This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 15 | ← | Archive 20 | Archive 21 | Archive 22 | Archive 23 | Archive 24 | Archive 25 |
We need to resolve certain disputes. Based on the past week's discussion and comments from vaious people, I am singling out two specific disputes, which is more than enough to work on for now. Once we resulve these, people are free to raise other issues which we can work on, one at a time, until we resolve them.
Well, there are currently two "quality of sources" locations: the actual guideline article and the NOR policy page. Much of the current controversy arises from having both a guideline and a separate policy section. Apparently the guideline doesn't have enough compelling force (but that's a personal interpretation that I'd welcome having replaced with a better characterization.) That dual-location situation is in part what creates a "can of worms." -- Minasbeede 14:51, 31 August 2007 (UTC)
I have now slept on it and feel even more strongly than before that the NOR policy must make a distinction between primary and secondary sources. I think Slim Virgin is right (and I do not always agree with her and long questioned her on this). NOR needs to explain what "original research" is, and what it isn't and I do not think there is any way to do this without discussing primary and secondary sources.
Some people do not understand the distinction or think it opens up a can of worms but I now believe that is because they think too literally. What I mean is, some people think that things are what they are. I do not mean to go into a long philosophical debate but Wittgenstien and Pierce and Dewey - generally acknowledged to be among the most important philosophers ever in the US and Europe - argued that the meaning of things including words depended on how they are used. To a large degree, what makes something a primary or secondary source is, how it is used. Primary sources can be used to establish facts without violating NOR, but they cannot be used to make analytic, synthetic, interpretive, or explanatory or evaluative claims without violating NOR. Secondary sources by contrast are the only sources we can use to make analytic, synthetic, interpretive, explanatory, or evaluative claims (as long as they are the claims explicitly made by the source itself).
We can, did, and I think still should provide examples of primary and secondary sources, but they are just examples. The essence is in how they can or cannot be used.
This by the way means it is conceivable for a text to be either a primary or secondary source - depending on how it is used. Let's take On the Origin of Species for example - I grant this may not be the best example, but I am just looking for a handy hypothetical. In it Darwin makes all sorts of analytic, synthetic, and explanatory claims about natural history. As a document concerning natural history, it is a secondary source ("According to Darwin, variation in finches on the Galapagos is best explained by ..."). But in it he also expresses his own beliefs and judgements. As a document about Darwin, it is a primary source ("According to Janet Browne, when Darwin wrote On the Origins of Species he was responding to ...."). In the article on Evolution we can use it freely as a secondary source (although evolutionary theory has gone far beyond Darwin, evolutionary scientists still consider the case he built to support his theory valid). In the article on Charles Darwin it is a primary source and while we can use it for facts, if we want to make arguments about what Darwin was like or what he meant, we should draw on appropriate secondary sources by historians of science, intellectual historians, and biographers of Darwin. I do not see this as a can of worms. I just see this as a more sophisticated way of thinking that understand things in terms of the ways they are used, and understands that appropriate uses depend on the context. Slrubenstein | Talk 10:43, 31 August 2007 (UTC)
Well it was chosen in haste and perhaps a bad choice, although I think you misunderstand. If someone wanted to make a claim about variation among finches, I would argue that Darwin's book is a valid secondary source. But there is no need to belabor this one example. Can you think of any texts that could be a primary source in one context, and a secondary source in another? That is the issue. If you can't think of any, I would just take that as a sign that the boundary between primary and secondary source is even clearer than I thought! What is your answer to the question that heads this section? You do not have to take a position on what I say; the question is your position on the current policy. Slrubenstein | Talk 17:37, 31 August 2007 (UTC)
Folks, for some this is a very contentious issue. I think we first need to resolve the basic question of whether the policy must or should make this distinction first. Then discuss editorial issues. Sambc, if you want to, creat a policy proposal page and invite discussion of your idea there. But here, the question is simple: should NOR distinguish between primary and seconday sources? That's all! Slrubenstein | Talk 12:38, 31 August 2007 (UTC)
I think the basic issue of NOR is that it's difficult to explain. Personally, I feel like it's kind of a "I can't really define it, but I know it when I see it," kind of thing. I've previously indicated that I think the idea that novel synthesis is banned by NOR is pretty ridiculous - an encyclopedia article is by definition a synthesis, and doing things like summarizing books, or trying to describe the state of the literature on a subject, are bound to involve some degree of original synthesis. It is novel analysis that we need to be careful about, and this is what we shouldn't be able to use primary sources for. The classic OR using primary sources is essentially to go to, say, the text of some primary source and then to quote that text, out of context, to support some new argument which is not explicitly made by the primary source. But I don't know that this is something that is unique to primary sources - one can easily do the same thing with secondary sources. I think the key issue here is not that primary sources are forbidden for any particular purpose. It's that, in evaluating primary sources, we need to stand by the judgments of reliable secondary sources. john k 17:36, 31 August 2007 (UTC)
Well, so far it appears that everybody is in agreement that a distinction should be made in regards to the types of sources. I didn't see any objections above that a distinction should not be made. The only objections I see (maybe I'm biased or blind), is in the wording of how to label the types of sources, and whether or not the "Policy" page should attempt to offer a definition of the types of sources.
Personally, I feel that the details of what the various "source types" are should be defined on the separate pages already created, and this "Policy" page merely point out that different types exist and here's the link for the definition. This page is solely about NOR and should not expand (rightly or wrongly) into the realms of other subjects, no matter how contentious or how badly needed. When it's necessary to delve into those realms, link to that subject appropriately and say something like "For more clarification or discussion, see ...". Keep each "Policy" and/or "Guideline" focused only on that subject without blurring the lines of distinction. wbfergus 17:49, 31 August 2007 (UTC)
Can the (Wikipedia) definitions of "primary" and "secondary" be made complete enough and thorough enough that they suffice to define "primary" and "secondary," as used/needed by the NOR policy? -- Minasbeede 20:29, 31 August 2007 (UTC)
I think the first order of business would be to stop using the words "primary" and "secondary" in a way different from how they are used outside Wikipedia. Choose/develop a different way to distinguish the usages from the Wikipedia point of view.
(Something about NOR I decided to move to my talk page, where it may languish and die, was here.) Minasbeede 13:32, 31 August 2007 (UTC)
There's no one simple usage outside of Wikipedia, which is part of the problem. But for our purposes, can't we just focus on the terms "primary" and "secondary" and just say that a primary source is a source that is "primary"? In other words, an original source? A secondary source is a source that is "secondary". In other words, not primary. Of all the ways the terms are used outside of Wikipedia, that seems to be the core idea. The original source of an idea, conclusion, or set of data is the primary source, while any source of that same idea, conclusion, or set of data that isn't the primary source, is a secondary source.
We could also switch to neologisms, such as "reliable source" and "auxiliary source", but I don't see the point, since reliable source is already well defined at WP:RS, and the new term auxiliary source would just be something that isn't a reliable source. COGDEN 21:15, 31 August 2007 (UTC)
How about "sources of facts" and "sources of interpretations"? -- Coppertwig 22:52, 31 August 2007 (UTC)
I would say, not exactly. It should give advice on the appropriate use of primary vs secondary sources, as primary sources are much more likely to lead to inadvertent synthesis, but can be used without it being OR. SamBC( talk) 12:34, 31 August 2007 (UTC)
No or not without major exceptions. I noted earlier that the Communist Manifesto is a primary source on Marx's and Engel's beliefs, and a secondary source on their various opponents' beliefs. It is a trustworthy description of their own beliefs, and a rather untrustworthy description of their opponents' beliefs. I think many such works are reliable when used as primary sources and unreliable when used as secondary sources, not the other way around. Jacob Haller 20:32, 31 August 2007 (UTC)
No. The policy should not privilege either primary or secondary sources. The pros and cons of each are to complicated to be found in a policy. Maybe some distinctions can be introduced first in an essay, then possibly in a guideline, but there should not be any blanket preference or discouragement of either in this most important of policies. The archived discussions brought out plenty of instances where primary sources are far preferable to secondary sources. But the main reason I don't think there should be a preference is that we can't yet agree on what the difference is between a primary and a secondary source. It makes no sense to say that one or the other should be "rare" or "relied upon" when we don't agree on what either means. Also, any stated preference of one over the other should describe widespread Wikipedia practice, not lead it (see WP:Policies and guidelines). This policy cannot change Wikipedia practice. It is bound by that practice. COGDEN 21:02, 31 August 2007 (UTC)
The policy cannot distinguish between primary and secondary sources unless there is a consensus as to the difference. In my view, it's pretty simple: a primary source is something original, right from the horse's mouth, while a secondary source is a re-telling of a primary source. In another sense in common usage, a secondary source is a source that discusses a primary source; however, any new conclusions by the secondary source is also a primary source as to those new conclusions. On this basis, there is absolutely no reason to favor secondary sources, unless the primary source is unpublished, in which case it would not be a reliable source.
Moreover, since polices must describe widespread Wikipedia practice, not lead it (see WP:Policies and guidelines), the fact that primary sources (such as journal articles, interviews, and fictional works) are widespread, and even more likely to exist in featured articles than in non-featured articles, would seem to show that if any distinction is to be made between the two, primary sources should actually be favored. Personally, I don't think WP:OR should make a decree either way. COGDEN 20:49, 31 August 2007 (UTC)
No, no privileging of one type of source over another. It has been pointed out that many FA use a lot of primary sources. The key thing is that any "analytical, synthetic, interpretive, explanatory, or evaluative claims" come from sources, not from Wikipedians. -- Coppertwig 22:57, 31 August 2007 (UTC)
Agree with Coppertwig. Get guidance from guidelines. If people come to the policy statement for guidance refer them to the guidelines from within the policy. That way there'd also be just one set of descriptions, not two. There could even be an examples page (that provides further guidance) of what meets the criteria and what doesn't, with an explanation of why or why not. These could be real cases, if that doesn't offend any of those involved. The examples could be slightly (or greatly) altered to avoid offense, if offense to someone is a problem. -- Minasbeede 03:26, 1 September 2007 (UTC)
There are already phenomena for which Wikipedia itself is a primary source, such as the Bogdanov affair, and since the page histories record edits as they occur, they are thus a reliable source of which editor accounts were used to write what things. The problem with OR or usage of primary sources might be biases or the creation of hoaxes. In some of those cases, the policy is justified. In some other cases, however, in which a phenomenon develops that numerous bloggers, message board users, etc., notice before the general or mass media does, OR might be the best way of describing the phenomenon, at least as the bloggers or message board users see it. I would suggest the following: if something is quite controversial and argued over by numerous people, the controversy should be discussed comprehensively from all sides; if something such as the cancellation of a TV show provokes an enormous campaign with a rapid accumulation of signatures that suggests numerous contributors, Wikipedia should take note of that as well. Try to give the benefit of the doubt when you can. — Rickyrab | Talk 14:51, 7 September 2007 (UTC)
Let's review. As I understand what I've read here, Wikipedia policies should be descriptive, not proscriptive. Yet so far nobody has been able to accurately describe or define the primary/secondary distinctions (that are made in a special Wiki way and aren't the same as occurs outside Wiki), which looks to me like the policy section that is so very important can't even be enunciated. Surely if it can't be said it doesn't describe anything.
So let me ask for a few examples of articles where this distinction is so important: if it can't be described maybe it can be illustrated. I looked at the Wiki article on "Evolution" last night and found that, to the level I checked, it looked like it was indeed a long listing of facts with nary a word of explanation, analysis, or an any other of the things that Wiki editors are forbidden to do unless they're quoting someone else. I think the Wiki "Evolution" article is a very fine article but apparently (unless I missed something, which is possible) it cannot serve to at all illustrate why the emphasis on the superior nature of secondary sources is so necessary, as is asserted here by some. I'd think that in the over 100 sources cited there are both primary and secondary sources but one might suspect that if all the content is simple facts that all of the content could and did appear first in primary sources. So, what are a few articles that illustrate why the emphasis on secondary (as the term is used in Wikipedia) sources should be given special emphasis? -- Minasbeede 11:26, 1 September 2007 (UTC)
Who asked for a debate? I said that the Evolution article seems to be a fine article yet it also seems to be a serial recitation of facts, something that some assert here is undesirable. Not only does it seem to be a serial recitation of facts I can see why that particular article needs to be that and must strenuously avoid "analytic, synthetic, interpretive, explanatory, or evaluative claims" - so probably it is indeed a bad article to look to for what is claimed. Still, the Evolution article seems to me to be an article that illustrates that the claim made here by some (that you have to use secondary sources in order to include analytical material) isn't universal. So I asked for examples of articles that do show what is claimed.
The title "regroup" is what I came up with. the real motive was to split the section so editing would be easier. Any title is fine - and if a different title is inserted this can be removed. -- Minasbeede 14:09, 1 September 2007 (UTC)
I've been asked to contribute on this topic and don't think that I can't really add any new arguments, as they have been all repeated many times. Instead, I think it would be useful to look at scientific papers - one thing I know well - to illustrate that it is impossible to classify some sources as either primary or secondary.
Take this 2004 paper as an example A trypanothione-dependent glyoxalase I with a prokaryotic ancestry in Leishmania major.
In the fist paragraph of the Results section I quote some dissociation constants I measured. If a Wikipedia article cited these results, it would be using the paper as a "primary source" - a source that presents novel, previously-unpublished data.
In the first paragraph of the Discussion on the other hand, I discuss previously-published data on Methylglyoxal synthase in the light of my results and bring these data together and interpret them. Here, if a Wikipedia article cited the paper on the possible problems caused by methylglyoxal synthase, it would be using the paper as a secondary source.
Scientific papers are neither primary nor secondary sources, they contain elements of both. Tim Vickers 01:28, 3 September 2007 (UTC)
Above, Tim write, "Why not just say that all material must be cited, with facts backed by citations to reliable sources that contain these facts, and interpretations and syntheses attributed to reliable sources that make these interpretations and syntheses?" Every bolicy starts with a box explaining the policy "in a nutshell." I'd have no objection to this being the nutshell, or part of it. Slrubenstein | Talk 15:01, 5 September 2007 (UTC)
At bottom "No original research" means that Wikipedia should not be written as a primary source. This policy used to state this explicitly as: "Research that creates primary sources is not allowed." until this significant rewrite during the time that the controversial depreciation of primary sources was being integrated into the policy.
I think that we should go back to basics and focus on the problem--editors generating new not-previously-published primary source material in WP articles, rather than trying to limit the re-use of existing published primary source material. The arguments about whether basing an article on primary or secondary source material creates laudable or deplorable articles is not relevant to this policy. We are only concerned here with whether the material is new or novel, not whether it is good or bad. Trying to stretch this policy to cover these types of subjective value judgments has only caused confusion and contention over what should be a very simple concept. Dhaluza 17:59, 3 September 2007 (UTC)
I think there should be a page that could be entitled something like "Wikipedia:Wikipedia policy for Dummies" (or possibly something more original) that gives a clear and consice version of the main policies and guidelines that avoids what is known as instruction creep. MrMurph101 03:13, 5 September 2007 (UTC)
We do not need a policy for dummies. What we need is patient editors who are willing to explain things to newbies - that is one of the things that makes this a collaborative project. Slrubenstein | Talk 14:59, 5 September 2007 (UTC)
Many articles on law, science and other complicated stuff have examples to help the reader understand. The examples are not referenced. Are they original research? (For example, "Concepts" section of mens rea). -- Kaypoh 06:26, 5 September 2007 (UTC)
I think the Concepts section of mens rea was not a good example of this type of "example used to explain" (don't know what to call it). Below are a few examples of what I really mean.
See this from Fallacy of many questions:
An example of this is the question "Are you still beating your wife?" Whether the respondent answers yes or no, he will admit to having a wife, and having beaten her at some time in the past. Thus, these facts are presupposed by the question, and if it has not been agreed upon by the speakers before, the question is improper, and the fallacy of many questions has been committed.
And this from Intention (criminal):
For example, suppose that A, a jealous wife, discovers that her husband is having a sexual affair with B. Wishing only to drive B away from the neighbourhood, she goes to B's house one night, pours petrol on and sets fire to the front door. B dies in the resulting fire. A is shocked and horrified. It did not occur to her that B might be physically in danger and there was no conscious plan in her mind to injure B when the fire began. But when A's behaviour is analysed, B's death must be intentional. If A had genuinely wished to avoid any possibility of injury to B, she would not have started the fire. Or, if verbally warning B to leave was not an option, she should have waited until B was seen to leave the house before starting the fire. As it was, she waited until night when it was more likely that B would be at home and there would be fewer people around to raise the alarm.
And this from Mistake (criminal law):
For example, if a defendant goes into a supermarket and places eight items in a basket which is presented to the cashier for payment in the usual way. Both honestly believe that all eight items have been scanned, and the defendant pays the sum shown on the bill. A store detective, however, notices that a mistake was made by the cashier so that only seven items were actually priced. This detective arrests the defendant after leaving the store. Since the defendant honestly believes that he has become the owner of goods in a sale transaction, he cannot form the mens rea for theft (which is usually dishonesty) when he physically removes them from the store. Accordingly, he should be acquitted.
And this from Self-defense (theory):
For example, a person who unknowingly chances upon two actors practising a fight would be able to defend his restraint of the one that appeared to be the aggressor.
-- Kaypoh 14:48, 5 September 2007 (UTC)
I have looked at articles that had a couple of funky tables and then looked at the referenced sources as well as others, and wondered if constructing a single table that merged the data from the various sources would constitute OR.
In the case of oil wells, there is a list of them with data pulled from multiple sources which seems not be OR, while a I have a feeling that creating a similar kind of compilaton for Palestinian census data would be seen as OR. I've concluded for the time being that I will fill in existing tables with info I find, but I won't create any because I now see it as minefield - one's structuring of data is a POV.
But perhaps this should be handled with a policy more like that for images and other original artwork? The objective is to present the diverse facts and analysis in some coherent fashion, right? Mulp 01:53, 6 September 2007 (UTC)
Is it OR to say that X term, that some people describe as a neologism, is not found in the OED or [insert name here] other specific dictionary? Obviously the OED is a primary source in this case and the factual accuracy is easily verified. 136.152.153.36 22:21, 6 September 2007 (UTC)
This seems to have gotten off-track with respect to addressing the issue I was interested in. There is an article that says "X is a term that...". Some verifiable academics say this term is a neologism. Some Wikipedians want to mention that the word "X" was published much earlier in books A, B, etc. Other Wikipedians want to mention that "X" is a not included in the OED or any other dictionary. Both of these statements are based on easily verified facts discovered by Wikipedians, and potentially they both add useful context to whether or not X is a new term (or perhaps a rarely used old term that has taken on a new meaning). So, consistent with the OR policies regarding primary sources, is one allowed to mention both, neither, just one, etc.? 136.152.153.15 21:25, 7 September 2007 (UTC)
The arguments for and against the use of specific types of sources in specific situations need to be discussed more fully, and in the context of WP:OR, WP:NPOV, and WP:Notability. I think we can work toward a potential guideline for this. In the meantime, however, to get a consensus, original research policy that we can all live with, why don't we get back to basics. We can avoid the controversy for now by paring back to the fundamental ideas we all agree with, which I think are represented in the following text, which I would propose as a near-term replacement for the "Primary, secondary, and tertiary sources" section:
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.
When citing a source in support of a proposition in a Wikipedia article, the proposition should (1) be verifiable to the source by any reasonable, educated person without specialist knowledge, and (2) make no novel analytic, synthetic, interpretive, explanatory, or evaluative claims based on the source.
Even if people want there to eventually be more than this in the policy, maybe this is an acceptable step that we can build on. COGDEN 23:59, 6 September 2007 (UTC)
I would like to give an instance that possibly hones the edge of the point being made by the "example" discussion above.
I recently edited an article of a person X where I wanted to say that "X is often referred to on the web as Prof. X". Later I show that his official biography says he was not a prof. This I felt was important to the article, since many readers may come to the article with a view that he is a Prof. An encyclopedia is being written against the backdrop of this common perception, after all.
Now in stating that "X is often referred to on the web as Prof. X", I have no "source" that makes this full claim. However one can cite hundreds of webpages that say he is a prof. It is a claim that is immediately verifiable, and backed up by these sources, which are completely "reliable" for this purpose - ie. they are clearly referring to him as Prof. X.
We are not introducing a new theory / original idea / term/ argument - it is just the immediately verifiable statement that "X is often referred to on the web as Prof. X". My question is - should it constitute original research?
I am referring to the article on P. N. Oak, where I have made this claim recently. If there is a consensus here that this is OR, I will remove the statement. mukerjee ( talk) 09:05, 7 September 2007 (UTC)
David Irving has repeatedly pushed the number 135,000 as the number killed in the bombing of Dresden in World War II. This number like the man is now throughly discredited. See
The trouble is that the number still appears in tertiary sources like encarta Dresden. This is not a problem at the moment because of the reliance on secondary sources. But how would one handle this if the new wording for sources is used? See Talk:Bombing of Dresden in World War II#Population statistics and use of napalm.
In a second case take an example from a primary source like Protocol I that says "The Parties to the conflict shall take all feasible measures in order that children who have not attained the age of fifteen years do not take a direct part in hostilities and, in particular, they shall refrain from recruiting them into their armed forces." Unless one is an expert on international law the nuance of the words "feasible measures" will be lost unless one had read the ICRC commentary on this that makes clear, this is not a complete ban on the use of children in conflict. The ICRC had suggested that the Parties to the conflict should "take all necessary measures", which became in the final text, "take all feasible measures" which is not a total prohibition on their doing so because feasible should be understood as meaning "capable of being done, accomplished or carried out, possible or practicable". See Military use of children#International humanitarian law for sources on this.
It seems to me at first glance that doing away with primary, secondary and tertiary sources may open up a can of worms and I would like others who are considering these radical changes to comment on how the new proposed policy will deal with these types of issues. -- Philip Baird Shearer 20:43, 20 September 2007 (UTC)
Editorial judgement is a problem, is Encata not as reliable source as Götz Bergander in this case? Under the current rules the answer is no. But under the new rules who says? Just look at the section on the bombing of Dresden given above to see the potential problems this opens up. -- Philip Baird Shearer 21:04, 20 September 2007 (UTC)
As to the second point about OR yes under the current rules this is clear (because it is a primary source), but under the new proposed rules, how do you tell the difference between a summary of a source and interpretation of a source given that there is no specific difference between a primary and a tertiary source? -- Philip Baird Shearer 21:12, 20 September 2007 (UTC)
Blueboar, I understand the arguments that both Vassyana are presenting, and as I said above it may be that the current section should be moved into WP:V, but if it is removed and replaced with the suggested text then it will also alter the reading of other sections of the content policy and guidelines. For example I think the section WP:V#Sources would need a rewrite, because quite obviously a treaty is a more reliable text on that treaty than any "peer-reviewed journals and books published in university presses" as are court transcripts and UN Security Council resolutions, all of which are quoted in articles other than those about themselves. I also think that WP:RS#What is a reliable source? is affected in a similar way. So before any change is made I would like an analysis made of how such a change would effect the reading of the whole content policies and guidelines.-- Philip Baird Shearer 11:35, 21 September 2007 (UTC)
This discussion about changing the emphasis on secondary and tertiary sources rather than primary sources is somewhat misplaced and misguided, IMO. Primary sources that report different empirical or historical calculations than widely reported, as was mentioned in the example at the beginning of this thread (whether 135,000 is = or ≠ to the number killed in Dresden), fall under WP:NPOV#Undue_weight and WP:RS, which work in tension with one another. Primary soucces are already stated in WP:NOR to be validly used to report plain facts without interpretation or synthesizing of ideas that may arguably follow from the plain facts. Moreover, WP:NOR, in its emphasis on secondary and tertiary sources for interpretive and synthetic issues, works in tension and in tandem with WP:VER and WP:RS, along with WP:NPOV or course. These tensions are natural and intentional, and the founding Wikimedia Board would appear to have clearly understood this when the three basic editorial policies were first put forward. WP:Consensus, of course, is the method by which these tensions are balanced or resolved in the process of writing the articles. The push to alter this balance by substantially rewriting one or more of the three most basic editorial policies neglects that the solutions to the dilemma already exist within present WP policy. The dog should move the tail here, not the other way around. ... Kenosis 13:54, 21 September 2007 (UTC)
As to my "let's be at least halfway real..." comment, this refers to what appears to be the present notion that this is just a minor proposal that can be implemented without further scrutiny because several WP users have worked on it and believe it is nearing consensus.My review of the situation indicates that this is in fact intertwined with an effort to significantly change the policy, up to and including rewriting the rules on "original synthesis" and other major changes. As to the "process" being referred to by Sambc, that process has been a discussion among several WP users, indeed properly regarded as a preliminary discussion, advocating one or more significant changes in WP policy w.r.t. WP:NOR. The "tail attempting to wag the dog" here is my reference to the exceptional cases driving attempts to change the basic, fundamental balances of WP:NPOV, WP:VER and WP:NOR. The use of anecdotal examples of users having problems in interpreting and resolving issues in individual articles, such as, e.g. the Dresden bombing example in this section, are properly regarded as a {{sofixit}} issue at the individual articles, not valid reasons to change the policy across the wiki. ... Kenosis 16:47, 21 September 2007 (UTC)
If we're writing an article about, say, Encarta, we have to treat what it or its authors and publishers say about it as primary material, and aim to base interpretations or opinions on that material on third party secondary sources. That's widely understood from the current policy, and should not be lost in this push to avoid "source typing". ... dave souza, talk 21:25, 20 September 2007 (UTC)
WP:NOR is one of WP's core content policies and has stood in essentially the same form for a very, very long time, and benefits from broad acceptance. It is uncontroversial and broadly accepted. I can tell you for a fact that any attempts to alter it's original meaning and formulation, such as diminshing the distinction between primary and secondary sources, are going to fail regardless of any agree between the handful of editors participating in the discussion here. FeloniousMonk 14:31, 21 September 2007 (UTC)
In most cases, Wikipedia articles include material on the basis of verifiability, not truth. That is, we report what other reliable secondary sources have published, whether or not we regard the material as accurate. In order to avoid doing original research, and in order to help improve the quality of Wikipedia articles, it is essential that any primary-source material, as well as any generalization, analysis, synthesis, interpretation, or evaluation of information or data has been published by a third-party reputable publication (that is, not self-published) that is available to readers either from a website (other than Wikipedia) or through a public library. It is very important to cite sources appropriately, so that readers can find your source and can satisfy themselves that Wikipedia has used the source correctly.
The first stub for the policy was December 21 2003. After a few edits no one worked on it until February 2004. Reddi, the third person to work on the stub, first introduced the term "primary source" in the twenty-first edit to the policy, in an edit summary, on February 13 2004. Late that same day, a few edits later, he introduced the distinction between primary and secondary sources. I think that one of the first twenty edits to the policy still falls under "original" and February 2004 was more than three and a half years ago - and thousands of edits ago. And we are talking about the 23rd or 24th edit. Really. Slrubenstein | Talk 16:16, 21 September 2007 (UTC)
Just to be clear, the above is not a neutral summary, but the case against. Spenny 10:20, 21 September 2007 (UTC)
Some counterarguments as to why categorizing sources is important -
- Wikidemo 12:06, 21 September 2007 (UTC)
I totally disagree with Spenny's characterizations of the policy and the comments in support. In actuallity:
The NOR policy is not going to so significantly change that editors will not be required to consider sources based on their relationship to the view or concept. That's simply never going to fly. FeloniousMonk 14:49, 21 September 2007 (UTC)
<undent> Thanks for listing the concerns, Spenny. Right at the outset the first assertion really has to be substantiated:
> If evidence for this is well presented in the archives, it needs to be linked from here so that newcomers don't spend days hunting for it. There are issues with the other points which I'll try to come back on shortly, but time doesn't permit just now. .. dave souza, talk 18:08, 21 September 2007 (UTC)
Stepping back from the specific issues:
Blueboar notes, "but fails to explain how the classification of sources relates to the concpet of "No Original Research"." And asks for an explanation. Wikipedia has an NPOV and a V policy which incouraged editors to cite sources as away of includin diverse views in articles in an unbiased way. Several editors became aware that some people were citing sources to introduce material in articles in biased ways. Mere compliance with V was not enough. And biased edits often seemed to comply with NPOV because they clearly identified the point of view of the cited material. How were people seeming to comply with NPOV and V, while violating them, at least in spirit? By selective quoting, by duxtapposing different quotes in order to make the editor's point, and by using primary sources to argue (the editor's argument) that a primary source was wrong. So it became evident that it was not enough to insist that editors had to use reliable sources. We needed to emphasize that editors' views shouldn't go into articles, and we needed to show how one could seem to comply with V and CS and still be forwarding one's own arguments, which meant we need to show some dangers in how editors could use sources, and encourage other ways to use sources. Slrubenstein | Talk 22:24, 21 September 2007 (UTC)
(To replace the section entitled: "Primary, secondary and tertiary sources"...)
Citing the Right Source Materials (alternative suggested section title: 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 the right sources to back those different types of statements. Statements of fact should be cited to reliable sources that clearly demonstrate that fact. Statements of interpretation, analysis or conclusion should be cited to reliable sources that contain the same interpretation, analysis or conclusion.
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.
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.
above is the most recent as of today. Blueboar 15:55, 21 September 2007 (UTC)
Where are you going to propose this, so it can be scrutinized by the community? Or perhaps, if sufficient opposition is expressed, it will not even be proposed?-- Filll 17:13, 21 September 2007 (UTC)
Reject NOR by definition is a proscriptive policy. While we should be clear about what is allowed and what is not allowed, the emphasis is on what is not allowed - e.g. original synthesis. The Roosevelt example is unhelpful because, even if it is permissable, what readers of this policy (which says "no") are going to need most clarity on and explanation of is what they should not do. We have "Cite sources" to encourage people to cite sources. This policy is about the different ways sources (which in general we have to use) can be misused and discussion of sources shoudl be with this purpose in mind. Slrubenstein | Talk 17:37, 21 September 2007 (UTC)
Comment there was widespread disagreement about what the current written policy, particularly the source-typing sections, mandates, but there was more agreement about what it aims for; there was also disagreement about how it is used in practice, which may vary from page to page. The proposed text is supposed to describe the current policy (in its aims and in practice) more clearly than the current text does. The old source-typing can be improved on another page, to remove discrepancies between Wikipedia descriptions, and to address differences between Wikipedia and various scholarly fields. Can anyone point to one place where the old text discouraged a practice, which was generally rejected, and the proposed one encourages it, or vice-versa? Jacob Haller 18:37, 21 September 2007 (UTC)
Okay, either I missed something in the wording above or I need new glasses. Somehow I missed that this (today's) version of a "proposed" substitute for the current "Sources | Primary, secondary, and tertiary sources" section was being put to a vote. It solely appears to me that it is simply todays version for people to comment on or suggest additional additions or other changes.
In regards to the various comments from some new names I see up above, let me try to clarify a couple things again. For at least a month now, there has been much discussion primarily focused on the "Sources" section of this policy. Many of us (actually all of us) who have been in engaged in the 'limited' discussions on this agree that we do not want to water down the policy in way, shape, or form. But, many of us have also come to agree through many 'spirited debates', that including the "Primary, secondary, and tertiary sources" section confuses the policy to a point that it currently provides more confusion than guidance, especially when viewed in conjunction with the other policies and their rewording of the same principles. We (the many editors participating in this discussion over the last month or so), are not trying to back-door anything. We thought that it was simply best if we could get close to something that we could agree on, and then tell others what we are working on, so that maybe reaching some kind of concensus would have been simplified by many arguments already having been discussed and argued back and forth for a while. That is why we finally posted this 'proposal' over at Wikipedia:Village pump (policy)#Proposed change at WP:NOR. If we wanted to backdoor this, we definately wouldn't have advertised our intentions or asked for further input from others.
So again, I will ask you to think about why a policy (in this case one about 'No Original Research') needs to also include extraneous definitions and examples of what the various types of source material are? What does that have to do with "No Original Research"? We can clearly state that original research is definately not allowed without worrying about how to classify different kinds of source material. If not, then I think we should all go back to school for a few more years.
Including the "Primary, secondary, and tertiary sources" section in the policy is like having an article about Apples, and then having a sub-section devoted to disambiguations, so road-apples and horse-apples aren't included as well. It may be related, but nothing at all to do with the main subject of the article.
Instead of just saying Disagree, try offering some constructive criticism on how to make it better. wbfergus Talk 18:52, 21 September 2007 (UTC)
I think we have two questions. Both came from the source-typing debate, but each can be dealt with individually:
There were several attempts to (1) fix source-typing within NOR (to resolve different definitions) and (2) address various concerns regarding appropriate uses of primary sources, appropriate uses of secondary sources, etc. within NOR. I think we eventually agreed that RS could resolve the occasional issues with the sources themselves, and that source-typing didn't help with the more common issues with their misuse. One proposal referred to "primary material" and "secondary material" within each source, then "interpretative material" for secondary, and eventually that led to this. Jacob Haller 19:07, 21 September 2007 (UTC)
Does anybody actually have any specific criticism of the proposal? Because "it's not an improvement" is not a criticism. The question is, is it an improvement over saying absolutely nothing on the subject? Because that's our alternative. If we can't find consensus on this topic, we're going to have an empty policy that helps nobody. I'd encourage people to take a good look at the proposal and see if they can't improve it. It's the closest thing we've had yet to a consensus proposal. COGDEN 01:14, 22 September 2007 (UTC)
...... Admittedly there remain residual bits of confusion. One user mentioned Encarta and Britannica, which are also "tertiary sources" according to the delineation presented in WP:NOR. I pointed out somewhere in these rapid-fire multiple threads today there is a wealth of more specialized tertiary sources such as specialized encyclopedias, to which we can also add many textbooks as tertiary sources. Indeed one could readily also argue that there are numerous sources that are quaternary, quinary, senary, septenary, octonary, etc. etc. Categorizing sources as primary, secondary and tertiary has, nonetheless, served quite well in clarifying the basic types of sources from which material in WP is drawn. If there is a proposal to better clarify the distinction or explain that the delineations between them are not always sharply drawn, or whatever, that would, as I said in the section below, be quite welcome and potentially helpful. But the proposal to replace it with the language at the top of this talk section is, IMO, completely unacceptable on the justifications given thus far. I believe I used the words "summarily reject" to covey my summary opinion about the proposal. ... Kenosis 02:47, 22 September 2007 (UTC)
In an effort to minimize confusion regarding how this current discussion of replacing the "Sources" subsection on PSTS came about, please refer to the Archive for the last month or two. It's an awful lot of reading I know, but it aslo way to much to even begin to try to re-link back into here. This is not a "let's get this done today" proposal. We have been discussing this for a while and will continue to discuss even longer.
Take your time to read the Archives and get caught up on what you think the pertinent issues are before posting comments, especially derogatory or accusatory comments. Probably the very earliest this "proposed change" would or could (or even should) take effect would around the end of October. We need plenty of time to assimilate everybody's constructive feedback, comments, and suggestions. Thanks you. wbfergus Talk 19:16, 21 September 2007 (UTC)
Okay, a new day, starting to get my caffeine and nicotine levels back to normal, and I had about 6 hours of sleep, so let's see if I can coherently try to state what we are attempting here. This perticular subject of what we are attempting has been scattered all over this page the last few days, so maybe this section title will draw a few of the newer people to this discussion here.
First, we are not, in any way, shape or form, attempting to weaken the NOR policy. We are attempting to make the policy itself more coherent and concise, by having this policy only deal with NOR related issues.
Second, this mainly came about because of numerous edit-wars on many different kinds of articles where one of the parties stated something was "OR". Many of these also related directly to the interpretation of primary, secondary or tertiary sources that were based solely on this policy's brief definition of them, not on the more detailed explanations (and examples) offered on their respective pages at Primary source, Secondary source, and Tertiary source.
Third, "Sources", in one form or another, appear of multiple policies and/or guidlines. For an example, open each of these in a new window and compare:
These all attempt to describe the same issue, but in different ways. Especially in regards to source-typing, why should each policy even attempt to re-define these, when they are easily linked to the aforemention articles created solely for the purpose of defining the various types of sources?
Over the last month that I have been here, there have been numerous examples of how OR can be inserted into an article, either intentionally or unintentionally, regardless of the type of source used. As a matter of fact, as several other threads on this particular page (not even in the archives yet) have shown, secondary source material can be mis-applied just as easily (if not even more so) to create original research than just quoting the raw data from a primary source can. With secondary material, one could easily cherry-pick various sentences from the source(s) to build an argument promoting a POV that is not even contained within the source. This is usually much more difficult to achieve with data from primary sources, unless the editor tries to add their own conclusion or interpretation, which can easily be challenged to cite an appropriate source to verify the claim.
Trying define a type of source (some sources can also be multiple kinds of sources, depending on the context of what is being cited to it), and general blanket statements can easily be either misconstrued or otherwise misapplied by different editors at different times. The type of source has nothing at all to do with whether or not OR is being inserted into an article.
In short, many of us (around 20-30 different editors in the last month), though still not in complete agreement on wording or implementation yet, are starting to lean more towards this policy (and maybe the others) being more like an insurance policy. Everything is not defined just in the one policy. Much of what is in your insurance policy actually comes in whole or in part, from either another policy or other 'document'. The entire policy is the result of multiple "definitions" taken from various documents, and there is no overlap of those definitions when all the pieces are put together. With the current state of the policies, there is much overlap and redefinition of the "sources" subject over and over and over, with variations in those definitions. To alleviate confusion, we would like to remove this confusion from this policy and place it in a more logical place, that (hopefully) the other policies could link to similarly. We are not proposing to remove WP:PSTS, just move it out to a better place, so that PSTS would still be part of the policy, just not defined within this policy. Similar in wording to how this policy and the others, state "Since NPOV, V, and NOR complement each other, they should not be interpreted in isolation from one another, and editors should try to familiarize themselves with all three". Instead,, probably something along the lines of "Since NPOV, V, NOR, BLP and 'source typing' complement each other, they should not be interpreted in isolation from one another, and editors should try to familiarize themselves with all, as a good Wikipedia article will adhere to all".
I probably worded that last sentence poorly, but I think you should be able to get the idea of what we are trying to accomplish. If the "source typing" was moved into it's own article, having the common definition that the others can all share makes this concept easier to comprehend. I think more thought and effort needs to go into what and where this "source typing" article should be, like a guideline isn't strictly enforcable, but do we really need a policy to define the types of sources?
Thanks for reading this, hopefully this may alleviate some concerns about this being conducted behind closed doors as an attempt to weaken current policies. wbfergus Talk 11:02, 22 September 2007 (UTC)
Would it make sense to create a "proposal" page... explaining the rational for the change, the proposed language, and the various options on implimentation (such as a) moving PSTS to some other policy or guideline, or b) creating a new guideline out of PSTS)? I know something like that is really meant for entire page revisions or new proposals.... and it is rare for policy change discussions, but it might help people FIND what is being discussed, and get a sense of the basic issues (pro and con) without having to sort through all the chatter or the archives. If people like this idea, I would suggest such a page say prominently that that comments continue to be made here, on this page. (My concern would be that someone would come along later and the say... "gee, I didn't even know the issue was being discussed"... my suggestion is so that both the proposal and it's discussion are prominent. Blueboar 19:39, 21 September 2007 (UTC)
I would suggest a cautious approach. I do not think that any of these wide-ranging changes will have any traction. Work on improving rather than changing may be a more fruitful approach. ≈ jossi ≈ (talk) 00:00, 22 September 2007 (UTC)
"Right" sources? That's hovering dangerously near "truth" - "right" is inherently open to interpretation, and is subjective. Primary, Secondary, and Tertiary are clearly defined and not open to abuse of interpretation. I do not see how reducing clarity and confusing the issue and inviting arguments by using subjective phrasing is in any way anything but a recipe for disaster. KillerChihuahua ?!? 00:02, 22 September 2007 (UTC)
WP:V requires us to rely on reliable, third-party published sources, and the NOR policy at present gives an exemption in that primary (or first party) sources can be used for statements of fact that require no analysis, in the context of analysis by secondary (or third party) sources. It seems to me that this distinction is being lost in changing the emphasis to what's really an expansion of "However, care should be taken not to "go beyond" the sources or use them in novel ways." If we accept the analysis that the primary/secondary/tertiary analysis is so broken that we have to make the radical change proposed (and evidence of this breakdown is needed for those of us who can't find it in the archives), we have to ensure that any relevant requirements under the present policy are fully covered by the new policy, and also have to provide clarification for editors used to the current terminology. This applies both to experienced editors who haven't re-read the policy for a while, and also to new editors advised in discussion that use of a source goes against WP:NOR in the old terminology. Here's a suggested addendum to the current proposal:
Care must be taken to accurately represent the overall sense of the source, and not introduce our own interpretation in summarising or selectively quoting from the source. Statements of fact can draw on documents or people very close to the situation being written about, often described as primary source material. Statements of interpretation, analysis or conclusion must draw on third party material, often called secondary sources, which draw on primary sources to make generalizations or interpretive, analytical, or synthetic claims. For example, we can state the fact that when Charles Darwin was finishing proofs of The Origin of Species he wrote to a friend "So much for my abominable volume, which has cost me so much labour that I almost hate it." However, we can't use this source from the author to state the conclusion that "the book is abominable or almost hateful", but for an assessment must find a reliable third party source giving the opinion of a recognised current expert in the field.
Obviously the example given [1] is the first that came to hand, and no doubt a better one can be found. There may be other points from the current formulation that need to be covered. Aspects of source categorisation not closely related to NOR could be covered in WP:RS, or in the proposed new guideline. If the "third party" definition needs clarification, that has to be reflected across all policies and guidelines that use the term. .. dave souza, talk 10:00, 22 September 2007 (UTC)
Back in another discussion taken from here, someone implied that primary/secondary source conflicts were "rare". I have now had to treat yet another case of a mistake in a secondary source where appeal had to be made to a primary source. This particularly seems to be a issue with BLPs of people who are notable enough for media sources, but who lack conventional biographies. These media sources make lots of mistakes; fact-checking (or at least cross-checking) is obligatory when questions arise.
As I commented over in the discussion of WP:RS, our policies don't treat this issue well. The problem in WP:RS is that the issue of source errors isn't addressed at all, so that being correct isn't a criterion for being reliable. The problem here is that the constant hammering on avoiding primary sources is ratifying the notion that secondary source errors are to be preferred over primary sources.
One point I consider non-negotiable: when a primary source is quoted in the article, the primary source should be cited if possible. In this case, a secondary source should be never, ever, be preferred over a primary source. Likewise, when a secondary source misrepresents a primary source, whether by accident or design, the secondary source statement must not appear in Wikipedia unless it is necessary to explicitly deny its accuracy. These are the absolute standards of non-fiction writing, and we cannot credibly contravene them. Whatever other rules we establish, these must not be contravened. The problem I've consistently seen in the discussion here is that the discussion of sources is so abstract that it doesn't consider the ways that secondary sources use primary sources. Some of these uses it is not our purpose to argue with, but some of them, we have to review. Mangoe 14:46, 22 September 2007 (UTC)
This issue of verifiability and reliable sources is one thing. The issue of how those sources are used in creating an article is another. Of course they interact. All three of the basic editorial policies interact -- WP:NPOV, WP:VER and WP:NOR. WP:NOR requires dealing further with the issue of what kinds of syntheses are generally available among the sources. The issue of whether they're "right" or "wrong", as has been explicitly stated in WP's policy growth and development from the very beginning of WP, is not the issue. Odd as it may seem at first blush, the original mandate from the founding Board (read that:"from Jimbo Wales") that the standard is "verifiability, not truth", still controls in Wikipedia. What's required are sources that show a published weight of opinion about a particular assertion of fact or conclusion that can be drawn from published information. But that's more for WP:VER and its offshoot WP:RS. For WP:NOR, the issue centers around relying on published sources and not formulating new syntheses or drawing new conclusions that are not already available in the published literature. When the published literature conflicts about the basic facts or conclusions, WP:NPOV#Undue_weight controls the method by which the material is to be presented. As I said earlier here and as others have stated too (not necessarily in this discussion), the "big three" basic editorial policies work in tandem and in tension with one another. That interaction and tension among the "big three" is mediated by consensus process at the local article level. For the purposes of this policy page, WP:NOR, it is useful and appropriate to provide a framework or basic breakdown of the range of sources available, which the primary, secondary, tertiary delineation does quite effectively. ... Kenosis 15:51, 22 September 2007 (UTC)
(To replace the section entitled: "Primary, secondary and tertiary sources"... which is to be moved elsewhere)
Citing Appropriate Source Materials (alternative suggested section title: 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.
I tried to address the concerns that had been expressed and to incorporate some of the wording from dave souza's proposal. Further discussion and proposed changes are welcome. -- Coppertwig 15:19, 22 September 2007 (UTC)
How about combining the titles and dropping the "right" (vs. wrong) issue (and using wiki-caps) as follows: Citing appropriate source materials, and sticking to those sources -- Dhaluza 16:02, 22 September 2007 (UTC)
..... The section on Primary, secondary and tertiary sources is a vital section that puts WP:NOR into perspective for users around the wiki. The proposal to replace the section or move it elsewhere is a major change in this policy page that I find unacceptable, at least without something that represents a clear improvement on it. Since well over three years ago, a distinction had been made between primary and secondary sources and what uses constituted original research or original synthesis. A significant number of WP users remained confused about the distinction, e.g., "well, if WP is not supposed to synthesise material directly from primary sources (except for purely factual matters), and if we're supposed to use secondary sources, then how is WP a secondary source?" The extra note about tertiary sources was inserted into WP:NOR nearly a year ago, and suddenly it made better sense to most other participants and much of the confusion diminished substantially. The description of encyclopedic matter as primarily tertiary helped to resolve this confusion right on the policy project page. Since then, it's tended to serve us well in resolving confusion about what is original synthesis, or original research, as distinguished from original wording and editorial decisionmaking.
...... Admittedly there remain residual bits of confusion. One user mentioned Encarta and Britannica, which are also "tertiary sources" according to the delineation presented in WP:NOR. I pointed out somewhere in these rapid-fire multiple threads today there is a wealth of more specialized tertiary sources such as specialized encyclopedias, to which we can also add many textbooks as tertiary sources. Indeed one could readily also argue that there are numerous sources that are quaternary, quinary, senary, septenary, octonary, etc. etc. Categorizing sources as primary, secondary and tertiary has, nonetheless, served quite well in clarifying the basic types of sources from which material in WP is drawn. If there is a proposal to better clarify the distinction or explain that the delineations between them are not always sharply drawn, or whatever, that would, as I said in the section below, be quite welcome and potentially helpful. But the proposal to replace it with the language at the top of this talk section is, IMO, completely unacceptable on the justifications given thus far. I believe I used the words "summarily reject" to covey my summary opinion about the proposal. ... Kenosis 16:52, 22 September 2007 (UTC)
<undent> Plain English or not, this now refers to secondary sources without making it clear that sources close to the subject, which for most of us is plainly and simply expressed by the term "primary sources", can be used for facts but not for interpretation etc. This is still less clear than the existing policy. The Darwin example I showed above was designed to bring out the point that a fact about an openly attributed opinion can be stated, but that opinion from, eh, a primary source can't be used as analysis, conclusion or whatever. That's a point which still seems to be lacking, and some thought is needed as to whether this proposal would have other unintended consequences. .. dave souza, talk 23:34, 22 September 2007 (UTC)
This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 15 | ← | Archive 20 | Archive 21 | Archive 22 | Archive 23 | Archive 24 | Archive 25 |
We need to resolve certain disputes. Based on the past week's discussion and comments from vaious people, I am singling out two specific disputes, which is more than enough to work on for now. Once we resulve these, people are free to raise other issues which we can work on, one at a time, until we resolve them.
Well, there are currently two "quality of sources" locations: the actual guideline article and the NOR policy page. Much of the current controversy arises from having both a guideline and a separate policy section. Apparently the guideline doesn't have enough compelling force (but that's a personal interpretation that I'd welcome having replaced with a better characterization.) That dual-location situation is in part what creates a "can of worms." -- Minasbeede 14:51, 31 August 2007 (UTC)
I have now slept on it and feel even more strongly than before that the NOR policy must make a distinction between primary and secondary sources. I think Slim Virgin is right (and I do not always agree with her and long questioned her on this). NOR needs to explain what "original research" is, and what it isn't and I do not think there is any way to do this without discussing primary and secondary sources.
Some people do not understand the distinction or think it opens up a can of worms but I now believe that is because they think too literally. What I mean is, some people think that things are what they are. I do not mean to go into a long philosophical debate but Wittgenstien and Pierce and Dewey - generally acknowledged to be among the most important philosophers ever in the US and Europe - argued that the meaning of things including words depended on how they are used. To a large degree, what makes something a primary or secondary source is, how it is used. Primary sources can be used to establish facts without violating NOR, but they cannot be used to make analytic, synthetic, interpretive, or explanatory or evaluative claims without violating NOR. Secondary sources by contrast are the only sources we can use to make analytic, synthetic, interpretive, explanatory, or evaluative claims (as long as they are the claims explicitly made by the source itself).
We can, did, and I think still should provide examples of primary and secondary sources, but they are just examples. The essence is in how they can or cannot be used.
This by the way means it is conceivable for a text to be either a primary or secondary source - depending on how it is used. Let's take On the Origin of Species for example - I grant this may not be the best example, but I am just looking for a handy hypothetical. In it Darwin makes all sorts of analytic, synthetic, and explanatory claims about natural history. As a document concerning natural history, it is a secondary source ("According to Darwin, variation in finches on the Galapagos is best explained by ..."). But in it he also expresses his own beliefs and judgements. As a document about Darwin, it is a primary source ("According to Janet Browne, when Darwin wrote On the Origins of Species he was responding to ...."). In the article on Evolution we can use it freely as a secondary source (although evolutionary theory has gone far beyond Darwin, evolutionary scientists still consider the case he built to support his theory valid). In the article on Charles Darwin it is a primary source and while we can use it for facts, if we want to make arguments about what Darwin was like or what he meant, we should draw on appropriate secondary sources by historians of science, intellectual historians, and biographers of Darwin. I do not see this as a can of worms. I just see this as a more sophisticated way of thinking that understand things in terms of the ways they are used, and understands that appropriate uses depend on the context. Slrubenstein | Talk 10:43, 31 August 2007 (UTC)
Well it was chosen in haste and perhaps a bad choice, although I think you misunderstand. If someone wanted to make a claim about variation among finches, I would argue that Darwin's book is a valid secondary source. But there is no need to belabor this one example. Can you think of any texts that could be a primary source in one context, and a secondary source in another? That is the issue. If you can't think of any, I would just take that as a sign that the boundary between primary and secondary source is even clearer than I thought! What is your answer to the question that heads this section? You do not have to take a position on what I say; the question is your position on the current policy. Slrubenstein | Talk 17:37, 31 August 2007 (UTC)
Folks, for some this is a very contentious issue. I think we first need to resolve the basic question of whether the policy must or should make this distinction first. Then discuss editorial issues. Sambc, if you want to, creat a policy proposal page and invite discussion of your idea there. But here, the question is simple: should NOR distinguish between primary and seconday sources? That's all! Slrubenstein | Talk 12:38, 31 August 2007 (UTC)
I think the basic issue of NOR is that it's difficult to explain. Personally, I feel like it's kind of a "I can't really define it, but I know it when I see it," kind of thing. I've previously indicated that I think the idea that novel synthesis is banned by NOR is pretty ridiculous - an encyclopedia article is by definition a synthesis, and doing things like summarizing books, or trying to describe the state of the literature on a subject, are bound to involve some degree of original synthesis. It is novel analysis that we need to be careful about, and this is what we shouldn't be able to use primary sources for. The classic OR using primary sources is essentially to go to, say, the text of some primary source and then to quote that text, out of context, to support some new argument which is not explicitly made by the primary source. But I don't know that this is something that is unique to primary sources - one can easily do the same thing with secondary sources. I think the key issue here is not that primary sources are forbidden for any particular purpose. It's that, in evaluating primary sources, we need to stand by the judgments of reliable secondary sources. john k 17:36, 31 August 2007 (UTC)
Well, so far it appears that everybody is in agreement that a distinction should be made in regards to the types of sources. I didn't see any objections above that a distinction should not be made. The only objections I see (maybe I'm biased or blind), is in the wording of how to label the types of sources, and whether or not the "Policy" page should attempt to offer a definition of the types of sources.
Personally, I feel that the details of what the various "source types" are should be defined on the separate pages already created, and this "Policy" page merely point out that different types exist and here's the link for the definition. This page is solely about NOR and should not expand (rightly or wrongly) into the realms of other subjects, no matter how contentious or how badly needed. When it's necessary to delve into those realms, link to that subject appropriately and say something like "For more clarification or discussion, see ...". Keep each "Policy" and/or "Guideline" focused only on that subject without blurring the lines of distinction. wbfergus 17:49, 31 August 2007 (UTC)
Can the (Wikipedia) definitions of "primary" and "secondary" be made complete enough and thorough enough that they suffice to define "primary" and "secondary," as used/needed by the NOR policy? -- Minasbeede 20:29, 31 August 2007 (UTC)
I think the first order of business would be to stop using the words "primary" and "secondary" in a way different from how they are used outside Wikipedia. Choose/develop a different way to distinguish the usages from the Wikipedia point of view.
(Something about NOR I decided to move to my talk page, where it may languish and die, was here.) Minasbeede 13:32, 31 August 2007 (UTC)
There's no one simple usage outside of Wikipedia, which is part of the problem. But for our purposes, can't we just focus on the terms "primary" and "secondary" and just say that a primary source is a source that is "primary"? In other words, an original source? A secondary source is a source that is "secondary". In other words, not primary. Of all the ways the terms are used outside of Wikipedia, that seems to be the core idea. The original source of an idea, conclusion, or set of data is the primary source, while any source of that same idea, conclusion, or set of data that isn't the primary source, is a secondary source.
We could also switch to neologisms, such as "reliable source" and "auxiliary source", but I don't see the point, since reliable source is already well defined at WP:RS, and the new term auxiliary source would just be something that isn't a reliable source. COGDEN 21:15, 31 August 2007 (UTC)
How about "sources of facts" and "sources of interpretations"? -- Coppertwig 22:52, 31 August 2007 (UTC)
I would say, not exactly. It should give advice on the appropriate use of primary vs secondary sources, as primary sources are much more likely to lead to inadvertent synthesis, but can be used without it being OR. SamBC( talk) 12:34, 31 August 2007 (UTC)
No or not without major exceptions. I noted earlier that the Communist Manifesto is a primary source on Marx's and Engel's beliefs, and a secondary source on their various opponents' beliefs. It is a trustworthy description of their own beliefs, and a rather untrustworthy description of their opponents' beliefs. I think many such works are reliable when used as primary sources and unreliable when used as secondary sources, not the other way around. Jacob Haller 20:32, 31 August 2007 (UTC)
No. The policy should not privilege either primary or secondary sources. The pros and cons of each are to complicated to be found in a policy. Maybe some distinctions can be introduced first in an essay, then possibly in a guideline, but there should not be any blanket preference or discouragement of either in this most important of policies. The archived discussions brought out plenty of instances where primary sources are far preferable to secondary sources. But the main reason I don't think there should be a preference is that we can't yet agree on what the difference is between a primary and a secondary source. It makes no sense to say that one or the other should be "rare" or "relied upon" when we don't agree on what either means. Also, any stated preference of one over the other should describe widespread Wikipedia practice, not lead it (see WP:Policies and guidelines). This policy cannot change Wikipedia practice. It is bound by that practice. COGDEN 21:02, 31 August 2007 (UTC)
The policy cannot distinguish between primary and secondary sources unless there is a consensus as to the difference. In my view, it's pretty simple: a primary source is something original, right from the horse's mouth, while a secondary source is a re-telling of a primary source. In another sense in common usage, a secondary source is a source that discusses a primary source; however, any new conclusions by the secondary source is also a primary source as to those new conclusions. On this basis, there is absolutely no reason to favor secondary sources, unless the primary source is unpublished, in which case it would not be a reliable source.
Moreover, since polices must describe widespread Wikipedia practice, not lead it (see WP:Policies and guidelines), the fact that primary sources (such as journal articles, interviews, and fictional works) are widespread, and even more likely to exist in featured articles than in non-featured articles, would seem to show that if any distinction is to be made between the two, primary sources should actually be favored. Personally, I don't think WP:OR should make a decree either way. COGDEN 20:49, 31 August 2007 (UTC)
No, no privileging of one type of source over another. It has been pointed out that many FA use a lot of primary sources. The key thing is that any "analytical, synthetic, interpretive, explanatory, or evaluative claims" come from sources, not from Wikipedians. -- Coppertwig 22:57, 31 August 2007 (UTC)
Agree with Coppertwig. Get guidance from guidelines. If people come to the policy statement for guidance refer them to the guidelines from within the policy. That way there'd also be just one set of descriptions, not two. There could even be an examples page (that provides further guidance) of what meets the criteria and what doesn't, with an explanation of why or why not. These could be real cases, if that doesn't offend any of those involved. The examples could be slightly (or greatly) altered to avoid offense, if offense to someone is a problem. -- Minasbeede 03:26, 1 September 2007 (UTC)
There are already phenomena for which Wikipedia itself is a primary source, such as the Bogdanov affair, and since the page histories record edits as they occur, they are thus a reliable source of which editor accounts were used to write what things. The problem with OR or usage of primary sources might be biases or the creation of hoaxes. In some of those cases, the policy is justified. In some other cases, however, in which a phenomenon develops that numerous bloggers, message board users, etc., notice before the general or mass media does, OR might be the best way of describing the phenomenon, at least as the bloggers or message board users see it. I would suggest the following: if something is quite controversial and argued over by numerous people, the controversy should be discussed comprehensively from all sides; if something such as the cancellation of a TV show provokes an enormous campaign with a rapid accumulation of signatures that suggests numerous contributors, Wikipedia should take note of that as well. Try to give the benefit of the doubt when you can. — Rickyrab | Talk 14:51, 7 September 2007 (UTC)
Let's review. As I understand what I've read here, Wikipedia policies should be descriptive, not proscriptive. Yet so far nobody has been able to accurately describe or define the primary/secondary distinctions (that are made in a special Wiki way and aren't the same as occurs outside Wiki), which looks to me like the policy section that is so very important can't even be enunciated. Surely if it can't be said it doesn't describe anything.
So let me ask for a few examples of articles where this distinction is so important: if it can't be described maybe it can be illustrated. I looked at the Wiki article on "Evolution" last night and found that, to the level I checked, it looked like it was indeed a long listing of facts with nary a word of explanation, analysis, or an any other of the things that Wiki editors are forbidden to do unless they're quoting someone else. I think the Wiki "Evolution" article is a very fine article but apparently (unless I missed something, which is possible) it cannot serve to at all illustrate why the emphasis on the superior nature of secondary sources is so necessary, as is asserted here by some. I'd think that in the over 100 sources cited there are both primary and secondary sources but one might suspect that if all the content is simple facts that all of the content could and did appear first in primary sources. So, what are a few articles that illustrate why the emphasis on secondary (as the term is used in Wikipedia) sources should be given special emphasis? -- Minasbeede 11:26, 1 September 2007 (UTC)
Who asked for a debate? I said that the Evolution article seems to be a fine article yet it also seems to be a serial recitation of facts, something that some assert here is undesirable. Not only does it seem to be a serial recitation of facts I can see why that particular article needs to be that and must strenuously avoid "analytic, synthetic, interpretive, explanatory, or evaluative claims" - so probably it is indeed a bad article to look to for what is claimed. Still, the Evolution article seems to me to be an article that illustrates that the claim made here by some (that you have to use secondary sources in order to include analytical material) isn't universal. So I asked for examples of articles that do show what is claimed.
The title "regroup" is what I came up with. the real motive was to split the section so editing would be easier. Any title is fine - and if a different title is inserted this can be removed. -- Minasbeede 14:09, 1 September 2007 (UTC)
I've been asked to contribute on this topic and don't think that I can't really add any new arguments, as they have been all repeated many times. Instead, I think it would be useful to look at scientific papers - one thing I know well - to illustrate that it is impossible to classify some sources as either primary or secondary.
Take this 2004 paper as an example A trypanothione-dependent glyoxalase I with a prokaryotic ancestry in Leishmania major.
In the fist paragraph of the Results section I quote some dissociation constants I measured. If a Wikipedia article cited these results, it would be using the paper as a "primary source" - a source that presents novel, previously-unpublished data.
In the first paragraph of the Discussion on the other hand, I discuss previously-published data on Methylglyoxal synthase in the light of my results and bring these data together and interpret them. Here, if a Wikipedia article cited the paper on the possible problems caused by methylglyoxal synthase, it would be using the paper as a secondary source.
Scientific papers are neither primary nor secondary sources, they contain elements of both. Tim Vickers 01:28, 3 September 2007 (UTC)
Above, Tim write, "Why not just say that all material must be cited, with facts backed by citations to reliable sources that contain these facts, and interpretations and syntheses attributed to reliable sources that make these interpretations and syntheses?" Every bolicy starts with a box explaining the policy "in a nutshell." I'd have no objection to this being the nutshell, or part of it. Slrubenstein | Talk 15:01, 5 September 2007 (UTC)
At bottom "No original research" means that Wikipedia should not be written as a primary source. This policy used to state this explicitly as: "Research that creates primary sources is not allowed." until this significant rewrite during the time that the controversial depreciation of primary sources was being integrated into the policy.
I think that we should go back to basics and focus on the problem--editors generating new not-previously-published primary source material in WP articles, rather than trying to limit the re-use of existing published primary source material. The arguments about whether basing an article on primary or secondary source material creates laudable or deplorable articles is not relevant to this policy. We are only concerned here with whether the material is new or novel, not whether it is good or bad. Trying to stretch this policy to cover these types of subjective value judgments has only caused confusion and contention over what should be a very simple concept. Dhaluza 17:59, 3 September 2007 (UTC)
I think there should be a page that could be entitled something like "Wikipedia:Wikipedia policy for Dummies" (or possibly something more original) that gives a clear and consice version of the main policies and guidelines that avoids what is known as instruction creep. MrMurph101 03:13, 5 September 2007 (UTC)
We do not need a policy for dummies. What we need is patient editors who are willing to explain things to newbies - that is one of the things that makes this a collaborative project. Slrubenstein | Talk 14:59, 5 September 2007 (UTC)
Many articles on law, science and other complicated stuff have examples to help the reader understand. The examples are not referenced. Are they original research? (For example, "Concepts" section of mens rea). -- Kaypoh 06:26, 5 September 2007 (UTC)
I think the Concepts section of mens rea was not a good example of this type of "example used to explain" (don't know what to call it). Below are a few examples of what I really mean.
See this from Fallacy of many questions:
An example of this is the question "Are you still beating your wife?" Whether the respondent answers yes or no, he will admit to having a wife, and having beaten her at some time in the past. Thus, these facts are presupposed by the question, and if it has not been agreed upon by the speakers before, the question is improper, and the fallacy of many questions has been committed.
And this from Intention (criminal):
For example, suppose that A, a jealous wife, discovers that her husband is having a sexual affair with B. Wishing only to drive B away from the neighbourhood, she goes to B's house one night, pours petrol on and sets fire to the front door. B dies in the resulting fire. A is shocked and horrified. It did not occur to her that B might be physically in danger and there was no conscious plan in her mind to injure B when the fire began. But when A's behaviour is analysed, B's death must be intentional. If A had genuinely wished to avoid any possibility of injury to B, she would not have started the fire. Or, if verbally warning B to leave was not an option, she should have waited until B was seen to leave the house before starting the fire. As it was, she waited until night when it was more likely that B would be at home and there would be fewer people around to raise the alarm.
And this from Mistake (criminal law):
For example, if a defendant goes into a supermarket and places eight items in a basket which is presented to the cashier for payment in the usual way. Both honestly believe that all eight items have been scanned, and the defendant pays the sum shown on the bill. A store detective, however, notices that a mistake was made by the cashier so that only seven items were actually priced. This detective arrests the defendant after leaving the store. Since the defendant honestly believes that he has become the owner of goods in a sale transaction, he cannot form the mens rea for theft (which is usually dishonesty) when he physically removes them from the store. Accordingly, he should be acquitted.
And this from Self-defense (theory):
For example, a person who unknowingly chances upon two actors practising a fight would be able to defend his restraint of the one that appeared to be the aggressor.
-- Kaypoh 14:48, 5 September 2007 (UTC)
I have looked at articles that had a couple of funky tables and then looked at the referenced sources as well as others, and wondered if constructing a single table that merged the data from the various sources would constitute OR.
In the case of oil wells, there is a list of them with data pulled from multiple sources which seems not be OR, while a I have a feeling that creating a similar kind of compilaton for Palestinian census data would be seen as OR. I've concluded for the time being that I will fill in existing tables with info I find, but I won't create any because I now see it as minefield - one's structuring of data is a POV.
But perhaps this should be handled with a policy more like that for images and other original artwork? The objective is to present the diverse facts and analysis in some coherent fashion, right? Mulp 01:53, 6 September 2007 (UTC)
Is it OR to say that X term, that some people describe as a neologism, is not found in the OED or [insert name here] other specific dictionary? Obviously the OED is a primary source in this case and the factual accuracy is easily verified. 136.152.153.36 22:21, 6 September 2007 (UTC)
This seems to have gotten off-track with respect to addressing the issue I was interested in. There is an article that says "X is a term that...". Some verifiable academics say this term is a neologism. Some Wikipedians want to mention that the word "X" was published much earlier in books A, B, etc. Other Wikipedians want to mention that "X" is a not included in the OED or any other dictionary. Both of these statements are based on easily verified facts discovered by Wikipedians, and potentially they both add useful context to whether or not X is a new term (or perhaps a rarely used old term that has taken on a new meaning). So, consistent with the OR policies regarding primary sources, is one allowed to mention both, neither, just one, etc.? 136.152.153.15 21:25, 7 September 2007 (UTC)
The arguments for and against the use of specific types of sources in specific situations need to be discussed more fully, and in the context of WP:OR, WP:NPOV, and WP:Notability. I think we can work toward a potential guideline for this. In the meantime, however, to get a consensus, original research policy that we can all live with, why don't we get back to basics. We can avoid the controversy for now by paring back to the fundamental ideas we all agree with, which I think are represented in the following text, which I would propose as a near-term replacement for the "Primary, secondary, and tertiary sources" section:
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.
When citing a source in support of a proposition in a Wikipedia article, the proposition should (1) be verifiable to the source by any reasonable, educated person without specialist knowledge, and (2) make no novel analytic, synthetic, interpretive, explanatory, or evaluative claims based on the source.
Even if people want there to eventually be more than this in the policy, maybe this is an acceptable step that we can build on. COGDEN 23:59, 6 September 2007 (UTC)
I would like to give an instance that possibly hones the edge of the point being made by the "example" discussion above.
I recently edited an article of a person X where I wanted to say that "X is often referred to on the web as Prof. X". Later I show that his official biography says he was not a prof. This I felt was important to the article, since many readers may come to the article with a view that he is a Prof. An encyclopedia is being written against the backdrop of this common perception, after all.
Now in stating that "X is often referred to on the web as Prof. X", I have no "source" that makes this full claim. However one can cite hundreds of webpages that say he is a prof. It is a claim that is immediately verifiable, and backed up by these sources, which are completely "reliable" for this purpose - ie. they are clearly referring to him as Prof. X.
We are not introducing a new theory / original idea / term/ argument - it is just the immediately verifiable statement that "X is often referred to on the web as Prof. X". My question is - should it constitute original research?
I am referring to the article on P. N. Oak, where I have made this claim recently. If there is a consensus here that this is OR, I will remove the statement. mukerjee ( talk) 09:05, 7 September 2007 (UTC)
David Irving has repeatedly pushed the number 135,000 as the number killed in the bombing of Dresden in World War II. This number like the man is now throughly discredited. See
The trouble is that the number still appears in tertiary sources like encarta Dresden. This is not a problem at the moment because of the reliance on secondary sources. But how would one handle this if the new wording for sources is used? See Talk:Bombing of Dresden in World War II#Population statistics and use of napalm.
In a second case take an example from a primary source like Protocol I that says "The Parties to the conflict shall take all feasible measures in order that children who have not attained the age of fifteen years do not take a direct part in hostilities and, in particular, they shall refrain from recruiting them into their armed forces." Unless one is an expert on international law the nuance of the words "feasible measures" will be lost unless one had read the ICRC commentary on this that makes clear, this is not a complete ban on the use of children in conflict. The ICRC had suggested that the Parties to the conflict should "take all necessary measures", which became in the final text, "take all feasible measures" which is not a total prohibition on their doing so because feasible should be understood as meaning "capable of being done, accomplished or carried out, possible or practicable". See Military use of children#International humanitarian law for sources on this.
It seems to me at first glance that doing away with primary, secondary and tertiary sources may open up a can of worms and I would like others who are considering these radical changes to comment on how the new proposed policy will deal with these types of issues. -- Philip Baird Shearer 20:43, 20 September 2007 (UTC)
Editorial judgement is a problem, is Encata not as reliable source as Götz Bergander in this case? Under the current rules the answer is no. But under the new rules who says? Just look at the section on the bombing of Dresden given above to see the potential problems this opens up. -- Philip Baird Shearer 21:04, 20 September 2007 (UTC)
As to the second point about OR yes under the current rules this is clear (because it is a primary source), but under the new proposed rules, how do you tell the difference between a summary of a source and interpretation of a source given that there is no specific difference between a primary and a tertiary source? -- Philip Baird Shearer 21:12, 20 September 2007 (UTC)
Blueboar, I understand the arguments that both Vassyana are presenting, and as I said above it may be that the current section should be moved into WP:V, but if it is removed and replaced with the suggested text then it will also alter the reading of other sections of the content policy and guidelines. For example I think the section WP:V#Sources would need a rewrite, because quite obviously a treaty is a more reliable text on that treaty than any "peer-reviewed journals and books published in university presses" as are court transcripts and UN Security Council resolutions, all of which are quoted in articles other than those about themselves. I also think that WP:RS#What is a reliable source? is affected in a similar way. So before any change is made I would like an analysis made of how such a change would effect the reading of the whole content policies and guidelines.-- Philip Baird Shearer 11:35, 21 September 2007 (UTC)
This discussion about changing the emphasis on secondary and tertiary sources rather than primary sources is somewhat misplaced and misguided, IMO. Primary sources that report different empirical or historical calculations than widely reported, as was mentioned in the example at the beginning of this thread (whether 135,000 is = or ≠ to the number killed in Dresden), fall under WP:NPOV#Undue_weight and WP:RS, which work in tension with one another. Primary soucces are already stated in WP:NOR to be validly used to report plain facts without interpretation or synthesizing of ideas that may arguably follow from the plain facts. Moreover, WP:NOR, in its emphasis on secondary and tertiary sources for interpretive and synthetic issues, works in tension and in tandem with WP:VER and WP:RS, along with WP:NPOV or course. These tensions are natural and intentional, and the founding Wikimedia Board would appear to have clearly understood this when the three basic editorial policies were first put forward. WP:Consensus, of course, is the method by which these tensions are balanced or resolved in the process of writing the articles. The push to alter this balance by substantially rewriting one or more of the three most basic editorial policies neglects that the solutions to the dilemma already exist within present WP policy. The dog should move the tail here, not the other way around. ... Kenosis 13:54, 21 September 2007 (UTC)
As to my "let's be at least halfway real..." comment, this refers to what appears to be the present notion that this is just a minor proposal that can be implemented without further scrutiny because several WP users have worked on it and believe it is nearing consensus.My review of the situation indicates that this is in fact intertwined with an effort to significantly change the policy, up to and including rewriting the rules on "original synthesis" and other major changes. As to the "process" being referred to by Sambc, that process has been a discussion among several WP users, indeed properly regarded as a preliminary discussion, advocating one or more significant changes in WP policy w.r.t. WP:NOR. The "tail attempting to wag the dog" here is my reference to the exceptional cases driving attempts to change the basic, fundamental balances of WP:NPOV, WP:VER and WP:NOR. The use of anecdotal examples of users having problems in interpreting and resolving issues in individual articles, such as, e.g. the Dresden bombing example in this section, are properly regarded as a {{sofixit}} issue at the individual articles, not valid reasons to change the policy across the wiki. ... Kenosis 16:47, 21 September 2007 (UTC)
If we're writing an article about, say, Encarta, we have to treat what it or its authors and publishers say about it as primary material, and aim to base interpretations or opinions on that material on third party secondary sources. That's widely understood from the current policy, and should not be lost in this push to avoid "source typing". ... dave souza, talk 21:25, 20 September 2007 (UTC)
WP:NOR is one of WP's core content policies and has stood in essentially the same form for a very, very long time, and benefits from broad acceptance. It is uncontroversial and broadly accepted. I can tell you for a fact that any attempts to alter it's original meaning and formulation, such as diminshing the distinction between primary and secondary sources, are going to fail regardless of any agree between the handful of editors participating in the discussion here. FeloniousMonk 14:31, 21 September 2007 (UTC)
In most cases, Wikipedia articles include material on the basis of verifiability, not truth. That is, we report what other reliable secondary sources have published, whether or not we regard the material as accurate. In order to avoid doing original research, and in order to help improve the quality of Wikipedia articles, it is essential that any primary-source material, as well as any generalization, analysis, synthesis, interpretation, or evaluation of information or data has been published by a third-party reputable publication (that is, not self-published) that is available to readers either from a website (other than Wikipedia) or through a public library. It is very important to cite sources appropriately, so that readers can find your source and can satisfy themselves that Wikipedia has used the source correctly.
The first stub for the policy was December 21 2003. After a few edits no one worked on it until February 2004. Reddi, the third person to work on the stub, first introduced the term "primary source" in the twenty-first edit to the policy, in an edit summary, on February 13 2004. Late that same day, a few edits later, he introduced the distinction between primary and secondary sources. I think that one of the first twenty edits to the policy still falls under "original" and February 2004 was more than three and a half years ago - and thousands of edits ago. And we are talking about the 23rd or 24th edit. Really. Slrubenstein | Talk 16:16, 21 September 2007 (UTC)
Just to be clear, the above is not a neutral summary, but the case against. Spenny 10:20, 21 September 2007 (UTC)
Some counterarguments as to why categorizing sources is important -
- Wikidemo 12:06, 21 September 2007 (UTC)
I totally disagree with Spenny's characterizations of the policy and the comments in support. In actuallity:
The NOR policy is not going to so significantly change that editors will not be required to consider sources based on their relationship to the view or concept. That's simply never going to fly. FeloniousMonk 14:49, 21 September 2007 (UTC)
<undent> Thanks for listing the concerns, Spenny. Right at the outset the first assertion really has to be substantiated:
> If evidence for this is well presented in the archives, it needs to be linked from here so that newcomers don't spend days hunting for it. There are issues with the other points which I'll try to come back on shortly, but time doesn't permit just now. .. dave souza, talk 18:08, 21 September 2007 (UTC)
Stepping back from the specific issues:
Blueboar notes, "but fails to explain how the classification of sources relates to the concpet of "No Original Research"." And asks for an explanation. Wikipedia has an NPOV and a V policy which incouraged editors to cite sources as away of includin diverse views in articles in an unbiased way. Several editors became aware that some people were citing sources to introduce material in articles in biased ways. Mere compliance with V was not enough. And biased edits often seemed to comply with NPOV because they clearly identified the point of view of the cited material. How were people seeming to comply with NPOV and V, while violating them, at least in spirit? By selective quoting, by duxtapposing different quotes in order to make the editor's point, and by using primary sources to argue (the editor's argument) that a primary source was wrong. So it became evident that it was not enough to insist that editors had to use reliable sources. We needed to emphasize that editors' views shouldn't go into articles, and we needed to show how one could seem to comply with V and CS and still be forwarding one's own arguments, which meant we need to show some dangers in how editors could use sources, and encourage other ways to use sources. Slrubenstein | Talk 22:24, 21 September 2007 (UTC)
(To replace the section entitled: "Primary, secondary and tertiary sources"...)
Citing the Right Source Materials (alternative suggested section title: 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 the right sources to back those different types of statements. Statements of fact should be cited to reliable sources that clearly demonstrate that fact. Statements of interpretation, analysis or conclusion should be cited to reliable sources that contain the same interpretation, analysis or conclusion.
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.
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.
above is the most recent as of today. Blueboar 15:55, 21 September 2007 (UTC)
Where are you going to propose this, so it can be scrutinized by the community? Or perhaps, if sufficient opposition is expressed, it will not even be proposed?-- Filll 17:13, 21 September 2007 (UTC)
Reject NOR by definition is a proscriptive policy. While we should be clear about what is allowed and what is not allowed, the emphasis is on what is not allowed - e.g. original synthesis. The Roosevelt example is unhelpful because, even if it is permissable, what readers of this policy (which says "no") are going to need most clarity on and explanation of is what they should not do. We have "Cite sources" to encourage people to cite sources. This policy is about the different ways sources (which in general we have to use) can be misused and discussion of sources shoudl be with this purpose in mind. Slrubenstein | Talk 17:37, 21 September 2007 (UTC)
Comment there was widespread disagreement about what the current written policy, particularly the source-typing sections, mandates, but there was more agreement about what it aims for; there was also disagreement about how it is used in practice, which may vary from page to page. The proposed text is supposed to describe the current policy (in its aims and in practice) more clearly than the current text does. The old source-typing can be improved on another page, to remove discrepancies between Wikipedia descriptions, and to address differences between Wikipedia and various scholarly fields. Can anyone point to one place where the old text discouraged a practice, which was generally rejected, and the proposed one encourages it, or vice-versa? Jacob Haller 18:37, 21 September 2007 (UTC)
Okay, either I missed something in the wording above or I need new glasses. Somehow I missed that this (today's) version of a "proposed" substitute for the current "Sources | Primary, secondary, and tertiary sources" section was being put to a vote. It solely appears to me that it is simply todays version for people to comment on or suggest additional additions or other changes.
In regards to the various comments from some new names I see up above, let me try to clarify a couple things again. For at least a month now, there has been much discussion primarily focused on the "Sources" section of this policy. Many of us (actually all of us) who have been in engaged in the 'limited' discussions on this agree that we do not want to water down the policy in way, shape, or form. But, many of us have also come to agree through many 'spirited debates', that including the "Primary, secondary, and tertiary sources" section confuses the policy to a point that it currently provides more confusion than guidance, especially when viewed in conjunction with the other policies and their rewording of the same principles. We (the many editors participating in this discussion over the last month or so), are not trying to back-door anything. We thought that it was simply best if we could get close to something that we could agree on, and then tell others what we are working on, so that maybe reaching some kind of concensus would have been simplified by many arguments already having been discussed and argued back and forth for a while. That is why we finally posted this 'proposal' over at Wikipedia:Village pump (policy)#Proposed change at WP:NOR. If we wanted to backdoor this, we definately wouldn't have advertised our intentions or asked for further input from others.
So again, I will ask you to think about why a policy (in this case one about 'No Original Research') needs to also include extraneous definitions and examples of what the various types of source material are? What does that have to do with "No Original Research"? We can clearly state that original research is definately not allowed without worrying about how to classify different kinds of source material. If not, then I think we should all go back to school for a few more years.
Including the "Primary, secondary, and tertiary sources" section in the policy is like having an article about Apples, and then having a sub-section devoted to disambiguations, so road-apples and horse-apples aren't included as well. It may be related, but nothing at all to do with the main subject of the article.
Instead of just saying Disagree, try offering some constructive criticism on how to make it better. wbfergus Talk 18:52, 21 September 2007 (UTC)
I think we have two questions. Both came from the source-typing debate, but each can be dealt with individually:
There were several attempts to (1) fix source-typing within NOR (to resolve different definitions) and (2) address various concerns regarding appropriate uses of primary sources, appropriate uses of secondary sources, etc. within NOR. I think we eventually agreed that RS could resolve the occasional issues with the sources themselves, and that source-typing didn't help with the more common issues with their misuse. One proposal referred to "primary material" and "secondary material" within each source, then "interpretative material" for secondary, and eventually that led to this. Jacob Haller 19:07, 21 September 2007 (UTC)
Does anybody actually have any specific criticism of the proposal? Because "it's not an improvement" is not a criticism. The question is, is it an improvement over saying absolutely nothing on the subject? Because that's our alternative. If we can't find consensus on this topic, we're going to have an empty policy that helps nobody. I'd encourage people to take a good look at the proposal and see if they can't improve it. It's the closest thing we've had yet to a consensus proposal. COGDEN 01:14, 22 September 2007 (UTC)
...... Admittedly there remain residual bits of confusion. One user mentioned Encarta and Britannica, which are also "tertiary sources" according to the delineation presented in WP:NOR. I pointed out somewhere in these rapid-fire multiple threads today there is a wealth of more specialized tertiary sources such as specialized encyclopedias, to which we can also add many textbooks as tertiary sources. Indeed one could readily also argue that there are numerous sources that are quaternary, quinary, senary, septenary, octonary, etc. etc. Categorizing sources as primary, secondary and tertiary has, nonetheless, served quite well in clarifying the basic types of sources from which material in WP is drawn. If there is a proposal to better clarify the distinction or explain that the delineations between them are not always sharply drawn, or whatever, that would, as I said in the section below, be quite welcome and potentially helpful. But the proposal to replace it with the language at the top of this talk section is, IMO, completely unacceptable on the justifications given thus far. I believe I used the words "summarily reject" to covey my summary opinion about the proposal. ... Kenosis 02:47, 22 September 2007 (UTC)
In an effort to minimize confusion regarding how this current discussion of replacing the "Sources" subsection on PSTS came about, please refer to the Archive for the last month or two. It's an awful lot of reading I know, but it aslo way to much to even begin to try to re-link back into here. This is not a "let's get this done today" proposal. We have been discussing this for a while and will continue to discuss even longer.
Take your time to read the Archives and get caught up on what you think the pertinent issues are before posting comments, especially derogatory or accusatory comments. Probably the very earliest this "proposed change" would or could (or even should) take effect would around the end of October. We need plenty of time to assimilate everybody's constructive feedback, comments, and suggestions. Thanks you. wbfergus Talk 19:16, 21 September 2007 (UTC)
Okay, a new day, starting to get my caffeine and nicotine levels back to normal, and I had about 6 hours of sleep, so let's see if I can coherently try to state what we are attempting here. This perticular subject of what we are attempting has been scattered all over this page the last few days, so maybe this section title will draw a few of the newer people to this discussion here.
First, we are not, in any way, shape or form, attempting to weaken the NOR policy. We are attempting to make the policy itself more coherent and concise, by having this policy only deal with NOR related issues.
Second, this mainly came about because of numerous edit-wars on many different kinds of articles where one of the parties stated something was "OR". Many of these also related directly to the interpretation of primary, secondary or tertiary sources that were based solely on this policy's brief definition of them, not on the more detailed explanations (and examples) offered on their respective pages at Primary source, Secondary source, and Tertiary source.
Third, "Sources", in one form or another, appear of multiple policies and/or guidlines. For an example, open each of these in a new window and compare:
These all attempt to describe the same issue, but in different ways. Especially in regards to source-typing, why should each policy even attempt to re-define these, when they are easily linked to the aforemention articles created solely for the purpose of defining the various types of sources?
Over the last month that I have been here, there have been numerous examples of how OR can be inserted into an article, either intentionally or unintentionally, regardless of the type of source used. As a matter of fact, as several other threads on this particular page (not even in the archives yet) have shown, secondary source material can be mis-applied just as easily (if not even more so) to create original research than just quoting the raw data from a primary source can. With secondary material, one could easily cherry-pick various sentences from the source(s) to build an argument promoting a POV that is not even contained within the source. This is usually much more difficult to achieve with data from primary sources, unless the editor tries to add their own conclusion or interpretation, which can easily be challenged to cite an appropriate source to verify the claim.
Trying define a type of source (some sources can also be multiple kinds of sources, depending on the context of what is being cited to it), and general blanket statements can easily be either misconstrued or otherwise misapplied by different editors at different times. The type of source has nothing at all to do with whether or not OR is being inserted into an article.
In short, many of us (around 20-30 different editors in the last month), though still not in complete agreement on wording or implementation yet, are starting to lean more towards this policy (and maybe the others) being more like an insurance policy. Everything is not defined just in the one policy. Much of what is in your insurance policy actually comes in whole or in part, from either another policy or other 'document'. The entire policy is the result of multiple "definitions" taken from various documents, and there is no overlap of those definitions when all the pieces are put together. With the current state of the policies, there is much overlap and redefinition of the "sources" subject over and over and over, with variations in those definitions. To alleviate confusion, we would like to remove this confusion from this policy and place it in a more logical place, that (hopefully) the other policies could link to similarly. We are not proposing to remove WP:PSTS, just move it out to a better place, so that PSTS would still be part of the policy, just not defined within this policy. Similar in wording to how this policy and the others, state "Since NPOV, V, and NOR complement each other, they should not be interpreted in isolation from one another, and editors should try to familiarize themselves with all three". Instead,, probably something along the lines of "Since NPOV, V, NOR, BLP and 'source typing' complement each other, they should not be interpreted in isolation from one another, and editors should try to familiarize themselves with all, as a good Wikipedia article will adhere to all".
I probably worded that last sentence poorly, but I think you should be able to get the idea of what we are trying to accomplish. If the "source typing" was moved into it's own article, having the common definition that the others can all share makes this concept easier to comprehend. I think more thought and effort needs to go into what and where this "source typing" article should be, like a guideline isn't strictly enforcable, but do we really need a policy to define the types of sources?
Thanks for reading this, hopefully this may alleviate some concerns about this being conducted behind closed doors as an attempt to weaken current policies. wbfergus Talk 11:02, 22 September 2007 (UTC)
Would it make sense to create a "proposal" page... explaining the rational for the change, the proposed language, and the various options on implimentation (such as a) moving PSTS to some other policy or guideline, or b) creating a new guideline out of PSTS)? I know something like that is really meant for entire page revisions or new proposals.... and it is rare for policy change discussions, but it might help people FIND what is being discussed, and get a sense of the basic issues (pro and con) without having to sort through all the chatter or the archives. If people like this idea, I would suggest such a page say prominently that that comments continue to be made here, on this page. (My concern would be that someone would come along later and the say... "gee, I didn't even know the issue was being discussed"... my suggestion is so that both the proposal and it's discussion are prominent. Blueboar 19:39, 21 September 2007 (UTC)
I would suggest a cautious approach. I do not think that any of these wide-ranging changes will have any traction. Work on improving rather than changing may be a more fruitful approach. ≈ jossi ≈ (talk) 00:00, 22 September 2007 (UTC)
"Right" sources? That's hovering dangerously near "truth" - "right" is inherently open to interpretation, and is subjective. Primary, Secondary, and Tertiary are clearly defined and not open to abuse of interpretation. I do not see how reducing clarity and confusing the issue and inviting arguments by using subjective phrasing is in any way anything but a recipe for disaster. KillerChihuahua ?!? 00:02, 22 September 2007 (UTC)
WP:V requires us to rely on reliable, third-party published sources, and the NOR policy at present gives an exemption in that primary (or first party) sources can be used for statements of fact that require no analysis, in the context of analysis by secondary (or third party) sources. It seems to me that this distinction is being lost in changing the emphasis to what's really an expansion of "However, care should be taken not to "go beyond" the sources or use them in novel ways." If we accept the analysis that the primary/secondary/tertiary analysis is so broken that we have to make the radical change proposed (and evidence of this breakdown is needed for those of us who can't find it in the archives), we have to ensure that any relevant requirements under the present policy are fully covered by the new policy, and also have to provide clarification for editors used to the current terminology. This applies both to experienced editors who haven't re-read the policy for a while, and also to new editors advised in discussion that use of a source goes against WP:NOR in the old terminology. Here's a suggested addendum to the current proposal:
Care must be taken to accurately represent the overall sense of the source, and not introduce our own interpretation in summarising or selectively quoting from the source. Statements of fact can draw on documents or people very close to the situation being written about, often described as primary source material. Statements of interpretation, analysis or conclusion must draw on third party material, often called secondary sources, which draw on primary sources to make generalizations or interpretive, analytical, or synthetic claims. For example, we can state the fact that when Charles Darwin was finishing proofs of The Origin of Species he wrote to a friend "So much for my abominable volume, which has cost me so much labour that I almost hate it." However, we can't use this source from the author to state the conclusion that "the book is abominable or almost hateful", but for an assessment must find a reliable third party source giving the opinion of a recognised current expert in the field.
Obviously the example given [1] is the first that came to hand, and no doubt a better one can be found. There may be other points from the current formulation that need to be covered. Aspects of source categorisation not closely related to NOR could be covered in WP:RS, or in the proposed new guideline. If the "third party" definition needs clarification, that has to be reflected across all policies and guidelines that use the term. .. dave souza, talk 10:00, 22 September 2007 (UTC)
Back in another discussion taken from here, someone implied that primary/secondary source conflicts were "rare". I have now had to treat yet another case of a mistake in a secondary source where appeal had to be made to a primary source. This particularly seems to be a issue with BLPs of people who are notable enough for media sources, but who lack conventional biographies. These media sources make lots of mistakes; fact-checking (or at least cross-checking) is obligatory when questions arise.
As I commented over in the discussion of WP:RS, our policies don't treat this issue well. The problem in WP:RS is that the issue of source errors isn't addressed at all, so that being correct isn't a criterion for being reliable. The problem here is that the constant hammering on avoiding primary sources is ratifying the notion that secondary source errors are to be preferred over primary sources.
One point I consider non-negotiable: when a primary source is quoted in the article, the primary source should be cited if possible. In this case, a secondary source should be never, ever, be preferred over a primary source. Likewise, when a secondary source misrepresents a primary source, whether by accident or design, the secondary source statement must not appear in Wikipedia unless it is necessary to explicitly deny its accuracy. These are the absolute standards of non-fiction writing, and we cannot credibly contravene them. Whatever other rules we establish, these must not be contravened. The problem I've consistently seen in the discussion here is that the discussion of sources is so abstract that it doesn't consider the ways that secondary sources use primary sources. Some of these uses it is not our purpose to argue with, but some of them, we have to review. Mangoe 14:46, 22 September 2007 (UTC)
This issue of verifiability and reliable sources is one thing. The issue of how those sources are used in creating an article is another. Of course they interact. All three of the basic editorial policies interact -- WP:NPOV, WP:VER and WP:NOR. WP:NOR requires dealing further with the issue of what kinds of syntheses are generally available among the sources. The issue of whether they're "right" or "wrong", as has been explicitly stated in WP's policy growth and development from the very beginning of WP, is not the issue. Odd as it may seem at first blush, the original mandate from the founding Board (read that:"from Jimbo Wales") that the standard is "verifiability, not truth", still controls in Wikipedia. What's required are sources that show a published weight of opinion about a particular assertion of fact or conclusion that can be drawn from published information. But that's more for WP:VER and its offshoot WP:RS. For WP:NOR, the issue centers around relying on published sources and not formulating new syntheses or drawing new conclusions that are not already available in the published literature. When the published literature conflicts about the basic facts or conclusions, WP:NPOV#Undue_weight controls the method by which the material is to be presented. As I said earlier here and as others have stated too (not necessarily in this discussion), the "big three" basic editorial policies work in tandem and in tension with one another. That interaction and tension among the "big three" is mediated by consensus process at the local article level. For the purposes of this policy page, WP:NOR, it is useful and appropriate to provide a framework or basic breakdown of the range of sources available, which the primary, secondary, tertiary delineation does quite effectively. ... Kenosis 15:51, 22 September 2007 (UTC)
(To replace the section entitled: "Primary, secondary and tertiary sources"... which is to be moved elsewhere)
Citing Appropriate Source Materials (alternative suggested section title: 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.
I tried to address the concerns that had been expressed and to incorporate some of the wording from dave souza's proposal. Further discussion and proposed changes are welcome. -- Coppertwig 15:19, 22 September 2007 (UTC)
How about combining the titles and dropping the "right" (vs. wrong) issue (and using wiki-caps) as follows: Citing appropriate source materials, and sticking to those sources -- Dhaluza 16:02, 22 September 2007 (UTC)
..... The section on Primary, secondary and tertiary sources is a vital section that puts WP:NOR into perspective for users around the wiki. The proposal to replace the section or move it elsewhere is a major change in this policy page that I find unacceptable, at least without something that represents a clear improvement on it. Since well over three years ago, a distinction had been made between primary and secondary sources and what uses constituted original research or original synthesis. A significant number of WP users remained confused about the distinction, e.g., "well, if WP is not supposed to synthesise material directly from primary sources (except for purely factual matters), and if we're supposed to use secondary sources, then how is WP a secondary source?" The extra note about tertiary sources was inserted into WP:NOR nearly a year ago, and suddenly it made better sense to most other participants and much of the confusion diminished substantially. The description of encyclopedic matter as primarily tertiary helped to resolve this confusion right on the policy project page. Since then, it's tended to serve us well in resolving confusion about what is original synthesis, or original research, as distinguished from original wording and editorial decisionmaking.
...... Admittedly there remain residual bits of confusion. One user mentioned Encarta and Britannica, which are also "tertiary sources" according to the delineation presented in WP:NOR. I pointed out somewhere in these rapid-fire multiple threads today there is a wealth of more specialized tertiary sources such as specialized encyclopedias, to which we can also add many textbooks as tertiary sources. Indeed one could readily also argue that there are numerous sources that are quaternary, quinary, senary, septenary, octonary, etc. etc. Categorizing sources as primary, secondary and tertiary has, nonetheless, served quite well in clarifying the basic types of sources from which material in WP is drawn. If there is a proposal to better clarify the distinction or explain that the delineations between them are not always sharply drawn, or whatever, that would, as I said in the section below, be quite welcome and potentially helpful. But the proposal to replace it with the language at the top of this talk section is, IMO, completely unacceptable on the justifications given thus far. I believe I used the words "summarily reject" to covey my summary opinion about the proposal. ... Kenosis 16:52, 22 September 2007 (UTC)
<undent> Plain English or not, this now refers to secondary sources without making it clear that sources close to the subject, which for most of us is plainly and simply expressed by the term "primary sources", can be used for facts but not for interpretation etc. This is still less clear than the existing policy. The Darwin example I showed above was designed to bring out the point that a fact about an openly attributed opinion can be stated, but that opinion from, eh, a primary source can't be used as analysis, conclusion or whatever. That's a point which still seems to be lacking, and some thought is needed as to whether this proposal would have other unintended consequences. .. dave souza, talk 23:34, 22 September 2007 (UTC)