Battle of Porto Bello (1739)

Tertiary and secondary

I just reverted a edit that would make tertiary and secondary sources equal in weight which may or may not be the case but the wording supports that they be equal. This is a small bit of wording but pretty big impact so should be discussed as to how to word it(olive (talk) 22:02, 30 September 2008 (UTC)) My edit summary should say do not necessarily have same weight...To clarify.(olive (talk) 22:07, 30 September 2008 (UTC))

RE this edit: If the wording is going to be "Wikipedia articles should rely mainly on published reliable secondary sources, and to a lesser extent, on tertiary sources." it should indeed be discussed. The issue of "weight" is completely irrelevant to WP:NOR. I think it's not appropriate to be making any kind of judgment about "weight" of secondary vs. tertiary sources, or any assertion about how often tertiary sources should be used compared to secondary sources. That's an editorial decision that depends in large part on an assessment of reliability w.r.t. what's being presented in some particular place in an article. Basic textbooks are perfectly acceptable sources if they're assessed to be reliable sources, as are encyclopedias. ... Kenosis (talk) 22:24, 30 September 2008 (UTC)

actually, that's not quite correct (granting that there is some ambiguity in the language, here). the distinctions being made are quite real. there are three different types of writing we have to consider:

writing intended to advance a position

This is mostly what we refer to by the term Primary Research. basically it amounts to an author or authors trying to present something new or innovative for the consideration of scientific peers. this has an inherent point of view, and can't be trusted to present an objective view of the topic (though it can be trusted to give a coherent description of its own perspective). this is experts in the field talking to other experts in the field

writing intended to explore or expound upon a topic

This is mostly what we mean by the term Secondary Sources. this is an author or authors analyzing and synthesizing a breadth of material about a given topic in order to give a general perspective, or to put it to use in a particular case. they are not trying to be innovative or new, but simply to sort through the various perspectives available and come to some sort of conclusion about them. this is experts in the field talking to both lay readers and other experts in the field.

writing intended merely to describe

this is Tertiary Sourcing. neither the authors nor the intended readership need to be experts in the field; this is merely synopses or collections of information about the field, without analysis. it's useful in some cases, but it is never as clear a picture of the topic as is found in secondary sources.

--Ludwigs2 22:56, 30 September 2008 (UTC)

No. This presentation doesn't accurately summarize primary, secondary and tertiary sources. Please review the sources given in Footnote 4 of the policy page. ... Kenosis (talk) 22:59, 30 September 2008 (UTC)

I did include, here, above, the words "necessarily" to correct my edit summary. The point is, that the wording in place when I reverted implies equal "billing" (so as not to use the word weight), (I wasn't using weight in the Wikipedia sense - sorry for that confusion) and this is also a judgment call. I would suggest,however, that the use of an encyclopedia, a tertiary choice, is a poorer choice than a secondary choice in the writing of another encyclopedia, if both are available. I think the wording here is critical and neither what was in place or what I reverted are ideal, but to my mind the version in place now, more clearly points an inexperienced editor in the right direction.(olive (talk) 23:04, 30 September 2008 (UTC))

Understood. Encyclopedias are not limited to general encyclopedias like Britannica and such. There are numerous highly reliable specialized encyclopedias that are invaluable for getting a handle on various difficult topics. Examples in, say, philosophy, would be the eight volume Macmillan Encyclopedia of Philosophy (both the 1967 and 2006 issue). It is sufficiently reliable that most professional academics regard it as a credible resource on any topic in it. Same with the Routledge Encyclopedia of Philosophy . The online Stanford Encyclopedia of Philosophy too is, on the whole, a credible online resource. Point being, these tertiary resources are sufficiently reliable that I'd have to disagree with the inclusion of any statement in this policy page that made an implicit demand that editors necessarily give priority to divergent secondary sources over reliable tertiary sources. I do realize it's likely not of major importance to the general thrust of the policy. ... Kenosis (talk) 01:41, 1 October 2008 (UTC)

Kenosis - actually, that footnote refers to sourcing in historiographic works, not to sourcing in general. if you read through what I wrote, you'll see it's just a generalization of what's given in footnote 4 (or rather, f4 is version of what I wrote applied to the needs of a particular field). --Ludwigs2 00:01, 1 October 2008 (UTC)

Obviously it's been changed along the way. Typical. My apology. ... Kenosis (talk) 01:46, 1 October 2008 (UTC)

Ludwigs, some broader academic perspectives that go beyond historiography can be seen here, here, and here. ... Kenosis (talk) 14:13, 1 October 2008 (UTC)

I don't have an opinion on this really, but FWIW, sourcing one encyclopedia off another sounds like concentrating poisons in the food chain. ——Martin^phi ☎ Ψ Φ—— 04:32, 1 October 2008 (UTC)

Wikipedia is an encyclopedia incorporating elements of general encyclopedias, specialized encyclopedias, and almanacs.

@Kenosis (and others)—those links which you provided just above make for worthwhile reading. They discuss distinguishing between primary secondary and tertiary sources and are very much worth a look at. -- NewbyG ( talk) 22:36, 4 October 2008 (UTC)

Those links are indeed worth a look-see. We had at one point included a couple of them, but they were removed/replaced. ... Kenosis (talk) 22:43, 4 October 2008 (UTC)

Refer directly

Re this edit [1], basically the change made it easier to do OR. I think that is something which we need to avoid, completely. I'll respond above, but let's not make the policy less strict, even by tiny degrees. "Refer directly" is different from "directly relate." The latter allows an editor to interpret what "relates." The brain is a relationship engine, and that gives way too much liberty. ——Martin^phi ☎ Ψ Φ—— 21:45, 1 October 2008 (UTC)

Support—Martinphi, strictly speaking, you are correct, "Original research is research that is not exclusively based on a summary, review or synthesis of earlier publications on the subject of research." (Original research) Since the policy is called "No original research", I cannot deny that such research is prohibited according to the policy. I conclude that to avoid "original research", one must use sources that "refer directly" to the topic of the article, defined broadly to include relevant subtopics. We can then rely on "Wikipedia:Ignore all rules" for the exceptions. --Phenylalanine (talk) 23:28, 1 October 2008 (UTC)

this makes no sense. "Broadly defined to include relevant subtopics" is another way of saying "refer indirectly" - you can then say it also has to refer directly. The key word here is relevant. Sources must be directly relevant, but they do not have to refer directly. Slrubenstein | Talk 23:46, 1 October 2008 (UTC)

Yes, they have to refer directly, in any case where there is any controversy: otherwise, the authors are doing original research by saying "I think this is relevant." For example, it might seem that spinal manipulation is directly relevant to Chiripractic practice, since chiropractors use SM. But unless the studies on SM refer directly to Chiropractic, it's OR to use them. When I say refer, I mean, they say the word "Chiropractic," as in "SM relates to Chiropractic in YYYY way." Then we can paraphrase, saying "According to X, YYYY relates to Chiropractic." If they mention Chriopractic as something that might be relevant, or cite a study of chiropractic as relevant to their conclusion, we can also mention those bare facts- but we can't go further than that. In other words, not broadly interpreted: broad interpretation is exactly what NOR is here to avoid.

Original research is putting together sources on the subject or relevant to the subject to reach new knowledge. In the case above, the "new knowledge" is that SM is relevant for drawing conclusions about Chrirpractic, or SM is relevant to Chriropractic.

I'm sorry to not agree, I don't want to make enemies here... ): ——Martin^phi ☎ Ψ Φ—— 00:10, 2 October 2008 (UTC)

Phenylalanine, the loophole is being closed. Jayjg ^(talk) 01:47, 2 October 2008 (UTC)

I see that once again, claim is made that we have consensus for the term 'refer directly'. And once again I see no consensus from this page. I suggest we all agree to follow the guidelines in WP:Consensus. When any editor reverts a change, that is prima facie evidence that consensus doesn't exist. Reverting it back with the claim that 'yes it does', is just ignoring plain facts (WP:IDIDNTHEARTHAT). lk (talk) 02:10, 2 October 2008 (UTC)

Lawrence, we aren't going to allow a loophole in the in the policy so that you can carry out a campaign against what you view as fringe theories. WP:NOR only works when it actually means something, and that's the case even if we really, really, really want to disprove some silly theory or claim which hasn't actually been disproved by reliable sources. Jayjg ^(talk) 02:15, 2 October 2008 (UTC)

Wait a min, weren't you the one proposing drafts which basically made the policy less strict? ——Martin^phi ☎ Ψ Φ—— 02:30, 2 October 2008 (UTC)

No, I've been the one trying to tighten up the policy from the start. Jayjg ^(talk) 02:32, 2 October 2008 (UTC)

Oh, right, that was Jayen466 :P ——Martin^phi ☎ Ψ Φ—— 02:34, 2 October 2008 (UTC)

• Actually, Martin, you got that wrong as well. You need to read the just-archived section Wikipedia_talk:No_original_research/Archive_37#.22Directly_related.22_vs._.22Refers_directly.22 – we've been discussing this for weeks. The people arguing for a loosening of the policy were, primarily, lawrencekhoo (lk) and Kotniski. Jayjg and myself were in favour of tightening it, and using the "refer directly" wording. Phenylalanine and Kenosis were somewhere in the middle. But all of us here did try to work out a compromise wording, taking on board everyone's concerns, and that is the one you saw. Jayen466 08:10, 2 October 2008 (UTC)

I'm not stuck one way or the other on the particular language here, so long as it reasonably well reflects the general intent of WP:NOR in a way that can continue to be dealt with by users across the wide range of topic areas on the wiki. I call attention to Slrubenstein's comment above at 23:46, 1 October 2008. To that comment I would add that the words "refer directly" needn't necessarily be taken in the strictest literal sense that sources must all refer explicitly to the topic of the article. A quick lookup in any dictionary shows that this wording does allow some reasonable degree of flexibility as to what precisely is meant by "refer directly". I point this out because it appears necessary for the policy language to be capable of accommodating the numerous instances wherein a source is unquestionably germane to the WP content at issue but where it may not explicitly refer to the chosen title of the WP article. I'm interested in seeing where this goes w.r.t. this choice of language. ... Kenosis (talk) 02:46, 2 October 2008 (UTC)

Yes, there would be cases where either use of a name does not really imply relevance, or when non-use does not mean it is not talking about the subject. I think the current wording is no different from the traditional understanding of NOR, but closes what people were trying to take as a loophole (else there would be no dispute here). Certainly, if I were to source something in an article about New York City to a book on the Indians of Manhattan, or something, who would complain? Similarly, I don't think anyone is going to stop giving necessary context when the context and relevance is not disputed. The only time NOR comes into play is when there is controversy. ——Martin^phi ☎ Ψ Φ—— 03:03, 2 October 2008 (UTC)

For example, this article by Fred Bauder the former Arbitrator is OR Luis Maria Baca Grant No. 4 [2] ——Martin^phi ☎ Ψ Φ—— 03:26, 2 October 2008 (UTC)

RE " The only time NOR comes into play is when there is controversy" : Well, yes, indeed. Same with WP:V and WP:NPOV. If a user sees something they assert is OR and deletes it or nominates the article for deletion, and someone else says "no, the material should stay", there's a controversy. Otherwise it's a non sequitur. It's somewhat analogous to the old conundrum: "If a tree fell in a forest and no one was there to hear it, was there a sound?" ... Kenosis (talk) 03:36, 2 October 2008 (UTC)

Yeah, (: ——Martin^phi ☎ Ψ Φ—— 04:56, 2 October 2008 (UTC)

I will support "refer to" only if it allows me to use a source referring to the "environmental effects of intensive livestock farming" in the article "environmental effects of meat production" and if I am allowed to update the schedule of the first LHC particle collisions in the Lhc safety article (see my example in the "Is this any good?" section). --Phenylalanine (talk) 03:33, 2 October 2008 (UTC)

Also, why not just: "you must summarize and cite the reliable sources available on the topic of the article." ? --Phenylalanine (talk) 03:40, 2 October 2008 (UTC)

Because that's just as big a loophole as the other wording. How does one decide if something is "on the topic of the article"? At least with "refers directly" there are some concrete parameters. Jayjg ^(talk) 03:56, 2 October 2008 (UTC)

Phenylalanine this should offer no problem with the meat production thing. On the face of it as you describe it, there shouldn't be any trouble with the LHC situation either. You'd have to have a really nasty POV pusher on your tail before it would, and a POV pusher is going to do something or other no matter what. So don't worry.

Agree with Jayjg on the loophole thing. ——Martin^phi ☎ Ψ Φ—— 04:02, 2 October 2008 (UTC)

Lawrence, we aren't going to allow a loophole in the in the policy so that you can carry out a campaign against what you view as fringe theories. No personal attacks please. And are you using the royal 'we' perhaps? Or is that 'Jayjg is Wikipedia'? lk (talk) 05:41, 2 October 2008 (UTC)

Lawrence, you showed up at this page insisting that you needed an even looser version of NOR so that you could use OR to battle "fringe theories". Your examples were roundly shown to be violations of the OR policy even under the old wording, much less the new; see Wikipedia_talk:No_original_research/Archive_37#Proposed_addition_to_WP:SYN. Meanwhile, most of the rest of us commenting here have been attempting to close the existing loopholes, rather than allowing the current abuses to go on. Now, if you can come up with other wording that helps close the loopholes, please suggest it. Regardless, the loopholes need to be closed, and will be. Jayjg ^(talk) 05:50, 2 October 2008 (UTC)

Ah, once more with the personal attacks, and the claim of consensus. No one can blame you for being inconsistent. lk (talk) 06:00, 2 October 2008 (UTC)

We're going round in circles here, repeating the same arguments. See the discussion higher up the page. We've already established that in some (probably quiet frequent) cases we actually need loopholes. If we say "refer directly" then we must explain somewhere else what exceptions from the literal interpretation of the phrase are tolerated. Until we do that there is clearly no consensus for any tightening up of the policy. Please, good people, STOP making these substantial edits to the policy for which you KNOW there is no consensus (it hasn't just been Lawrence opposing, far from it). If the policy keeps changing or gets locked down in a random state, it loses all credibility. --Kotniski (talk) 05:57, 2 October 2008 (UTC)

We have not established, in any way, shape or form, that "in some (probably quiet frequent) cases we actually need loopholes". There have been no examples provided where they were required, or even desirable. What makes a policy lose credibility is when its wording is so vague as to be meaningless; which is, in fact, the current state. The NOR policy is intended to mean something - in fact, what it says - and the wording will be changed to reflect the intent of the policy. You can assist in that process, or not, but please stop disrupting it. Jayjg ^(talk)

Well, the policy is NOR, the name says it all. I explained above why no clarification is needed. Also, we already explain what we mean by "refer directly":

"This page in a nutshell:

• Wikipedia does not publish original thought: all material in Wikipedia must be attributable to a reliable, published source.
• Articles may not contain any new analysis or synthesis of published material that serves to advance a position not clearly advanced by the sources."

Note the bolded word.

"If the sources cited do not explicitly reach the same conclusion in direct connection to the topic of the article, then the editor is engaged in original research."

"Even with well-sourced material, however, if you use it out of context or to advance a position that is not directly and explicitly supported by the source used, you as an editor are engaging in original research."

No original research means none at all, except as Kenosis notes above everything is relative to objections. ——Martin^phi ☎ Ψ Φ—— 06:07, 2 October 2008 (UTC)

And please stop fighting, I say that to a few of you. Ain't going to help. Also, don't incite people by wielding the sword of might and saying what will be done- it just makes people mad. ——Martin^phi ☎ Ψ Φ—— 06:11, 2 October 2008 (UTC)

maybe (if we're going to consider a change here at all) we just need to re-place the focus. something like: "to demonstrate that you are not presenting original research, you must cite reliable sources who are specifically presenting the argument or opinion with respect to this topic" would that work, or does that make it too narrow? --Ludwigs2 06:17, 2 October 2008 (UTC)

I can understand the desire to tighten the criteria for using sources. I have some personal experience, though, with using sources that don't, arguably, refer directly to the subject of the article. I wrote much of this [3] section on official apologies in the Japanese war crimes article, primarily in response to two editors who were edit-warring over its content. Some of the sources I used cover the general topic of the nature of apology, but not specifically as to how it relates to Japanese war crimes. Under the proposed verbiage, would all sources have to mention the subject of the article to be used? Cla68 (talk) 07:06, 2 October 2008 (UTC)

• I am sorry, but everything from "Some in Japan have asserted ..." in the [4] section is OR, and moreover, it does not even adequately reflect what the cited sources state.
• "Some in Japan have asserted that what is being demanded is that the Japanese Prime Minister and/or the Emperor perform dogeza, in which an individual kneels and bows his head to the ground" is not backed up by the cited source. The controversy about the dogeza statement should simply be described, without adding the (weasel-worded) OR sentence "Some in Japan have asserted."
• "Some point to an act by German Chancellor Willy Brandt, who knelt at a monument to the Jewish victims of the Warsaw Ghetto, in 1970, as an example of a powerful and effective act of apology and reconciliation similar to dogeza, although not everyone agrees." This is not backed up by the source, which does not mention Japan or anyone in Japan or elsewhere pointing to Brandt's act, but simply discusses Brandt.
• The paragraph "Citing Brand's action as an example ..." is again OR, because none of the sources cites Brandt as an example in the Japanese context. Jayen466 10:09, 2 October 2008 (UTC)

(after ec): That sounds fine, in the case of arguments and opinions. As long as it's made clear that established facts which serve to add context to the article or statements made in it, thus improving the reader's understanding (see examples above), are not excluded by this policy just because we don't have a source for them that specifically mentions the "topic" (whatever that means) of the article. An excess of off-topic facts is likely to be undesirable (because it distracts readers from what they came looking for), but that's an issue to be dealt with primarily through other policies (guidelines) than this one, because: (a) however undesirable, it's not original research in any normal meaning of the phrase; (b) it's just as undesirable if you can source it to on-topic references; (c) it's a delicate matter for editorial judgement and good sense, and therefore better dealt with through guidelines than an important core policy whose rules need to be enforced very strictly.

My other concern with the "refer directly" wording is that we don't know what the "topic" of the article means. It might be taken to be the title, word-for-word, which is clearly not wanted - we have examples above that show that. In fact the topic of a particular section might be not even a synonym of the article title, but a subset of that topic, or a closely related topic that has been merged into the article.

So to sum up, I don't mind what exact wording we use as long as it can't be interpreted as meaning either of (1) sources that don't mention the topic are abolutely prohibited even for well-established facts (2) the "topic" is always the article title or synonym thereof. These are my only concerns; they don't relate to a desire to push any point of view, but to prevent POV-pushers from hijacking this important policy to thwart editorial common-sense. --Kotniski (talk) 07:09, 2 October 2008 (UTC)

I agree with Martin, Jayjg, Kenosis and other's comments in support of a strict reading of the policy. I think one problem is that a lot of people write content to articles and then look for sources to support it. I don't think this process will change (so this policy needs to take it into account) but I think a much better process is (1) identifying relevant sources and (2) building up an article that provides accurate accounts of the views expressed in those sources. This raises the question of what sources does one use? And I have to emphasize that my perspective is not what sources are appropriate to support a sentence or paragraph that has already been written, but rather, what sources to use to build up new content in an article. I guess that is why the word "relevant" or "related" to the topic is so important. When I worked on the article on the historical and cultural context for Jesus, some sections were not directly about Jesus, but were about the historical context (e.g. who were the Pharisees? What was the Temple?). Now, I know "context" or "background" is there in the title of the article, but there are many articles where words like context or background are not in the title but still relevant to the article and editors should be looking for sources on the relevant context or background. When it comes to the second part of this process I agree completely in strictness: once we have agreed that a source is relevant, it has to be represented accurately. It has to be used as a source for claims that the source itself makes and not for claims not made by the source. In this case the word "direct" is sensible and appropriate. I hope this clarifies my position. Articles should be based on sources relevant to the topic. Specific content in the article has to be directly linked to the sources. Does this make sense? Slrubenstein | Talk 14:21, 2 October 2008 (UTC)

I think I agree with you entirely.--Kotniski (talk) 17:23, 2 October 2008 (UTC)

Whatever language is used will need to allow for the kind of sourcing Slrubenstein is referring to here. Note that I used the words "refer-ing to" in describing what Slr is talking about. It appears to me there is some degree of editorial flexibility afforded users by the statement that sources must "refer directly to the topic (or subject) of the article." If, however, WP users are going to be widely interpreting this language to mean that anything with sources that don't explicitly say they're talking about the title of a given article is original research, I would imagine the "refer directly to" language won't be very sustainable. ... Kenosis (talk) 21:22, 2 October 2008 (UTC)

Cla68, you are demonstrating what I've been saying: the speed-limit model which we have now works. The speed limit is an absolute which everyone breaks to some degree. If there were no objections to your OR, that is because no one got to it, or because everyone agreed. At any time, they could take it out as OR. But, because it has consensus it stays. And that model works. Making it clear, however, will open the door to all sorts of screwing around. We do not want this policy to be clear. We want an absolute bedrock which is only dug down to in an emergency, but nevertheless remains solid. No, we don't have to make it clear. It's worked fine for years without being clear that people do a lot of OR.

Now, "topic": "The subject matter of a conversation or discussion". "Subject": "The subject matter of a conversation or discussion" Those definitions give all sorts of room. There is no need for any more. Yet, it is not so inclusive that if the source may not be applicable, we nevertheless let it in, as in the example of Chriopratic and spinal manipulation sources. The reason that SM doesn't work for Chiro is that there is no guarantee that SM is the same as what Chiro does, especially in context (chiro might do something to ruin the effectiveness of SM, who knows). But, like I said, there will be times when the same thing under another name is fine. Every time someone here has said "well can I do this reasonable thing," I've been able to answer yes, as with the meat production example. However, for example Electronic voice phenomena, do you use sources from the recording industry? No, you don't, because you have no guarantee that the sources are really relevant to the subject. There might be all sorts of ways in which what the recording industry says about anomalies in recording media are not relevant to the particular anomalies which are said to be EVP- and you don't know, even when it seems relevant. So, because the source does not mention EVP, it is out. Yet, if you had a source, say, for the history of hypnosis, and it was about "mesmerism" and never said anything about hypnosis, you use it because the two are directly related in a known way. By "in a known way," I mean that it is connected without any possible flaws in the connection. The same goes for different wording which is clearly relevant, again as with the meat production example.

So, in my belief we already have a policy which allows enough room to use the sources which we know and can demonstrate without doubt are relevant, yet is strong enough that it excludes anything doubtful (per a reliable encyclopedia). It even allows OR, as with Cla68, if no one objects (as they very often don't). ——Martin^phi ☎ Ψ Φ—— 21:58, 2 October 2008 (UTC)

(undent) allow me to try to clarify the issue (don't know if I'll succeed, but it's a start...):

• what we do want to allow
  • sources that are writing about the article topic in a general way
  • sources that are writing about a subject that is a noteworthy and important element of the article topic
  • sources that are writing noteworthy criticisms of the main subject(s) of the article
• what we don't want to allow (forms of wp:syn)
  • sources writing about different topics entirely, that happen to mention or make reference to the article topic
  • sources writing about related subjects, where their results might be applied to the article topic, but weren't explicitly
  • sources writing about broader topics, which might be construed to apply to the article topic itself, but weren't explicitly
  • sources writing general criticisms of a broader topic, or writing specific criticisms of a sub-element of the article topic, unless those criticisms are noteworthy with respect to the article itself

basically, the kind of synthesis we are trying to avoid here is what they call in syllogistic logic the Fallacy_of_the_undistributed_middle or the Fallacy_of_four_terms where an unspoken and potentially unwarranted inference is drawn between two ideas, producing a questionable result. does that capture it? and if so, how do we word it to reflect this? --Ludwigs2 22:16, 2 October 2008 (UTC)

Yes, that's perfect. ——Martin^phi ☎ Ψ Φ—— 22:46, 2 October 2008 (UTC)

Excellent Ludwigs. Unfortunately, in discussion, to argue that there are "jumps" in logic, or that a certain kind of writing creates implied hidden and illogical conclusions meets with very little respect. Including something about logic here seems an excellent idea. This seems very good to me: "where an unspoken and potentially unwarranted inference is drawn between two ideas, producing a questionable result." (olive (talk) 23:10, 2 October 2008 (UTC))

RE Ludwigs' post: In other words, we do want sources to "refer directly" to the topic or subject of the article. ;) ... Kenosis (talk) 00:05, 3 October 2008 (UTC)

I'm not sure Ludwig's got it exactly right, but it's close. I don't follow Kenosis' interpretation of his words though - Ludwig seems explicitly concerned with what a source is "about" than what it "refers to". Most of all I don't agree with Martin's statement that "we don't want this policy to be clear". I guess this is just a difference of fundamental philosophy, but making things deliberately opaque seems quite foreign to what we are trying to do here. It assumes very bad faith on the part of fellow editors to imply that if you tell them accurately what you know about how things are done, they will abuse that information to do something else. I can't accept the idea of a priestly caste who know the truth deliberately withholding that truth from the sinful masses. (Perhaps we should write the whole thing in Latin to make it even less accessible?) No, we should work towards finding wording (as Ludwig has attempted) that makes it very clear what is acceptable, and very clear what isn't. --Kotniski (talk) 07:51, 3 October 2008 (UTC)

It seems clear to me that much of this ongoing debate over the words "refers to" vs. "in direct connection" or "directly related", etc., has to do with how various participants interpret the meaning of those phrases. TBH, the points being made here are all, or mostly all, fairly reasonable. But I'd like to discourage any major makeovers unless and until there's some broad consensus for some particular significant change. ... Kenosis (talk) 14:30, 3 October 2008 (UTC)

Kotniski - I think part of the concern about making instructions too precise and detailed is that it starts to make editing a legalistic process rather than a cooperative one. editors might start pushing on the letter of the law, rather than discussing what makes sense in the context of particular articles. to that extent, I think a loose, interpretable rule is more in line with wikipedian philosophy than a precise and specific one. --Ludwigs2 02:44, 4 October 2008 (UTC)

Watering down WP:SYN

Semitransgenic, this edit is a very significant dilution of WP:SYN that goes well beyond the main focus of the debates in the talk sections above. I've reverted back to Newbyguesses' last version. We can't remove the explicit statement about what editors cannot synthesize without proper sourcing-- at least not without some indication that there is broad consensus to do so. ... Kenosis (talk) 14:16, 3 October 2008 (UTC)

I understand your point but the wording I provided was pretty clear about what constitutes synthesis. It also clearly stated that synthesis is OR, so it spells it out in such a manner that I see the word explicitly as being redundent. The additional verbiage in this section really is not required if we can simply and plainy state what synthesis is. Also, the heading says advance a position not reach a conclusion, why the chopping and changes of words? keep it consistent, it gets the point across better. The previous Using sources section establishes the importance of verifiable sources so in keeping with this, it was mentioned again in the first sentence, instead what we get now is published reliable sources, rather than a reinforcement of verifiable sources. You and I may understand the distinction, but how about viewing all of this with the eyes of someone who is figuring all of this out for the first time? The wording is simply all over the place, so it was an attempt to tidy it up, not water it down, one pedantic editor can slow things down, but discussion here is mired by the thoughts of multiple pedantic editors. Let's try and consider that some people reading these guidelines may not have the same level of apprecaition for legalistic verbiage as some here. They may also be of varying ages and with different aptitudes for such matters so how about working on giving the facts in a clear and concise format? Semitransgenic (talk) 16:45, 3 October 2008 (UTC)

No Original Logic Leads into SYNTH

I think the proposed "No Original Logic" should precede the section on SYNTH as it helps to set the stage for how SYNTH is a form of logic that goes beyond the sources. SYNTH, in other words, is a subset and specific example of people going beyond the content, logic, and conclusions of a source. With this as a preceding section, we can beging SYNTH with something like this, using the current version as a model:

Synthesis is a form logic which occurs when an editor puts together multiple sources to reach a conclusion that is not itself in any of the sources used.

Clearly the SYNTH section is under close review now, and may evolve a bit more. But I'd like to go ahead and put the NOL (no orignal logic) section in ahead of it to (1) get more community response and (2) to see if it helps us to make the section on SYNTH more concise. Clearly the "A and B, therefore C" example is great, but by preceding it with both the NOL section and SYNTH as a specific example of NOL may help a lot of editors as they discuss the use of sources.--SaraNoon (talk) 15:40, 3 October 2008 (UTC)

I personally would love to see that happen. The usual method of introducing something new on a policy page might be to put it here until a consensus has been reached, but I have found that to be ineffective most of the time . Putting the section into place really allows us to see it, and read it as it would appear on the actual policy page. If it doesn't work editors could agree to remove it. The other possibility is to create a sandbox, and to work on this part of the policy there. Thanks Sara. I feel that we are close to have something really comprehensive and well written in this area of the policy.(olive (talk) 16:01, 3 October 2008 (UTC))

No, it shouldn't go on a policy page and I note a number of editors above have expressed this view too. To quote:

Article pages are reserved solely to "parroting" the logic of experts as reported in verifiable sources.

This isn't grammatical. We don't parrot, putting the word in scare-quotes doesn't help. An encyclopaedia contains statements of fact, not the logic used to ascertain them. Not everything on WP needs to have expert sources (popular culture == 99% of WP). The term "verifiable sources" doesn't make sense. Colin°^Talk 16:44, 3 October 2008 (UTC)

If in reporting what a source says you make any inference or draw any deductions, you have crossed the line into engaging in original research.

This is already handled by SYNTH. Colin°^Talk 16:44, 3 October 2008 (UTC)

You must be especially aware of this when seeking to contribute to any article that may be controverisial and involves a variety of opinions. While you may "get away" with a bit of "simple logic" when contributing to a noncontroversial (and little watched) article, the barrier against original research is much higher in regard to even modestly controversial topics.

Policy affects all articles. We don't set out to document how unloved articles "get away with it". Colin°^Talk 16:44, 3 October 2008 (UTC)

This is one of the keys to avoiding original research: make sure that the logic you report reflects the same logical path and interpretation of the authors you cite.

Don't highlight bits of policy as "one of the keys". Just state the guideline. I'm afraid I don't follow what this is trying to prevent. Can you give an example where an article "reports logic" (both correctly and incorrectly in your opinion)? Colin°^Talk 16:44, 3 October 2008 (UTC)

That does not mean that the logic of the authors cited by editors must be impeccable, or even defensible. A notable, verifiable source may in fact be riddled with any number of logical fallacies. But if the opinions are verifiable and relevant to the topic, they may be included, and in some cases may be important to include.

This is an issue for WP:V and WP:NPOV and does not belong here.

For the purpose of both avoiding original research, spotting original research of others, and discussing content with other editors, it is highly recommended that you become familiar with the many forms of logical fallacies. This will help you be a better editor and a better collaborator with other editors.

This isn't relevant to WP:NOR. Editors aren't allowed to make unsourced deductions, whether logically correct or fallacious. Colin°^Talk 16:44, 3 October 2008 (UTC)

I have to agree, it's really just more clever bullshit. Semitransgenic (talk) 16:56, 3 October 2008 (UTC)

Clearly, I and others disagree with Colin and Semitrangenic. As you note yourself, "Editors aren't allowed to make unsourced deductions, whether logically correct or fallacious." That is precisely the point that NOL articulates in a manner underscoring that even what a person considers to be an "obvious piece of deduction" is not allowed, and if called to the mat on it in a controversial article (where such things will no doubt raise cries of OR) he or she will need to back down. Moreover, while original logic is not allowed in articles, we know people are striving to make logical arguments on talk pages for how to interpret and represent sources, and that is why I refer people to look at logical fallacies so they better guage their own and each others arguments over what a source "really" says, which is very relevent to NOR

Most notably, the present policy, and even the current argument over SYNTH, does not actually address lines of logical argument at all. Do a search on the page for logic, and you won't even find it. You say that the statement "If in reporting what a source says you make any inference or draw any deductions, you have crossed the line into engaging in original research" is already covered by SYNTH, but not explicitly, only implicitly. Deductions and inferences are different than facts and conclusions.

Instead, it is heavy on sticking to the "conclusions" of sources, which is fine. But part of the struggle over whether a source refers to, directly refers, explicitly refers, etc to the topic lies in the fact that the reference may not be logically connected by the author of the source to the point being made in the article.

By pointing out that sources have a logic underlying their conclusions, we are reminding editors that if the conclusions of sources are presented in a fashion that presents a logical argument which is not the same as the logical argument used in the source, then that is OR. As MartinPhi noted, it reminds editors that "You can't paraphrase without following the logic, however, so you have to have logic to write a sourced encyclopedia," and that line of logic should flow from the source, not the editor.

Colin, I surely encourage you to make suggestions to improve this, but there is clearly a significant number of editors who have weighed in with strong support, and surely enough that it justifies putting it into the policy for at least a 24 hour period, for example, to elicit more input from the community. While a few have already argued that it is unnecessary, I have seen no arguments that it would actually be harmful to the project, which suggests to me that it would be "safe" to add it to poicy for at least a short time to get more community input.

I'll wait a while longer for any more constructive suggestions, but judging by the comments of support from olive, Martinphi, Vassyana, Blueboar, Fullstop, Ludwig, and Jayen that something of this sort would be beneficial either in policy or at least in an essay associated with this policy, I plan to put it into policy and will ask other editors to at least "tolerate" it for 24 hours in order to invite a more vibrant discussion from the community regarding whether it would be beneficial to keep it in policy, put it in a guideline or essay, or to just bury it in the archives.--SaraNoon (talk) 02:10, 4 October 2008 (UTC)

essay yes, policy no. It's a case of KISS. Semitransgenic (talk) 02:35, 4 October 2008 (UTC)

I'd like to see it in the policy for 24 hours. This is an approach I haven't seem before and I'd like to see how this addition reads, and what happens with method of introducing new material. I can't see how we could lose anything. As with any material in the policy I would think edits could be made.(olive (talk) 02:46, 4 October 2008 (UTC))

Sorry, Olive and Sara, but that's not how it works. And Vassyana wasn't supportive of including this in the policy. We don't have many policy pages, and generally editors aren't keen on making them longer. You'll get more support by tightening what is there than expanding it with new material. If you feel WP really needs a new section in its policy, then propose it properly at the Village Pump and elsewhere. Read about how policy and guidelines are formed. Make it clear why the new material is needed, what makes WP dysfunctional without it, what arguments would be handled better with it, etc. Be prepared to revise the text many times or to find that not enough people think the addition is required. Too many people read NOR every day for you to experiment with adding what is a rough draft of a section. Colin°^Talk 09:42, 4 October 2008 (UTC)

Explicit logic is only one of the methods our sources use to arrive at their conclusions. Other methods include mathematical proof, statistical analysis, personal experience, committee discussion, gut feeling, and inspiration. All of these might be misrepresented by editors, or might be flawed in some people's opinion. NOR is a fairly simple concept and applies to any research method whether that includes formal logic or not. Colin°^Talk 09:42, 4 October 2008 (UTC)

You're partly right. That's not how it usually works. My thought was to give it a try. There's nothing to lose. We could also create a sand box. Rather than looking at this as new material I see it as a better version of what we have. We won't be able to tell anything unless we can see it in place, either here or in a sandbox.(olive (talk) 18:07, 4 October 2008 (UTC))

Use a sandbox if you like. Plus, you need to invite more folk. Not everyone has this page on their watchlists, but lots of folk would be concerned if text was significantly removed or added without their knowledge. Colin°^Talk 20:47, 4 October 2008 (UTC)

Community input-That's a given.I believe Sara said she intended to inform the community, and to ask for input.(olive (talk) 21:36, 4 October 2008 (UTC))

Logic

I've inserted a new thread here for ease of reading and discussion...

As per Colin's comments on Logic. I think that you are using logic in a different way than is being used in Sara's text. I believe that Sara's text refers specifically to OR and the kind of synthesis of material that occurs when "jumps" in logic are made, so that material that is not connected within the sources gains connection by the act of the editor making the connection . This combination of information creates conclusions, some of it implied, some not, but in either case new material . Because the material is new or original it is not compliant in an encyclopedia. You, I think are talking about the actual kinds of logic/analysis the source itself may display, and this is of course permissable to relate in an article if it is sourced and if it complies with other Wikipedia standards.

Sara's text describes in a sense what goes on behind OR and synthesis... why these are not only poor for an encyclopedia that doesn't publish new thought, but is often flawed in terms of the conclusions reached. The point, ultimately, isn't just whether the conclusion is correct, or not, although inaccuracy and speculative reasoning are a by product of this kind of research, but rather that because there is this gap in the logical unfolding of the information, that can only be "filled" by the editor, the material cannot be considered a neutral point of view, and risks being erroneous in the bargain. Its poor research and is non-compliant here.

Explaining and understanding, the "behind" workings of OR and synthesis in some cases could extremely useful to editors, in my experience, especially in heated discussion. That's why I'd like to see us seriously consider this text and find a way to use it to its best advantage.(olive (talk) 19:27, 4 October 2008 (UTC))

No understanding of logic is required to follow WP editorial policy. The threshold question for WP:NOR is whether content provided by WP users reflects the sources put forward in support of that content. Among other things, this is why when a primary source is used, the content must be able to be compared to the source by anyone with a general education, and that source should be able to be readily discerned to "say" the same thing as the WP content at issue, even though the content may be expressed in the editor's own words. No knowledge of logic or logical fallacies required, only an understanding of something very basic like the general principle of A+B=C that is given as an example of synthesis. A user can't take water and CO² and say it makes club soda, unless there's a reliable source for that conclusion. A user can't say that combining turmeric, coriander and cumin makes curry powder, nor that sulphur, charcoal and saltpeter make gunpowder, unless there are reliable sources for such syntheses. ... Kenosis (talk) 20:39, 4 October 2008 (UTC)

I would agree in principle. If you've been in a discussion on material that becomes contentious, you might have found editors sometimes do not understand the rationale behind the A+B=C, and will argue in circles endlessly refusing to budge on an OR or synthesis point. An explanation might be useful. I didn't see this as difficult to understand.(olive (talk) 21:43, 4 October 2008 (UTC))

Sure, an essay page on this would appear to be perfectly appropriate. Perhaps a review of the archives will point up some examples of common sources of confusion? ... Kenosis (talk) 22:04, 4 October 2008 (UTC)

Everyone uses logic, whether formal or informal. Indeed, formal logic is just an attempt to understand how we arrive at logical conclusions, and in the process, helps us to identify when our conclusions are not based on logically consistent thoughts. Pointing out to editors that the sources they use employ some logic in stating their conclusions is not a startling revelation but it does underscore that any summary of material presented in a source must resplect and reflect the logic of the source...not the editor contributing the material.

I strongly feel this concept should be noted in policy and believe that we will find that it will eventually prove to be a more solid basis for editor discussions than some of the arguments over sources, ie whether the source "refers" or "directly refers" to the subject or is a primary or secondary source. NOL is a more fundamental, and if an edit can't past the NOL test than there is no need to move on to PSTS questions, for example.--SaraNoon (talk) 18:27, 6 October 2008 (UTC)

I'll repeat my question: Can you give an example where an article "reports logic" (both correctly and incorrectly in your opinion)? In addition, can you give an example of source -> text (made up, if you want) that breaks your "NOL test". Lastly, none of my criticisms of the proposed section text have been addressed. Instead, Sara and olive seem to think saying "this is important, it should be policy" often enough will make it happen. People will always "argue in circles endlessly refusing to budge", and adding to policy only ensures they have more to argue about! Colin°^Talk 19:06, 6 October 2008 (UTC)

• Okay, here is a made-up example of "Original Logic". Imagine the following edit to the lead sentence of Blackbird:

The Blackbird, Common Blackbird or Eurasian Blackbird (Turdus merula) is a species of true thrush dinosaur<ref>Source stating that all birds are theropod dinosaurs</ref> which breeds in Europe, Asia, and North Africa, and has been introduced to Australia and New Zealand.

• The editor's source is correct – birds are indeed classified as dinosaurs today – but it does not directly refer to the blackbird. The logical progression from "birds are dinosaurs, blackbirds are birds, hence blackbirds are dinosaurs, therefore this article should call the blackbird a dinosaur" is the editor's logic, and one that is not borne out by reliable sources on blackbirds, which do not refer to them as dinosaurs. Jayen466 19:42, 6 October 2008 (UTC)
• P.S. Note that the editor could and would claim that his source on the dinosaur nature of birds is "directly related" to the blackbird, since blackbirds are birds. But the problem is, his source does not "directly refer" to blackbirds. Jayen466 19:52, 6 October 2008 (UTC)

That example can be defeated on a number of fronts, without the suggested NOL section:

1. The statement "birds are theropod dinosaurs" is a (on its own) misleading oversimplification of a complex problem in classification.
2. Wikipedia's animal articles use a particular classification system, grouping by living categories rather than evolutionary ones. (waffling a bit here, but you get the point -- there's more than one way to classify things).
3. The convention is "species of <genus>" so the statement is as bizarre as "a species of vertebrate".
4. "Chickens are dinosaurs"[5] (pretend it is a reliable source). Just because some palaeontologist says something doesn't make it useful to repeat in an encyclopaedia. Common sense applies.

Let me clarify my second question above: an example that requires the addition of the proposed NOL section. And clarify which sentence in that section outlaws it. Colin°^Talk 21:38, 6 October 2008 (UTC)

Well I can't speak for Sara , but my desire to engage in discussion on this, like every other editor here means that I see this as important and worth talking about, in the same way you don't. Its acceptable if not desirable for opinions to be discussed. That is the nature of collaboration.

Logic is part of a "thinking" process, and is not a thing. There is importance in understanding how a conclusion is reached. If the process is flawed logically, then the conclusion may well be flawed. With logic, the mind moves sequentially from one step to the next in terms of trying to understand something. It builds one point onto the next until the conclusion has been reached. If it makes jumps and misses a step then chances are the conclusion will be incorrect. An article doesn't report logic as a thing. An article can indicate, however, if studied, if the writer moved from one step to the next in a sequential way so that no steps were missed, no jumps made. Synthesis as defined by Wikipedia is a situation in which the sequences for placing one sourced fact after another misses a step and the editor invents as it where a step to combine material not referenced to one particular source. This is not sequential, and so is illogical, that is, that the editor in not seeing how these combine in an article, does the combining himself. That involvement of the editor in this aspect of the process is considered OR and can create a POV. Describing how this process works can help editors understand why something is not compliant as per Wikipedia. Logic is what happens behind the scenes as the mind processes information, and provides further information if and when there is concern about synthesis and OR. Describing this logical process is a simple tool for further understanding.(olive (talk) 19:53, 6 October 2008 (UTC))

But it isn't "being discussed". There's a lot of waffle about the importance of logic but absolutely nothing about the proposed text, which I have criticised. This is the talk page of a policy, for suggesting and discussing changes to policy text. You can't just propose some text and then spend three sections discussing the importance of the theory behind it as though that makes up for its deficiencies. Colin°^Talk 21:38, 6 October 2008 (UTC)

WP:OUTING vs. WP:COI & WP:NOR

While WP:OUTING addresses the serious issue of stalking, as currently written, it renders WP:COI and aspects of WP:NOR (specifically WP:COS) unenforceable. Any good-faith attempt to identify a user as having a COI usually necessitates some degree of real-life identification. Typically, the user name is a give-away, but otherwise only an intentional or inadvertent admission by a registered or anonymous user is usable evidence – and the problems are mostly with editors who do not want to have their COI edits exposed as such or are unfamiliar with WP:COI in the first place. I feel that the community needs to discuss whether WP:Outing trumps WP:COI and WP:COS or else needs to accommodate legitimate, good-faith enforcement of these policies and guidelines. While the issues have been raised before, there has been no resolution, and that lack of resolution is hampering the work of enforcing WP:COI. If you have an interest in helping resolve this problem, please comment at WP:OUTING vs. WP:COI & WP:NOR. Thank you, Askari Mark (Talk) 15:47, 3 October 2008 (UTC)

Sticking closely to sources

The Smith/Jones example of synthesis misrepresents the Harvard "Writing with sources" manual; i.e. it implies that this manual says things it does not actually say. These are:

• "...the practice recommended in the Harvard Writing with Sources manual, which requires citation of the source actually consulted". The requirement to "cite the source actually consulted" is "Smith's" interpretation of the Chicago Manual of Style, not the Harvard Writing with Sources Manual.
• "plagiarism is defined as using a source's information, ideas, words, or structure without citing them.".. The origin of this is a quotation from "Smith's" writings (Norman Finkelstein [6]); here he appears to be paraphrasing the Harvard Writing with Sources manual, although it is not actually mentioned.

This also means that these quotations were not the Wikipedia editor's invention; instead, they were giving the point of view of Smith without attribution (I would see this as more a violation of NPOV than NOR; the article should have made it clear that these statements were Smith's point of view).

Confusingly, the explanation to the example given on the page describes things very differently. What actually happened is that an editor was quoting "Smith's" point of view without proper attribution. But the explanation describes it as the editor doing original research to support "Jones's" point of view.

Given that the whole point of this page is about sticking closely to sources, I suggest that this example is changed so as to properly represent what the sources say; at the moment, it misrepresents both the Harvard manual and the original case covered by the Wikipedia article.

Enchanter (talk) 17:48, 3 October 2008 (UTC)

I'm not familiar with the original WP disagreement referred to here. But it's supposed to be an object lesson in original research/synthesis, not sourcing per se which is more the province of WP:V. TBH, I think the example is fairly confusing in that it includes a discussion of both plagiarism and synthesis. As to where "Smith" (said to be Finkelstein) got the material from, it's plainly the Harvard manual. Either way, I'm not at all sure it's the best possible illustration of original synthesis. ... Kenosis (talk) 20:45, 3 October 2008 (UTC)

I always thought the example served its purpose well. The article on the dispute described is here: Dershowitz–Finkelstein affair. Jayen466 20:55, 3 October 2008 (UTC)

The change was made quite recently, here. The reference to "citing the source consulted" does seem to come from the Chicago Manual of Style, rather than the Harvard guide. (The Finkelstein document does not paraphrase the Harvard manual's definition of plagiarism, it quotes it verbatim.) Jayen466 22:19, 3 October 2008 (UTC)

Darn right it's a [nearly] exact quote of the Harvard manual. Which is why I changed it, in response to another user having pointed out the discrepancy on this page. As to the use of the words "cit[ing] the source consulted", they do seem to come out of the CMS, though I'm very much at a loss to understand why it might be a big deal either way. If it's that disorienting, align the wording with the Harvard manual or the Chicago manual, and copyedit the given example accordingly. ... Kenosis (talk) 23:38, 3 October 2008 (UTC) ... An additional note: Finkelstein ("Smith") says: "It is left to readers to decide whether Dershowitz ("Jones") committed plagiarism as defined by Harvard University -- "passing off a source's information, ideas, or words as your own by omitting to cite them." (Here's the link again.) So "Smith" never used the words ""cit[ing] the source consulted", but rather, it appears the additional language was a WP user's contribution, thus resulting in a synthesis consisting of material from both the Harvard and Chicago manuals on this project page. There might be something vaguely ironic about that in the context of the WP:SYN section of WP:NOR, except that this is the WP policy page and not a WP article. The original editing disagreement to which the example refers appears to be this, amid a number of edits leading up to as well as following the one I link to. The article text, in the section on Dershowiz' response, stated:

If Dershowitz's [Jones's] claim that he always consulted the original sources is false, this would be contrary to the practice recommended in the Chicago Manual of Style as well as Harvard's student writing manual, but neither of these sources call it "plagiarism."

Obviously this quote of the WP user's edit has since been modified several times since it was placed on the policy page. ... Kenosis (talk) 03:02, 4 October 2008 (UTC)

Follow-up: The original insertion of WP:SYN on 11 April 2006 can be found [7]. The text of the then-newly-placed section can be seen here. After basic copyediting, it read like this on 12 April 2006. The original example of text asserted to be OR, as it was first used in the section that is now WP:SYN, is:

If Jones's claim that he always consulted the original sources is false, this would be contrary to the practice recommended in the Chicago Manual of Style as well as Harvard's student writing manual, both of which require citation of the source actually consulted. Neither manual calls violations of this rule on citing original sources "plagiarism." Instead, plagiarism is defined as using a source's information, ideas, words, or structure without citing them.

... Kenosis (talk) 17:20, 4 October 2008 (UTC)

• Thanks, and well done. Jayen466 20:21, 4 October 2008 (UTC)

VisionThing's revert

How do editors here feel about this revert? My impression was that the "in direct connection to the topic" wording (in the version prior to VisionThing's revert) has had the most support here on the talk page, with editors on all sides of the debate on this page, incl. Kenosis, Lawrencekhoo and Martinphi, preferring it. Actually, I am not off-hand aware of anyone supporting the version VisionThing reverted to. Shall we revert, or do you think VisionThing has a point? --Jayen466 19:55, 3 October 2008 (UTC)

Vision Thing's version is syntactically stronger. This wording replacing "topic" with "subject" is not accurate seems, but could be easily be replaced.

"or if the sources cited are not directly related to the subject topic of the article." Better syntax while maintaining meaning seems like a good deal! (the preceding comment was posted by olive (talk) 20:23, 3 October 2008 (UTC))

Well, the downside with "directly related" in this sentence is – as has been pointed out here by various editors over the past few weeks – that things may well be "directly related" in the mind of the editor wishing to add the material, without having been related to the article topic in the mind of the source's author. Only the latter's intentions count; saying that the source itself has to make the argument "in direct connection to the article topic" takes care of that. That's why I prefer the version prior to VisionThing's revert. Cheers, Jayen466 20:29, 3 October 2008 (UTC)

I haven't kept track of every edit, but I think NOR has basically been fine for ages. I think that "refer directly to the topic of the article" is a better choice than "sources that are directly related to the topic of the article". It's just a bit stronger, as it does not allow editors to decide what is related and what isn't. I know of lots of cases where editors would like to say "well, this is related," but that is just their opinion. The source has to say it is related. Agree with Jayen here. We could say something like

If the sources cited do not explicitly reach the same conclusion in direct connection to the topic of the article, then the editor is engaged in original research. The source itself must state what relationship it has to the article topic, since otherwise connecting the source to the article's topic is itself original research. ——Martin^phi ☎ Ψ Φ—— 22:15, 4 October 2008 (UTC)

• Well, note that what the SYN lede currently states is, If the sources cited do not explicitly reach the same conclusion, or if the sources cited are not directly related to the subject of the article, then the editor is engaged in original research.
• This, to my mind, opens the loophole of allowing the editor to establish that the source is "related" to the article. Jayen466 07:45, 5 October 2008 (UTC)

Which, of course, it does, which is what we have been trying to put an end to. Jayjg ^(talk) 07:47, 5 October 2008 (UTC)

“Directly related” — Towards a viable alternative

Ludwigs pointed out a fundamental flaw with the "refer directly" wording [8], mentioning as an example of what we don't want to allow, "sources writing about different topics entirely, that happen to mention or make reference to the article topic". If "refer directly" is taken to mean "mention", it does not take into account this critical distinction. To "properly" source the information add to a WP article, one would have to cite sources that mention the topic of the WP article and that present that information in direct connection with the article topic. Another shortfall is that "refer" is used to mean several things including "1. allude to, mention, cite, speak of, bring up, invoke, hint at, touch on, make reference to, make mention of" and "2. relate to, concern, apply to, pertain to, be relevant to" [9]. So, "refer to" could be interpreted to mean "relate to" and we are back to square one.

Furthermore, "refer directly to", if used to mean "mention" would, in principle, prevent editors from adding useful and uncontroversial contextual information to an article (see examples provided above by Slrubenstein [10] and I [11]). Therefore, as long as we adhere to WP:SYN, "original research" should be defined more broadly to mean "any information that could shed new light on an article topic by making connections that are not mentioned in the literature on the topic and that might not be immediately obvious to researchers working on that topic, such that these connections could suggest new research conclusions or approaches." In my opinion, the Lawrencekhoo's "Inflation and Gold standard" example is a case in point.

In this regard, I think we will all agree that information added to an article should generally (preferably always) be based on reliable sources that present this information in direct connection with the topic defined by the article's name. Now we must determine what sort of supplementary information, from generic sources that cover a broader subject area than the specific article topic, is acceptable. In order to do so, we must address these two points:

1) We need to decide and state in the policy whether (or in which cases) information that serves to advance a position or point of view with respect to any aspect of an article (without breaching the WP:SYN rule) would be considered acceptable contextual information, if it is based on generic sources that cover a broader subject area than the specific article topic (see Kotniski's hydrogen atom example and Lawrencekhoo's "Inflation and Gold standard" example).

It was previously proposed that one should be allowed to "add a contrasting mainstream view from a generic source." According to Vassyana, "[C]ontrasting mainstream view from a generic source in an article on a fringe topic" is anything but "uncontroversial and acceptable". Looking at the alternative medicine areas of the wiki, for example, will make it clear that the assertion made in the proposed addition is quite simply untrue." [12] According to Someguy1221, "most reasonable readers won't attempt to learn physics from an article on a nutter, especially when that nutter's ideas have been properly framed as his own personal claims. The interested reader will follow the handy bluelinks provided for him to see how retarded the theory was he had just been reading. And sure enough, the harder it is to find a reliable response to a theory, the less notable it probably is, and the less likely anyone will come to the page anyway. So basically, I see no fault in dropping the fact that hydrogen has two protons, or that there are 60 seconds in a minute (even when the article's subject says 42)[...] What had been a simple explanation that the source is off topic now becomes an argument over whether a scientific fact is sufficiently established - a debate that should never happen when writing an article." [13][14] This proposal could also arguably violate the NPOV (see my previous proposal [15] and Vassyana's and Someguy1221's responses [16][17]. I am persuaded by these strong arguments that we should not allow "contrasting mainstream view" from generic sources that cover a broader subject area than the specific article topic. I do not, however, accept the argument that this proposal will lead to massive WP:SYN violations. As long as the SYN rule is clearly layout, I am not worried about this.

2) I think we would all agree that, in some cases, supplementary information from generic sources that cover a broader subject area than the specific article topic may be deemed to add value to an article, in order to clarify places, people or things mentioned in the article—when the "WP:SYN-compliant" supplementary information does not serve to advance a position or point of view with respect to an aspect of the article topic (see the following examples: [18] [19]). But, as some have pointed out, in certain articles, e.g. political articles, such supplementary information can sometimes create a certain bias or unduly cast a favorable or unfavorable light on a some aspect of the topic, even though this information does not strictly advance a position or point of view with respect to an aspect of the article topic (I'm not talking about WP:SYN here, see above). In this regard, Jayden and I gave some examples of "acceptable" subtopics (of the article topic) to which the sources could connect this information (see [20] and [21]). However, what generalizations can we make based on these specific examples that would enable us to establish guidelines applicable to all Wikipedia articles? It seems to me that the best way (probably the only way) to ensure that such biased and undue information does not make its way into articles is to state in the policy that "such information should only be introduced in an article as per consensus".

--Phenylalanine (talk) 01:01, 5 October 2008 (UTC)

Just be aware that this discussion is being echoed at WP:V, and may have an impact there. Think carefully. Blueboar (talk) 01:33, 5 October 2008 (UTC)

RE the statement above "I think we will all agree that information added to an article should generally (preferably always) be based on reliable sources that present this information in direct connection with the topic defined by the article's name. Now we must determine what sort of supplementary information, from generic sources that cover a broader subject area than the specific article topic, is acceptable" :
..... Actually, no, it's not necessary to determine in advance, on a policy page, "what sort of supplementary information..." is acceptable. Thus far in WP's existence that has been decided via the WP:Consensus process at the local article level, on a case by case basis. To attempt to decide such editorial policy en masse from a policy page would be an unprecedented centralization of a "local" editorial decision that quite reasonably plays itself out in widely varying ways across the entire wiki. ... Kenosis (talk) 13:57, 5 October 2008 (UTC)

I appreciate your opinion, but I disagree that it "quite reasonably plays itself out" most of the time. I believe we need certain guidelines on the use of supplementary information: see point 1) and 2) above. --Phenylalanine (talk) 14:22, 5 October 2008 (UTC)

I saw the points. I was specifically responding to the statement I quoted from "Point 2". This is an an attempt to standardize common editorial decisions across the wiki, decisions made by local consensus about how to write articles. IMO, even with a long runway it won't fly, at least not at this stage of the wiki's growth. ... Kenosis (talk) 14:50, 5 October 2008 (UTC)

If I understand you correctly, you're saying that we should restrict (to some extent at least) "supplementary points of views" (point 1), but not "supplementary facts" (point 2)? --Phenylalanine (talk) 15:14, 5 October 2008 (UTC)

No. I was specifically responding to Point 2. In response to Point 1, my reaction is a bit stronger. Citing to lk's and Kotniski's examples of situations they thought were relevant to WP:NOR in which there was difficulty in resolving via local consensus certain issues relating to WP:NOR, you said: "1) We need to decide and state in the policy whether (or in which cases) information that advances a position or point of view with respect to any aspect of an article (without breaching the WP:SYN rule) would be considered acceptable contextual information, if it is based on generic sources that cover a broader subject area than the specific article topic..."
..... My response is: No, we don't need to decide whether and in what cases original synthesis is permitted. It's not permitted, period. What you're referring to, I imagine, is not original synthesis, but rather standard editorial decisionmaking by consensus as to what relevant content is appropriate to include in a given article. That's a community process to be decided by consensus, as it has been to date in WP. In other words, as the WP:SYN policy already indicates, there's a difference between writing and editing an article on the one hand, and original synthesis on the other. This is left to the local consensus to decide, and in cases where there are unresolvable disagreements, WP already has other avenues already in place in order to mediate or arbitrate such disagreements. To attempt to legislate such a thing wiki-wide would, IMO, not work. IMO, it simply will not fly. The reason it won't fly is that you've defined standard WP article writing as "synthesis", and that's not what the community means by "synthesis" w.r.t. this policy page. ... Kenosis (talk) 15:31, 5 October 2008 (UTC)

There is a misunderstanding here, I use "synthesis" as it is used in the WP:SYN section. WP:SYN simply restates the verifiability requirement as it applies when you combine properly sourced information. What I'm saying has no bearing on WP:SYN. Again, I disagree that we should let editors determine what "WP:SYN-compliant" information that serves to advance a point of view is acceptable, for the reasons outlined in point 2. --Phenylalanine (talk) 15:48, 5 October 2008 (UTC)

What you're referring to as "synthesis" is standard WP article writing and editing. And, "information that advances a position or point of view" is the province of WP:NPOV, not WP:NOR or WP:V. ... Kenosis (talk) 15:56, 5 October 2008 (UTC)

What you're referring to as "synthesis" is standard WP article writing and editing. — Again, my proposal has no influence on WP:SYN. Compliance with the current WP:SYN policy is a requirement.

And, "information that advances a position or point of view" is the province of WP:NPOV, not WP:NOR or WP:V. — Not necessarily (see my previous proposal [22] and Vassyana's and Someguy1221's responses [23][24])

--Phenylalanine (talk) 16:18, 5 October 2008 (UTC)

As I've indicated before, I think your points are reasonable. I also think this proposal to specify in WP:NOR what precisely is the scope of peripheral material that can be included in WP articles won't fly, no matter how long the runway is. I don't think one can successfully redefine article writing as "synthesis". And even if we do redefine it, "original synthesis" remains distinguished from article writing. Articles are written by local consensus, within the three core content policies. That includes the decision about what peripheral material is sufficiently relevant to include in a given topic. Any attempt to specify the breadth of individual WP articles from this policy page will, in my estimation, almost certainly fail to gain acceptance by the WP community at this stage of the wiki's growth, because that's presently the domain of other policy pages. ... Kenosis (talk) 16:31, 5 October 2008 (UTC)

Thank you for your opinion, and I respectfully disagree. By the way, you have again misrepresented my position by speaking of "synthesis". My proposal in no way redefines the WP:SYN policy, just as it does not redefine the WP:VER policy, rather it requires compliance with these policies. I trust that the other editors will know the difference [25]. --Phenylalanine (talk) 17:09, 5 October 2008 (UTC)

RE the use of the word "synthesis" to describe something other than original synthesis: OK, thanks for clarifying. I drew that particular usage from my interpretation of an earlier post of yours. ... Kenosis (talk) 16:22, 6 October 2008 (UTC)

• Just a quick note on this: Ludwigs pointed out a fundamental flaw with the "refer directly" wording [26], mentioning as an example of what we don't want to allow, "sources writing about different topics entirely, that happen to mention or make reference to the article topic". I believe Ludwigs went too far here. Examples:
• I think it is alright to cite an article summarising this year's Oscars, which happens to mention that Director X was nominated in the best foreign-language film category, in our article on Director X, as published evidence that he received such a nomination.
• If a scholarly, historical work on Göttingen University mentions that Georg Christoph Lichtenberg became a professor there on such and such a date, then I think it is fine to cite that in the article on Georg Christoph Lichtenberg, even if the cited work is mainly about Göttingen University and only mentions Georg Christoph Lichtenberg in passing.
• If notable minor author A exchanged letters with famous poet B about a literary controversy that they were both involved in, and both halves of this correspondence have been published in "The Collected Letters of B", then facts mentioned in this correspondence that are relevant to the article on A can be sourced to "The Collected Letters of B" in the article on A, even though correspondence with A only represents a minor part of "The Collected Letters of B".
• If an academic study surveying a whole range of writers working in a specific field has a chapter – or just a page – on writer X, this is a good source for our article on writer X, even though the source is not mainly on writer X. We have to bear in mind that in many cases, minor writers may not have entire works devoted to them, although they may still feature in various "round-ups". In such cases, these round-ups are all we have available to us.
• What is common to all the examples I give above is that the sources do make direct reference to the article topic. Demanding that sources be mainly about the article topic may be fine with people such as Goethe or Shakespeare, where there is a wealth of dedicated literature, but it would be too restrictive for one-hit wonders, less well-known actors, minor writers etc.
• As far as I am concerned, "directly refer to" might as well be "directly mention". Jayen466 12:30, 5 October 2008 (UTC)

Jayen466, I don't think Ludwigs was suggesting that the sources be mainly about the article topic. The idea is to cite sources that make a connection between the article topic and the information added to the article. If a source mentions the article topic, but does not present this information in direct relation to the article topic... you see what I'm saying. For example, take an article on the benefits of a "nutrient X" and suppose that the article mentions, en passant, that "nutritional program Y" (the topic of the WP article) is one of the most popular diets today. Even if "nutritional program Y", in fact, happens to have lots of "nutrient X", if the source does not make a clear connection between "nutritional program Y" (WP article topic) and "nutrient X" (for example, suggesting that the dietary program provides lots of this nutrient), you should not (at least in general) make a connection between "nutrient X" and "nutritional program Y" is the WP article. This does not require that the editor divine the intent of the source author, it requires simple editorial judgment. --Phenylalanine (talk) 14:03, 5 October 2008 (UTC)

• Still no We will often need sources for miscellaneous or background facts needed to be documented as part of the article, and they will need to be taken from whatever is the most reliable and accessible and appropriate source. This may well be a general work. If an article on an author should mention a disease, and we need to document a basic fact about the disease itself, it should be the best source on the disease, not a source on the author & the disease necessarily. Just an illustration, but i continue to wonder exactly what evils this proposal is meant to address. For the examples above, if the letters between A & B have been published in a book about C, as does happen, then that's the source; if as more likely they've been published in a book about the town one of them comes from, then that's the source, even if the book itself simply reprints them for its own purposes, and does not specifically have much text related to them. And, if an article saying D took a course at some university, and says the course had just been introduced that year, that fact can be documented from a source that never even mentions him. There is no appropriate fixed rule, except to use the best sources available. [[User:|DGG]] (talk) 03:00, 6 October 2008 (UTC)

If you had actually read my proposal, instead of responding to the title, you would realize that we are in agreement. --Phenylalanine (talk) 10:36, 6 October 2008 (UTC)

I apologize for my rough reply. I will clarify my position in another posting. Cheers, Phenylalanine (talk) 03:53, 7 October 2008 (UTC)

DGG, I have nothing but the deepest respect for you, but I could hardly disagree more. For example, if reliable sources about the author feel no need to mention some "basic fact" about the disease, we should not do so. We already have a way to provide further in-depth and contextual information about various passing topics and concepts mentioned in an article in the form of blue links. Beyond that, I strongly feel that letting editors decide what is pertinent, relevant and/or related in the absence of clear support from reliable sources is counter to the fundamental principles of Wikipedia. We're not even supposed to include information only covered by a small minority within relevant sources, let alone information not covered at all by topical sources. Relevant to this policy, I honestly do not understand how editors drawing connections or deciding that certain information is pertinent, relevant and/or related in the absence of clear support from reliable sources can be claimed to not be original research. For better and for worse, we should be simply reporting what the best available sources say about the topic we are covering, no more or less.

Opening to door to allowing editors to draw their own connections and fill in their own "background" bits is not only (in my opinion) against the very fundamentals of Wikipedia, it is the path to sheer madness. Eastern European articles will become even more mired in soapboxing as editors compete to detail past aggressions, international law, treaties and other "contextual" information. Fringe articles will become even more contentious wavering between the extremes of Debunkopedia and Fringepedia, as people will struggle to include (respectively) "contextual" counterpoints and "related" historical detail. Religious articles will become bogged down with "pertinent" cult analysis and "supporting" scriptural commentaries. Those are just a few blatantly obvious examples that already struggle with the comparatively small amount of wiggle room already allowed. There are several other similarly broad areas of the wiki that also have some similar difficulties, if on a lessor scale of disruption. Opening the kind of door that you and others would advocate not only runs counter to the very fundamentals of Wikipedia, it would lead to whole new levels of disruption and edit warring when we can barely manage (and inadequately correct) the issues already endemic to the project. It would additionally bring areas that currently only simmer to a full boil by giving editors a tool to wield in pushing their views, drastically expanding the scope of the project affected by soapboxing and extreme disruption. While I understand the position that we should formally allow sensible contextual and/or historical information, such a position will do far more harm than good to the project. I cannot emphasize enough that opening the door for editors to decide what goes in an article in the absence of topical supporting reliable sources is a Pandora's box that we should never open. Vassyana (talk) 05:50, 6 October 2008 (UTC)

Have to agree with DGG that this appears to be a blunt instrument for solving specific issues editors here may have encountered, whatever they are. It will break too much perfectly reasonable sourcing on uncontentious but minor facts. The issues Vassyana raises are all handled by WP:WEIGHT and are nothing to do with NPOV. As Somedumbyankee says below, there's a danger in trying to make one policy outlaw all evils. Colin°^Talk 08:32, 6 October 2008 (UTC)

WP:WEIGHT only covers how to handle viewpoints already present in reliable sources about the topic. WP:NOR seems like the appropriate place by definition to address information not covered by reliable sources in relation to the topic. Vassyana (talk) 12:39, 6 October 2008 (UTC)

(responding to Vassyana and Phenylalanine above). This discussion is far from clear, it is very difficult to know what is being proposed and perhaps editors are arguing at cross purposes. If the opening statement [we don't want to allow, "sources writing about different topics entirely, that happen to mention or make reference to the article topic".] is not what you are talking about, could you strike it. Colin°^Talk 16:17, 6 October 2008 (UTC)

I agree that the opening paragraphs may be confusing, sorry about that. I will post a clearer summarized version of my proposal and rationale below. Cheers, Phenylalanine (talk) 03:43, 7 October 2008 (UTC)

As I argued in the previous discussions above, the problem is not with the sources, but with how they are used. Sources that mention the article topic in passing can be used inappropriately, and more general sources may need to be used for a valid reason. We have to tackle this problem by beefing up a policy like WP:TOPIC, rather than a strict rule like 'a source must mention the name of the topic somewhere'. LK (talk) 03:24, 6 October 2008 (UTC)

We can't have WP:NOR trying to be all policies and the final decision on inclusion or exclusion. A lot of the "slippery slope" of information that is not directly related is covered more by WP:POINT, WP:COATRACK, or the universal warrior-whacker, WP:NPOV. Straitjacketing editors into not including relevant context because their source does not explicitly mention it will make technical articles completely useless, since the most authoritative (i.e. "best") sources rarely include the context, they assume that the reader knows it. Tendentious editing can be addressed with many current policies and guidelines, and turning WP:SYNTH into a rule that will be more honored in the breach than the observance will not stop it. In short, the editor's job is to insert relevant context, otherwise people may as well read the primary source instead of wikipedia. If that context is not neutral, we have a far more authoritative policy than SYNTH that we can use to remove it. SDY (talk) 06:43, 6 October 2008 (UTC)

No one, even those editors supporting the most restrictive interpretations, are trying to have this policy "be all policies and the final decision on inclusion or exclusion". We're only discussing the limited issue of inserting information not covered by reliable sources in relation to the article topic. I have provided an example of how technical topics can be handled through forking article structure here. We already have the means to provide the context and background not covered by topical reliable sources without running counter to the fundamentals of Wikipedia in the form of wikilinks and article hierarchy. Vassyana (talk) 12:39, 6 October 2008 (UTC)

I think part of the problem, then, is that the revised language is becoming arcane to the point where the words alone do not clearly state the policy. This is considered harmful. Forced balkanization of information to the point where it takes seventeen articles to explain a single scientific topic doesn't make a lot of sense to me. This changed policy essentially forces us into using tertiary sources (i.e. no original determinations of appropriate context) rather than secondary sources to write articles, which doesn't seem right to me. SDY (talk) 16:02, 6 October 2008 (UTC)

I don't follow you here. The way I read it, WP:SYN – even if it did have the "refer directly" wording advocated by Jayjg, myself and others – states that to support conclusions about the article topic, you can use any source that refers directly to the topic and comes to the same conclusion about it. Why do you think that would exclude secondary (or primary, for that matter) sources? Or are you referring to something different altogether? Jayen466 16:41, 6 October 2008 (UTC)

I wholly endorse the idea of "no original logic" but the "refer directly" language seems to be very subjective. The "refer directly" wording to me seems to indicate, in the literal interpretation, that I couldn't use a source that talks about food-borne illness in general when writing an article specifically about salmonellosis (obviously related), and I don't think that was the intent. Are most non-direct conclusions about a topic original research? I would hold that most are and would strongly endorse an essay covering "refer directly" but don't think that it is such an ironclad association that it should be part of a core policy. SDY (talk) 22:25, 6 October 2008 (UTC)

Well, would you endorse "directly refer" or "reach the same conclusion in direct connection to the topic of the article" in the SYN section's lede, if we keep "directly related" elsewhere? That would prevent you from drawing conclusions like, "A: Foodborne illnesses are rarer in France than in the US", "B: Salmonellosis is a foodborne illness", ergo "C: Salmonellosis is rarer in France than in the US" and putting C in the article, but it would still allow you to mention that foodborne illnesses generally are rarer in France. Jayen466 22:45, 6 October 2008 (UTC)

If "directly related" semantically equals "unquestionably related" I have no problem with it. I just don't want to see an arbitrary "magic words" test for use of sources. Given your example, if the source stated that "all foodborne illnesses are rarer in France than in the US" and salmonellosis is a foodborne illness, then making the step that salmonellosis is rarer in France than the US is not original research. SDY (talk) 23:49, 6 October 2008 (UTC)

<outdent>

• Call me old-fashioned, but from the point of view of WP:Verifiability, I would like to see such a statement, in an encyclopedia, tied to a source that specifically says, "Salmonellosis is rarer in France than the US", and actually gave some figures. Because in practice, conclusions of this type have an unhappy habit of going awry every so often, simply due to non-expert editors' good-faith misunderstandings or mistaken assumptions. Such errors lead the public to doubt the reliability of Wikipedia as a reference source.
• As for the "magic words" test, do you actually agree with the Smith & Jones example? Because this already says a reliable source is needed that specifically comments on the Smith and Jones dispute. So the "magic words" test has been policy for 2½ years, and has informed discussions on thousands of talk pages.
• Last December, the SYN section still looked essentially unchanged. Then a lede paragraph was inserted in the SYN section on December 29, 2007, which stated that in order to avoid OR, sources must be directly related to the topic of the article.
• The intent of this lede was not to change the Smith & Jones rule, but simply to have a smoother intro to it. The didactic intent of the Smith & Jones case is clear and has always been the same: it requires a source that specifically comments on the Smith and Jones dispute (and that means nothing but a direct reference to the Smith & Jones case, or presenting the argument in direct connection with the Smith & Jones case).
• If I understand you correctly, it seems to me you ought to argue in favour of the whole SYN section being scrapped. Because as I understand your reasoning, you would equally argue that since Dershowitz was a Harvard scholar, the Harvard definition of plagiarism was directly related to the Dershowitz–Finkelstein affair, and therefore the inclusion of this material in no way constituted Original Research, but merely provided appropriate context. If I got you wrong there, please explain why the Smith & Jones case is different. Jayen466 10:50, 7 October 2008 (UTC)

I would also "like" to see it directly from an explicit source, but that's very different from demanding that it be from an explicit source. Ultimately, I believe that policy should place a certain limited trust in the authors to make intelligent decisions. This policy essentially "assumes stupidity" on the part of people writing articles, which doesn't seem productive. Saying that articles cannot make the most basic of logical conclusions (in this case, a simple syllogism) is akin to saying that unit conversions are original research. Are there theoretical possibilities where it could go awry? Surely. Would it be better to have a source that talks about the topic specifically rather than in general? Definitely. Should the editor who inserts such a statement be ready to defend it? You betcha. The Smith & Jones case is different in that a specific definition of plagiarism is used, and there is no indication that that specific definition applies. S&J also fails WP:NPOV in that the article is taking sides in the dispute, and no "direct reference" rule is needed to show that it is inappropriate.

Let's say, hypothetically, that Mr. Jones commits a crime and someone writes a wikipedia article about it. It meets notability requirements, but doesn't receive an exhaustive legal analysis in the press because Mr. Jones is WP:NPF. The available sources state that the case is tried in Federal court instead of in the State of New York because of sentencing rules but gives no further information. Would it be wholly inappropriate to provide (and cite) the difference in the sentencing rules? I would find it, as a reader, to be important and useful context, but as I'm understanding the "direct reference" rule this would not be allowed. SDY (talk) 14:03, 7 October 2008 (UTC)

I do understand where you're coming from. But how is what you propose different from "helpfully" listing what the Harvard manual has to say about plagiarism – given that Dershowitz is a Harvard scholar and thus "evidently" bound by Harvard rules? I think we are better off putting our faith in well-targeted wikilinks, rather than duplicating such factual information in each article where it might be useful. Jayen466 19:56, 7 October 2008 (UTC)

• I don't see why we should rewrite WP:NOR to explicitly allow something which it does not forbid. I think we are all agreed that the Smith & Jones example is a valid example of problematic behaviour that policy should discourage. The policy does that, and we should leave it at that. Jayen466 09:28, 6 October 2008 (UTC)

I am relieved to see Vassyana's points since they bring, for me at least, a simple clarity to a long, involved discussion. Bringing any kind of information into an article that is not explicitly referenced in a source, is a crack in the door of this policy that not only could, but will allow huge amounts of misuse to enter the writing of the encyclopedia. Already with the policy worded to prevent this kind of problem, arguments and attempts to introduce OR and synthesis into contentious articles is occurring. I can't imagine what would happen if the policy was loosened in any way,and I personally can't support any efforts, for whatever reason, to do so. I'm rigid about this because I've already seen what is possible, and it isn't pretty.(olive (talk) 15:38, 6 October 2008 (UTC))

I second that, Olive. Jayen466 16:41, 6 October 2008 (UTC)

Nobody is proposing that we do not require specific referencing to sources. The question is in what way to limit our sources. I (and phenylalanine and Colin) want to keep the [present restrict, to RSs that are pertinent to the matter at issue. You and Vassyana for some reason want to limit it further, to the sources that deal specifically and directly with the general subject of the article, and reject those that deal primarily with the matter at issue in a broader contest. I think that's laying the way to defeat NPOV, be preventing what may well be the best sources from being used in any particular circumstance. If we';re discuss a battle, and need to get the name of a place correct, we'd be prevented fro musing the best sources for that, and limited to the one or two that happen to describe it in the specific context of that battle. If we need to refer in a bio to a political scandal, and there';s an article dealing with the scandal & giving the best background, but not mentioning the individuals--and multiple sound unquestionable sources for the person's involvement that are less clear on the background, we'd have t use the unclear ones. If we are discussing who a movie star performed with, and we have sources for him in a movie and sources for someone else in the movie, we couldnt say they both appeared in the same movie--indeed, we could even mention the source that there was someone else in the movie--or any source at all about the movie-- unless it specified him as well. In practice, this is going to promote pseudo science and pseudo medicine and miscellaneous absurdities. If an article is about treatment X for a disease, and we had excellent sources for X not working, we couldn't mention the real treatment, unless an article happened to say so at the same time. Indeed if the sources for X being a treatment, good or bad, for Cirrhosis did not mention that this was a liver disease, we couldn't say so. I can not think Vassyana intended to do this, but such would be the effect. I await a clear statement of what harmful use of non-obvious synthesis the proposed restriction would prevent that we cannot prevent otherwise? By experience, those who wish to push a particular POV in an article, normally try to do so by trying to find some reason for removing sources not to their liking. DGG (talk) 20:19, 6 October 2008 (UTC)

DGG, I appreciate your examples, and have been mentally chewing on similar issues. But it seems to me we're at risk of forgetting in this discussion that SYN is not about using sources in an article, but about drawing conclusions about the article topic from such sources. Taking your examples, by correcting the spelling of a place where a battle took place (if that is what you meant), would we be using the source giving the correct spelling to reach a conclusion about the battle? No. If we insert a one- or two-sentence summary of a scandal that the article subject was, according to multiple RS, involved in, is that reaching a conclusion about the subject? No (at least not if the description of the scandal is neutral). Likewise, if we have multiple sources saying that treatment X is not working, and have reflected these in the article, we are not reaching a conclusion about treatment X by adding, Actually, what does cure this disease is antibiotics. I think the only example that would cause a problem is the actor/co-star one – that the article subject "co-starred" with actress X would be a conclusion reached by using a source that does not mention the article subject – but I daresay sources could easily be found in such a case that would list the entire cast.

On the whole, I am in favour of retaining the "directly related" wording in the policy lede (i.e. the lede at the very top of the policy page), to account for examples of the type that you give, but isn't there merit in the argument that the SYN lede (following the "Synthesis of published material which advances a position" subheading) should state that conclusions about the article topic should be based on sources that directly refer to the article topic, or, alternatively, reach these same conclusions in direct connection with the article topic? I find "directly related" too elastic there.

As for examples of harmful synthesis, Jayjg gave one here, in the earlier part of this discussion. Jayen466 21:11, 6 October 2008 (UTC)

I think the above does clarify things for me but am rather concerned about the focus on "article topic" and even more concerned about "topic defined by the article's name" that Phenylalanine suggests below. This assumes that articles focus on one topic the whole way through and never describe conclusions about related aspects. WP:NOR needs to be applicable and work for a section or paragraph within an article, and handle the idea that the topic focus has shifted temporarily. So I suggest the proposed sentences try to drop "article" and certainly "article title". As an example, WP:NOR should not be affected if a section is excised and moved to a daughter article per Summary Style, or if the reverse happens (perhaps as a result of an AFD saying the stub article should be incorporated into parent X). So can we phrase something about "drawing conclusions about topic X" coming from sources that "directly refer to topic X" without X necessarily being the article topic or the article title. Colin°^Talk 11:08, 7 October 2008 (UTC)

Thanks, I took note of your suggestion and modified my proposal accordingly below. I don't see the relevance of WP:SYN to our discussion (see proposal below). Also, in my opinion, "directly refer" is not a viable alternative as I explained above. --Phenylalanine (talk) 12:32, 7 October 2008 (UTC)

The example mentioned above does not convince me. It's one fairly routine case of clearly inaccurate summarising on a political article, and easily enough dealt with. Changing fundamental policies for something as trivial as that does not make sense. DGG (talk) 01:51, 9 October 2008 (UTC)

Proposal and rationale revised

I don't see the relevance of WP:SYN to our discussion. One violates WP:SYN when one "synthesizes" two sources in such a way that neither source properly verifies the new statement that is made (one draws a new conclusion not present in the individual sources). Whether the sources DO or DO NOT refer/relate to the article topic makes absolutely no difference. Also, in my opinion, "directly refer" is not a viable alternative as I explained above.

A better alternative would be: "cite a reliable source that presents the information in direct connection with the article topic" (that's what I would call an "on-topic source"). In this regard, Jayden and I gave some examples of "on-topic sources" (see [27] and [28]). However, what generalizations can we make based on these specific examples that would enable us to establish a guideline applicable to all Wikipedia articles which would define more precisely an "on-topic source"? It seems to me that the best approach would be to state in the policy that a source presenting the information in direct connection with a subject that does not strictly correspond to the topic defined by the article title should only be used in an article as per consensus".

This is a general guideline, of course, and I believe that we must allow for the occasional use of valuable and unobjectionable contextual information from "off-topic sources" (i.e. reliable sources that DO NOT present the information added to the WP article in direct connection with the article topic), but I will explain how we might word the policy in order to best ensure that editors will use, when necessary, the sort of contextual information from "off-topic sources" that we deem valuable and unobjectionable and avoid the kind of unwanted "off-topic source" information that several editors here have been denouncing (Jayden and Jayj have provided several examples in previous discussions):

(1) There are several strong arguments (see [29][30]) against the use information from "off-topic sources" that serves to advance a position or point of view with respect to any aspect of an article (without breaching the WP:SYN rule) (examples of this sort of information: [31][32]), but I believe the strongest argument is with respect to the NPOV policy. "The neutral point of view is a means of dealing with conflicting verifiable perspectives on a topic as evidenced by reliable sources". NPOV thus involves the neutral handling of "on-topic sources" (i.e. reliable sources that DO present the information added to the WP article in direct connection with the article topic).

So, editors must be careful, when using, "off-topic sources" not to misrepresent the Neutral point of view as established by "on-topic sources". (see my previous proposal [33] and Vassyana's and Someguy1221's responses [34][35]). If an editor uses "off-topic sources" in this regard, there is a great risk of misrepresenting the Neutral point of view established by "on-topic sources". Since the editor must rely on "on-topic sources" to determine the Neutral point of view, the editor may directly cite these sources in a WP article. I conclude that it is wholly unnecessary and risky to cite "off-topic soures" when adding information in an article which serves to advance a position or point of view with respect to any aspect of the article (without breaching the WP:SYN rule of course).

(2) Supplementary information from "off-topic sources" may be deemed to add value to an article, in order to clarify places, people or things mentioned in the article when the editor adheres to the WP:SYN policy and the information does not serve to advance a position or point of view with respect to any aspect of the article topic (examples of this sort of information: [36] [37]). In certain articles, e.g. political articles, such information can however sometimes create a certain bias or unduly cast a favorable or unfavorable light on an aspect of the topic, thus misrepresenting the Neutral point of view as established by "on-topic sources", even though the editor adheres to the WP:SYN policy and the information does not serve to advance a position or point of view with respect to any aspect of the article topic.

Furthermore, useful "off-topic source" information may sometimes shed new light on an article topic by making connections, not mentioned in the literature on the topic, that might not be immediately obvious to researchers working on that topic, such that these connections might suggest new research conclusions or approaches (IMO, this would be an example: [38]). This is what I would consider truly unacceptable "original research" and I would propose that we also state this in the policy. Therefore, it seems to me that the best way to ensure that such unwanted information does not make its way into articles is to state in the policy that useful "off-topic source" information intended solely to clarify places, people or things mentioned in the article should only be introduced in an article if there is consensus that this information does not misrepresent the Neutral point of view as established by "on-topic sources" and does not represent the form of truly objectionable original research described above. It should be clearly stated that "ideally" an editor will "cite a reliable source that presents the information in direct connection with the article topic", and only, exceptionally, may an editor use such useful "off-topic source" information. Also, WP:TOPIC should be mentioned somewhere in the policy.

--Phenylalanine (talk) 12:31, 7 October 2008 (UTC)

I agree with this. More or less, I believe the policy should say that non-"magic words" sources (aka "off topic sources" aka "sources that do not directly refer") are akin to primary sources: They aren't ideal, they should be used with great care, they should be replaced if a better source is available, but they are not explicitly forbidden. SDY (talk) 15:27, 7 October 2008 (UTC)

I recommend exercising some restraint in developing strong expectations that any major changes to WP editorial policy will occur here which would gain acceptance by the community. WP policy is driven not only by policy pages; but policy pages are also very much driven by WP community practice. See, e.g., WP:Policies and guidelines generally, and the section on "sources of WP policy" in particular. It's an interactive process that cannot, in general, be radically altered based solely upon a theoretical discussion in a talk page such as this one.
..... At this stage of the now very lengthy discussion about the relationship of sources to article topics, it's pretty clear to me that countless featured articles and good articles would fall outside a new policy mandate such as is being proposed here. This policy page is not the place to specify the breadth of article content via a predetermined limit on how closely sources should be related to the article topic. Sources obviously need to be directly relevant to the material presented in any given place in an article, and the presented material obviously should be relevant to the topic. The determination of what relevant material can be included in an article is a local editorial decision that falls well outside of WP:NOR, but instead is the domain of WP:TOPIC, WP:NOT and other pages that deal with appropriate scope of individual WP articles. ... Kenosis (talk) 16:30, 7 October 2008 (UTC)

Comments:Proposal and Rationale revised

For me this is a concern:

"off-topic source" information intended solely to clarify places, people or things mentioned in the article should only be introduced in an article if there is consensus that this information does not misrepresent the Neutral point of view as established by "on-topic sources" and does not represent the form of truly objectionable original research described above."

I follow Phenylalaline's reasoning and can see what he's trying to do, and I wish I thought it could work. In my experience, the complexity of the reasoning is the very aspect that will giving rise to problems. Already in contentious articles what is a neutral point of view is the cause of huge edit wars. There's a lot room for interpretation in this addition, and that's a problem on any policy. I think an editor earlier made a comment about trusting editors. I think what we can trust is that people all come into Wikipedia with their own individual POVs -just human nature, and the policies have to provide strong, crack free guides for all of us, so we can't get in the way of ourselves, and our best editing practices.(olive (talk) 16:06, 7 October 2008 (UTC))

I have a problem with assuming that all editors are POV warriors at heart. Local consensus can decide whether a source or statement is inappropriate original research and will continue to do so even if this becomes policy because of WP:IAR. It's how all other article content is adjudicated, and it seems to work OK. There are some rough spots, I admit, but the real POV warriors are going to continue no matter what policy we try and implement. SDY (talk) 18:46, 7 October 2008 (UTC)

lol....No, not what I meant. I meant that as human beings we all have predispositions that we can't escape just because of who we might be, not that we are all warriors. We all hopefully make efforts to edit in a way that transcends our personal identities and the Policies/guidelines remind us of how we can do that.(olive (talk) 18:57, 7 October 2008 (UTC))

I'll bring up the parallel situation of primary sources again: they can be used in inappropriate ways, but we allow editors and consensus to make the decision about what is or is not appropriate. Would you also ban the use of primary sources because editors cannot be trusted to use them appropriately? SDY (talk) 19:38, 7 October 2008 (UTC)

You misunderstand me... I wouldn't ban anything, but I do believe signposts that point to best editing practices as determined by the community are necessary. Its not my business to trust anyone or not here in terms of their ability to edit, nor is that what I said, but it is my business along with every other member of this community to have input into how those signposts are derived and lettered. Simple really (olive (talk) 19:52, 7 October 2008 (UTC))

If this were a guideline or essay, I would wholly agree. It's policy, and one of the core policies of the project, so it is fundamentally "do this" or "do not do this." It gets back to the point of "I would like it" as opposed to "it must be done this way." I would also prefer not to see people using unrelated sources, but I recognize that this is simply my preference and not fundamentally necessary. Making it mandatory is WP:CREEP. SDY (talk) 19:58, 7 October 2008 (UTC)

Additionally, I think the wording above would be more used to keep good and necessary material out in order to establish an non-neutral POV, than to keep bad material out. "off-topic" is a concept too vague to be of any good use. Frankly, i think this will be used as an excuse for including one-sided material found in a particular source without providing the necessary background to show it's one-sided. I don't want to mention specific articles here, but i can see this on all sorts of fringe topics, notably in pseudomedicine (where it can be used very effectively to keep out the bulk of reliable medical information.) Looking at the history, the wording has been sponsored by good people whom I often agree with and i suppose they must not realise how extremely insidious it could become. The best way of censoring at Wikipedia is to object to the sources. DGG (talk) 00:50, 10 October 2008 (UTC)

You're taking my words "off-topic" out of context. --Phenylalanine (talk) 10:07, 10 October 2008 (UTC)

I will be back with a formal text addition proposal, summarizing the informal proposal and rationale outlined above, when I get the time. --Phenylalanine (talk) 10:14, 10 October 2008 (UTC)

I agree with DGG's observation: "The best way of censoring at Wikipedia is to object to the sources." NOR is about not adding logic or interpretations to an article that is not found in the sources used. It should not be provide verbiage for yet another test of whether a source (though reliable and verifiable) should be "allowed" into an article. Objectors to the material added should develop their objections based on the subject and NOR. But I don't want to add more fodder for wikilawyering arguments built on a particular phrase of policy such as "directly related", or even "primary or secondary."

As mentioned earlier, I agree that in some cases unrelated material is being snuck in for SYNTH. For example, in an article about A, the valid observation that A worked for B may be followed by the invalid observation that B did lots of bad thing, which is invalid because the soruces about B doing bad things never mention A. Such guilt by association tactics clearly do not follow the logic of the sources. But I think it better to explain and prohibit such bad logic (more elegantly than above) than try to sum this up within a single phrase such as "directly refers to the subject of the article." Sometimes being too concise creates as many problems as it solves. --SaraNoon (talk) 15:30, 10 October 2008 (UTC)

Original research in plot summaries

I'm somewhat bothered by the fact that Wikipedia's policy is to automatically reject original research, but almost never has citations for the plot summaries in movies or books or television shows (radio shows, too?). If Wikipedians are made to cite external sources in order to show that they are not coming up with falsified or original research, then why are plot summaries exempt? Certainly some people don't seem to think they are, from the number of {{cn}} templates I've found in plot summaries, and possibly for good reason, because Wikipedia itself does not claim that it is reliable, and by extension (or perhaps because) its editors are not guaranteed to be reliable (this page is ample evidence, given how stringently it stops editors from trying to insert their own uncited experiences, which would include watching a series and then filling out an article on what happened in it, into articles, because that would be personal and original, and not many people who can be considered reliable also have static websites—blogs are banned as sources, last I checked—to put up the entire summary).

One might say there are plot summaries on IMBD, and they'd be right. But how reliable are they, with an open-door policy on writing them, much as Wikipedia has? Sometimes, there simply isn't any way to fact-check a summary except by watching it by one's self and then correcting or adding to the summary as needed, something which by definition is original research—sometimes, summaries simply don't exist for a particular title. Rotten Tomatoes also has overviews, which are quoted directly from a studio description, but the studio has a vested interest in not telling information, in order to sell their product: something which Wikipedia's mission and studios are fundamentally at odds with. What about fan sites? Do we—can we—honestly consider fanatics a reliable source, people who have a vested interest in the subject?

What should be done to harmonize practice and law? If we suggest that some editors should be allowed to insert uncited information, by what metric should we judge their reliability—do they have a college degree? In what they're talking about? Do they teach it? Do they just have a good vibe? Should we check out the source material ourselves? Or what? The hell if I know, but I don't want to add even more policies, for sure. Octane ^{[improve me?]} 10.10.08 0556 (UTC)