argue that the body of coverage is collectively a secondary source, but I think the Notability guidelines are really using secondary in the sense of third-party or independent, i.e. non-close sources as in item #1 above.
“Continuing news coverage over a period of time would be evidence of Notability”. Yes. I consider that to be true regardless of the source typing of the news coverage. --SmokeyJoe (talk) 20:31, 7 January 2008 (UTC)
As to your analogy, estimating the volume of a cloud might be used to derive its mass, which may have some practical application. Measuring the surface area of a cloud seems to be more of a useless pedantic folly, and that is like the problem we are dealing with here. Dhaluza (talk) 13:39, 7 January 2008 (UTC)
A useless pedantic folly? Yes, that’s part of what I had in mind. [Source typing / determining surface area] may have application, and may be interesting, but take the analysis too far and it becomes useless pedantic folly. --SmokeyJoe (talk) 20:31, 7 January 2008 (UTC)
When a journalist reports an eyewitness account, even as a verbatim quote, it implies that the journalist believes what the eyewitness has to say has both credibility and relevance. This makes reporting the quote from the journalist different from quoting the eyewitness directly for Wikipedia purposes. One is WP:OR, the other is not. And this is the case even though for many other purposes both could be classed in the same PSTS source class. Some aspects and traditional uses of the PSTS typology are suitable for Wikipedia's needs, others perhaps not. Best, --Shirahadasha (talk) 23:34, 7 January 2008 (UTC)
Reporting an eyewitness account does not mean the journalists "believe it," it just means that is what the eyewitness said. Newspaper articles contain conflicting eyewitness acounts all of the time. Amaltheus (talk) 23:37, 7 January 2008 (UTC)
A good journalist won't report what they believe, but they will assess credibility and relevance as Shirahadasha said. This is the second-hand form of secondary source above. We look to journalists to be a filter for Notability, and in this respect the source is secondary for that purpose. It may be primary from a first publication sense, but for WP, there is no general requirement that we only publish something if it has been published more than once, so that is not relevant. For NOR purposes, we only need to ensure it has been published once in reliable sources. We may assign more weight to opinions that have been republished more than once per NPOV and relative weight, but that's not an OR issue. Dhaluza (talk) 00:52, 8 January 2008 (UTC)


I would like to propose a limiting standard on application of this policy, because its enthusiasts can be a bit draconian. Saying "use common sense" is insufficient, because the policy there is so vague and weak that it really doesn't provide any defense. Specifically:


Complete description of a topic often requires a basic process of deduction. Data from reliable sources may often be available in incomplete or inconsistent form because readers are assumed to have the skill to work out additional meanings for themselves. Wikipedia encourages editors to use good judgment to make basic scientific deductions, provided that they are obvious to one of reasonable skill in the art. For example, editors may:

Editors making deductions must be careful not to make mathematical errors or to overstate their conclusions -- for example, the theoretical values predicted for a polypeptide may bear no resemblance to the experimental values for the posttranslationally modified protein. The data used to make the deduction should be provided and appropriately referenced, and the rules, assumptions, and approximations used to make any deductions (beyond basic arithmetic) should be explicitly mentioned and Wikilinked. As conversions are done to allow comparison of data from more than one reliable source, they should cast data into units that are prevalent among the sources cited.

Editors removing such unsourced deductions do not need to prove that they are truly incorrect, but only that their accuracy or applicability is not obvious to someone readily verifiable by anyone familiar with the science.

What do you think? Wnt (talk) 17:34, 17 January 2008 (UTC)
This assumes that a person editing an article is skilled in the art. This is not necessarily true for someone cleaning up after a vandal. Converting units is a widespread skill that does not even require skill in the art, but beyond that, it could be quite difficult to clean up after vandals. --Gerry Ashton (talk) 17:42, 17 January 2008 (UTC)
I agree. "Skilled in the art" may be good for patents, but not for Wikipedia articles. Our threshold is "reasonably educated person", or the common sense of the Clapham Bus rider. Anything more invites trouble. If the material is correct and important, someone will have published it in the desired form. Crum375 (talk) 17:50, 17 January 2008 (UTC)

It may be useful to go a bit beyound the concept of "deduction" and add some additional common examples where editors routinely summarize and explain in ways that are essential to making specialized knowledge comprehensive to a lay reader but which are technically OR. For example:

  • Summaries in lay English can explain or translate common terminology, jargon, and foreign-language terms used in sources which are ordinarily known to people in the field but which would not be known to a non-expert. However, preferred practice is to create and link to reliably-sourced articles explaining the concepts underlying these terms.

Best, --Shirahadasha (talk) 18:03, 17 January 2008 (UTC)

The problem I have always had with these deductions is that often there are little important details left out by the publishers (or merely those who put it on the Wiki, unbeknownst to the person who is making the deduction). To take the polypeptide sequence as an example: While the primary sequence of a polypeptide could be used to logically deduce the molecular mass of the primary sequence, this does not necessarily imply the molecular mass of the polypeptide. For all of the posttranslational modifications it can undergo, the calculated molecular mass could be completely off-base. Deductions should really be limited to those cases in which the deduction can be made without any assumptions. If you find yourself making assumptions, you're probably conducting original research. Someguy1221 (talk) 21:04, 17 January 2008 (UTC)
I appreciate the feedback, but I'm having some trouble reconciling myself to it. The point of appealing to skill in the art rather than the "Clapham Bus rider" is to plug holes in more specialized literature. People may publish the pKa of thousands of acids and give an example of pH for different concentrations of hydrochloric acid, but no one is going to provide the pH of every concentration of every acid known to man because there's a formula for it. So if you want to edit an article about a known concentration of a certain acid featured in some particular news event this policy asks you to wait for the hit-and-miss event of whether a mainstream media with little interest in science chooses to comment on it, and do without the figure permanently if that's what they decide. I don't think that's right, if anyone who has taken a chemistry course can agree that the pH should be a certain value. I know that the policy describes a "reasonable educated person without skilled knowledge" in regard to primary sources, but that is different. A primary source is used cautiously because it is an isolated and perhaps debatable result, whereas the general laws that I've suggested people Wikilink while using them for calculation are the best-known and most widely accepted principles of the science, as evidenced by their familiarity to and obvious use by the consensus of skilled readers favoring the edit.
Where "summaries" of jargon and terminology are concerned, I find myself resisting in the opposite direction. If there is no calculation involved, but simply explanation, there should be a source available somewhere. If a neuroanatomist can tell you where the lateral geniculate nucleus is, then believe me, there's a ridiculous amount of literature out there that defines that term. So there's no need for off-the-cuff explanations except as a WP:IAR stopgap measure.
Lastly, my polypeptide example was chosen to illustrate that the conclusion of a calculation easily can be overstated, but the fact that a reader can repeat and concur with that within 12 hours in a non-biology article should serve as proof that someone skilled in the art can spot and revert an overstatement when one is improperly made. Yes, caution is needed, but not as much caution as it takes, say, to edit an article subject to an ethnic edit war without getting reverted. There's no need to toss the baby with the bathwater. (talk) 02:46, 18 January 2008 (UTC)
In general, you will get less resistance to "OR" if you are not trying to "advance a position." That normally means doing things that are uncontentious. So if there is a clear formula for something, and you can point to the formula in a good textbook (ideally with examples) and then run some numbers through it, that others with no specialized skill or knowledge can repeat, odds are the edit will stick as "verifiable". If someone comes along and says the formula is wrong, or is being misapplied, be ready to defend it, or it will all go out pending published proof. Crum375 (talk) 03:14, 18 January 2008 (UTC)
That may be true, but the reason why I think it needs to be spelled out in the policy is that I've seen a few situations where people start reverting calculations simply because they are "original research" with no other reason to disagree. One I remember in particular was when I was doing some editing in Bode's Law before because I thought that an unsourced statement that the law resulted from orbital resonance was misleading (the law predicts irrational ratios while the orbital resonances are all different rational numbers). Maybe my table comparing the numbers from the two articles would run afoul of the policy anyway, but the point is, under the policy as it is I didn't even have the right to work out the orbital period ratios from the Bode's Law distances let alone to make the comparison, and it all simply got reverted immediately. (talk) 03:37, 18 January 2008 (UTC)
I would think you could have argued in favor of your text, especially if it is known or implicit in the literature. The OR policy is not a stark line, and isn't intended to prevent people adding uncontroversial things that are well known to all in a particular field. — Carl (CBM · talk) 04:00, 18 January 2008 (UTC)
Since we have no formal peer reviewers here, we can allow just the bare minimum: a clear and uncontentious textbook formula, that can be used by non-experts, which requires no special assumptions to be made, could be considered "verifiable". But your case sounds like it was more complex. If something is correct and important enough, it will be published elsewhere, and we can then refer to it. Conversely, if it hasn't been published, it's either incorrect or not important enough. Crum375 (talk) 04:03, 18 January 2008 (UTC)
"Provided that they are obvious to one of reasonable skill in the art" is a really loaded phrase. I see the argument used to justify unsourced additions to articles like List of musical works in unusual time signatures or the various "List of music in this key" on the different key signature articles. That to me seems like a clear case of interpretation of the primary source. How about if you substitute "obvious to one" with "objectively verifiable by anybody with". I don't even like that, but it's a compromise. Maybe the musical ones should be mentioned specifically in the article? Torc2 (talk) 04:36, 18 January 2008 (UTC)
Your change is good, but I prefer "readily" to "objectively" which to me sounds like it would attract disputes. I don't understand your point about the musical list - I'm no expert, but I assume anyone can tell whether a score is written in 5/8 time? Wnt (talk) 07:56, 20 January 2008 (UTC)
Thank you for that "readily". I like that. It points out that, if we are skilled in the art, we should not have to use a lot of effort to verify a deduction that we as editors are making. The word "readily" eliminates people pushing the envelope to do original research and claim it's "objectively verifiable! you're just ignorant in this art!" And of course we want to minimize conflicts based on interpretation as much as possible. Wjhonson (talk) 09:24, 20 January 2008 (UTC)
If they have a printed score, they can. If they're interpreting the meter from listening to the song, that's definitely OR. They're interpreting something from the original source. How do they know it's 5/8 and not 5/4, or 10/8 or 15/8, or 10/16. With key signatures, it can be even more of a stretch, and is OR even with a score - two sharps in a key signature can indicate either D major, B minor, or a number of possible modes (E dorian, F# phrygian, etc.). Torc2 (talk) 09:24, 21 January 2008 (UTC)
But is that correct? A source is not simply printed material. We have source references that are audio, and sources that are videos. If we watch a video and say in our summary "Bob took out a hammer and hit Mike with it.", do we need a printed transcript or stage directions to state that this is what the episode shows? If a typical viewer would state the summary is essentially the same way, then this is not original research rather it is source-based research. We are using the source to paraphrase in text, what the source states. We are not creating the fact of what the episode shows, we are only restating it. In this particular case, we are saying that if a person "proficient in the art" would state essentially the same thing about the source, then we can as well. So if a trained musician listens to an audio tape and would state "this is allegro", then we can state that as well without the need for it to first be printed. Not all sources allowed here, are print sources. Wouldn't you agree ?Wjhonson (talk) 14:27, 21 January 2008 (UTC)
No, not at all. It takes no specialized knowledge or skill, or any interpretation to see a video of somebody being hit with a hammer and say they were hit with a hammer. Can a well-trained musician figure out a moderately complex time signature from hearing it? Maybe (they can also get it wrong pretty easily), but Wiki has no way of verifying the credentials of the person who added the information, and no way for the average editor to verify the material itself. Can they say "this is allegro"? No, because allegro is an interpretive term; how do you know the score isn't marked presto and they just took it too slow? WP:OR is exceptionally clear about this, saying information based on primary sources should be descriptive only, and: "make no analytic, synthetic, interpretive, explanatory, or evaluative claims about the information found in the primary source." How do you figure out a key? Harmonic analysis. Looking at a score and seeing a meter marking is a description of the primary source; listening to the song and figuring out the meter from that is an evaluation of the primary source. Torc2 (talk) 18:57, 21 January 2008 (UTC)
It sounds like these musical decisions are neither obvious nor readily verifiable by anyone skilled in the science, so the policy would prohibit them. It is of course possible that an editor could make a mistake and overestimate his own abilities, but that is always a possibility. Wikipedia assumes good faith and trusts the editors. The proposed policy is not asking for contributions based on unique world-class expertise, just skill - skill that can be verified. There are probably more skilled biologists, chemists, physicists who edit Wikipedia now than there were editors of all varieties the first year of its operation. That was no catastrophe, and neither would this be. Wnt (talk) 19:43, 22 January 2008 (UTC)

Interpretation of images

I noticed a possible loophole/room for misinterpretation in the section on original images, in that it did not cover requirements on the interpretation of images. I suggest adding a short paragraph to emphasize two points that should be uncontroversial:

  • The interpretation of images must be based on specifically related verifiable, reliable sources. (cf. WP:V)
  • Images must not be used out of context or in contexts for which reliable sources do not exist. (cf. WP:NOR#Sources).

I suggest the following wording:

Editors should note that the interpretation of images, particularly depictions of allegorical or symbolic works, is subject to this policy. Descriptions in articles and captions of the meaning and context of an image must draw on reliable sources specifically relating to that image. Images must not be used out of context or in contexts that are not supported by reliable sources.

This is a straightforward reiteration of long-standing existing policy - there's nothing new about either of these principles. It is, however, useful to make it clear in this section that the text that accompanies images is subject to the same requirements as any other text anywhere else in an article.

I've had some personal experience of this issue. Some years ago, I rewrote the article on the Vinča signs to which an editor had added a series of images to support a fringe theory from an unreliable source that the signs were an early version of the Etruscan alphabet (see [12]). The article has been much improved since then, but the issue of the misuse of the images is something that's stuck in my mind for a long time. It's worth emphasizing to editors that our content policies apply to all material, not just text. -- ChrisO (talk) 19:01, 21 January 2008 (UTC)

Why the rush to push this through undiscussed, Chris? SlimVirgin (talk)(contribs) 18:55, 21 January 2008 (UTC)
If you have a substantive objection to the proposal, could we please hear it? As I've said above, there's nothing novel here. -- ChrisO (talk) 19:01, 21 January 2008 (UTC)
What do you mean by "images must not be used in contexts that are not supported ..."? What is the difference between that and "out of context," and what do you mean by "out of context"? SlimVirgin (talk)(contribs) 19:05, 21 January 2008 (UTC)
I'll give you a hypothetical example by way of illustration. This image depicts the US Capitol Building. If it was added to Fascist architecture, it would be displayed out of context, as there's no conceivable reliable source that would describe it in that way. On the other hand, if it was added to List of buildings where people have been assassinated that might be an appropriate context, as it's conceivable that someone was indeed assassinated there (I have no idea if that's the case or not). However, if there's no reliable source to establish that context the image shouldn't be used. It's the difference between a patently inappropriate context and a context that might be appropriate but is lacking sourcing. -- ChrisO (talk) 19:18, 21 January 2008 (UTC)
That kind of thing is already covered by this or V. SlimVirgin (talk)(contribs) 19:27, 21 January 2008 (UTC)
If you want to use that example again, use this image, depicting San Francisco City Hall. Mayor Moscone and Harvey Milk were assassinated there. --Hugh7 (talk) 01:19, 27 January 2008 (UTC)

I tend to think that this is covered already by the inclusion of "photographs" in the PSTS section. That section explains how primary sources, like photographs, should only be described, rather than interpreted. The section on original images, on the other hand, is about diagrams that users create for the express purpose of putting on WP. It isn't meant to cover existing photographs. — Carl (CBM · talk) 19:07, 21 January 2008 (UTC)

The section on original images does actually specifically refer to photographs, including an entire paragraph about photomanipulation - something that usually affects photos of people or events. So it's plainly about more than just users' diagrams. -- ChrisO (talk) 19:21, 21 January 2008 (UTC)
It's about original images, rather than text. What you're talking about seems to be the interpretation of primary-source material, which is already covered. But in addition you want to add material that's very unclear -- out of context, and contexts not supported by sources. You'll have to say what you mean so we can judge the implications. SlimVirgin (talk)(contribs) 19:25, 21 January 2008 (UTC)
I just explained it above. What do you consider unclear about it? Perhaps you can suggest an alternative form of words? -- ChrisO (talk) 20:58, 21 January 2008 (UTC)
I think it would end up being used as a stick to beat people with. (I'm not suggesting you intend this, of course, but I think it would be a byproduct.) For a start, no sources are going to be available to supply a context for many of the images taken by editors themselves, which won't have been published anywhere. We want to encourage original work with images, not discourage it. SlimVirgin (talk)(contribs) 21:04, 21 January 2008 (UTC)
ChrisO, you're right it does mention photos. For example, if I write an article about a hiking trail, I am permitted to take a photo of the trail to use in the article, and the OR policy is not meant to prohibit that. But if I manipulate my image (say,to remove some trees to make the view look better), that would be discouraged. I am not claiming that this section is worded perfectly already - it seems that rewording it may help. But I also don't think your concerns, about interpreting images, belong in this particular section of the policy, about creating images specifically for WP. — Carl (CBM · talk) 19:28, 21 January 2008 (UTC)
And if you did add a photograph of a hiking trail, according to the proposed addition, you might need to find a source that supported its use in that context.
The issue of original work and image use on Wikipedia is very complicated -- too complex for this page, I think. SlimVirgin (talk)(contribs) 19:33, 21 January 2008 (UTC)
Oh no. My entire understanding of the purpose of this section was to point out our practice that I do not generally need another source to add a picture I take of a hiking trail to the article about the hiking trail. There is still the issue of consensus, but not of verifiability. — Carl (CBM · talk) 19:37, 21 January 2008 (UTC)
As it's written now, yes. SlimVirgin (talk)(contribs) 19:43, 21 January 2008 (UTC)
That's not the intention of my proposal (perhaps it needs modifying to clarify this?). The aim is to ensure that where the meanings of images are not immediately obvious, or are capable of being disputed, descriptions and interpretations should be based on reliable sources. -- ChrisO (talk) 21:05, 21 January 2008 (UTC)

I would say, if, for an article on William Shatner, I take a picture of William Shatner, I can say "This is a picture of William Shatner"Wjhonson (talk) 19:47, 21 January 2008 (UTC)

To clarify: one would not need a third source to verify that the person on the photo is indeed William Shatner. If it's William Shatner dressed as a bear, we might need verification, however. I can see cases where there will be disagreements. If I submit a picture of a person who has a passing resemblance with Shatner apparently passed out in a ditch, we'd probably need verification. But I suggest we take this on a case by case basis, rather than try to see how many distinctions between "interpretation" and "description" we can fit on the head of a needle. --Leifern (talk) 20:29, 21 January 2008 (UTC)
Of course, there's no need to cite a third source where the subject of an image is obvious - be it William Shatner or a hiking trail. But it's often the case that the subject of an image isn't obvious, particularly where it concerns an allegory or symbolism of some sort (which is why I mentioned those aspects specifically in the proposal). I'll give you two examples. Consider the painting of the Raft of the Medusa. It's a very well-known image, but it's not immediately obvious from the picture that it's a political allegory - the captain of the wrecked ship was a supporter of Napoleon Bonaparte, and it was painted the year after Waterloo. The image's context gives it a specific meaning, but documenting the context and meaning requires reliable sources that specifically relate to the picture. Reliable sourcing is particularly important where multiple interpretations have been advanced or there's some controversy over the meaning, as in the case of The Last Supper (Leonardo) (anyone who's read The Da Vinci Code will be aware that there are some crackpot theories about the painting!). -- ChrisO (talk) 20:54, 21 January 2008 (UTC)
I don't know... I mean I do agree in principle with you Chris. I'm just not yet convinced there is a pressing need to state this. I am open to being more convinced. It would probably take an actual ongoing dispute to illustrate the problem in dramatic relief. We do still have problems with people not understanding a lot of things, but so far I haven't encountered picture captions as being a subject of revert-warring. Wjhonson (talk) 21:11, 21 January 2008 (UTC)
Raft of the Medusa is not an original work created for WP. Maybe that section needs to explain that it doesn't refer to digitalization of other works, but to works created specifically for use on WP, such as photographs taken by WP editors and diagrams created to illustrate articles. — Carl (CBM · talk) 21:18, 21 January 2008 (UTC)
I don't mean to nitpick, Chris, but both the examples you mention are not original work, as CBM points out, and I think the caveat you want to include is pretty well covered in the clause that original work does "generally ... not propose unpublished ideas or arguments." So to the extent that the caption does assert something that is truly original ("an alien that has disguised itself to look just William Shatner," or "The US Capitol building, built at night by leprechauns enslaved by Mrs. Coulter," we have an issue. There's actually a recent example where I uploaded a photograph that I thought I took of Qiryat Shmona in Israel, which turned out to be somewhere else. An editor pointed out my mistake, and I'll correct it, but the presumption was that I wasn't lying about the caption. We generally assume that image descriptions include the plain meaning of the image. --Leifern (talk) 21:30, 21 January 2008 (UTC)
Wjhonson, I'll ask a few editors I know who're active in contentious areas of Wikipedia to see if they can furnish any examples of active ongoing disputes. CBM, I think the scope needs to go wider than that; most of the images on WP aren't in fact taken by WP editors. I was thinking more of images from external websites and sources such as Flickr. Leifern, if you don't mind me saying so, I think that's a bit of a fanciful example. :-) I had in mind something more mundane - such as a scenario where the meaning of an image is disputed by editors because it isn't reliably documented. That was essentially the problem in the article I first mentioned above, Vinča signs, until I and other editors were able to find reliable sources on the images on question. -- ChrisO (talk) 22:02, 21 January 2008 (UTC)
But when you write "meaning of an image" and "image...isn't reliably documented," the distinction gets lost again. The fanciful example involving leprechauns and Mrs. Coulter was an overstatement to make a point. Virtually all images, and certainly all original images, may have incorrect labeling for one reason or another. But when I uploaded a picture of the Madison, NJ city hall, nobody challenged my assurance that it was what I claimed it to be. That's one extreme case on one hand; on the other is the US Capitol Building/leprechauns/Mrs. Coulter example. I think the distinction is nicely captured in the clause that the caption or description should not contain unpublished ideas or arguments. What "meaning" is in this context is a philosophical discussion we should refrain from. --Leifern (talk) 01:29, 22 January 2008 (UTC)

(Outdent) Thanks Chris, I think it would be helpful to have concrete examples. Btw, are you certain we can even *use* images from Flickr? I was under the perhaps false impression that all images here, not taken/created by ourselves, had to be so old they were out-of-copyright. That otherwise we could not guarentee they were copy-free. Wjhonson (talk) 22:24, 21 January 2008 (UTC)

Yes, we can use Flickr images, provided that the copyright is compatible with Wikipedia's GFDL license - this basically means they have to be licensed under the Creative Commons. There's even a bot called FlickrLickr running on the Wikimedia Commons that identifies and automatically uploads compatible images. According to the FlickLickr page, there are about 2 million such images on Flickr. You can search through them yourself from this link - simply enter your search query and check "Only search within Creative Commons-licensed photos" and the two options below to find compatible images. I've found it to be a very useful resource for illustrating articles. -- ChrisO (talk) 22:34, 21 January 2008 (UTC)

I don't follow any objection to Chris's proposal other than the same anti-elitist sentiment which desires for WP:OR, WP:FRINGE, and WP:RS to either be deleted or be as unclear, undetailed, and utterly confusing as possible. Several points:

  • Response to SlimVirgin: "Why the rush to push this through undiscussed, Chris?" See WP:BOLD and WP:Assume Good Faith.
  • Even if Chris's specific examples are incorrect, that's irrelevant. Images are forms of information, no different than text, and can still be used for original research and it should be clarified that it isn't allowed, to discourage people from saying, "WP:OR doesn't apply to images" in discussions.
  • Response to Wjhonson: Wikipedia policy is developed in order to help Wikipedia run more effectively, not to protect it from total apocalyptic collapse from massive edit-warring or a mass exodus of good editors. If a proposal to clarify policy makes sense, why object to it? You acknowledge it makes sense, so why the obstruction? The precautionary principle? Or are you a fan of Edmund Burke, perhaps? It doesn't matter whether there are hordes of people already misunderstanding policy in that way, nor is there an inherent necessity to find concrete examples, because even addressing the potential existence of 3% of users who might use images for WP:OR (though we just might not have found them yet, considering Wikipedia's size) would help build consensus and make Wikipedia closer to being worthy of being called an "encyclopedia." If you acknowledge that's what following this policy would seem to do, it's not clear why it should be opposed. To summarize your argument, "This is a good idea, but until I see examples of where Wikipedia has failed in this way, I'm going to vote no." Imagine if you had made this same argument regarding vandalism of certain pages before Wikipedia developed a protection policy. Should we wait until we have large flame-wars and users getting blocked over unclear policy before clarifying it?   Zenwhat (talk) 22:35, 21 January 2008 (UTC)
I simply don't care for hypotheticals whose end-result creates a change in our policy wording. I find that, in general, people use hypothetical cases to effectually argue for a pet project, and then turn and use the letter-of-the-policy to push for something which violates the spirit-of-the-policy. Which then leads to another battle and more instruction-creep to account for the new discrepancy. In general, I'm not opposed to instruction-creep to alleviate real issues, but I'm not a fan of it in the case of a presented hypothetical case. You're wrong Zen that I'm voting no. You, as a person named Zen, should be the ultimate arbiter that it's possible to vote "no answer yet, please come back tomorrow"Wjhonson (talk) 22:44, 21 January 2008 (UTC)
I'm certainly sympathetic to your concerns, Wj. As I said, I'll try to document at least one real-world examples of such a case so that we can discuss the proposal in that light. I would say, though, that one of the key functions of Wikipedia policy is not simply to fix problems when they arise but to prevent problems from arising in the first place. That's the reason why, for instance, I devised Wikipedia:Naming conflict, to provide a template for how to deal with disputed names and head off future disputes. But I'd also reiterate that my proposal isn't actually a new policy at all - it's merely a brief restatement of an existing requirement, clarifying a point of WP:NOR that I felt was ambiguous. -- ChrisO (talk) 22:54, 21 January 2008 (UTC)
Well one obvious case springs to mind. The Zombietime image you would like removed:
I think it's a very bad idea to attempt to alter policy as a means of winning a content dispute or WP:CANVASSing support for the change. Why don't you let the mediation on the subject go ahead, rather than trying to pre-empt the decision with this policy change? Perhaps you should add yourself to the list of involved parties. <<-armon->> (talk) 01:19, 22 January 2008 (UTC)
I see that I've arrived a bit late to this discussion, but it may still be worth noting that the Zombietime image in question has been the subject of extensive discussions in recent weeks. Several editors have raised objections to the current placement of the image, with some of the discussion revolving around the issue of Original Research. Readers are encouraged to review the relevant discussion at Talk:New antisemitism for additional context. 06:46, 22 January 2008 (UTC)
??? What are you on about, Armon? I'd completely forgotten the first discussion you cited, but I never sought to remove the image - I simply questioned its sourcing. I've had no participation in the second discussion other than seeking to clarify its copyright status (which SlimVirgin kindly did by e-mail) and the third URL you cited is simply the request I posted to a number of editors of my acquaintance who've had involvement with controversial articles. I'd appreciate it if you withdrew your remarks above, since they're untrue and uncivil, and a gross violation of WP:CIV and WP:AGF. -- ChrisO (talk) 08:03, 22 January 2008 (UTC)
And what mediation? I'm not involved in any that I know of (unless someone has added my name to one without telling me). I came here from Talk:An Inconvenient Truth#Error lists, where I've been involved in a discussion with an editor for the last two days about why we can't cite his personal website. For obvious reasons, I've been reviewing WP:NOR in connection with that. -- ChrisO (talk) 08:45, 22 January 2008 (UTC)

Eh, I generally avoid policy debates and I'm not clear what underlying content issue is being discussed here. However, I received a note from ChrisO asking if I had any examples of images being used to push original research. As it happens, I do. Months back, an editor insisted on adding an image to Terri Schiavo, as a counterpoint to her CT scan. The image was a CT scan of a man with Dandy Walker syndrome, a congenital brain anomaly. This man's scan had some features in common with Terri Schiavo's, and he was living a relatively normal life as a civil servant in France with a near-normal IQ (no jokes please; they've all been made). The implication, by placing the scans side-by-side, was that Terri Schiavo's scan did not preclude near-normal brain function. This ignores medical reality - Dandy Walker is a congenital anatomical defect, while Terri Schiavo was neurologically normal until a devastating neurologic event; in addition to the disparate clinical scenarios, the CT scans are only superficially similar; the man with Dandy Walker was not typical even of that syndrome, but so remarkable as to be written up in the Lancet and the popular press; and needless to say, no reputable medical or lay source (actually, no source at all) attempted to juxtapose the two situations or images. Discussion is archived here. Needless to say, I think this violated WP:NOR. Does that mean we need to amend policy? I dunno. Outside opinion ended up shooting down this particular editor, who was problematic enough that he's currently under ArbCom sanction, with the existing NOR policy. And this is actually the only case I can think of offhand, so this may not be a particularly widespread problem. I'm just offering it in response to the request for case studies in original-synthesis-via-image. MastCell Talk 04:53, 22 January 2008 (UTC)

Thanks for the example, but in your particular case I don't even think you need to invoke OR, you can just use undue weight. A scan of someone else's brain has almost no relevance to an article about Schiavo. I appreciate the long description though, very interesting. Wjhonson (talk) 05:59, 22 January 2008 (UTC)
That's a very interesting example - thanks for the input. In the case you cite, an image was being used out of context - i.e. one medical complaint being used to illustrate another one, without any reliable sources to make a valid connection or comparison between the two. It's the kind of scenario I had in mind for the line in the proposal that reads "Images must not be used out of context or in contexts that are not supported by reliable sources." I'm not quite sure it's covered by WP:UNDUE, but it's certainly an example of implicit OR. -- ChrisO (talk) 08:50, 22 January 2008 (UTC)
I see two problems with your rational for this change. 1) Current policy would seem more than adequate to address the problem that this change is designed to address. If current policy is not sufficient, then you have failed to articulate in any clear manner why it is not sufficient. 2) Far more problematic is your claim that this change is not an organic evolution, but the result of a hypothetical situation. Policy should be based upon actual experience and should address those situations that have occurred on at least a semi-regular basis. If this change addressed a real need or closed an actual loophole, it should be easy to show multiple clear examples to substantiate the need for a change. Baring that, we do not amend core policy to better address hypothetical situations -we do not supply solutions that are in search of a problem. We fix those things that are broken; doing otherwise creates more problems. Brimba (talk) 18:14, 22 January 2008 (UTC)
Agree. Are there any examples of discussions where the existing policy has proved to be insufficiently clear? If not, let's keep it as it is. Andrewa (talk) 21:30, 22 January 2008 (UTC)

Wjhonson: Your claim seems to assume bad faith.

As for the repeated invocation of wikidarwnism by multiple users above: I'm not a piece of DNA, I'm not a mindless chunk of living tissue to be pushed around or mutated by violent aggregate forces (flamewars, edit-wars, etc..) -- I'm an individual human being who deserves respect, courtesy, to be listened to carefully, and the benefit of the doubt and because of the inherent irrationality in human behavior, none of that is going to be provided through irrational, pseudoscientific social darwinist theories applied to Wikipedia policy. Herbert Spencer, if he were alive today, would not be able to edit Wikipedia because he was a dick. Let's not revive his ideas here, please.

If a policy proposal makes sense, it's totally illogical for you to say, "But I want evidence that we need it," before supporting it. As noted above, it was this obfuscation that prevented there from being a protection policy for such a long time and it's basically a form of Conservatism. Because humans are so irrational, the more explicit we make policy, the better. That's partially the reason why policy pages exist to begin with, not simply "empirical observations," because human behavior is so complex and there is so much stuff on Wikipedia that there can be a multitude of totally different problems which all arise from one single inappropriate comma on a policy page.

When one attempts to remove this comma, they are reverted and asked, "Why did you remove this comma? Can you PROVE that removing this comma actually benefits anyone?"

It would be like the same fallacious argument that we should wait for people to die before we do anything about global warming. I mean can anyone show me anyone who has been hurt by global warming, such that we need the Kyoto protocol? No. Because the problems of global warming are based largely on inference, which are just as valid as direct observation. We don't need to wait until we have thousands of people spamming Wikipedia with tubgirl before making "Don't spam Wikipedia with tubgirl" a policy.   Zenwhat (talk) 07:24, 24 January 2008 (UTC)

Sorry I'm not that person. I am assuming the change is requested in good faith, but I simply disagree with requesting changes based on hypothetical situations. I'm not in control of this board, I'm just voicing my own personal opinion. I need clarification with actual examples. Btw I don't revert people who make grammar corrections. Wjhonson (talk) 09:12, 24 January 2008 (UTC)