Good point, Kla'quot. It's clarified a little in the FAQ, but it should be tweaked here too. SlimVirgin ^(talk) 08:57, 16 February 2007 (UTC)[reply]

It generally seems like a bad idea for Wikipedia to prejudice itself against a whole medium of publication, namely blogs. After all, we would not want others to be prejudiced against Wikipedia solely because it is online. Individuals' or group blogs should be considered similarly to other published sources of similar scale. For some blogs, that means treating them like a fanzine; for others, it means treating them like a newspaper. For that matter, many small-town newspapers are on about the same level of reliability as many news-oriented blogs .... --FOo 07:44, 20 February 2007 (UTC)[reply]

Blogs are not a medium. Reliability as a published source for the purposes of Wikipedia is based on factors involving fact-checking and lack of conflict of interest and past history of accuracy and objective editorial mechanisms. Blogs generally do not qualify as reliable under these standards. WAS 4.250 08:21, 20 February 2007 (UTC)[reply]

Talking Points Memo is a blog. It is frequently cited by major newscasts, sometimes without attribution (aside from the inadvertent logo found on the picture they use). Josh Marshall is a respected journalist and it is his own privately funded publication. Just pointing out that not all blogs are created equal. -Sparaig 19:39, 22 February 2007 (UTC)[reply]

The current text states that, "Wikipedia articles should rely on reliable, published secondary sources wherever possible." In general this is certainly true, but there is a particular type of text where I strongly disagree with it. When talking about the content of a primary source (such as a book, play, transcript, et cetera), with no interpretations or conclusions, it seems to me to make more sense to cite that primary source directly than to cite a secondary source which itself quotes or references the primary. For interpretations you need a secondary source, but for simple citation the primary is the place to go. For example, a statement that "Call me Ishmael" is the first sentence of the story, should be sourced directly to Moby Dick... not some secondary source which says the same thing. This would seem self evident, but 'wherever possible' would argue for taking a secondary source first. In the past that wording has been 'interpreted' to argue that primary sources should be avoided. There is a sentence in the previous text explaining that the 'wherever possible' wording, "means that we present verifiable accounts of views and arguments of reliable scholars, and not interpretations of primary source material by Wikipedians"... but that clarifying sentence (that 'wherever possible' is meant for views/arguments/interpretations - not simple factual citations) has not been copied over.

Thus, if we want to keep it short I'd suggest changing the wording to something like, "Wikipedia articles should rely on reliable, published secondary sources for analysis and interpretation wherever possible." That maintains the intent of the existing policy without the additional sentence.

Another bit of lost text is, "However, research that consists of collecting and organizing information from existing primary and/or secondary sources is, of course, strongly encouraged." I think it is important to include this for the frequent cases where people claim, incorrectly, that such collections are 'original research'. Per this lost sentence, pages that collect information from multiple reliable sources into a 'unique' (not previously published) list are NOT original research. Which is good... because otherwise almost all of our Featured lists would have to go. --CBD 12:13, 16 February 2007 (UTC)[reply]

I understand your first point, but think that the change would be an innovation and should be held-off on. And "Wikipedia articles should rely on reliable, published secondary sources for analysis and interpretation wherever possible" accidentally leaves the barn door open. All analysis and interpretation should be attributed to reliable, published secondary sources, not merely when possible (if it's not possible, don't add it).

The second point is, I think , a corollary to "original prose, not original research". Articles ought to bring sources together in a unique way; they just shouldn't advance any unique ideas. I'm wondering if there should a section for this (note there's one at the bottom of the FAQ now) but am worried about bulk here. Marskell 13:10, 16 February 2007 (UTC)[reply]

My interpretation of "Wikipedia articles should rely on reliable, published secondary sources wherever possible" is that articles ordinarily should state conclusions and interpretations about the topic, which in turn requires secondary sources. The dumping of raw facts, which is what you get from primary sources, does not usually lead to worthwhile articles (but there can be exceptions). That's why it says "should" and not "must".

As for "original prose, not original research", my attitude is good, free, information is the main goal, and if we can achieve that by copying a public-domain or other free source word-for-word, do it. So the policy should not be worded in a way that frowns upon the copying of free information. --Gerry Ashton 15:02, 16 February 2007 (UTC)[reply]

"My interpretation of "Wikipedia articles should rely on reliable, published secondary sources wherever possible" is that articles ordinarily should state conclusions and interpretations about the topic, which in turn requires secondary sources." So is mine; I was pointing out that inserting "analysis and interpretation", with "wherever possible" as an adverbial, creates a loophole. Marskell 17:20, 16 February 2007 (UTC)[reply]

The loophole is real, and should be there. Occasionaly it is appropriate to base an article on primary sources, if the article is a recital of bare facts without interpretation or analysis. Whether it is sufficient to say "should" rather than "must" to allow this loophole, or whether we should also say "whenever possible" is a matter of editorial style. --Gerry Ashton 17:25, 16 February 2007 (UTC)[reply]

"Occasionaly it is appropriate to base an article on primary sources, if the article is a recital of bare facts without interpretation or analysis." We're agreeing with one another. "Interpretation and analysis" doesn't need to be inserted because the exception is for cases where none is present. Marskell 17:28, 16 February 2007 (UTC)[reply]

If I were to tweak the section, I would leave "whenever possible" but I'd change the last sentence: "This means that, as a rule, we publish the opinions of reliable authors, and not the opinions of Wikipedians who have read and interpreted primary source material for themselves." --Gerry Ashton 17:45, 16 February 2007 (UTC)[reply]

I agree. We don't need that loophole. Septentrionalis PMAnderson 03:27, 17 February 2007 (UTC)[reply]

Apologies. I don't have a lot of time at the moment to study this closely, but a cursory glance suggests that a few of points I've relied on in the past will no longer be there. (I do, however, think this merger is excellent and really reallly appreciate the clarity and organization.)

I've always liked this definition of secondary sources: "The informed and expert interpretation, interpolation, extrapolation or corroboration of primary sources to synthesize a conclusion." It seems like the new policy opens up the use of tertiary sources more than previously. Here's how it was constrained before: "Tertiary sources can be used for names, spellings, locations, dates and dimensions." Also, I thought this section in wp:nor was important: "Synthesis of published material serving to advance a position."[1]. Also, I don't see the stipulation that scientific research that's referenced should have been published in an academic journal and peer-reviewed.

Thanks to all those who work so diligently to create these policies and guidelines. TimidGuy 16:52, 16 February 2007 (UTC)[reply]

Partial reply: " However, that would be an example of an unpublished synthesis of published material serving to advance a position, and it constitutes original research" is still in. Marskell 17:01, 16 February 2007 (UTC)[reply]

In reply to TimidGuy, using a good tertiary source is better than having no article on a subject. I think it's better to provide free access to information that is good enough than to ignore any topic unless some editor can produce a superb article on the topic.

As for scientific research, it is not feasible to limit ourselfs to academic peer-reviewed journals. We would have to eliminate sources such as Scientific American and the science section of the New York Times, the web sites of high-tech companies, and other useful sources. We would also have to eliminate textbooks, so we couldn't write about the fundamentals of science, only the topics that have been the focus of research within the time-span of the editor's journal collection. --Gerry Ashton 17:12, 16 February 2007 (UTC)[reply]

Thanks, Marskell. Good to know that's there. And good points, Gerry. The examples you cite would fall under the guideline for non-schoarly sources, and hence be admissible. But wp:rs is fairly clear in requiring peer review for scholarly sources: "The material has been thoroughly vetted by the scholarly community. This means published in peer-reviewed sources, and reviewed and judged acceptable scholarship by the academic journals." The merger notice said there wouldn't be major revisions in the merger, just a reorganization. So I assume that the intent is still to maintain this standard of peer review for published scientific studies.

I see in a thread above that the FAQ is still a work in progress, so maybe this is yet to be incorporated. TimidGuy 19:48, 16 February 2007 (UTC)[reply]

TimidGuy wrote at the beginning of this secton "Also, I don't see the stipulation that scientific research that's referenced should have been published in an academic journal and peer-reviewed." I don't see any stipulation in [[WP:RS] about scientific research having to come from scholarly sources, only a description of scholarly and non-scholarly sources. There are some recommendations on the inactive page Wikipedia:Reliable sources/examples, but they don't carry to much weight considering the page is inactive.

Also, remember that "Reliable sources" is a guideline, while "Attribution" is a policy. When writing a policy, care should be taken not to forbid anything that ought to be acceptable, even under unusual conditions. Guidelines can be written to focus on the usual situation, and the best practices, because people will be more willing to violate a guideline when the circumstances call for it. --Gerry Ashton 20:27, 16 February 2007 (UTC)[reply]

Thanks, Gerry. Good points. Appreciate your thoughts. Note that Blueboar just today removed the "inactive" template from the Examples page. But it's hard to know whether it will have a place post-merger. I hope it does. TimidGuy 21:59, 16 February 2007 (UTC)[reply]

Oops. SlimVirgin just put back the inactive tag. TimidGuy 22:08, 16 February 2007 (UTC)[reply]

The issue about Scholarly sources is a section re-inserted by Slim-Virgin and jossi into RS, who then refused to engage in any discussion about the value. I am aware that the whole ATT/ RS issue appears to be considered as a political p!ssing match in some areas and tbh lack of engagement doesn't help ones confidence about future progress. The wording does not mandate the use of peer reviewed sources, indeed it's an area which has a great many weaknesses.

The examples page had the potential to be quite useful, but things moved on. It's guidance, rather than a guideline.

ALR 10:11, 17 February 2007 (UTC)[reply]

I would like to see WP:RS incorporated as a subpage of this policy to reduce as much redundancy as possible, but I'm not sure about the FAQ format, although the symmetry with NPOV is nice. I think we should wait until SlimVirgin is done with the FAQ, and then see how much RS material is still needed. --Merzul 07:34, 17 February 2007 (UTC)[reply]

Thanks, ALR. Interesting to hear that background of the situation. I guess Merzul is right -- we can wait until the FAQ is done and then see what's still needed. I agree that the Examples page could be very useful. TimidGuy 12:58, 17 February 2007 (UTC)[reply]

Centrx, could you say what you're trying to achieve with your edits, please? The problem is that some of the edits actually change the policy (as in V or NOR), and we've been extremely careful not to do that here. WP:ATT is simply a merger of V and NOR. Although there's a consensus to merge them, there's no consensus to change them. SlimVirgin ^(talk) 23:12, 16 February 2007 (UTC)[reply]

There are three separate sorts of changes:

• The third paragraph states that this is one of two core content policies. As a policy, this is not established without broader advertisement that this merger is in fact happening, now, and consensus for it.
• There are several purely formal changes to spelling and grammar, which do not change the meaning and in regard to tone are more similar to the tone of WP:V and WP:NOR. These changes are much less significant than the many minor textual changes that have been made in the merger of the pages.
• There are a few somewhat less minor changes which can be delayed in the interest of making the purity of the merge clearer, but which should be changed in the future and discussed if there are problems with them.

—Centrx→talk • 23:22, 16 February 2007 (UTC)[reply]

"Wikipedia:Attribution is one of Wikipedia's two core content policies. The other is Wikipedia:Neutral point of view." seems like a straightforward combination of "Wikipedia:Verifiability is one of Wikipedia's three core content policies. The other two are Wikipedia:No original research and Wikipedia:Neutral point of view." and "Wikipedia:No original research (NOR) is one of three content policies. The others are Wikipedia:Neutral point of view (NPOV) and Wikipedia:Verifiability (V)." Am I missing something? --Akhilleus (talk) 23:30, 16 February 2007 (UTC)[reply]

That's fine, but there should be widespread knowledge that the merger is happening and the opportunity for everyone to verify that it is sound before it is officially implemented as policy. It is also inconsistent to have both WP:V and WP:NOR state that each is "one of Wikipedia's three core content policies" while at the very same time this one states that it is "one of two core content policies". This cannot be the case until Wikipedia:Verifiability and Wikipedia:No original research are either tagged as {{historical}} or redirected here, and it sticks, and before that can happen it must be widely known that it is going to happen. Separately, there should possibly be some message put at the top of the front-facing WP:V and WP:NOR pages, but a regular {{merge}} tag does not seem appropriate. —Centrx→talk • 23:39, 16 February 2007 (UTC)[reply]

This was already announced on WP:VPP, WP:AN, and has already been at RFC; during the 4 months it was mentioned many times on various policy related talk pages. I notice you are re-announcing it everywhere, but please don't make too big of a deal out of this, as we don't want a slashdot article about this! --Merzul 09:02, 17 February 2007 (UTC)[reply]

It was announced as being worked on and in progress as a proposal, not that it was certainly going to be implemented and not at a final version; many proposals get such announcements but fail. Regardless, even if WP:ATT were absolutely and without exception perfect such that there would be no need or benefit from wider input before its implementation, it would still need to be widely advertised so that people actually know about it and not continue to refer to Wikipedia:Verifiability. And this still has the problem where "attributable" is less clear than "verifiable". —Centrx→talk • 15:11, 17 February 2007 (UTC)[reply]

I haven't read too much into the history of this effort, but a quick search of the word "Foundation" in the talk archives hasn't really shown me what the Foundation's reaction to this has been. If I recall correctly from my quick glances, some people said it wasn't technically a Foundation issue. Even if it is not technically a Foundation issue it's still a big change (well, not really "change", but update), and I'd really like to know what their thoughts are on this. I know a lot of this kind of stuff is up to us to figure out, and we don't have to "ask mommy and daddy" every time we want to change something, but I am surprised at the lack of talk about the Foundation's involvement in the past discussion. (Maybe I'm just missing it or searching for the wrong keywords?) Would anyone like to fill me in on this? -- Ned Scott 23:33, 16 February 2007 (UTC)[reply]

I don't think that combining/merging the two existing policies, WP:NOR and WP:V, is a fundamental or substantive change that requires the Foundation's blessing. It is more a formatting change, which makes the content policy easier to work on consistenly and coherently. As of now, it is essentially a reflection of the two policies and hence a non-issue from their perspective, IMO. Crum375 23:39, 16 February 2007 (UTC)[reply]

I agree that a merger wouldn't require the Foundation to be involved; even so, I e-mailed them about it when this proposal was initiated. SlimVirgin ^(talk) 23:42, 16 February 2007 (UTC)[reply]

Who? You mean the wikipedia-en discussion? —Centrx→talk • 01:17, 17 February 2007 (UTC)[reply]

I know, and I'm saying I agree that we don't need to ask them first, but knowing their input wouldn't hurt. -- Ned Scott 23:50, 16 February 2007 (UTC)[reply]

To be very clear, this page is intended to supercede WP:NOR, WP:V and WP:RS, correct? Presumably those pages would then redirect here?

If that's the intent, I think you need to restart the clock at the top of this page. My initial reactions are supportive but this hasn't been widely enough advertised yet. Rossami (talk) 23:58, 16 February 2007 (UTC)[reply]

Mmm... there are different opinions here, but it is widely advertised now, and the proposal has been developing for over four months during which it has been mentioned many times on the WT:V and WT:NOR discussions. So most "policy discussion" people have been long aware of it. Anyway, if your initial reactions were positive, and only object to it's suddenness, then I don't really think it's a big issue, because this isn't really a big policy change, and there is still time to discuss any discrepancies between this and previous policy. --Merzul 07:13, 17 February 2007 (UTC)[reply]

I like this. Only one query. Some of the articles I write are based on one source, there being only one secondary source available. The subjects are non-contentious, and I just place the reference at the bottom (see for example here). I would hate for a picky editor to go through and place {{fact}} templates throughout. It would seem pointless to place the same citation throughout the article. Maybe a mention on how this type of article might be handled? --Michael Johnson 00:36, 17 February 2007 (UTC)[reply]

This issue of whether inline citations are needed is not changed by this policy, so the balance shouldn't have shifted on way or the other. It is still up to the quality assurance projects like GA or FA to decide how detailed or fine-grained attribution they require. In my opinion, the phrase "challenged or likely to be challenged" should ensure that you don't need to cite every statement. --Merzul 06:49, 17 February 2007 (UTC)[reply]

Eventually every fact on Wikipedia should have an inline citation. Many articles start with one general reference - I do that too - but it is bad practice, because in the end the article will have to be brought up to the inline citations for every fact standard. Therefore I would actually advice to discard the entire phrase 'challenged or likely to be challenged'. Everything is 'like to be challenged' - with some commons sense in it (nobody should demand sources for styllistical changes). 'Editors should provide attribution for quotations and for any material' [they add] is much better then 'Editors should provide attribution for quotations and for any material' [challenged or likely to be challenged].-- _{Piotr Konieczny aka Prokonsul Piotrus | talk}  22:43, 17 February 2007 (UTC)[reply]

I'm liking the page, but is "attribution" the best name? When I first saw a link here, I thought it was going to be about GFDL-necessitated edit attribution. Picaroon 00:39, 17 February 2007 (UTC)[reply]

Yes, this is acknowledged, and the solution was to rename those "attribution" templates to "copyright attribution" templates. I don't remember, but this issue will be dealt with. --Merzul 06:52, 17 February 2007 (UTC)[reply]

Personally I think we should just keep using Wikipedia:Verifiability as the main title, as the.. dominate (for a lack of a better word) policy. -- Ned Scott 09:37, 17 February 2007 (UTC)[reply]

Hello, I just read through this, and I find it kind of confusing. I guess the thing that threw me off in the beginning was the usage of "attribution". The way I see it, the difference between "verifying" and "attributing" is that when you verify something, you just list your source somewhere (preferably inline), and are free to state whatever it is as a fact. On the other hand, attribution (as explained in WP:NPOV) involves actually giving credit to the source in the structure of the sentence (e.g. "According to X, Y", "X said Y"), thus avoiding stating Y as a fact.

The current version of WP:RS lists attributability as an aspect of reliability and gives another explanation:

Attributability—The more we know about the originator, either organisation or individual, of source material, the better. This helps us measure the authority of the content.

The other thing that throws me off is that WP:V and WP:OR don't seem that closely related to me. WP:V says you need sources and talks about sources are acceptable, and WP:OR is more about what you can do with those sources.

Anyways, maybe it will make more sense if I read through it on another day.

— Armedblowfish (talk|mail) 00:55, 17 February 2007 (UTC)[reply]

A more accurate name might be "Attributability". The page explicitly states that not all edits need to be explicitly attributed, so "attribution" is not quite accurate. —Centrx→talk • 01:21, 17 February 2007 (UTC)[reply]

I guess I am confused about the meaning of "attributability" as it is used in the context of this proposal. Perhaps someone could write a short definition in the intro? Also, attribution, as I understand it to be used in WP:NPOV, is something you generally do when the source is biased or less than ideally reliable, or the statement is disputed. So, yeah, still somewhat confused. Sorry. — Armedblowfish (talk|mail) 01:36, 17 February 2007 (UTC)[reply]

Attribution definitions:

• Indicates the source of information
• assigning some quality or character to a person or thing; "the attribution of language to birds"; "the ascription to me of honors I had not earned"
• assigning to a cause or source; "the attribution of lighting to an expression of God's wrath"; "he questioned the attribution of the painting to Picasso"
• In journalism attribution is the identification of the source of reported information. Journalists' ethical codes normally address the issue of attribution, which is sensitive because in the course of their work journalists may receive information from sources who wish to remain anonymous. In investigative journalism important news stories often depend on such information.

... etcetera.

In the context of Wikipedia, attribution refers to whether material is attributable to a reliable published source. ≈ jossi ≈ (talk) 02:35, 17 February 2007 (UTC)[reply]

From the 1st and 3rd definitions you gave, attribution sounds a lot like verification, although it is possible I am not picking up on the fine distinctions. In the final definition you gave, however, wouldn't it be better to avoid using a variation of the word in the definition? It is also entirely possible that this entire thing will make more sense to me when my head is clearer. — Armedblowfish (talk|mail) 03:14, 17 February 2007 (UTC)[reply]

I think you are such an experienced Wikipedia contributor that when you think about verification, you are actually talking about attribution. My definition of attribution is the following.

Attribution is the means to ensure "that any reader should be able to check that material added to Wikipedia has already been published by a reliable source."

The above is WP:V's definition of verifiability, and while attribution might have some problems, it is nowhere as confusing as calling the above process verifiability. Again, for experienced editors, the term "verifiability" simply means attribution, so this renaming may simply take time to get use to; but I think using "attribution" to describe this process is more accurate. In the world outside Wikipedia verifiability doesn't mean "that any reader should be able to check that material added to Wikipedia has already been published by a reliable source." In the world outside Wikipedia, verifiability has strong connections to notions of absolute truth, and this causes much confusion. So this policy is simply using the right dictionary word to describe what we, Wikipedians, mean when we talk about verification. --Merzul 06:22, 17 February 2007 (UTC)[reply]

I don't see a big problem with the point, given that I wrote that line into RS. Attributability applies at different levels; for an individual point within a discussion and for the discussion as a whole.ALR 10:14, 17 February 2007 (UTC)[reply]

Merzul - Thank you for that definition. It makes more sense, I think it should be included in WP:ATT.

ALR - So, providing a nice m:Cite.php citation, knowing about the author of the source, and saying "According to [source]" in the sentence in addition to a full m:Cite.php citation are all different levels of attribution? Can we use different words to describe these different levels of attribution?

Anyways, WP:ATT is starting to make more sense to me.  : ) However, given the controversy over whether reliability is on/off and can be determined with a checklist, or whether it is a flexible thing to be determined relative to the material being cited, I am not sure this much information on reliability should be covered in a policy. But, WP:V seems to have that problem too, now.

— Armedblowfish (talk|mail) 15:13, 17 February 2007 (UTC)[reply]

I'm glad it's starting to make sense to you, so yes how explicit an attribution is can vary from discrete footnotes to very strong prose attributions that distances Wikipedia from the statement. So in this terminology, those "according to..." statements are called "prose attributions". Regarding how much reliability information should be included, I tend to agree with you because most editors I've met have had good intuitive judgment on what are reliable sources; but some people feel strongly about this part. --Merzul 16:46, 17 February 2007 (UTC)[reply]

Whilst I'm not a fan of the according to... form of words I think you've got it. A single item can be sourced adequately and is attributable although it may not be reliable, an article made up of appropriately sourced material is attributable but may or may not be inherently reliable. Essentially source quality is what is missing from the proposal at the moment.

I'd agree with Merzul that in the vast majority of fairly vanilla, non-contentious, topic areas then people have a gut feel for what is reliable and what is not. In the more contentious areas, where source reliability is an issue, then the current wording is weak; gut feel is wholly inappropriate where different people will feel differently about source quality. Weaker quality sourcing needs a bit more caveating around it in the text though.

ALR 17:52, 17 February 2007 (UTC)[reply]

My own concern is that the word Attribution is non-intuitive. For newcomers to Wikipedia, it's always been very easy to remind them about the Wikipedia policies of "Verifiability" and "No original research". Those terms are easy to understand. But to remind people about "Attribution" isn't something that most people will immediately get. --Elonka 19:32, 18 February 2007 (UTC)[reply]

I think it might be best to just keep the name it "Verifiability". —Centrx→talk • 19:33, 18 February 2007 (UTC)[reply]

I think the following statement (2nd para., last sent.) should be emphasized in italics: "The burden of evidence lies with the editor wishing to add or retain the disputed material." It's an important way to extinguish arguments and, unhighlighted, easy to miss when skimming the material. Askari Mark (Talk) 03:07, 17 February 2007 (UTC)[reply]

Yes, I also think this is very important, and I strongly support this suggestion. The only reason I don't immediately apply it, is that being conservative during the transition week is a good idea. Otherwise, there are good arguments for this because compared with WP:V it has lost a bit of prominence, so I think we can make a case for italicizing it. Let's just wait a bit. --Merzul 06:31, 17 February 2007 (UTC)[reply]

Actually, I think this is the best time to do it, when a great many people will be looking here to see what may have changed. It does not in itself make a substantive change, but highlights it commensurate to its importance. However, I'm reluctant to be bold at this point in time without further consensus. Askari Mark (Talk) 18:19, 17 February 2007 (UTC)[reply]

By the strong definition of OR here, this doesn't have to be original research, merely unsourced. We don't know that someone hasn't connected X's citation habits with the Chicago Manual of Style until someone unleashes Lexis and other tools and looks. I'm not disagreeing with the argument here; but it should be rephrased for consistency. Septentrionalis PMAnderson 03:40, 17 February 2007 (UTC)[reply]

Ok, I see your point, but your objection seems to apply to any piece of original research, because we can never prove that it "cannot be attributed to a reliable source". Do you essentially agree there's difference between "not yet sourced" and "not reliably sourceable", or do you find this distinction not helpful at all? --Merzul 07:03, 17 February 2007 (UTC)[reply]

I think it is an important point -- the material is probably original research. We can't ever know, though, unless the author admits it, that any paticular material is original research. It may have been taken from a published document that we're unaware of and the author forgot to cite. JulesH 10:04, 17 February 2007 (UTC)[reply]

I think providing someone makes a reasonable stab at looking for a source, and can't find it, they are entitled to remove the statement as 'apparent OR' then the onus is on those wishing to retain it to source it.--Doc^g 10:28, 17 February 2007 (UTC)[reply]

I agree; but that's not what the present phrasing says. We have a useful distinction here; let's not muddy it. Septentrionalis PMAnderson 19:02, 17 February 2007 (UTC)[reply]

I have tried a rephrase for both objections in this section; if someone wants to go back, fine; but it would be better to try another draft. Septentrionalis PMAnderson 19:11, 17 February 2007 (UTC)[reply]

I think the whole plagiarism example is ambiguous. It's meant to be illustrating the point that given points A and B which have reliable sources, you can't a new point C if this argument has not been made in the sources. However, it's not at all clear from the example what points A, B and C are meant to be, particularly as the plagiarism case it is based on is quite complicated. As a result, different people have repeatedly been coming to different conclusions about what the example is trying to say.

I propose replacing the example in the synthesis section with a more straightforward example along the following lines:

Example:

"In Northtown there were 1000 violent crimes last year^{Reliable source 1}, whereas in Southtown there were 2000 violent crimes in the same period^{Reliable source 2}. Therefore there is more violent crime in Sorthtown than Northtown^{No source}."

Under the attribution policy, the conclusion that there is more violent crime in Southtown than Northtown is not permitted, where it is not backed up by an appropriate source. This is because coming to a conclusion like this requires careful consideration of the source data; for example, the two statistics might have been calculated in different ways, or using different definitions of which crimes are violent or not. Wikipedia is not in a position to verify this background research itself. Instead, a conclusion must be attributed to a reputable source, unless it is very straightforward and uncontroversial.

Lots of alternative examples along similar lines could also be considered. What do you think? Enchanter 12:13, 17 February 2007 (UTC)[reply]

This seems much better than the Dershowitz example. Ken Arromdee 15:34, 17 February 2007 (UTC)[reply]

I'm thinking of creating Wikipedia:Synthesis as a guideline/guidance/example page, and to focus discussion on SYNT. More examples of synthesis would be very helpful. However, there will be no changing of the synthesis section here until the merger is complete. --Merzul 17:08, 17 February 2007 (UTC)[reply]

I support waiting for the merger to complete before we make any changes - I'll raise the issue again later. Enchanter 17:49, 17 February 2007 (UTC)[reply]

It would make everything much cleaner to do a merge, then apply any changes afterwards per regular consensus finding. HighInBC ^{(Need help? Ask me)} 17:51, 17 February 2007 (UTC)[reply]

If you do that, it'll never get changed. This is really our one chance to fix problems and codify existing practice. If we can't do that, what's the point? Ken Arromdee 16:24, 19 February 2007 (UTC)[reply]

If I understand Centrx's concerns correctly, the issue was that the paragraph implied that WP:V and WP:NOR no longer existed as policy, which was incorrect. Does the current version work better? JulesH 10:02, 17 February 2007 (UTC)[reply]

If WP:ATT ends up actually replacing WP:V and WP:NOR then the old version was fine. If it doesn't, then the new version would be rather inaccurate. —Centrx→talk • 15:07, 17 February 2007 (UTC)[reply]

OK, for the week of its debut, this contradiction is a minor issue. It should read as we would want it to read when NOR and V are redirects. Marskell 15:19, 17 February 2007 (UTC)[reply]

This "week of debut" is a novel concept. What is it's purpose? People can view the proposal regardless of whether it is concomitantly listed as policy, and once made policy it need not be listed alongside WP:V and WP:NOR. —Centrx→talk • 15:23, 17 February 2007 (UTC)[reply]

Right, and people have had five months to view the proposal. The purpose with running concurrently should be obvious: we couldn't just up and redirect V and NOR. You're asking for more exposure before the final decision, which is exactly what we're trying to do. Marskell 15:30, 17 February 2007 (UTC)[reply]

First, the proposal has changed. If people have had five months to view the same proposal, it could have been implemented several months ago. Second, in those five months there are numerous people who thought either "I don't care to become deeply involved in the formation of this; I will review it when it is actually done." or "This is going to fail anyway, as many proposals have failed in the past, so there is no reason to bother with it." Running concurrently gains no extra exposure unless the proposal is advertised, which is being done, and, even after acceptance for the page, there is no need for them to run concurrently. WP:V and WP:NOR can be tagged with {{historical}} for a while, for at least the week in which you proposed to run them concurrently. Doing that would gain more exposure even purely on these three pages; currently, there isn't even any message about a merger at the top of the WP:V and WP:NOR front pages, which is the only way you could transition people over. Without a notice there, a week in which they are concurrent is useless. —Centrx→talk • 16:36, 17 February 2007 (UTC)[reply]

The proposal has changed to mimic current policy exactly; there are no innovations to review. The principal reason to have the policy tag on while it's being announced on WT:V etc., is that people will be less likely to try and introduce changes. The policy tag forces stability on the page. But it would've been too early to redirect V and NOR at the point of announcement. Hence the three at once.

Numerous people have thought "I don't care to become deeply involved in the formation of this; I will review it when it is actually done." or "This is going to fail anyway, as many proposals have failed in the past, so there is no reason to bother with it." Well, good for numerous people. We can't read the minds of those who aren't here. Marskell 16:49, 17 February 2007 (UTC)[reply]

They must be notified, so they are here now. —Centrx→talk • 16:59, 17 February 2007 (UTC)[reply]

Yes I agree with Marskell on this stability issue, there is already 9 pages of archives. So Centrx, what is the best way of letting people review and answer "do you think this is an appropriate merger" without suggesting improvements to policy, such as changing the synthesis example? If there is a proposal tag on there, we're going to spend another 5 months on this. --Merzul 16:59, 17 February 2007 (UTC)[reply]

I doubt it. People will make changes regardless of whether that tag is on there, and new things that are not found in WP:V, WP:NOR, or WP:RS were added to this page even by people that were involved here prior to the announcement. In any event, policy pages are not static. —Centrx→talk • 17:04, 17 February 2007 (UTC)[reply]

For Centrx

Centrx, please get on board with this. The merge of V and NOR has widespread consensus; in over two years of editing, it's the thing that attracted the second largest number of e-mails I've had over any single issue, all of them very postive, very enthusiastic. And on-wiki, I don't think I've seen any negative responses, except for you and one or two others; and yet, I feel you probably also support it in principle.

All that has happened here is that, instead of having the words spread out over two pages, they are now on one; the writing has been tightened; we've gotten rid of the word "verifiability," a word that caused a lot of confusion, because it was the wrong word to describe checking source material; we've found a word instead — attribution — that incorporates the idea "don't add OR but not every single sentence needs a source," which is "material must be atttributable, but need not actually be attributed."

That's it. There are no actual changes to policy. Everything that mattered in NOR and V has been carried over to ATT (and if you feel that a principle you relied on while editing has been left out, please tell us). Policy pages do have to remain stable, especially the core content policies; we can't have them being constantly tweaked. Therefore, instead of arguing about issues of detail here and there, can you say overall what your fundamental objection is? SlimVirgin ^(talk) 17:19, 17 February 2007 (UTC)[reply]

I have no objection to merging WP:V, WP:NOR, and WP:RS. The problems are laid out above, at #Changes. Aside from the one about future minor but substantive changes to the page, this cannot be a "core content" policy without, in fact, replacing WP:V and WP:NOR; replacing WP:V and WP:NOR requires broad announcement and agreement, and the successful tagging of those pages as {{historical}} or redirecting them here. Second, you should not make blanket reverts of purely textual changes which do not change the meaning of this page. If you think that none of these are purely textual changes then, insofar as they are substantive, WP:ATT is a deviation from the originals; there are no contractions in WP:V, WP:NOR, or WP:RS, nor is there second-person addressing of the reader (e.g. "We encourage"), nor a redundant repetition of "Edits that rely on primary sources..." (or even that sentence at all), etc. —Centrx→talk • 17:34, 17 February 2007 (UTC)[reply]

Centrx, forget about RS for the time being, because this page isn't seeking to incorporate RS. Everything in RS that matters was taken from V and NOR. (I know this because I was the one who originally added them to ensure RS, V and NOR were consistent with one another, and I've made sure the consistency issue didn't deteriorate too much.) And everything that matters from V and NOR is in ATT. So RS is a separate issue and will hopefully be merged eventually into ATT/FAQ.

Your point about reverting "purely textual changes": some of them changed the policies in small but subtle ways — your addition of "novel" theory, for example.

NOR and V won't become historical, because ATT *is* V and NOR. The pages will simply be redirected.

Can you give me a concrete example of ATT deviating in a substantive way (in a way that's meaningful) from NOR and V? I'm sorry if I'm being dense, but I don't see one. What do you mean by "there are no contractions" in V and NOR? I've removed "we encourage," because you're right: we don't have second-person language in V and NOR. Not sure what you mean by "edits that rely on primary sources ..." SlimVirgin ^(talk) 17:59, 17 February 2007 (UTC)[reply]

Or we could be bold, ignore all rules, and improve Wikipedia. HighInBC ^{(Need help? Ask me)} 17:36, 17 February 2007 (UTC)[reply]

If that is what were going on, grammatical corrections (and other changes too) would be perfectly welcome and we could just tag WP:V and WP:NOR with {{historical}}. Try it. —Centrx→talk • 17:42, 17 February 2007 (UTC)[reply]

Ok, she has this habit of monolithic reverts, considering the amount POV pushing she has to deal with on a daily basis, please just have some oversight with occasional to-big-reverts. Now, what are your objections, again? I will look into "Edits that rely on primary sources", but could you make a bulleted list of "diffs" that disturb you? --Merzul 17:45, 17 February 2007 (UTC)[reply]

There is nothing about the current text that disturbs me, though it is not a pristine merge with no changes in meaning and it does not include some things that were in the originals. It should, however, be the subject of wider consideration, or at least be more widely known and there is no reason why all changes whatsoever, including the most mundane textual changes, should be disallowed. —Centrx→talk • 17:51, 17 February 2007 (UTC)[reply]

Centrx, the proposal has been worked on by 84 editors in four months; 157 editors have taken part on the talk page; the proposal has been posted several times to V, NOR, RS, the village pump, AN, AN/I, the mailing list, the Foundation, and there has been an RfC. I've received a ton of e-mails about it — and even a three-man barnstar. It's a good proposal! That doesn't mean it can't be tweaked, but any tweaking (a) must improve the writing; (b) must be completely consistent with NOR and V; and (c) mustn't inadvertently introduce any new issue. SlimVirgin ^(talk) 18:05, 17 February 2007 (UTC)[reply]

And there are 10,000 editors who refer to Wikipedia:Verifiability and Wikipedia:No original research on a regular basis. There are many people who do not personally contribute to every proposal, but who nevertheless should at least know about the actual implementation of such a major change. You would have found people coming and objecting when you redirected WP:V and WP:NOR even if there had been no other advertisement. —Centrx→talk • 18:19, 17 February 2007 (UTC)[reply]

I'm well aware of the importance of NOR and V — I'm one of the editors who maintains the pages — and I'm not suggesting people shouldn't be told about this, obviously, so that's a bit of a red herring. SlimVirgin ^(talk) 18:33, 17 February 2007 (UTC)[reply]

I have to admit that I was caught unawares by this merge proposal as well. I try very hard to stay up-to-date on policy changes, but Wikipedia's a big place, and it's difficult to keep track of everything. If I don't have my watchlist set on the right page, something can completely slip by. And more often than not, I'm not spending time at the policy talkpages, because I'm spending time actually writing articles.  ;) For something this big, I guess I would have expected an announcement in the Signpost, but I scanned through all the headlines over the last year, and haven't seen anything related to "Attribution." So, I'd recommend making a big deal about this merge in an upcoming issue of the Signpost, to ensure that everyone's aware of it. Also, could you please provide links to some of the above discussions that you mention, so that I can get caught up? Thanks. --Elonka 19:11, 19 February 2007 (UTC)[reply]

{{Policylist}} at the top of the page will need to be updated to initially include WP:ATT then remove WP:NOR and WP:V Gnangarra 14:27, 17 February 2007 (UTC)[reply]

This template needs more changes, see Wikipedia_talk:Verifiability#three_core_content_policies.-- _{Piotr Konieczny aka Prokonsul Piotrus | talk}  22:47, 17 February 2007 (UTC)[reply]

The popular culture exception or anything like it isn't included.

WP:RS is a guideline, not a policy, and as such is easier to work around. By creating this policy this way, you're turning one of the worst parts of WP:RS into established policy. Ken Arromdee 15:42, 17 February 2007 (UTC)[reply]

The pop cult exception was specifically removed to avoid us arguing over it. No actual change/innovation to policy, was felt to be the best way to make the merger happen. When the dust settles, the first thing you can do is raise a thread regarding it. Marskell 16:12, 17 February 2007 (UTC)[reply]

If something was removed to avoid argument then that should probably be a red flag as to the maturity of the proposal. What else was removed because the active participants here couldn't reach a consensus?ALR 17:45, 17 February 2007 (UTC)[reply]

What? If things were removed, that is certainly a change to policy. How in the world can you then claim that it's just a merge, with no changes, and nobody should dispute it? - unsigned

The removed item in question (type-of-article based criteria for reliable source, like pop culture) was not and is not part of V or NOR but instead is currently handled by IAR. There is difficulty in getting a concensus on how to word as policy the kinds of things that in actual practice are let "slide by" in terms of sourcing, especially when IAR is working just fine (or so some think). Some wanted one wording, others another wording, others perfered IAR. Some wanted tight wording while others wanted loose guidance and principles. We all wanted to steamline two policies into one. We are now doing that. This is a good thing. All the discussion is in the archives. Go ahead and read it. WAS 4.250 20:54, 17 February 2007 (UTC)[reply]

Thank you WAS. I did not mean something was removed from V or NOR as they stand. A new innovation (the pop cult exception) was removed so that the merge could go forward. Marskell 21:16, 17 February 2007 (UTC)[reply]

Ignoring the question of why certain things shouldn't have to be verifiable, why is that worse than what we had originally with WP:V and WP:OR? -Amarkov moo! 15:59, 17 February 2007 (UTC)[reply]

Because use of popular culture sources falls under WP:RS, not WP:V or WP:OR. WP:RS is a guideline, not a policy, and is therefore easier to ignore. Putting it in here transforms a guideline into a policy, making its problems worse. Ken Arromdee 16:27, 19 February 2007 (UTC)[reply]

What's the problem? Nothing has been added to this that isn't in WP:V or WP:NOR. If it has, please show us so we can rectify it if needs be. Steve block Talk webcomic warrior 16:39, 19 February 2007 (UTC)[reply]

Doc, your point about the difference between something being attributed and being attributable is well taken and lies at the heart of this policy. But in the sections you changed, the word "supported" doesn't mean it must actually be attributed in the text, simply that the source, whatever it is, must actually support the edit that is being made. That is, the edit mustn't go beyond the source material. Whether the source is added is a separate matter. But a source that supports the edit must exist, otherwise it's OR. To add "capable of supporting it" says the same thing, but it seems to weaken it somehow, and as the extra words aren't necessary, I'm hoping we can leave them out. SlimVirgin ^(talk) 17:06, 17 February 2007 (UTC)[reply]

This page merely codifies what is already expected of editors and told them across a number of other policies. WP:ATT has long been enforced on the project. Since WP:ATT is already de facto policy, it needs to be noted as such here. FeloniousMonk 17:54, 17 February 2007 (UTC)[reply]

Then you should tag WP:V, WP:NOR, and perhaps WP:RS with {{historical}} and add a note about it there. All forks and mergers could be called "de facto policy", that does not mean that the old pages are also kept as policy with an endless multiplication of them. —Centrx→talk • 18:02, 17 February 2007 (UTC)[reply]

Please forget about RS because it has nothing to do with this page. NOR and V will be turned into redirects. What would your objection to that be? SlimVirgin ^(talk) 18:43, 17 February 2007 (UTC)[reply]

Cute, but no. FeloniousMonk 19:01, 17 February 2007 (UTC)[reply]

How are the advocates of this as policy intending on dealing with precedent established in the extant policies? DEspite the protestations it's clear that the content is different, with some material taken out and some having a different emphasis, which changes the implications of any previous decisions made.

Has there been any thought put in to handling that? I can see a risk in some, mainly contentious, areas an increase in effort to exploit the new loopholes that have been created vice the old loopholes.

ALR 17:55, 17 February 2007 (UTC)[reply]

I think that's a risk with any change in policy, no matter how small. And this is quite a large undertaking. But problems like this can be dealt with as they appear, by minor tweaks to the phrasing. JulesH 19:12, 17 February 2007 (UTC)[reply]

Big edits going on... some changes are easy to resolve, but here is an edit discussed above (end of section). It's a good edit I think, but I'm not sure. There are some small fixes now as well, such as html code for '--' I'm not sure what is preferred. --Merzul 18:36, 17 February 2007 (UTC)[reply]

What do you mean by "big edits going on"? SlimVirgin ^(talk) 18:41, 17 February 2007 (UTC)[reply]

I mean they include too many fixes at the same time. --Merzul 18:47, 17 February 2007 (UTC)[reply]

As the editor, I don't see this. These were two defects in the same sentence brought up in the same discussion. One is an unwarranted loophole, the other mere redundancy; but if anybody wants to support one change and reverse the other, they need merely edit, rather than simply undoing. This is a good thing; they may come up with an improvement on both texts. Septentrionalis PMAnderson 19:10, 17 February 2007 (UTC)[reply]

PMA, I haven't looked at your edits, but this can't be used as an opportunity to make changes to the policies, because the consensus was gained with the assurance that ATT is a summary without changes; therefore, if you make changes, this will not become policy. NOR and V are widely supported. SlimVirgin ^(talk) 19:33, 17 February 2007 (UTC)[reply]

PMA, I think the second phrase removed, "read and", was not just a redundancy; since the sentence is stating what should not be done, it should not mention desireable behavior such as reading sources. --Gerry Ashton 20:12, 17 February 2007 (UTC)[reply]

That is not in the slightest just a merger of verifiability and original research. And RS is not only just a guideline, but a highly disputed one. It should not be a core policy. -Amarkov moo! 18:39, 17 February 2007 (UTC)[reply]

Amarkov, this doesn't include RS. Everything from RS that looks similar to this is from V and NOR in the first place. Do you have any specific examples of what you mean? SlimVirgin ^(talk) 18:42, 17 February 2007 (UTC)[reply]

On WP:V, we see that "Because this is the English Wikipedia, for the convenience of our readers, English-language sources should be used in preference to foreign-language sources, assuming equal quality, so that readers can easily verify that the source material has been used correctly." On this page, the language reads, "English-language sources should be used whenever possible, because this is the English Wikipedia. Sources in other languages are acceptable when there are no English equivalents."

I would like to see the phrase "assuming equal quality" restored by changing this sentence to "Sources in other languages are acceptable when there are no English equivalents of equal quality." I fear that the current wording on WP:ATT will cause editors to favor (or push for) inferior English-language sources — often translations of the originals — over foreign-language sources which can be more accurate and/or reliable. Any thoughts or objections to this? Dekimasuが... 18:44, 17 February 2007 (UTC)[reply]

I would say that "equivalent" implies "equal quality", as if the quality of the sources is not the same, they are clearly not equivalent. JulesH 19:13, 17 February 2007 (UTC)[reply]

Rephrase to "no equivalent English sources" to avoid this confusion? Septentrionalis PMAnderson 19:16, 17 February 2007 (UTC)[reply]

I think that "whenever possible" tends to deprecate the use of foreign-language sources regardless of their quality prior to the ending qualifier. Equivalence could mean simply the ability to act as a citation for a certain statement, as opposed to being a description of the quality of the sources. This is more a question of clarity than anything else; I fear an increase in the number of editors who, for example, say that WP:V or WP:N have not been met when an article contains only foreign-language sources. Dekimasuが... 03:39, 18 February 2007 (UTC)[reply]

I think the policy should recognize that different editors have different sources available to them, and different abilities to search for sources. A phrase like "when there are no English equivalents of equal quality" can be interpreted to place on each editor the burden of searching the entire English-language literature to make sure there is no source as good as the foreign language source that the editor has convenient access to. I would like to see that section restated to say that when several sources of equal quality have been found, the English language version should be used. I have no objection to waiting until the merger issue is settled for such a change. --Gerry Ashton 20:20, 17 February 2007 (UTC)[reply]

You are right about the burden on the editor. How about, "If English sources of equal quality have not been cited (found? referenced?), sources in other languages are acceptable"? Dekimasuが... 03:39, 18 February 2007 (UTC)[reply]

Agreed. English language sources are prefferable, but non-English are acceptable, particulary when no English ones are available - that's the bottom line. And sometimes non-English may be of higher quality then English (particulary when dealing with history issues).-- _{Piotr Konieczny aka Prokonsul Piotrus | talk}  22:35, 17 February 2007 (UTC)[reply]

Yes, I like that wording, "If English sources of equal quality have not been cited (found? referenced?), sources in other languages are acceptable" ^HighInBC ^{(Need help? Ask me)} 14:20, 18 February 2007 (UTC)[reply]

Located? Uncovered? Which word would you suggest? This seems uncontroversial, so I don't see much reason to wait to make a change if we can agree on the wording. Dekimasuが... 17:08, 19 February 2007 (UTC)[reply]

My attempts to change the wording here have been reverted three times now, twice referring to the idea that this represents a "change in longstanding policy" and once referring to a "lack of consensus". I brought the issue up here, and received positive feedback before adding the changes. As per my first post above, the changes were meant to better represent the longstanding formulation of WP:V than what is currently on ATT. Meanwhile my requests for further comment on the reversions have gone unanswered. It seems, frustratingly, like the longstanding policy is actually to prevent any edits from being made to ATT. Please present feedback here. Dekimasuが... 15:23, 21 February 2007 (UTC)[reply]

I have restored Dekimasu's edit. I suspect Jaygj did not notice this section of the talk page when reverting the change. --Gerry Ashton 18:54, 21 February 2007 (UTC)[reply]

"Equal quality" are not the same as "English equivalent". For example, we can have two books of equal quality wriiten by two equally distinguished scholars, but with opposing POVs. One book is in English, another one in French. The "equal quality" clause will rule out the French source because an English one of the same quality is available. However, the "English equivalent" phrase will allow both sources because there is no English book written by a similarly distinguished scholar and representing the same POV. "Whenever possible" are also crucial words here. Many editors of the English Wikipedia do not speak languages other than English and often cannot possibly verify the quality of sources in languages other than English unless the sources have reviews in English. Beit Or 19:29, 21 February 2007 (UTC)[reply]

I wouldn't interpret the sentence as Or does; to me, every significant point of view should be included, and the fact that one point of view has an English language source and another has a French source in no way prevents the latter POV from being included. The phrase "English equivalent" could be interpreted in a variety of ways too; if two sources convey the same information in a science article, but the one in French is published in a peer-reviewed scientific journal, while the one in English is a popular newspaper, some people might argue the newspaper is an English equivalent and the French version should not be cited.

My real preference, when both English and non-English sources are available, is to cite both of them, unless the non-English source is merely a translation of the English source. More information is better.

In any case, the emphasis on sources that have been found, rather than sources that exist, should remain in the policy, because we can never know if we've found all the soruces. --Gerry Ashton 19:58, 21 February 2007 (UTC)[reply]

A very frequent mistake made by editors is thinking that you can't do O.R. on Talk pages. I believe it would be useful to include the following paragraph:

Original research is allowed on Talk pages

A large amount of Wikipedia activity takes place on the Talk pages, and users sometimes believe that the prohibition to use original research applies to those pages as well, which is not the case. Actually, original research is not only allowed, but also necessary on Talk pages, particularly when an editor argues for or against the use of a certain source. In those cases, the editor may do some research about the source, find that it is reliable or unreliable, and argue for using it or not based on his research. Other users participating in the Talk page may accept or reject his arguments, but may not refuse to consider them on the grounds that they constitute original research. --Abenyosef 19:57, 17 February 2007 (UTC)[reply]

Talk pages are for discussing the article and not for discussing the subject. ≈ jossi ≈ (talk) 20:57, 17 February 2007 (UTC)[reply]

For what it's worth, the Wikipedia:Talk page guidelines clearly says "A talk page is research for the article, and the policies that apply to articles also apply to talk pages. Research and debate should meet the same standards of verification, neutral point of view and no original research." Friday (talk) 21:07, 17 February 2007 (UTC)[reply]

I believe you misunderstand wikipedia's idiosyncratic use of "original research". We are supposed to do research to understand the subject of an article, the nature of the sources and the qualifications of the authors of those sources. This helps us in making editorial decisions such as what is or is not "original reseach"; such as what is a neutral way of phrasing something; such as what not to include as if it were a fact because the author is a liar or fraud or an incompetant. In short, research is how we know what is original research or not. WAS 4.250 21:14, 17 February 2007 (UTC)[reply]

Sure, but note that talk pages are an integral part of an article, and they are indexed by Google and other search engines. We can discuss the subject, assess sources, etc., but we should not encourage "forum discussion" type of engagements between editors. Rather, talk-page discipline focused on how to improve the article should be the modus operandi. ≈ jossi ≈ (talk) 21:21, 17 February 2007 (UTC)[reply]

Well, your answer is proof of why my proposed paragraph is necessary. Talk pages are not (NOT!) an integral part of an article. That's a big confusion many (if not most) editors have. In talk pages we discuss the articles and we may do as much research as we like for discussing what gets and what doesn't get into the article. What we don't do is publish that research in the article itself. I may do a lot of research to conclude that Fulano de Tal is not a reliable source on Rioplatense Spanish, and thus convince my fellow editors that he mustn't be used as a source. What I can't do is write "Fulano de Tal is not a reliable source" into the article on Rioplatense Spanish.--Abenyosef 15:46, 18 February 2007 (UTC)[reply]

Please don't add confusing sections to the policy that have no consensus and contradict existing policy. Jayjg ^(talk) 03:42, 21 February 2007 (UTC)[reply]

Interestingly enough, I found myself involved in a discussion on this very subject not long ago, after I decided to challenge a source that was being used in the "New antisemitism" article. I held that the source was inappropriate, as it included material that was both demonstrably false and irrelevant to the subject matter. (Click here for details. The relevant sections are those marked "Source Credibility" and "Demonstrably Poor Scholarship").

I thought I had a pretty strong argument: after all, I was able to show that one of the statements included in the article was outright false. Most contributors seemed to agree with my position. Nevertheless, two editors chose to advance the view that my criticism of the source was irrelevant, as it constituted a form of Original Research. (I've responded to this accusation, but my interlocutors have not revised their initial position.)

It's now been almost a month since my original post appeared on the discussion page. The "New antisemitism" article continues to include factually inaccurate information, and efforts to remove it have been met with resistence. The matter is currently before informal mediation, although neither of my interlocutors have contributed to this process as of yet.

I remain confused by the adamantine efforts of certain editors to retain false information in this article. I can only imagine that these editors have concluded that (i) "false, but verifiable" is an acceptable standard for the inclusion of information, and (ii) any efforts to demonstrate the falsehood of such information are contrary to Wikipedia policy. In the back of my mind, I cannot help but wonder if the "No Original Research" policy is being abused in some way. CJCurrie 00:18, 22 February 2007 (UTC)[reply]

It is WP's core policy that we strive for "attributability, not truth". It is not a view held by 'certain editors', but by WP as a whole, by consensus and policy. Regarding the specific issue of OR on talk pages, as I replied to Abenyosef on his Talk page, the key point is that we don't vet authors or their work - we vet the publishers and publications. Once we deem a publication/publisher as acceptable, we leave it to the editors and publishers at that publication to vet the authors and provide an editorial oversight, while we summarize and cite the work. Trying to vet authors and their work in Talk pages is OR and is wrong, we need to focus on the publication. Crum375 00:44, 22 February 2007 (UTC)[reply]

The policy you're referring to is, "The threshold for inclusion in Wikipedia is verifiability, not truth", and it's a policy with which I strongly agree. This policy, however, was designed to ensure that "accurate, but unverifiable" information would be excluded from Wikipedia, not to allow "demonstrably false, but attributable" information to be included.

Beyond this, I cannot agree with your arguments on Abenyosef's page concerning "tests of verifiability". If the NY Times or a major scientific journal prints a statement which can shown to be false, there is surely no obligation on our part to convey the statement as fact.

In any event, I raised concerns about the method of publication in my initial post on the subject. CJCurrie 00:56, 22 February 2007 (UTC)[reply]

If you disagree with a policy, which of course is perfectly legitimate, you can certainly work to revise it. But as long as the policy is in force, we must all respect it and abide by it. I think that WP's strength is in not trying to be 'accurate' or 'truthful' but instead to present attributable information, in a neutral and balanced fashion. An encyclopedia is not a religious document that we accept as gospel - it is a collection of summarized published information, properly balanced and attributed to its sources. If the NYT reports that the moon is made of green cheese (to use SlimVirgin's example), once we deem the NYT as an acceptable source, we could hypothetically report that the NYT said that, and perhaps with due weight show that it is in a small minority (and possibly even exclude it if it is on the fringe). In a less extreme example, we would include information published by an otherwise reliable source, and cite it to that source, even if some of us at WP suspect it is false. Crum375 01:09, 22 February 2007 (UTC)[reply]

What if we know that it is false?

The "New antisemitism" article currently includes the following statement: "[Edward Flannery] cites historian Zosa Szajkowski, who writes that he could not find a "single word on behalf of Jews" in the entirety of French socialist literature from 1820 to 1920."

There's a serious problem with this statement. Edward Flannery does, in fact, write that Zosa Szajkowski could not find a "single word on behalf of Jews" from 1820 to 1920. The problem is that Szajkowski himself did not make this statement. I looked up the Szajkowski article cited by Flannery, and found that Flannery had misstated Szajkowski's position: ZS's survey ended in 1886, and he spoke of French socialists opposing antisemitism in 1889.

There's no ambiguity on the matter: Flannery's statement is not merely suspect or falsifiable, but false. Why then are we still presenting it as a statement of fact?

I would submit that any Wikipedia policy that allows retention of this demonstrably false statement is fundamentally flawed, and is at the very least in need of clarification. CJCurrie 01:57, 22 February 2007 (UTC)[reply]

Just to clarify my initial remark about 'truth', I believe that your interpretation of the policy that it implies rejecting "false but attributable", is your own novel take on it, which I don't see in the text. The way I read the policy is very simple - we don't look for truth - only attributability. Crum375 01:39, 22 February 2007 (UTC)[reply]

The purpose of WP:NOR was to seek assurances of verifiability, not simply of attributability. The policy was never designed to permit "false, but attributable" statements to be presented as fact, even if this was never explictly clarified. CJCurrie 01:57, 22 February 2007 (UTC)[reply]

WP:NOR was designed to prohibit Wikipedians to perform their own research and/or present their own results or opinions. Perhaps you meant WP:V? Crum375

I'll restate myself: The purpose of WP:NOR was to *ensure* compliciance with WP:V. Returning to the specific issue, could you address my concerns regarding one author's misattribution of another's statement? CJCurrie 02:18, 22 February 2007 (UTC)[reply]

We may have to disagree on the purpose of NOR - to me it explains that we may not create any original research, and that OR can be as simple as putting two well sourced facts together to advance a position, if those facts have not been previously juxtaposed by others. This point, called synthesis, is not covered directly by WP:V. Regarding your example of source A misquoting source B, I accept that when we present a statement that appears to be self inconsistent or contradictory, we may carefully mention that fact to prevent reader confusion, bending over backwards to be extra neutral and use minimal verbiage. Of course if a third party already remarked on that inconsistency, that would be infinitely better. Crum375 18:55, 22 February 2007 (UTC)[reply]

Reliable sources section does not link to WP:RS (like other sections reffering to subpolicies do), and it has a shortcut WP:V (which redirects to Wikipedia:Verifiability which is being merged here...). Also some shortcuts are ilinked (and point to themselves - i.e. this page), others are not. This is confusing.-- _{Piotr Konieczny aka Prokonsul Piotrus | talk}  22:33, 17 February 2007 (UTC)[reply]

The plan was that WP:RS would be replaced by a combination of the text here and at WP:ATTFAQ. I'm not sure if this has changed, and if so what to...? JulesH 22:44, 17 February 2007 (UTC)[reply]

Many newcomers may think this policy describes how Wikipedia editors receive attribution for their work. At the very least we should address this with a relevant link in see also. Or we should consider changing the name of this policy to avoid confusion.-- _{Piotr Konieczny aka Prokonsul Piotrus | talk}  22:46, 17 February 2007 (UTC)[reply]

I agree, that is what I innitially thought when I found this proposal. While the "nutshell" explains everything, people won't see the nutshell when it is being linked to from another page. I think it needs a clearer title. Perhaps something like "Attributability"?--Konstable 00:17, 18 February 2007 (UTC)[reply]

Why not stick to the old name most people are familiar with (i.e. verifiability)?-- _{Piotr Konieczny aka Prokonsul Piotrus | talk}  01:44, 18 February 2007 (UTC)[reply]

The main argument was that when an editor checks a source they are not "verifying" info as such, but simply checking that the info has been attributed to a reliable source. I'm not in a position to verify the content of a Nature article; I can only demand that "the Sun is hot" be attributed to one. Attributability? Bit of a clunker, but maybe better. Marskell 07:19, 18 February 2007 (UTC)[reply]

I like the name, people will figure it out. ^HighInBC ^{(Need help? Ask me)} 14:18, 18 February 2007 (UTC)[reply]

I agree. I think the name is good, and any confusion will be cleared away in a short time, as people get used to it. Verifiability is somewhat of a misnomer, as has been pointed out. "Attributability, not truth" sounds a lot less confusing to me. Crum375 21:55, 18 February 2007 (UTC)[reply]

Personally, I think the old policies were easier to explain, especially WP:NOR. But I suppose we don't actually have to stop talking about it. semper fictilis 04:58, 19 February 2007 (UTC)[reply]

How often have editors argued that people would "get used" to our private definition of "Verifiability"? Septentrionalis PMAnderson 16:08, 19 February 2007 (UTC)[reply]

I'm thinking more of WP:NOR: I think I've seen lots of times when someone is advancing some private theory. Telling them that wikipedia is not the place for such "original research" is the most direct way of dealing with that problem. Yes, we can tell them that they need to attribute it. But when that argument ultimately is challenged ("I don't need to attribute it because I know it to be true" or "it is obvious"), we ultimately have to invoke the principle of "no original research". semper fictilis 16:16, 19 February 2007 (UTC)[reply]

This is much weaker on the sourcing requirements than WP:V, it's especially wobbly on the removal of unsourced material. WP:V has a whole section on burden of evidence, and that unsourced material may be removed at any time, and the onus of sourcing is on people who want to include the information. The key principles here says "Any unsourced material may be removed," but this is undermined by the later section that says "not all edits need a source, and this policy should never be used to cause disruption by removing material for which reliable sources could easily or reasonably be found." It essentially encourages the inclusion of unsourced material, rather than recognizing it as something undesirable that we'd rather not have but need to keep around to avoid disruption. There needs to be much stronger language supporting those who remove unsourced information. People should not be able to point to a policy that says "material must be attributable" and claim it as justification for not actually attributing their material. Night Gyr (talk/Oy) 01:01, 18 February 2007 (UTC)[reply]

Yes, this seems to be a major change from current policy, not simply part of a merger. —Centrx→talk • 01:41, 18 February 2007 (UTC)[reply]

(edit conflict) I think it's a good thing to discourage people from WP:POINTedly going through Wikipedia and removing material that is likely true just because it doesn't have a footnote yet, but the current wording goes too far. As written it would discourage editors from removing dubious claims unless they are outright absurd, which is a very high standard.

Boldly moving questionable material to talk is highly desirable in some circumstances, particularly on little-edited pages, where misleading claims can sit around for months with a ^{[citation needed]} tag. Is there a reason that the sentence about that option from WP:V was not included here? —Celithemis 01:56, 18 February 2007 (UTC)[reply]

I believe that the goal is for the first two paragraphs here to clearly state that all material must be attributable. Would the addition of another sentence, reiterating this, help clarify the later part about not disrupting WP? I'll insert one to see how it fares. CMummert · talk 02:13, 18 February 2007 (UTC)[reply]

I think we need to change things from "You should probably ask for a source first" to "it shouldn't be here, but don't be too much of a dick about taking it out" Night Gyr (talk/Oy) 04:37, 18 February 2007 (UTC)[reply]

I think it is a question of tone rather than any real difference in poIicy. I added a sentence reiterating the burden of evidence, and added the option of moving material to the talk page; does this change the tone enough? CMummert · talk 12:40, 18 February 2007 (UTC)[reply]

We could just take the V wording verbatim for now. Marskell 13:16, 18 February 2007 (UTC)[reply]

That phrasing has been the subject of contention before. If we can come up with an alternative here that people are happy with, we might be able to solve one of the problems of WP:V. JulesH 14:14, 18 February 2007 (UTC)[reply]

Lets merge first, then change things, I like how strong WP:V was about removing unsourced material. ^HighInBC ^{(Need help? Ask me)} 14:16, 18 February 2007 (UTC)[reply]

I don't see that there is really much difference between the two policies. The following is from this policy:

Wikipedia is an encyclopedia, not a publisher of original thought. The threshold for inclusion in Wikipedia is whether material is attributable to a reliable published source, not whether it is true. Wikipedia is not the place to publish your own opinions, experiences, or arguments.

Although everything in Wikipedia must be attributable, in practice not all material is attributed. Editors should provide attribution for quotations and for any material that is challenged or likely to be challenged. The burden of evidence lies with the editor wishing to add or retain the material.

From WP:V

The threshold for inclusion in Wikipedia is verifiability, not truth. "Verifiable" in this context means that any reader should be able to check that material added to Wikipedia has already been published by a reliable source. Editors should provide a reliable source for material that is challenged or is likely to be challenged, or it may be removed.

...The burden of evidence lies with the editor who adds or restores material. Any material that is challenged or likely to be challenged needs a reliable source, which should be cited in the article. If an article topic has no reliable, third-party sources, Wikipedia should not have an article on it.

We could add "or it may be removed" to the fifth sentence in the quote from WP:ATT to make it more parallel, but the fact that it may be removed is already clearly implied by the "burden of evidence" sentence. CMummert · talk 21:40, 18 February 2007 (UTC)[reply]

The current one only says "should," the old one says "needs," implication isn't enough, the necessity of sources needs to be explicit. Night Gyr (talk/Oy) 23:56, 18 February 2007 (UTC)[reply]

To use the distinction ATT makes: Original research needs to be removed; unsourced material needs to be sourced. Septentrionalis PMAnderson 16:14, 19 February 2007 (UTC)[reply]

But this seems to say if you can't find a source, leave it and wait for someone else, instead of taking it out pending a source, which is better practice. Night Gyr (talk/Oy) 03:39, 20 February 2007 (UTC)[reply]

I agree with Night Gyr and have boldly made the change back to the WP:V emphasis on "source it or remove it". Sandstein 22:21, 20 February 2007 (UTC)[reply]

When the U.S. Constitution was first created they voted to approve a long detailed version with lots of nuances and caveats. Then a small group was appointed to go through and 'clean it up' a bit. Their mandate was to not change any of the meanings, but to shorten it. However, without all the details and nuances and caveats people inevitably take different interpretations of the text than they would if all that further explanation were still in there. Which is why, officially, the original text still holds and people still go back and quote it as the actual intent of the Constitution (e.g. the standards for impeachment of the President are given as "high crimes and misdemeanors", but in the original it was, "high crimes and misdemeanors against the state").

Given all of the above consternation about variations in wording from the 'merge' I'd suggest we do something similar here. The intent of this page was to merge the content of the 'Verifiability' and 'No original research' policies without changing their meaning at all. Thus, I'd suggest that we lock down those two policies and let this one go live with the understanding that 'interpretations' of the text need to square with the original policies that this one was intended to NOT change. Going forward there may be consensus to amend / alter this policy in ways that contradict the older two, but at present it should not be taken to mean anything different than the two policies it was derived from. They are still the 'original and officially approved text' which was established by community consensus. This page is then the 'working copy' which is intended to provide the same meaning as those more detailed texts in a 'cleaned up' and unified form. Make sense? --CBD 16:52, 18 February 2007 (UTC)[reply]

Yes, that makes a lot of sense. SlimVirgin ^(talk) 16:58, 19 February 2007 (UTC)[reply]

That would make any changes to this policy difficult, as they would have to be phrased as, well, amendments. This is exactly the kind of problem that the proposal to merge was intended to solve. So, no, I don't think keeping the old versions as an aid for interpretation helps. JulesH 16:55, 18 February 2007 (UTC)[reply]

Changes to the policy would require consensus... just as they always do. Barring a consensus to change the intent/meaning of the old policies they would still be held to apply. It needn't be done as amendments. I'm just suggesting that rather than wrestling over whether the phrasing of this page projects exactly the same nuances as the two originals (it doesn't and never could) we accept it as 'good enough' with the understanding that those nuances are all still in effect and citable from the old policies until/unless there is an actual consensus to change them. --CBD 18:27, 18 February 2007 (UTC)[reply]

Policy should always be changeable based on sufficient consensus. You could keep WP:Verify and WP:NOR as guidelines subordinate to WP:ATT, assuming WP:ATT does in fact replace them, and later demote them to historical when they get out-of-date. This would probably make the transition more smooth than simply redirecting them in a week. — Armedblowfish (talk|mail) 20:47, 19 February 2007 (UTC)[reply]

I like that approach. Rossami (talk) 21:11, 19 February 2007 (UTC)[reply]

I disagree with any retention of the old policies (as Guidelines or anything), because to my mind it would defeat the point of ATT. V + NOR = ATT is a reduction of two pages to one. Good. Keeping the two former as supplements only means increasing "link-creep"--I have to go here, then here, then here, to understand. At present our policies are stable but still not as easy as they could be. One does have to click on links too much: from V to RS back to V and V over to NOR and back to V, etc. WP:ATT, IMO, was to solve this, and this merge (while certainly not solving it completely) was the first step. Marskell 21:50, 19 February 2007 (UTC)[reply]

The basic point of this policy, as far as I can tell, is that anything that is not a book or a newspaper or some such is not a valid source. Congratulations, you've just made it impossible to source anything related to Internet-only stuff as well as most pop culture/anime/TV things and so forth. I would be willing to bet money that we'll shortly see a massive amount of PRODs. Jtrainor 02:45, 19 February 2007 (UTC)[reply]

This is not a new policy. It is a merger of WP:V and WP:OR, there is nothing (or very little) new here.--Konstable 02:55, 19 February 2007 (UTC)[reply]

Exactly. This policy merely merged WP:V and WP:NOR. ≈ jossi ≈ (talk) 03:02, 19 February 2007 (UTC)[reply]

The disputes on this talk page would seem to indicate otherwise. Jtrainor 03:43, 19 February 2007 (UTC)[reply]

In general, Jtrainor has a point. In the past, editors on the policy and guideline pages have debated what is or is not published but, if I recall correctly, never reached a consensus on a definition. That failure is reflected in the failure of this policy to define published, even though forms of the word are used more times than I care to count. There is a substantial danger that some readers will only think of books, newspapers, and magazines as being published, and will not think of DVDs, CDs, computer programs, and websites as publications. --Gerry Ashton 03:53, 19 February 2007 (UTC)[reply]

Well, it is technically true that there is nothing new. However, that excludes the multiple deletions from the policies (which were somehow supposed to be a compromise?), and removals change things just as well as additions. -Amarkov moo! 03:54, 19 February 2007 (UTC)[reply]

Published means published. Do we need to re-define what that means? Also note that we are not just discussing "published". We are discussing reliable, published sources. Add "reliable" to "published" and you have all what you need. See Wikipedia:Attribution#Key_principles. ≈ jossi ≈ (talk) 04:51, 19 February 2007 (UTC)[reply]

(<---)Migrating to this wording of policy will not change in the slightest the existing practice on Wikipedia concerning sourcing. It only makes clearer what we are trying to do. WAS 4.250 05:19, 19 February 2007 (UTC)[reply]

If the sources are reliable, I wouldn't think it makes a bit of difference online or off. On the other hand, if we started seeing a "massive number of prods" on minor fictional-character bios and Internet fads, the downside to this would be...? Seraphimblade ^{Talk to me Please review me!} 07:06, 19 February 2007 (UTC)[reply]

The downside would be that there's no point in removing stuff unless it's completely incomprehensible, malicious, or beyond improvement. This site takes up less than 2 gigs, after all, and pages no one visits don't use bandwidth. Yet, all too often, flavor of the month prods are done on all sorts of anime/game/pop culture related articles. Jtrainor 09:40, 19 February 2007 (UTC)[reply]

This policy isn't about deleting articles. Determining what kind of article we should have is the job of Wikipedia:What Wikipedia is not and its accompanying guideline Wikipedia:Notability. What we've changed here will not greatly affect the interpretation of those pages.

What this policy is about is ensuring that wikipedia presents the highest quality attainable. Now, I've argued in the past that we need to change the existing policy WP:V and its guideline WP:RS to allow a wider variety of sources in some articles. But that's a thorny issue, and a lot of people don't like it. So it can't happen here, not right now. But later, I'm sure it will be proposed again, and we can start gathering consensus to make that change. JulesH 12:17, 19 February 2007 (UTC)[reply]

I understand the point Jtrainor and others are making, but like others have also said, we're not changing anything anything with this merge. The cause of the problem is not the merge. Don't object to the merge because it doesn't fix every problem at once. -- Ned Scott 07:14, 19 February 2007 (UTC)[reply]

Shouldn't this page also obviate the need for Wikipedia:Reliable sources and Wikipedia:Citing sources? >Radiant< 12:04, 19 February 2007 (UTC)[reply]

Citing sources is a style guide, and should probably remain as such. I think WP:RS can be replaced by a combination of this page and its FAQ. Whether that means the FAQ needs to be a guideline or not, I don't know. JulesH 12:19, 19 February 2007 (UTC)[reply]

One step at a time. ^HighInBC ^{(Need help? Ask me)} 16:46, 19 February 2007 (UTC)[reply]

I was about to write "one bite at a time" but the edit conflict got in the way. Marskell 16:48, 19 February 2007 (UTC)[reply]

Great minds think alike, though fools rarely differ. ^HighInBC ^{(Need help? Ask me)} 17:00, 19 February 2007 (UTC)[reply]

I disagree. WP:RS contains a lot of guidance, much of it contentious, that would not be appropriate to include in a policy. As it is, I think a bit too much of WP:RS is already in WP:V and WP:ATT. As for Citing sources, that is a style guide, as JulesH said. — Armedblowfish (talk|mail) 18:27, 19 February 2007 (UTC)[reply]

ABF, which parts of RS are in ATT and V? SlimVirgin ^(talk) 20:53, 19 February 2007 (UTC)[reply]

The parts have obviously been changed some, but here's some examples:

• Wikipedia:Attribution#Exceptional_claims_require_exceptional_sources and Wikipedia:Reliable sources#Exceptional_claims_require_exceptional_sources Post-signing due to comment split, Armedblowfish (talk|mail)

• Yes, that has been added, after agreement, although as I recall it started life in V or NOR. SlimVirgin ^(talk) 22:37, 19 February 2007 (UTC)[reply]

• Wikipedia:Attribution#Living_persons and Wikipedia:Reliable sources#Biographies_of_living_persons Post-signing due to comment split, Armedblowfish (talk|mail)

• This is from V and/or BLP. SlimVirgin ^(talk) 22:37, 19 February 2007 (UTC)[reply]

• Wikipedia:Attribution#Questionable_or_self-published_sources and much of Wikipedia:Reliable sources#Using_online_and_self-published_sources. Also see WP:V#Sources Post-signing due to comment split, Armedblowfish (talk|mail)

• This is from V or NOR. SlimVirgin ^(talk) 22:37, 19 February 2007 (UTC)[reply]

• Maybe I misremember, but I think WP:V#Sources_in_languages_other_than_English used to be in WP:RS Post-signing due to comment split, Armedblowfish (talk|mail)

• From V. SlimVirgin ^(talk) 22:37, 19 February 2007 (UTC)[reply]

The first two are fine, in my opinion, although it might make more sense to summarise in WP:ATT, leaving the details to WP:RS, rather than duplicating and allowing forking. It is the third one I have a problem with because, having discussed it much on WT:RS, I disagree and have seen exceptions that seem to enjoy consensus. Well, you saw my suggestion on WT:RS recently, so I guess I don't have to explain here. If I were to suggest a part of WP:RS to be included in policy, it would be a summarised version of Wikipedia:Reliable sources#Aspects_of_reliability, but a lot of that material is new so it might not actually have enough consensus to become part of policy.

I hope that clarifies my position,

Armedblowfish (talk|mail) 22:32, 19 February 2007 (UTC)[reply]

The problem is that people editing RS are not familiar with the content policies, which is why the page has become such a problem, and so they don't realize that the only bits of RS that matter were copied from V or NOR. And therefore those bits will be carried over here. They don't come from RS. SlimVirgin ^(talk) 22:37, 19 February 2007 (UTC)[reply]

"The only bits of RS that matter"? Well, I guess the people who brought them from WP:RS to WP:V and WP:NOR thought they mattered the most. Post-signing due to comment split, Armedblowfish (talk|mail)

You've misunderstood. I added them to RS, because they were already in V and NOR, and RS had to be consistent with V and NOR. So they've now been added to ATT. But not because they were in RS. SlimVirgin ^(talk) 00:23, 20 February 2007 (UTC)[reply]

The community pays more attention to policies than guidelines, but guidelines are still cited by many, and WP:RS is probably among the guidelines that are most often cited. Those parts might not come directly from WP:RS, but indirectly, I think they do. Months ago, I don't remember WP:V saying much more than that if you used a dubious source, like the tabloid newspaper The Sun, you should use a prose attribution. In any case, I am disagreeing with what was chosen, at least the self-published sources part, however long ago that happened. Restoring original indentation due to paragraph split, and post-signing due to comment split, Armedblowfish (talk|mail)

It wasn't "chosen." This is a copy of V and NOR. The self-published thing is an important part of those policies. It has nothing to do with RS. SlimVirgin ^(talk) 00:23, 20 February 2007 (UTC)[reply]

Looking at the most recent WT:V talk page archive, there is a lot of discussion about self-published sources, suggesting dissatisfaction with the current wording on the matter.

— Armedblowfish (talk|mail) 00:13, 20 February 2007 (UTC)[reply]

The self-published thing has a lot of support. SlimVirgin ^(talk) 00:24, 20 February 2007 (UTC)[reply]

Refactoring note: I am sorry for refactoring a live conversation, but the replies in the middle of my comments are starting to get a bit confusing. — Armedblowfish (talk|mail) 01:02, 20 February 2007 (UTC)[reply]

On 16 December 2006 2005, WP:V did not contain the word "self-published", although it did mention a couple types of self-published sources. At that time, WP:RS talked about self-published sources rather extensively.

Regardless of history, there does seem to be consensus that use of self-published sourcs should be singificantly limitted, but the nature of that limitation appears to be contentious to me. More on this at WP:RS when I have time. — Armedblowfish (talk|mail) 01:59, 20 February 2007 (UTC)[reply]

I'm concerned that a couple of editors are using the merge as an excuse to change the policy. PMAnderson's change, for example: [2] the example is OR, which is why it's cited. Why would it only "probably" be OR?

Please, no more changes. There's a consensus for a merge, but not for change. SlimVirgin ^(talk) 16:57, 19 February 2007 (UTC)[reply]

Then don't change it; or preferably, be consistent about changing it. Septentrionalis PMAnderson 17:57, 19 February 2007 (UTC)[reply]

Nothing has been changed. SlimVirgin ^(talk) 18:00, 19 February 2007 (UTC)[reply]

I support the change to "Original research is material that cannot be attributed to a reliable source.", but I do not see it in either source. Septentrionalis PMAnderson 14:45, 20 February 2007 (UTC)[reply]

Because the distinction this page draws between OR and unsourced material, which is a good one - and sharper than it used to be- is that original research cannot be attributed to a reliable source. Phrasing it as at present destroys this distinction; it is probable, but not certain, that the example paragraph is some editor's own opinion; it is possible that he saw the CMS cited in some reliable source's comment on the Smith-Jones matter, and has forgotten to cite it. The existing text also leaves unstated that the implied conclusion is what may well be OR. Both of these have been discussed on this page before. I strongly oppose letting this policy go live with this confusion.Septentrionalis PMAnderson 17:54, 19 February 2007 (UTC)[reply]

I agree with PMA on this one. Clarity is key. --badlydrawnjeff talk 17:56, 19 February 2007 (UTC)[reply]

Okay, I see what you mean. In this case, however, it's clear from the writing that it's the editor's argument, and in a case like that, we assume that it's OR until the editor shows that it's someone else's argument. Otherwise, we'd never be able to call anything OR, because it's always possible that there might be a source out there somewhere, not yet found. SlimVirgin ^(talk) 17:59, 19 February 2007 (UTC)[reply]

I agree that it's a reasonable inference; but the existing text goes on to discuss the possibility of finding a source, when we have just defined OR as making that impossible. As far as I am concerned, this is a matter of phrasing, not of policy; it may be sufficient to state the assumption SV makes to avoid misleading those who will read this policy. Septentrionalis PMAnderson 18:06, 19 February 2007 (UTC)[reply]

I have modified this to explicitly make Slim Virgin's assumption. I believe it was true of the actual case from which this is adapted, but it was not certain of the hypothetical; now it is.

Would it be useful to add: "The conclusion would be OR if stated; it still is, as an implication"? Septentrionalis PMAnderson 14:43, 20 February 2007 (UTC)[reply]

The following material is a change; it is a good change

• Unsourced material is material not yet attributed to a reliable source.
• Original research is material that cannot be attributed to a reliable source.

but as long as it stands and is not used consistently on this page at a minimum, this page is not fit to be policy. Policy should, at a minimum, be coherent. Septentrionalis PMAnderson 16:07, 20 February 2007 (UTC)[reply]

I would appreciate it if Slim Virgin would explain what policy change she sees in this edit. To the best of my ability, I expressed present policy in terms of the definitions above, which are a change from WP:NOR. Septentrionalis PMAnderson 16:12, 20 February 2007 (UTC)[reply]

The wording in question is:

This entire paragraph is original research, because the editor has put two (citable) assertions together in order to imply the editor's opinion that, given the Chicago Manual of Style's definition of plagiarism, Jones did not commit it. If there were a reliable source that specifically comments on the Smith and Jones dispute and makes the same point about the Chicago Manual of Style and plagiarism, it would not be original research; but should express the conclusion as that source's opinion, not Wikipedia's.

Septentrionalis PMAnderson 19:37, 20 February 2007 (UTC)[reply]

Hello, on Wikipedia talk:Reliable sources#Self-published_sources:_Can_we_simplify_this.3F, I suggested changing the criteria for inclusion of self-published sources to something simpler, and, in my opinion, more to the point. Anyways, since WP:V and WP:ATT discuss this and say what WP:RS currently says on the subject, and policy overrides guideline, your input on this would be wecome. Thanks!

— Armedblowfish (talk|mail) 18:51, 19 February 2007 (UTC)[reply]

Cross-posting to WT:V.

The article layout has a serious redundancy problem. The key principles are stated twice in different formulations. This is not a good idea. The two iterations need to be combined so there is one definitive statement for each. I think the best way to do this is to refactor the headings from:

1 Key principles
2 No original research
3 Reliable sources

to:

1 Key principles
1.1 No original research
1.2 Reliable sources

Any objections? Dhaluza 01:27, 20 February 2007 (UTC)[reply]

I think it works as is. The first section "1 Key principles" is a summary of the following two sections. Good for these people that want just to read the basics. ≈ jossi ≈ (talk) 03:35, 20 February 2007 (UTC)[reply]

Exactly -- this is inverted pyramid structure, which is useful. JulesH 09:33, 20 February 2007 (UTC)[reply]

My background is in tech writing (especially specs) so I may have a different approach. But I don't think the news style is appropriate for a guideline. My experience is that if you define something twice, you don't get the sum of the two definitions, you get the least common denominator. Dhaluza 12:03, 20 February 2007 (UTC)[reply]

While I agree that a brief mention followed by a more detailed description would be useful, what we have is the opposite:

Wikipedia does not publish original research

Original research refers to material that is not attributable to a reliable, published source. This includes unpublished facts, arguments, ideas, statements, and neologisms; and any unpublished analysis or synthesis of published material that appears to advance a position. Material added to articles must be directly and explicitly supported by the cited sources.

followed by:

No original research

Wikipedia does not publish original research or original thought. This includes unpublished facts, arguments, and ideas; and any unpublished analysis or synthesis of published material that serves to advance a position.

--Dhaluza 11:58, 20 February 2007 (UTC)[reply]

I still think this needs to be addressed. Are you saying the definition should come in the first appearance or the second? Dhaluza 23:31, 20 February 2007 (UTC)[reply]

The WP:RS guideline is a seperate issue and no comments about it make any difference as to the acceptance of the wording of ATT replacing the wording of V and NOR with the understanding that it is only wording being changed and not actual policy or implementations of policy being changed. In short arguments about RS are not arguments about ATT. WAS 4.250 02:22, 20 February 2007 (UTC)[reply]

Invoking an understanding does not affect that changing the wording does change the policy for anyone who hasn't read it before. -Amarkov moo! 05:45, 20 February 2007 (UTC)[reply]

Nonsense. If they have not read it before it can not be a change from what they read before. You are simply trying to be clever and failing at it. WAS 4.250 07:08, 20 February 2007 (UTC)[reply]

This is readable, concise, well-organized, and easy to understand. Kudos to everyone that was involved. Now let's see if we can keep this from becoming a complete mess within a few months... --Spangineer^ws (háblame) 05:43, 20 February 2007 (UTC)[reply]

I dont like this. Wikipedia:No original research and Wikipedia:Reliable sources state their meanings in their titles alone, even if you don't bother to actually read the policy itself; and as such will make far more sense to a newbie that Wikipedia:Attribution. I don't like this at all. Glen 10:25, 20 February 2007 (UTC)[reply]

Surely we can't title our policies with the expectation that nobody will read the text. If anything, the title "No original research" is misleading, because the OR policy is not about research in the common sense of the term; if readers only look at the title, they will never understand the OR policy. CMummert · talk 12:54, 20 February 2007 (UTC)[reply]

I agree, a user cannot simply look at the title of the verifiability policy for example and be expected to understand it. I think the title Attribution explains a difficult concept more accurately, though perhaps less intuitively. ^HighInBC ^{(Need help? Ask me)} 13:05, 20 February 2007 (UTC)[reply]

(conflict -- response to CMummert) No, but on the other hand it is useful if the title is a convenient summary of what the policy requires.

I think, however, that that is achieved here: We can say, in the case of either "unverifiable" or "originally researched" contributions, the they "lack attribution". And this makes a lot more sense to somebody who hasn't read the policy than either of the other two terms do -- those names don't carry an implicit understanding with them of why they're bad. Attribution, I think, does. JulesH 13:08, 20 February 2007 (UTC)[reply]

It has always been a challenge to distinguish between original research and unverifiable edits. This policy resolves the confusion. Beit Or 15:35, 20 February 2007 (UTC)[reply]

I agree. I think that from the title significance alone, this is a major step forward. It clearly tells editors that our statements must be attributed or attributable (depending on the controversy level) to some other published source. I think that is the core concept of Wikipedia's content - that we are here to summarize what has been published elsewhere. None of the previous titles so clearly conveyed this critical point, IMO. Crum375 16:09, 20 February 2007 (UTC)[reply]

Our policies are a lot more nuanced than even the nutshell intros can explain. There is no way for that to make sense in the title. SchmuckyTheCat 20:08, 20 February 2007 (UTC)[reply]

Still, it would be nice if we could find a title that was more active. Perhaps Wikipedia:Everything must be attributable? Or even Wikipedia:is a tertiary source? Rossami (talk) 01:46, 21 February 2007 (UTC)[reply]

Wikipedia:is attributable? Septentrionalis PMAnderson 21:32, 21 February 2007 (UTC)[reply]

Per my comments above I think we should simply merge this into WP:V, that title is quite clear and already known to most Wikipedians.-- _{Piotr Konieczny aka Prokonsul Piotrus | talk}  00:46, 22 February 2007 (UTC)[reply]

I think the problem with 'verifiability' is that it is a constant source of confusion. Many editors interpret 'verification' as digging into the 'truth' of a source, as in verifying its veracity. OTOH, the term 'attributability' clearly conveys that we as WP are not interested in truth per se, only in faithfully and neutrally reporting what reliable sources are saying about an issue, and attributing it to them. I personally consider the term 'attributability' a major step forward in clearly explaining our core content policies. Crum375 00:58, 22 February 2007 (UTC)[reply]

Do people understand that if this keeps being changed, it will fail, or is that the point? Pmanderson, I've not seen you help to maintain V or NOR, and yet you've turned up to make changes to them here. Can you explain why, please? SlimVirgin ^(talk) 20:25, 20 February 2007 (UTC)[reply]

Not my primary interest, but I have maintained them: check the history. Septentrionalis PMAnderson 20:54, 20 February 2007 (UTC)[reply]

You asked above of this diff what the problem with it was. As follows: (1) What is a citable assertion? (2) It doesn't "imply" the editor's opinion, it expresses it, as the original says. (3) "but should express the conclusion as that source's opinion, not Wikipedia;s [sic],": yes, but it's poor writing, so why change the original? It's also confusing because you move from talking about the editor's opinion to WP's opinion with nothing to explain the transition.

1. One for which there could be a citation (if the example, as it might, cited the CMS, we could have said "cited"). The point of this much-worn example is that two statements can both be attributable, as here, but their conjunction constitutes OR.
2. No, it doesn't express the editor's opinion; or else the whole paragraph wouldn't be OR, just the words in which that opinion was expressed. I should have stated the opinion as "that Jones did not actually commit plagiarism."
3. Please try to improve poor writing, not remove it; that's the wiki method. Septentrionalis PMAnderson 20:54, 20 February 2007 (UTC)[reply]

The writing is fine, but you're causing it to deteriorate, so reverting does improve it.

A citable assertion is one that can be cited, not one for which there is a source. The whole paragraph is OR because the point of the paragraph is the conclusion, obviously, and the conclusion is OR. SlimVirgin ^(talk) 21:09, 20 February 2007 (UTC)[reply]

I agree that the whole paragraph is OR because the point of the paragraph is the conclusion, obviously, and the conclusion is OR; but the text you are defending doesn't actually say that. I think it should. Septentrionalis PMAnderson 21:17, 20 February 2007 (UTC)[reply]

The point you are trying to make — that if a source can be found even in principle, it's not OR — is true, but trivially so. Every single meaningful combination of words imaginable might have a reliable source for it, somewhere, some day, so we couldn't ever call anything OR if our definition of OR is that we know there's no source. So it's trivially true that, although OR is material that isn't attributable, the state of our knowledge is going to be imperfect.

• It will not be obvious to all readers what is OR here (the argument that Jones is not actually a plagiarist); nor is it obvious that the editor's argument is OR rather than unsourced; let's state that assumption. There will be people who read this page without Slim Virgin's long experience reading PoV-pushers' abuse of Wikipedia. Septentrionalis PMAnderson 20:54, 20 February 2007 (UTC)[reply]

Your changes make it murkier, and anyway, the point is: the policies can't be changed on this page. Please go to WP:V or WP:NOR if you want to propose changes. SlimVirgin ^(talk) 21:09, 20 February 2007 (UTC)[reply]

I repeat: I'm not trying to change policy, I'm trying to express the same policy. The old words were never clear, and this page has now changed the definition of OR. 21:15, 20 February 2007 (UTC)

You could change "a source is needed" to "a source would have to be found" to make exactly the point you're trying to make, emphasizing the conditional nature, but without messing up the rest of the paragraph.

I will consider that; I'm not sure it makes either point. Septentrionalis PMAnderson 20:54, 20 February 2007 (UTC)[reply]

SlimVirgin ^(talk) 20:37, 20 February 2007 (UTC)[reply]

Thanks for your response; it is this sort of dialogue which should be had over this, and over the question of "equivalent sources". Septentrionalis PMAnderson 20:54, 20 February 2007 (UTC)[reply]

You're welcome, but do you see how time-consuming it is, and how little is accomplished? Do you agree that that tweak would achieve the same point? SlimVirgin ^(talk) 21:09, 20 February 2007 (UTC)[reply]

I have asserted that this policy does require attribution to the source, when found. This is repetitious, but (I trust) harmless; we don't want to say that attribution on the talk page is "consistent with this policy". Septentrionalis PMAnderson 21:07, 20 February 2007 (UTC)[reply]

What does "attribution to the source, when found" mean? SlimVirgin ^(talk) 21:09, 20 February 2007 (UTC)[reply]

What I actually wrote was "provided the opinion is then attributed to that source": by adding to the sample paragraph: "as Robinson argues,¹³" or some such.

You mean a prose attribution? But that's a policy change; prose attributions aren't necessary in all cases. SlimVirgin ^(talk) 21:35, 20 February 2007 (UTC)[reply]

I agree, but they would be necessary in the hypothetical; the opinion that Jones didn't commit plagiarism can't be consensus of the sources. However, if both wordings could be misleading, we should stick with the old one. Septentrionalis PMAnderson 21:48, 20 February 2007 (UTC)[reply]

The present wording only seems clear to Slim Virgin because she's seen it so often; it is not. In combination with the new definition of OR, it actually changes present policy. Septentrionalis PMAnderson 21:26, 20 February 2007 (UTC)[reply]

What is the change from current policy exactly? There is no change in the definition of OR. SlimVirgin ^(talk) 21:35, 20 February 2007 (UTC)[reply]

The expression "cannot be attributed to a reliable source" does not occur in WP:NOR, and I see nothing equivalent to it. I join the consensus on making this change; I even agree it expresses best present practice; but it is a change. Septentrionalis PMAnderson 21:48, 20 February 2007 (UTC)[reply]

Pma, I apologize if a tone is creeping into my posts. Your questions are perfectly legitimate. I'm just frustrated because a lot of work has gone into this, and I'm not understanding your points, but that's almost certainly my fault, not yours. Please bear with me. SlimVirgin ^(talk) 21:37, 20 February 2007 (UTC)[reply]

And I don't mean to take up the tone; thank you for taking a breath. I think my last edit solves one problem; and I owe you the wording. Septentrionalis PMAnderson 21:48, 20 February 2007 (UTC)[reply]

""cannot be attributed to a reliable source" is the same as saying that there is no verifiable source to which we can attribute the material. WP:V speaks of "material that has been published by reliable sources", I do not see a change in policy, just a change in wording. ≈ jossi ≈ (talk) 21:53, 20 February 2007 (UTC)[reply]

I agree there has been no change of policy yet, just a shift in the range of what is OR; but therefore we must watch how we use "original research". In this case, we should, and I think the explanation now does, postulate that the conclusion that Jones isn't a plagiarist is OR in the narrow sense; or else say it was the case in the original dispute we are modifying. Septentrionalis PMAnderson 22:02, 20 February 2007 (UTC)[reply]

What would people think of giving the next sentence an active voice, thus: "If the same opinion were attributed to a reliable source that specifically commented on the Smith and Jones dispute and made the same point about the Chicago Manual of Style and plagiarism, that would be consistent with this policy."? That's the only possible improvement I can see. Septentrionalis PMAnderson 22:21, 20 February 2007 (UTC)[reply]

I still think this tweak is a good idea; and it would satisfy all my concerns. At this point, I'm content. If anyone still thinks the result unclear, or a change in policy, I don't intend to fight further, but I would like an explanation of the difference here. Septentrionalis PMAnderson 21:28, 21 February 2007 (UTC)[reply]

For every stereotype it is not hard to find a source attesting to it.

It is often not easy to find a source refuting it.01001 04:21, 21 February 2007 (UTC)[reply]

Do remember the sources have to be reliable. We're not just talking something someone posts on h(is|er) blog, we're talking fact-checked, editorially-controlled publications or scholarly peer-reviewed journals. And we certainly can use statistical information. So if a sociologist does careful research and publishes a peer-reviewed paper stating that a given stereotype really is correct 73% of the time, we use that. On the other hand, if they find it's totally bogus, we use that. Seraphimblade ^{Talk to me Please review me!} 04:38, 21 February 2007 (UTC)[reply]

Yes, it does. The reliability of sources will to some extent counteract that; but if we were doing Wikipedia in 1911, we'd wind up with the 1911 Britannica, more or less. But the solution is not to make Wikipedia a crystal ball, to forecast the science of the blessed, bias-free future; it's to go out and write the reliable sources that debunk stereotypes. Septentrionalis PMAnderson 05:13, 21 February 2007 (UTC)[reply]

We acknowledge that Wikipedia can't be used as a source. Howevery, do summary style paragraphs have to repeat sources made in primary articles? For instance, if we have a main article called "Religion in Fictional Country" and then a summary link about religion in the article about Fictional Country, do the two or three sentences in the summary need to be entirely resourced? I just watched someone do a wholesale wipeout of 30 or so entries from a summary list where each entry had it's own (presumably sourced) article because "the entries weren't sourced." I think the original articles sourcing should stand up when repeated in summary articles - or at least, if an existing summary exists with a link to the main article, it's bullying to delete it without checking the main article for a source. SchmuckyTheCat 22:43, 21 February 2007 (UTC)[reply]

This was discussed at WT:V a while back; the general consensus (IIRC) was that the risk of the two articles getting out of sync with each other was too high, so the policy shouldn't make any allowance for it. Each article must be sourced, rather than relying on the sources of other articles.

Deleting the information without making a good-faith effort to look for a source isn't particularly great editing, but it doesn't violate policy. JulesH 23:11, 21 February 2007 (UTC)[reply]

I have serious concerns about one particular aspect of the proposed WP:A policy.

The essence of the old WP:NOR policy was that "Articles may not contain any unpublished material, or any unpublished analysis or synthesis of published material that serves to advance a position." This policy established that verifiability, rather than truth, was the standard for determining whether or not information could be posted on an article page.

The purpose was to ensure that "true, but unverifiable" information would be excluded from the project, so as to prevent contributors from posting personal knowledge and/or novel theories and approaches.

I have concerns that the new WP:A policy could be misused, and invert the principle of WP:NOR to permit "false, but attributable" information to be presented as fact.

I would ask readers to consider the following questions:

• If a source which is generally considered to be reliable publishes a false statement, should we convey that statement? If so, how should we convey it?
• If a source of suspect reliability publishes a false statement, should we convey that statement? If so, how should we convey it?
• Does it violate Wikipedia's original research policies to show that a published statement is demonstrably false?

And,

• Should WP:A clarify that *verifiability*, rather than *attributability*, is the most appropriate standard for the inclusion of factual information?

I do not believe it was ever the intent of WP:NOR to permit false statements to be included on Wikipedia as "statements of fact", and I am concerned that the proposed policy change would lead to this very outcome. Comments welcome. CJCurrie 01:41, 22 February 2007 (UTC)[reply]

I think WP should not engage in deciding on the merit of issues. Deciding if something is 'true' or 'false' is definitely a value judgment, and we should refrain from it, as doing so would violate WP:NOR. Therefore, the only issue that we can address is "is it attributable to a reliable source or not". If yes, it's acceptable, else not. The remaining issues would be to find the proper balance required for a neutral presentation, but that also does not require us to make a true or false value judgment - only to determine the relative weight a given view point has in the overall published information about the subject. Crum375 01:48, 22 February 2007 (UTC)[reply]

With respect, I cannot entirely agree with this assessment.

This may come as a surprise in light of our disagreements earlier on the page, but I understand where you're coming from with your objections. I recognize the inherent dangers in evaluating a published source, and I understand the problems associated with "mission creep" (as some have described it). Notwithstanding this, I cannot accept the conclusion that all evaluation is inappropriate for this project.

Perhaps it would be worthwhile to compare the following situations ...

A.

A credible and respected newspaper asserts that the moon is made of green cheese. Should Wikipedia convey this information? If so, how?

Answer: Wikipedia should convey this information, as a simple statement of fact and without embellishment.

B.

"Respected Author X" attributes a statement to "Respected Author Y", citing one of "Respected Author Y"'s works. However, "Respected Author Y" does not actually make the assertion attributed to him/her in the citation provided.

Should Wikipedia convey "Respected Author X"'s assertion? If so, how? Is it a violation of WP:A to review "Respected Author Y"'s original statement?

I'll leave this for others to contemplate. CJCurrie 02:12, 22 February 2007 (UTC)[reply]

I basically agree with you about the intent of the policy, but I think that it's a relatively limited danger for a few reasons. For one thing, attribution is a criterion for inclusion, rather than exclusion. The burden of persuasion is always on the editor who wishes to include a statement to demonstrate that it represents the content of a reliable source; an editor who feels that the statement is insufficiently sourced, or that the source is unreliable, has the authority to delete it.

Secondly, I think that the reliability of a "reliable source" has to be adjudicated on a case by case basis, and implicit in that adjudication is an evaluation of whether the source is reliable at this time, on that subject. The possession of a PhD by the author of a book does not inherently confer reliability, particularly on subjects outside his specialty. Similarly, sources that were quite reliable at the time of their publication may become outdated. The 1911 Encyclopedia Britannica was authoritative at the time of its publication, and written by some of the foremost scholars of the time, but relatively little of its content should be deemed authoritative now. This is why the policy makes clear that it cannot be read in isolation from other policies, particularly not NPOV. NPOV's discussion of "undue weight" clearly presumes that an evaluation of "prominent adherents" and "significant viewpoints" will take place.

Third, given the inclusion rules, the "demonstrably" prong will generally mean that there are contrary sources to cite. If a statement can be demonstrated to be false using other reliable sources, then one of two things is likely to happen: 1) the contents of the debate will be included in the article, allowing the reader to observe that one side is plainly wrong, or 2) the debate will be on the talk page and the false statements will be removed from the article. Some debates among the experts are sufficiently notable for 1) to be appropriate, but I think 2) will be more common, particularly on minor points.DCB4W 02:21, 22 February 2007 (UTC)[reply]

A big problem happens when the statement can be (in the ordinary sense) demonstrated to be false, but the sources used to do so don't meet Wikipedia's standards for reliable sources. Then no "debate" can be included in the article.

Of course, you can say "if you haven't demonstrated it to be false using reliable sources, then you really haven't demonstrated it false at all", but that ignores the fact that Wikipedia's requirements are strict enough to exclude the kind of reasoning that we accept in almost any non-Wikipedia context. If the sources say a bridge has no traffic, and a person who visits it sees that it has traffic, then that person has, in any reasonable sense, demonstrated that the reliable source is wrong, even if Wikipedia policy says that we can't use his words because they aren't professionally published. Ken Arromdee 03:50, 22 February 2007 (UTC)[reply]

I don't see it as a problem. Wikipedia is simply a summary of published sources. This does not mean that all sources are always correct, in any sense. All we are trying to do is to present a balanced summary of what reliable sources are saying about a given topic, and if some are clearly wrong, so be it. We as WP are just objective witnesses of the published information, not judges of fact or truth. Crum375 03:57, 22 February 2007 (UTC)[reply]

Except we allow articles on topics that have relatively few, less reliable sources. If Wikipedia presents some information on Newtonian mechanics, there are thousands of highly reliable sources on this to corroborate it and with which to correct any errors. If Wikipedia presents some information that only has a handful of weak sources (perhaps even simply old sources) we have a situation where if any highly reliable source were to examine the issue they would correct the obvious errors in the weak sources—but the topic has so little notability that no publisher cares to do so. Consider another possibility. There is a published book from 1950 on some obscure physical theory that has "false" information, versus a physicist's blog today. The book is conventionally more "reliable", and the blog is a self-published website, but the book is demonstrably wrong. Also, note that even in topics with numerous highly reliable sources, Wikipedia is not simply a summary of published sources; in order for that to be the case "Wikipedia" would need to review all published sources in order to reflect them accurately. What we have instead is some editors who might only look at a couple published sources, and someone else comes along and points out that the layman wrongly copied the idea out of a wrong textbook (and hopefully points us to a more reliable source, but often not). —Centrx→talk • 05:24, 22 February 2007 (UTC)[reply]

I see it as a rather serious problem if WP:A is to be interpreted in such a manner as to give editors sanction to convey one author's misattribution of another as a simple factual statement. The very idea almost seems a reductio ad absurdum demonstration of the principles of NOR. CJCurrie 04:24, 22 February 2007 (UTC)[reply]

I'll respond to both Centrx and CJCurrie. I see WP as a diligent and intelligent librarian who is being asked by someone interested in a topic to neutrally summarize the published information about it. The librarian would try to only look at 'reliable' publications, would summarize them and balance their contributions to the summary based on their relative preponderence and acceptance (i.e. cross citations) in the published literature. The librarian is not a researcher beyond knowing how to summarize the existing materials. Yes, there could be topics as Centrx said where the sources are weak, where old books compete with modern blogs, and as CJCurrie says where some 'reliable' source are plain wrong (assuming a refutable type topic). But the librarian still has to come up with an optimal summary given the facts on the ground, and creates a reasonable common sense document. That should be our approach, where all of us collectively are that librarian, and we'll keep tweaking each entry to get it closer to the ideal optimum, given the constraints of poor or inconsistent sourcing etc. Bottom line: we want pure summary of attributed sources, no OR, we keep working on the balance and source quality, we get ever closer to the the 'sweet spot', but may never quite get there. Crum375 06:04, 22 February 2007 (UTC)[reply]

Policies and guidelines are not a substitute for thinking. There is no substitute for actual understanding. WAS 4.250 06:14, 22 February 2007 (UTC)[reply]

Thank you for restoring some sanity to this discussion.

If Wikipedia ever reaches the point where "false, but attributable" information is allowed to be presented as a simple matter of fact, then we'll have become the monstrosity our that enemies make us out to be. CJCurrie 06:32, 22 February 2007 (UTC)[reply]

I see a red herring above. "False, but attributable" information is only going to become a problem if people cease using common sense. Nowhere is it suggested that if you know a supposed fact to be wrong, you must assert it.

• "If a source which is generally considered to be reliable publishes a false statement, should we convey that statement?" If you know it's false don't assert it or do an in-sentence attribution: "The Times reports the moon is made of cheese.[1]"
• "If a source of suspect reliability publishes a false statement, should we convey that statement?" No, obviously, and the policy doesn't say we should.
• "Does it violate Wikipedia's original research policies to show that a published statement is demonstrably false?" If it's "demonstrably false" than you can find another reliable source saying so. "The Times reports the moon is made of cheese.[1] NASA disputes this.[2]" Marskell 08:30, 22 February 2007 (UTC)[reply]

Note that this entire issue isn't new. It was discussed on WT:V more than once. Why? Because verifiable is defined there in pretty much exactly the same way attributable is defined here. Only a misunderstanding of what the policy meant by verifiable let people think that whether something is true or not entered into it at all.

I still think this policy should include text that instructs editors to use common sense. Reliable sources publish incorrect information, and unreliable ones publish correct information. If there's a good way of telling the difference (which may be specifc to the particular situation), that should be enough for us. JulesH 09:35, 22 February 2007 (UTC)[reply]

Something can be demonstrably false when there is no other "reliable source" saying so. In Wikipedia articles, only statements supported by reliable sources are allowed. But on article talk pages, editors can express their opinions; present logical arguments; refer to blogs; report that they've emailed an institution and gotten a certain answer; encourage other editors to email the same institution; ask other editors to go and look at the sky right now to see whether an eclipse is happening, etc. If the editors on the talk page are convinced that a statement in a reliable source is false, or that there's a good chance it's false, they can and should simply delete the statement from the Wikipedia article. --Coppertwig 09:42, 22 February 2007 (UTC)[reply]

I agree with some parts - I agree that we should use common sense as we try to investigate the various sources. But our goal should still be to simply present a balanced and neutral summary of what's been reliably published. If something that was reliably published is 'demonstrably wrong', odds are extremely high someone else would have said so and published that, so obviously we can add this into the balance. We could also use time as a factor - if a modern publication disputes an older one, we should say that, and it of course would impact the presented balance. I see no need to send editors outside to look at the sky, or to otherwise do OR - but I do see it very reasonable to seek better sources, or seek input from sources via email. Often the source may give you via email a pointer to a publication you missed. But at the end of the day, all we do is present a balanced picture of the published reliable sources, period - we use our 'source research' only to seek more sources and vet the ones we have. And our 'vetting' should focus on the publisher/publication, not on the individual author or work, per my comments elsewhere on this page. Crum375 10:57, 22 February 2007 (UTC)[reply]

I agree with DCB4W, above, that "attribution is a criterion for inclusion, rather than exclusion". If a statement is false, and it isn't crucial to the article, we can and should simply exclude it, even if we only explain why on the talk page, rather than in the article itself. We ran into that sort of thing lots of times in writing Jenna Jameson, where several of our sources disagreed. We weighed the evidence, and simply excluded the ones we believed were false. Talk:Jenna_Jameson/Archive_2 has many instances. If a statement is crucial to the article, then we need to give it, then give the sources that describe it as false. For example, in Situation A, we have to ask, is the article about the newspaper, or about the composition of the moon? If about the newspaper, we can probably just exclude it, the newspaper publishes lots of things every day. If about the composition of the moon, we should include it, in an "alternative views" section; hopefully we have lots of more scientific views earlier in the article. In Situation B, similarly, we need to treat the assertion differently based on what the article is ostensibly about. --AnonEMouse ^(squeak) 16:14, 22 February 2007 (UTC)[reply]

Presenting a balanced summary of sources requires discretion and common sense. If some reliable source appears to represent a tiny fringe view then it can be excluded. OTOH, if there is a reliable source which appears to many Wikipedians to present 'false' information, yet is clearly widely accepted by many people and is relevant to the subject at hand, we cannot exclude it. Otherwise, someone can decide to start removing certain well-sourced religious views or historical interpretations they don't accept because they consider them 'false', etc. Of course when we do include something controversial, we attribute it to its source. But we can't just arbitrarily remove well sourced relevant items because editors on a given page decide they are 'false'. I think "attributability, not truth" sums it up nicely. Crum375 16:25, 22 February 2007 (UTC)[reply]

"If something that was reliably published is 'demonstrably wrong', odds are extremely high someone else would have said so and published that".

"Odds are extremely high" doesn't mean "it's certain". If the odds are extremely high, but not certain, sooner or later you will run into cases where it hasn't happened, like the covered bridge example. *Nobody* published anything saying the reliable sources are wrong, even though they are indeed wrong.

This tends to happen more with obscure topics, when there are relatively few sources and many of them just copy from each other.

"If some reliable source appears to represent a tiny fringe view then it can be excluded."

This would fail in the covered bridge example, since all the sources say the bridge is closed to traffic. It's not a fringe view--it's the majority view. Ken Arromdee 16:49, 22 February 2007 (UTC)[reply]

I think what many Wikipedians fail to understand is that we are not here to present the truth - simply a summary of reliably published sources. So if once in a while we'll have a case where all the published sources are plain wrong, we would still not be able to say so, since that would be OR. All we can do in such cases is search long and hard for the correct source, and make sure it is properly presented. I would not even rule emailing a local reporter to write a little blurb about a widely mis-reported phenomenon. Once that article is published, we can then use it. Until then, it would violate our policies to create OR or ignore the reliably published sources. Crum375 17:04, 22 February 2007 (UTC)[reply]

Exactly. This may be the most difficult aspect of WP for some to swallow. ≈ jossi ≈ (talk) 17:41, 22 February 2007 (UTC)[reply]

I don't disagree with you, though I suspect that it's exceedingly unlikely that we'd encounter a situation where Wikipedians would be aware of the "correct" information and all published, reliable sources would be plain wrong. Perhaps this might occur for very short period of time (a few hours) where the media is relatively slow to publish a story confirming what was already being reported by blogs or such. You're quite correct in your main point: that Wikipedians often see "truth" as the primary goal of the project. ChazBeckett 17:58, 22 February 2007 (UTC)[reply]

What do you mean it's exceedingly unlikely? I just gave you the covered bridge example. There's also Talk:Sacha_Baron_Cohen#include_Persian_descent. (I suppose technically the latter doesn't count because some published sources are silent on the guy's ancestry, but it's still along the same lines.)

Incidentally, I don't believe that saying that published sources are wrong is OR. We are allowed to do things to decide what goes in articles, that we couldn't actually *put* in the articles. The canonical example is that we're allowed to search Google to see if something is notable, but we can't put the results of a Google search in the article. Ken Arromdee 21:56, 22 February 2007 (UTC)[reply]

We are allowed to state our opinion in the Talk page as to the quality of various sources and our focus should be on finding the best possible sources and summarizing them in a neutral and balanced fashion in the article. The actual 'veracity' of the stated information, while interesting, is not our focus here. Crum375 22:08, 22 February 2007 (UTC)[reply]

Opinions seem to vary widely in Wikipedia regarding whether studies that are cited must necessarily be peer reviewed. It would be great if this could be clarified in WP:ATT. What do those here think? Should research that's referenced have met the standard of peer review? Yes or no. : ) Thanks.TimidGuy 16:31, 22 February 2007 (UTC)[reply]

No. Too simplistic, and as a peer reviewer I would not support the assertion that it is inherently reliable.ALR 16:39, 22 February 2007 (UTC)[reply]

Peer review is only one of a number of criteria. And journals have varying degrees of peer review, they can claim peer review but have pretty low standards for it, or only do it in some cases. The context of a study should be included in an article so readers can decide how much credibility to grant it. And results of a study should be presented as just that - things shouldn't be presented as "fact" unless there's extremely widespread and documented acceptance. I'm dealing with some major headaches right now over this with pseudoscientific topics. One study in a fringe journal with weak peer review and people want to say "scientists say..." or "experts agree..." --Milo H Minderbinder 16:48, 22 February 2007 (UTC)[reply]

The studies that TimidGuy is referring to include an unpublished study found in a published book consisting mostly of peer-reviewed studies, performed by a guy who never published a peer-reviewed study on the subject Otis. My own question is, if self-published studies are allowed, why not poster sessions/conference proceedings? -Sparaig 20:03, 22 February 2007 (UTC)[reply]

I've seen poster sessions, but I never tried to find a published source for one after the conference. Are poster sessions published?

As for conference proceedings, where does it say they are not allowed?

Going back to the original post, I am not aware of different criteria for "studies" compared to other forms of information. Certainly we cannot require that all scientific information come from peer-reviewed journals, because these journals only publish in fields of current research activity. Settled science is found in textbooks, not journals. --Gerry Ashton 20:15, 22 February 2007 (UTC)[reply]

But science textbooks themselves are peer-reviewed, IIRC, and only include the most widely accepted material, unless they explicitly state otherwise. There's a BIG difference between a textbook and a popular-press treatment, which can include anything the authors want want to say. -Sparaig 20:44, 22 February 2007 (UTC)[reply]

This page purports not to be creating policy innovations. However, WP:REF and WP:RS are guidelines, not policy. This document incorporates too much of the content of those documents not to be seen as creeping innovation. I cannot help but feel that this direction of creeping innovation has made Wikipedia worse over the past year or so (for instance the bazillion pages littered with citations whereby some self-proclaimed expert is trying to enforce his POV with what he deems to be reliable sources (and where nationalism gets involved, it just gets ugly)).

Aside from the innovation matter, this proposed policy is far less clear than the two it is trying to replace. The nationalist misunderstanding (and its variations) will be perpetuated and accentuated, and élitism will be re-inforced (there *is* a cabal), an élitism which I cannot reconcile with WP:DBN.

Two policies give two lines of attack against the bad-faith "but I've attributed everything" brigade. One policy gives just one, and this proposed policy would not be a very good one at that.

I am not advocating an "if it's not broke, don't fix it" approach to the two policies concerned. My opinion is that they are very much broke, but that this is absolutely not the way to fix them. A435(m) 22:29, 22 February 2007 (UTC)[reply]