Wikipedia talk:No original research (draft rewrite 5th December 2004 to 5th February 2005)/Archive 2

This page is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

An example

Social structure of the United States is an example of how original research can creep into an article, mixed with established fact but put together in a pattern which is unique to its author. While I would not apologize for the article, I'm not sure anyone could find substantial authority anywhere for the exact way it is put together. Fred Bauder 15:11, Dec 5, 2004 (UTC)

Here's a recent example, "Market based health care systems such as that used in the United States rely on private medical insurance." recently added to Medical insurance. It this just what we all know or original research? I would be at a loss to cite a reference. Fred Bauder 15:29, Dec 5, 2004 (UTC)

A good example, Fred, that reminds me of issues I often have with new college students. I have a vigorous policy against plagiarism and stress the need to cite. Students ask if very fact -- even facts from an assigned book -- need to be cited. For example, if they read a book on Indians in S. America, and in their paper they write, The Ajwe-Xavante live in Brazil," do they need a citation? This is of course an issue facing researches who write articles, and the general answer is "no, you don't need to cite this." But why not? How do I explain this to students? Usually I tell them that if a fact is essential to an argument, or an element of an argument, it must be cited/sourced. Is this adequate/appropriate to Wikipedia? I am not sure. What do you think? Slrubenstein

Well, how is it that "everybody knows" about the Ajwe-Xavante? Presumably if you're not living next door to them in Brazil, you got the info from somewhere. One of things that struck me when comparing modern EB articles to WP are the mundane nature of some of EB's references - they are not some mighty mystical tomes, but simply the standard textbooks that are already sitting on my shelf. My reaction was "gee, is that all there is to it?". I don't think every statement needs citation of a specific page of a specific work, save that for talk pages in the less-likely event that somebody challenges and is unable to find the specific statement for themselves. Stan 21:12, 5 Dec 2004 (UTC)

The example I was giving was context specific -- I meant that in a particular course I taught, there were some things that students did not have to reference in their papers (the question is, who is "everybody?" In my example, the "everybody" is everyone in my class). Obviously, I wouldn't have the exact same standard in an encyclopedia article. I only meant that my dilemma -- where to draw the line -- is analogous to the one Fred mentions (where "everybody" refers to all readers of Wikipedia). Slrubenstein

My mental model is of a high-school student in Botswana clicking on random page; context in the lead section is critical. On the plus side, we can exploit linkage, and most references for the "everybody knows"-type info are better located at the linked-to article. Articles like sociology should list the dozen or so best textbooks, so as to ensure that readers develop a common base of understanding. Stan 03:23, 6 Dec 2004 (UTC)

A key difference between Wikipedia and Encyclopedia Britannica is that EB authors are generally chosen by the Editorial Board from amongst recognized authorities in their fields, usually people whose reputation and future as academics depends on their accuracy and perceived lack of bias. The EB authors are identified by name. So when an EB article says the Ajwe-Xavante live in Brazil, you don't need a citation of the source. You can rely on its being true. If the same sentence appears in the Wikipedia, without a citation as to the source of the information, you only have the say-so of the anonymous editor. Indeed it might be based on primary reporting (he's been to the reservation in Brazil himself). Or the information may come from the editor's personal fund of knowledge for which he has forgotten the sources. Some extra digging might produce a citable source but is the editor obliged to do this digging on items that he "knows" to be so before including them in the article. I think the "no original research" rule should not prevent someone from including facts in articles from their personal knowledge. But it should prevent the inclusion of information which cannot, even in principle, be independently verified through existing and reasonably accessible literary sources. Personally, I don't want the articles to be weighed down by constant citing of sources. On the Wikipedia, authoritativeness and reliability emerge from the community process. You have to rely on the fact that if a statement is untrue or biased, that it will be challenged by other editors, and removed if the person putting forward the claim cannot cite acceptable sources. I don't think one needs a source cited for every fact and factoid. --BM 19:34, 9 Dec 2004 (UTC)

How can authoritativeness and reliability emerge if no one ever has to provide any basis for their claims? I've made a bunch of careless statements, realized some of them six months later, and went back to fix - not one person had challenged them in the meantime, although they made spelling fixes around the statement. Writing sloppily in the hopes that someone else will clean up is not a crime, but it's not admired either, and citation of sources should be viewed in the same way. Did you ever look closely at the 1911 EB articles? They all have plenty of citations. So apparently even the big-name authorities hold themselves to the standard of providing the backing for their material. Stan 04:50, 10 Dec 2004 (UTC)

Cite sources

Added section on citation. Used the intro from the style guide as a basis and I think improved it. I think citation should be part of the policy of No Original Research rather than in the style guide. The style guide should quote the policy. :ChrisG

Good work

This has made good progress. I think it's much better than what we hope it will replace. I hope to get back to it and contribute within a few days. Maurreen 01:11, 6 Dec 2004 (UTC)

Logical inferences

If published information is summarized and sourced, may not a non-partisan logical inference may be drawn that represents a sensible mainstream opinion, even if no specific published source can be found? To exclude natural conclusions seems artificially constraining. Look at the "name origin" of Mayonnaise and the Discussion at Talk:Mayonnaise. Does Wikipedia intend to exclude such material as "theory"?

Where several schools of thought are in conflict, logical inference may establish some parameters that a counter-position needs to refute, in order to be credible. For example, at Gospel of Peter an observable, confirmable character of the text is followed by a logical inference:

"Some characteristics of Peter suggest a place early in an oral tradition. To be specific, the developed apologetic technique typical of the final edition of the Gospel of Matthew and of Justin Martyr, which seek to demonstrate a correspondence between prophetic predictions in the Tanakh and their detailed fulfillments in the fate of Jesus, is quite lacking. A credible assessment of Peter as dependant on the synoptic gospels needs to account for this consistent omission of any reference to the fulfillments of prophecy."

These guidelines might address such cases. --Wetman 06:16, 6 Dec 2004 (UTC)

I think conclusions should generally be left to the readers. Maurreen 06:45, 6 Dec 2004 (UTC)

Questions

I have added several "invisible" comments or questions concerning points that I'm not sure of. Maurreen 06:53, 6 Dec 2004 (UTC)

Journalism?

Is anyone interested in discussing the case I mentioned above, in Wikipedia:Divulging personal details? From what I know of the case, it at least verges on journalism, and I have some uneasiness about the ethical questions involved (not inherently because of the journalism angle). Maurreen 07:12, 6 Dec 2004 (UTC)

I looked at that page and it only confirms my strong sense about "no original research." Wikipedians should be doing something akin to scholarly research -- we should not be detectives, and not journalists. Of course we will have articles on current events and the news, but this is not a place for eye-witness acounts or personal reflections (like Everything2) either. Do we just need to write this into the policy: Wikipedians are neither detectives nor journalists?" "Wikipedia is not a place for eye-witness accounts?" You'd think "no orignal research" was clear enough. To me the importance of the case Maurreen calls attention to is not that we need to answer this person's queries; rather, we need to ask why the "No Original Research" policy as written was inadaquate to answer his questions. Slrubenstein

I think the "No original research" as now written at least has a strong implication to just "scientific research."

A lot of this stuff, I'm unsure about.

But Wikipedia is a different animal than any other, and I'm not sure that's been fully considered.

A lot of issues and pages are overlapping in my mind. I started a related but broader discussion at Wikipedia needs special standards. Maurreen 07:59, 8 Dec 2004 (UTC)

I think the policy as it stands applies to other forms of research (social science and humanities). However, I think it needs to be strengthened. This is what I think the policy should be: if the article makes no interpretive, analytical, or synthetic claims, it can rely on primary sources (and thus, be a secondary source). But if the article is to make interpretive, analytical, or synthetic claims, it must rely on secondary sources (and thus, be a tertiary source). Slrubenstein

The above paragraph by Slrubenstein does accord with what I have been getting at. Writing history or biography should be purely factual - a recital of the facts, preferably in a readable manner! The question I have been posing is where does information from original sources, such as archives fit?

As an example I wrote the article Simon Goodrich, based on articles I had previously written for the Oxford Dictionary of National Biography and the Biographical Dictionary of Civil Engineers. These were based on a handful of printed sources published between about 1920 and 1950. Since then I have been undertaking extensive research in Goodrich's papers for another project, but have picked up oddments about his life which might flesh out the Wikipedia article. If it is acceptable to use this extra material, with proper referencing, without running foul of the 'no original research' rule, I see little point in attempting to get it into a journal as the prequisite of being able to use the information in the Wikipedia article.Apwoolrich 19:06, 8 Dec 2004 (UTC)

Just to pick on Simon Goodrich a bit, it says "The Simon Goodrich papers are an incomparable source of detailed information" about his age. Now if you're quoting from the Biographical Dictionary of Civil Engineers, then that's cool; I or somebody else who's dubious can go check it for ourselves. If, however, that is an unpublished opinion you've settled on while examining the papers, that would not be so good to put in here; perhaps an authority more distinguished than yourself will publish a paper next week mentioning that Goodrich's papers are "all rubbish and fabrication", and then you've put us all in the position of trying to weigh your un-peer-reviewed personal opinion against the published opinion of someone else. Suppose you've tired of WP and moved on? Would future editors have to send you letters verifying that you still hold that opinion? This is really all about limiting ourselves to a solid foundation that will outlast individual participation, and eventually even individual lifespans. Stan 05:11, 10 Dec 2004 (UTC)

Well, actually, I think the writing of history is much much more than a recital of facts; I happen to know many professional historians who would be insulted by this characterization. They think history requires interpretation and analysis, and I agree. BUT I do not think it is the place of contributors to Wikipedia articles to do this analysis and interpretation. There are books by historians, and journals that publish articles by historians, in which one finds much interpretation and analysis. A contributor to wikipedia should consult those books and articles and then summarize their interpretations and anaylses, and provide the appropriate sources. When writing history articles we should also seek out events the analysis or interpretation of which historians argue, and provide accounts of those arguments/debates. Slrubenstein 21:16, 8 Dec 2004 (UTC)

Apologies for sloppy writing, what I actually meant in my first para was writing history and biography for Wikipedia. Apwoolrich 08:57, 9 Dec 2004 (UTC)

Well, there are historians and biographers that have done a lot of interpretation and analysis; their research (available in secondary sources) should be made available in Wikipedia in an NPOV way, right? Slrubenstein

It seems to me that "original research" is not the appropriate title for this policy. It is more accurate to state that Wikipedia articles are not primary sources. For example, I doubt that it would be acceptable to insert information into articles whose only source was the editor's personal obvservations or reporting; that is, "original reporting". We would want this to have passed through the filter of a journalistic organization, first. For example, the reporting might be published by a newspaper, with the reliability and accuracy of the reporter being vouched for by the newspaper organization. --BM 19:03, 9 Dec 2004 (UTC)

Various problems

If this is to be reworked, rephrase the following:

Wikipedis is not a primary source. Specific factual content is not the question.

So what is the question and who asked it? The phrasing is awkward. Perhaps:

Wikipedia is not the place for any but the most minimal amount of primary source material, however factual that primary source material might be. Primary source material should appear only in short quotations and citations or in summarized form.

Concerning topics of scholarly concern, Wikipedia is a tertiary source ... books published by a legitimate publisher and journal articles ... Concerning any subject that is a matter of academic research, we should exclude unattributed (original) interpretive and synthethic statements.

On topics which have not attracted sufficient academic research, primary source (such as diaries, government documentation, etc.) may be used in addition. In such circumstances, the Wikipedia article is a secondary source.

This would be major change of policy, and a strange one.

What is a legtimate publisher? Is there a list somewhere, or perhaps an Index of publishers whose books are not to be referenced? This just pushes argument onto another issue, especially as reputable publishers (even university presses) sometimes publish assinine books and schlock publishers sometimes publish good material.

Why are there two sets of rules, one for matters of scholarly concern and one for everything else? Does this mean that I can mention that Achilles killed Hector in the film Troy, but not mention that Achilles killed Hector in the Iliad, only mention that in a certain secondary book or article someone claims Achilles killed Hector in the Iliad. What happens when the film Troy becomes subject to academic research somewhere?

What is not of scholarly concern? There are collector journals and hobby magazine and so forth that often contain scholarly articles on areas which are not normally the object of academic study? Books are published in such areas, often very good ones. Why, for example, would it be recommended that one cite a particular author for the information that Christopher Columbus discovered "America" in 1492 but be free in an unacademic article to omit a source for the information that comic artist Jerry Robinson created the cover of Daredevil #5 (Nov. 1941)? The source for the latter is more important to cite, if such a detail is mentioned, because the information is less easily found. All information comes from somewhere. See for example Mario. This article is filled with weasel words and vague claims. But it is no worse sourced than the article Lion (picked at random) and many other articles that some think more central to Wikipedia. But neither article contains any original research of the kind that is forbidden.

The proposed revision seems to miss that one can summarize primary sources and present uncontroversial interpretations, pointing out relationships to other articles and so forth, without doing any original research whatsoever, as original research is defined here. Many articles in Encyclopædia Britannica are just such summaries as anyone who knows the material could write, the kind of secondary material that also fits in Wikipedia. They are secondary material, just normal, standard, vanilla secondary material as found in numerous sources, the kind of thing that knowledgeable scholars dash off in without much thought. People don't write original encyclopedia articles about Napoleon, Joan of Arc and Lion. They mostly just include the secondary information that is found everywhere.

Also, again and again in the areas I mostly write in, I find that going back to the primary sources and creating an article that summarizes and discusses those primary sources is far better than basing my article on changing scholarship. I want to provide information from the original source material that allows a reader to understand scholarship the reader comes across based on such primary data rather than to get deeply into tertiary material. Once that foundation is laid, then discussion of tertiary material can be added to an article. But first provide the secondary material (and even primary material) on which the tertiary material is based, and as much as possible present the primary and secondary data unbiased and raw. Pausanias says this, but Apollodorus says that, and so forth. Write the secondary material based on the secondary material and be unintelligent about it. Then afterwards add what modern scholars agree on or debate about on the matter of provenance, meaning, relations to other material, possible historic truth, and so forth. The numerous new theories that come and go are often just grist for the publish-or-perish mills.

The current rules prohibits original research whether an article is secondary or tertiary, and that should be enough.

Diaries and government documents are forbidden to be cited, if they become the subject of academic research? That seems absurd to me. Does this mean I shouldn't cite Thucydides or Pepy's Diary or Caesar's Gallic War directly because I must be only tertiary. But I and any editor will still select among tertiary sources. Bias will find its way.

I have a feeling this draft is an attempt to solve some particular difficulties, but one which more rules won't solve at all. The current rules would be enough ... if enforced. But they are not enforced. It becomes too much trouble for most to prevent a POV warrior insisting on some original research being included in an article. If old rules are not enforceable then new ones won't help. The POV warriors who don't care about the old rules that aren't enfored won't care about new rules that aren't enforced. And secondary versus tertiary is not the issue in any case.

As to the new long section on citing sources, does it really belong here?

Ideally, there should be proper citation for all topics that are either:

Academic

Controversial

"Ideally" is a give away. When I see the word in a ontext like this I know it means that the ideal can be ignored. Bad. And why only for two kinds of articles if we are being "ideal"? This disagrees with the plea later that all material should be sourced.

Wikipedia is not the place for original ideas, therefore the material you contribute must have been learned from an external source or more probably various external sources.

"Therefore" is not used logically.

As a contributor, you should therefore list as many of those sources as you can.

Overkill. I should list major general sources and sources which provide unique information.

Depending on the details, lack of attribution can constitute plagiarism.

Cite or reference the details, otherwise this sounds like legal scare tactics. It is no secret that Wikipedia articles are gravely lacking in source references but the information in them must have come from somewhere. So far no-one is suing and accusations about semi-plagiarism are seldom heard, though edit histories make it clear enough who is adding material without sources. No-one wants to open the can of worms about where the information has come from in so many articles. Over a million articles .... over a million articles .... over a million articles ....

Ideally, you should actively search for authoritative references.

"Ideally" again. My mind immediately inserts the phrase "but in reality ..." following the sentence.

Jallan 05:13, 8 Dec 2004 (UTC)

Related discussion

I started a related but broader discussion at Wikipedia needs special standards. Maurreen 08:06, 8 Dec 2004 (UTC)

Copy edit

Hi, I did a bit of a rewrite of some of this draft today and added a couple of sections. I've also made a few other sentences "invisible" because I felt they contradicted some of the rest of the article.

I may have labored the verifiability/truth distinction with my long example, but I did this because, time and again, editors don't seem to understand that their firm belief in something being true is not a reason to have it in Wikipedia.

I also said something about the checks and balances a newspaper provides because that partly addresses the issue of what is to be regarded as a reputable publication. And I added material from Jimbo Wales for two reasons: (1) because he expresses the issue well and also (2) because he is the only person in Wikipedia who actually has the power to make these decisions, or decide upon these definitions. The rest of us can only recommend. Slim 18:19, Dec 8, 2004 (UTC)

Pop culture and primary sources

I've put the fourth paragraph back in. Pop culture topics, for instance, often have not attracted any secondary materials. In order to write those articles there needs to be less stringent standards. Exactly, how stringent remains a question. Are authors of pop culture articles allowed to non controversially analyse Star Trek? :ChrisG 13:16, 9 Dec 2004 (UTC)

Hi Chris, what is the fourth paragraph exactly? I'm also not sure what you mean by secondary materials.

I meant to say yesterday, by the way, regarding my edit, please feel free to change and/or question it, as I know I added rather a lot. Best, Slim 14:41, Dec 9, 2004 (UTC)

I think interpretations are best left to the reader, not the writer. I do agree some areas probably deserve looser standards. Maurreen 17:35, 9 Dec 2004 (UTC)

References and sources

Maurreen, I noticed you said in your edit that some editors use the word "references" differently. I thought maybe the article should say something about the definition. To me, a reference is an article, book, paper, website that is referred to the author while he or she is writing an article. References should be noted throughout the text, though there is no clear standard for how to do this, but we can recommend one. At the end, there should be a References section, which should list the full publication information, and of websites, ideally should list the date the information was retrieved. Personally, I find this latter to be unnecessary, but it's apparently a requirement of Featured Articles, as I've recently learned to my cost (objections have only just been dealt with: mostly to do with lengths of paragraphs (!) and various reference formalities (sigh).) Anyhow, I digress . . . We should explain that references are not the same as "bibliography" which I would regard as being a more comprehensive list of works about the subject, excluding the works referred to, as they are listed under "references". People can use the word "bibliography" or "further reading" or "external links" or whatever they choose, but they should stick to the word "references" for works actually referred to by the author in the course of the edit, as it's the references that will allow the reader to check the veracity of the article. Do you agree with my understanding? I don't think we need to labor these points, as they are dealt with elsewhere (though never dealt with clearly), but they do go to the heart of the "original research" issue, in that people must provide decent references.

Also, am I noticing a difference of opinion regarding the definition of primary and secondary sources. Chris has used it, above, in a way that's not clear to me, though I may have misunderstood. If there's a lack of clarity on that point, that might be worth getting into as well. Chris, what do you think?

Also Maurreen, you felt my use of "claims" in the article casts doubt. Would "assertions" be better, or "statements"? The thing about "information" or "facts" or "knowledge," is that, strictly speaking, for something to be any of these things, it must be true. People do use the expression "false information," but that's actually a contradiction, as is "false facts". But perhaps I'm being pedantic. Slim 14:56, Dec 9, 2004 (UTC)

I agree. :) Specifically, with:

Defining "references" as you say. Eventually, this should also be checked so all Wikipedia guidance uses terminology in the same way.

Using "assertions" or "statements" instead of "claims."

People in this discussion are using "primary source", etc., differently. I don't know whether the answer is to avoid using the terms or come to a consensus on what they mean here. Maurreen 17:31, 9 Dec 2004 (UTC)

The reference section should be documents used to create the article; but I also tend to think it should be any document used to verify the article. By implication that article has thus contributed to the article. But I offer that as a suggestion not a firm opinion.

With regard to primary sources we use the phrase differently; and it seems from discussion here and elsewhere that there are as many interpretations as people. Here's Webster's definition: "any document created by someone with first-hand knowledge of an event, person, place, or thing; any record created at the time of an event or something's existence" from any Dictionary.com.

Using this common definition we run into problems with saying Wikipedia doesn't use primary sources in current affairs; because often reporters do use first hand information. Maureen considers newspapers to be a secondary source; but that is really not so clear from a definitional point of view. And certainly there is a difference between a newspaper article and peer reviewed journal. Even within newspapers there are differences in quality between them; and within a specific newspaper differences in quality and seriousness between the articles.

We also have major problems with pop culture and other articles, because often there has been no scientific research into subjects like Star Trek, Tottenham Hotspur football club etc. etc. Therefore there are no reputable secondary sources or even any informed debate.

A last issue the rewrite doesn't fully address is when writing about fact. Generally I think it better to go back to the original source when describing a theorists point of view. For instance, with JS Mills On Liberty; when describing his theory it is better to summarise directly from the text, rather than pull together summaries from various authors saying what they think he is saying. In writing fact it is better to use the primary source if at all possible.

Developing on from that if there has been no academic research into a subject; to what degree is a contributor allowed to make non-controversial criticisms and logical deductions?

So to conclude the issue of primary sources is a red herring. The rewrite needs to take a stand on the specific issues above and has to recognise the existing status quo. I suppose as a general point of principle Wikipedia is about knowledge not science or academia. Science produces the most reputable knowledge; but it is not its only source. :ChrisG 11:11, 10 Dec 2004 (UTC)