Wikipedia talk:No original research

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Peacedove.svg The project page associated with this talk page is an official policy on Wikipedia. Policies have wide acceptance among editors and are considered a standard for all users to follow. Please review policy editing recommendations before making any substantive change to this page. Always remember to keep cool when editing. Changes to this page do not immediately change policy anyway, so don't panic.
Frequently Asked Questions (FAQ)
I disagree with the definition of secondary source.
Wikipedia mostly follows the definition in use by historians, which requires more than simply repeating information from some other source or rearranging information from the author's notes. The earliest definition of a secondary source in this policy was in February 2004 "one that analyzes, assimilates, evaluates, interprets, and/or synthesizes primary sources".
WikiProject Spoken Wikipedia
WikiProject icon This page is within the scope of WikiProject Spoken Wikipedia, a collaborative effort to improve the coverage of articles that are spoken on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.

Synthesis of published material that advances a position[edit]

I suggest to add after the first sentence:
"Similarly, do not combine different parts of one source to reach or imply a conclusion not explicitly stated by the author(s) of that source."

I see another problem in the title of the section, which suggests that SYNTH is allowed as long as it does not advance a position. As this page is titled Wikipedia:No original research, I think that is not correct. Just remove the last part that advances a position?
--Wickey-nl (talk) 09:24, 11 May 2014 (UTC)

S Marshall, you seem not to understand the difference between SYNTH and synthesis. If you are so quickly with your responses, why did you not respond here, first? --Wickey-nl (talk) 11:44, 16 May 2014 (UTC)
  • Missed the talk page edit, I'm afraid.—S Marshall T/C 11:56, 16 May 2014 (UTC)
You missed the talk page and not the page edit? Not a reason to avoid a substantial reaction. Actually, your statement that all Wikipedia articles are syntheses of published material sounds a bit stupid. Synthesis is not necessarily advancing a position. Essentially it is against the WP rules at all. It is OR. --Wickey-nl (talk) 12:08, 16 May 2014 (UTC)

Not sure which edits you two are talking about... but to address the suggested addition:
I think it makes sense and support adding it.
As to the language of the section header ... I think you may be over thinking it a bit... a policy section header is not the policy. We could change the policy section header to "NOR rule number 5" and the policy would not change. What outlines the policy is the text that follows the section header. That said... if you have a suggestion for a better section header, please do so. We are always looking for ways to make our policies clearer and more understandable. Blueboar (talk) 12:29, 16 May 2014 (UTC)

  • I disagree with Wickey-nl completely. On Wikipedia we use "WP:SYNTH" to mean going beyond what the sources say, but that's got nothing to do with what the English word "synthesis" really means. If we use words with their ordinary English meaning, "synthesis" means "put together", which is exactly what Wikipedians do: we put together several different sources to make an article. That's right and normal. It's only a problem if you use that technique to advance a position the sources do not advance.—S Marshall T/C 12:34, 16 May 2014 (UTC)
At Blueboar: Today, I already made the suggested edits. Your remark about the heading would make sense if it were not misleading.
At User:S Marshall: The section is all about the meaning of SYNTH in the context of WP. Indeed, it has nothing to do with what the English word "synthesis" really means. Puting together several different sources to make an article is not the same as SYNTH. The key is "synthesis" to reach an original conclusion. --Wickey-nl (talk) 13:02, 16 May 2014 (UTC)
  • Policies should be written in plain English and use words with their natural English meaning.—S Marshall T/C 14:20, 16 May 2014 (UTC)

How about replacing the two sentences by: "Do not combine material from multiple sources, or multiple parts of one source, to reach or imply a conclusion not explicitly stated by any of the sources." ? Zerotalk 14:33, 16 May 2014 (UTC)

@ S Marshal... I am confused... because I don't see what you are objecting so strongly to. If you "combine different parts of one source to reach or imply a conclusion not explicitly stated by the author(s) of that source" ... you "put things together" to go beyond the source (no?). Blueboar (talk) 23:04, 16 May 2014 (UTC)
I'm sympathetic, but I'm worried about this expansion. I had a conversation with someone last year who appeared to believe that different sentences in the same (short) article about the same subject could not possibly be related to each other, because there were other sentences in between them, and the source did not say, "Apples are a kind of fruit, kiwi are a kind of fruit, and yes, these two facts about kinds of fruit are related to each other". SYNTH has always been multi-source, and I haven't seen any good reason to make the multi-source NOR section include single-source NOR problems (which are amply covered elsewhere on the page). WhatamIdoing (talk) 23:23, 16 May 2014 (UTC)
Ah... Thanks WAID... that is a practical concern about the addition that I can understand. Blueboar (talk) 00:37, 17 May 2014 (UTC)
Leaving the question of the heading out of the discussion here, what is the difference between combining two parts from different sources and combining two parts from one source to create a conclusion? Are not both Original Research, OR? An article is not always so straightforward. It may contain different subjects/lines/views. It may have different interpretations. Sometimes a conclusion is obvious Wikipedia:What SYNTH is not#SYNTH is not mere juxtaposition, Wikipedia:What SYNTH is not#SYNTH is not obvious II; sometimes a conclusion is an interpretation, not given in the source, especially when it is a single source. --Wickey-nl (talk) 10:24, 17 May 2014 (UTC)
Yes, the multiple-source/single-source issue is a red herring. The key point about SYNTH is that two or more sourced facts are used to draw a conclusion that is not sourced. It doesn't make a difference where the individual facts come from. In principle they could even be parts of the same sentence in a single source. Zerotalk 10:37, 17 May 2014 (UTC)
Traditionally, exceeding the contents of a single source is a plain-vanilla NOR violation rather than a SYNTH violation. With SYNTH, you are combining two sources to claim (or even just to imply) something that is in neither of them. It's particularly undesirable, because you don't know if the two sources are really talking about the same thing: Alice says that apple consumption is X in Canada, Bob says it is Y in Mexico, and Chris says it is Z in Canada, so you can just add it up and get "North American", right? Except that Alice's numbers were apples purchased, and Bob's were apples actually eaten, excluding apple juice and food waste, and Chris's were both apples and russets, measured from when they left the farm, rather than from when they reached the store or someone's mouth. These are not comparable numbers.
When you are dealing with a single source, it requires much more careful assessment to decide whether those claims (or implications) are actually present in the source, because the plain fact is that the author of that source believed them to be related at some level (else the author wouldn't have put both of those facts in the same article, right?) and there are far fewer problems with comparabiility. WhatamIdoing (talk) 02:03, 18 May 2014 (UTC)
You mean that a single source already may contain SYNTH? You say SYNTH is traditionally associated with the use several sources and the use of a single source is not SYNTH. I wonder if you can prove such "tradition", but if so, I don't see a problem in mentioning both here. I am not talking of just exceeding the contents of a single source. --Wickey-nl (talk) 10:23, 18 May 2014 (UTC)
A caution... please don't conflate the section short-cut with the policy itself. "SYNTH" is simply a handy five letter short-cut pointing to a specific section of the policy. That section highlights a particularly egregious type of synthesis. It does not cover every possible situation in which synthetic Original Research can occur. I think what WhatamIdoing is pointing out is this... The fact that other forms of synthetic OR are not covered in the "SYNTH" section does not mean that these other forms of synthetic OR are somehow allowed... it simply means that they are covered elsewhere in the policy. Blueboar (talk) 13:13, 18 May 2014 (UTC)
I agree that SYNTH is actually just a short-cut, but I used it in broader sense in accordance with this essay. The point is that the single source issue was not covered by this project page. --Wickey-nl (talk) 10:22, 19 May 2014 (UTC)
Um... I would say that the single source issue is covered in WP:STICKTOSOURCE... no?. Blueboar (talk) 11:12, 19 May 2014 (UTC)
Agree, but the same is true for the multiple-source SYNTH. --Wickey-nl (talk) 11:40, 19 May 2014 (UTC)
Yes, but multiple-source synthesis is a more complex issue than single-source synthesis... and so (in my opinion) it needs to be discussed in more detail. I think what we say in STICKTOSOURCE is enough to deal with single source synthesis, but isn't always enough to deal with multi-source synthesis. If you can convince me otherwise (with some real examples of situations where it hasn't been enough), I might change my mind. Blueboar (talk) 12:41, 19 May 2014 (UTC)
This discussion is enough reason. WhatamIdoing disputed that a single source can produce SYNTH. --Wickey-nl (talk) 16:47, 19 May 2014 (UTC)
Confusion in terminology... Looking at the discussion, WhatamIdoing disputed that a single source can produce SYNTH... but admitted that a single source can produce a synthesis. An understandable confusion, since the five letter short cut "SYNTH" does not mean any synthesis... but refers to a specific type of synthesis (the multi-sourced synthesis highlighted in the SYNTH section of the policy). This is why WhatamIdoing noted that Single-source synthesis is covered in other parts of the policy (such as WP:STICKTOSOURCE). Blueboar (talk) 20:54, 19 May 2014 (UTC)
Are there objections to this carefully formulated edit? --Wickey-nl (talk) 10:22, 19 May 2014 (UTC)
Not a strong objection... I just don't think it's necessary. It's essentially instruction creep. Blueboar (talk) 12:41, 19 May 2014 (UTC)
I don't think it's an improvement. SYNTH has functioned well with its multi-source focus, and STICKTOSOURCE covers everything that could be considered original research in a single source. (I'm also starting to wonder what dispute triggered this.) WhatamIdoing (talk) 01:29, 20 May 2014 (UTC)
The dispute was triggered by S Marshall's revert of my change of the heading. He essentially stated that WP is nothing more than putting together several different sources to make an article, and suggested to explain WP policies from the perspective of the meaning of the English word "synthesis". Than, you focussed on the multi-source issue.
I consider the single source used to produce a synthesis in the context of SYNTH worth to mention, in spite of the fact that Blueboar considers this core issue a minor one, already covered by the general STICKTOSOURCE policy. --Wickey-nl (talk) 08:49, 20 May 2014 (UTC)
  • That's totally not what I said and I object to this misrepresentation of my position. I suggest you listen to WhatamIdoing.—S Marshall T/C 09:23, 20 May 2014 (UTC)
  • To be clearer: I said that an important part of the process of article-building is to assemble sources into articles. I certainly didn't say that Wikipedia is nothing more.—S Marshall T/C 09:25, 20 May 2014 (UTC)

"Advance a position" vs synthesis vs SYNTH[edit]

Probably the short cut "SYNTH" was not a good choice. Is it possible that the issue originally was related to the POV issue, producing the heading "Synthesis of published material that advances a position"? "advance a position" has an ambiguous meaning here, fortified by the given examples.
What do you think of the heading "Original Research by synthesis of sources"? --Wickey-nl (talk) 12:57, 20 May 2014 (UTC)

Wickey-nl... with due respect... I think you may be over-thinking things, and seeing problems that don't actually exist. Is there an article where this has become an issue? Blueboar (talk) 13:09, 21 May 2014 (UTC)
The phrase "advance a position" in the section title seems too narrow because it suggests to me that someone has an agenda that they are advocating. I think most Synth occurs because editors are simply trying to improve Wikipedia in good faith without understanding that their edit is Synth. --Bob K31416 (talk) 16:41, 1 June 2014 (UTC)
Nothing says the position that is being advanced has to be due to advocacy... it could be completely unintended. The point is that if you phrase something in a way that reaches or implies a conclusion... that's "advancing a position" on the topic. And if you advance a position on a topic that is not made by a source... that's a form of Original Research. Blueboar (talk) 21:49, 1 June 2014 (UTC)
Definition of position from a dictionary:
"6. A point of view or attitude on a certain question: the mayor's position on taxes."[1]
Along the lines of what I wrote previously, most Synth does not occur because an editor is advancing a point of view or attitude, i.e. advancing a position, but rather because they are simply trying to improve Wikipedia in good faith without understanding that their edit is Synth. --Bob K31416 (talk) 22:47, 1 June 2014 (UTC)
Here's an edit on the project page that would address this issue [2] . (I only made this edit for information purposes and I immediately reverted it.) --Bob K31416 (talk) 01:25, 3 June 2014 (UTC)
That's an interesting approach, but I'm unconvinced by it. I think that having a little diversity in our language is helpful. Some people will understand "advance a position" and others will understand "state or imply a conclusion", and all of them will understand the policy as a result. I'm not inflexible here; if other people like it, then I'll go along. If it actually causes confusion, we could always revert to the older formulation after trying this out. WhatamIdoing (talk) 03:04, 3 June 2014 (UTC)
I made the edit. If you or anyone else wants to revert it then that will be the end of it for me. --Bob K31416 (talk) 01:00, 7 June 2014 (UTC)

WP:SYNTHNOT says "If something is obvious to anyone who reads and understands the sources that are supposed to support it, then it's not SYNTH" thereby implying that one is generally OK if one is sticking to the obvious. "[M]ost Synth occurs because editors are simply trying to improve Wikipedia in good faith without understanding that their edit is Synth" , however, assumes that you can stick to the obvious and still be routinely committing SYNTH. Certainly it sometimes occurs when an editor doesn't realize he or she is "advancing a position", but if one believes it's routine from editors who are not trying to advance a position, the implication is that a great deal of Wikipedia content is problematic SYNTH. Once we start implying that SYNTH is all around us then the accusations that such-and-such is SYNTH start flying around even more and as I note in a new section below calling for the return of "advances a position" I don't think we need that.--Brian Dell (talk) 18:05, 26 July 2014 (UTC)

"Quaternary sources"?[edit]

In this post I refer to an ongoing content dispute on Emperor Jimmu. I am not looking for input on the content dispute. This post is about a proposed clarification on this policy page. I also (obviously) don't want to be told that specific OR concerns should go to ORN.

Some articles (the one I'm involved with right now is Emperor Jimmu) cover subjects (the main subject) for which there are hardly any primary sources, and a relatively small number of secondary sources that are all within a niche/specialist field, but the subject is loosely connected to one or more other subjects (the peripheral subjects) that are extremely famous/important and covered in literally millions of secondary sources. This results in a large number of sources that may be considered reliable with regard to a peripheral subject, but whose author(s) may have only a cursory knowledge of the main subject. It's possible that such a source's bibliography might not include a single primary or secondary source on the main subject. This would make the source a "quaternary source".

Such quaternary sources should be rejected as innately unreliable, especially if they contradict material in reliable primary, secondary or tertiary sources.

This statement is already policy, since the only way one can use such sources is by WP:SYNTHesizing them with other sources that contradict them, but should we clarify this? Perhaps add another subsection to the WP:PSTS part of this page? If a source specifies in its own bibliography that it is a sub-tertiary source for certain information, we should probably specify somewhere that this particular kind of source is not to be used.

I was gonna take this to WT:RS, but the discussion of WP:PRIMARY, WP:SECONDARY and WP:TERTIARY sources is on this page, so discussion of a proposed WP:QUATERNARY. I know editing policy pages is trickier than guideline pages, and I don't have much experience of either, so any advice would be welcome. (talk) 14:37, 7 June 2014 (UTC)

I this is a valid issue for us to discuss here on the talk page... but I not sure that we should mention it in the policy. So let me ask a few questions... How often does something like this occur? And if the misuse of "quaternary" sources can be corrected based on what the policy already says, is there really a need to extend the policy to mention them more explicitly? Would mentioning it in the policy simply amount to instruction creep? Blueboar (talk) 15:08, 7 June 2014 (UTC)
It happened on the Emperor Jimmu article, and in theory it could happen to every article related to any of the hundred-twenty or so Japanese Emperors before 1941, and any aspect of the Imperial Histories and mythology surrounding the issue. (I mean "mythology" literally here.) And with all non-Anglosphere topics, English-language sources are one step down the ladder. An English translation of a Japanese primary source is actually a secondary source, a "secondary source" that has only consulted the primary source in translation is actually a tertiary source, and so on. And there are a lot of otherwise "reliable" sources that would slip into the quaternary category if we look at it like this. Last autumn I brought The Japan Encyclopedia to RSN's attention. The book is an English translation of a French work whose original bibliography didn't include a single source written in Japanese, and the translator was a French-English translator who apparently also didn't speak Japanese. This made the work a quaternary source on just about everything. It contains accurate information from time to time, but it's random, almost accidental. At the time, the source was cited in around 1,000 of our articles, and that situation hasn't changed much. These quaternary sources may accidentally contain correct information, and if we repeat what they say then we are also accidentally passing on correct information, but we probably shouldn't be pretending like quaternary sources are like tertiary sources but "more tertiary". We certainly shouldn't use The Japan Encyclopedia to "evaluat[e] due weight, especially when primary or secondary sources contradict each other". Talk:Kujiki is another place where in the last year User:Shii and I encountered a situation where the primary and secondary sources often contradicted each other. Had we gone to The Japan Encyclopedia to evaluate due weight, it would have essentially told us to buzz off, because all of the primary and secondary sources are wrong and the only one who's right is the guy who wrote The Japan Encyclopedia. (talk) 15:51, 7 June 2014 (UTC)
Few people will agree with you that a translation of a primary source is a secondary source, at least in the general case. However, I see that as yet another example of the minefield caused by excessive invocation of the primary/secondary/tertiary division in making rules. I don't think we need an extra category of source to confuse the issue even further. What your example suggests is more advice on the reliability of sources, which belongs at WP:RS. For example, that page says "Articles which deal in depth with specific studies, as a specialized article on science, are apt to be of more value than general articles which only tangentially deal with a topic." but only in relation to news sources. It ought to say much the same thing in relation to academic sources. The word of a specialist expert trumps that of a tangential remark by a non-specialist, in most cases. If WP:RS doesn't make that clear, it should. Zerotalk 16:00, 7 June 2014 (UTC)
If by "few people" you mean "nobody", then I agree with you, Zero. WhatamIdoing (talk) 19:14, 7 June 2014 (UTC)
Types of source have nothing to do with reliability, but only with the distance to the original event. E.g. Primary = Eye witness report of event; secondary discusses primary sources often using a specific frame; tertiary sources consolidate primary and secondary sources.
If anything in my view a Quaternary source should be at least as much (if not more) integrative and building on all lower order sources. So that would be sources that consolidates knowledge from tertiary sources. But as tertiary sources are limited to sources that consolidate primary and secondary sources a quaternary source would be indirectly consolidating primary and secondary sources and hence by definition be a tertiary source. In other words I do not think quaternary sources exist.
I do recognize the problem though. A source that provides expert knowledge on one issue may introduce some other issue in a speculative way, or as example; and these utterances should not be interpreted at the same level as the keystone of the source. In my view there would be 2 ways to deal with this. First - Stating that this is not the key message of the source and hence a better source is needed. Second - Speculative statements may be considered primary opinion by the author of the source. Arnoutf (talk) 16:28, 7 June 2014 (UTC)
In other words I do not think quaternary sources exist. What about my Emperor Jimmu example? A book about World War II that makes a few peripheral references to the mythical first emperor of Japan, but that gets its information exclusively from other books about World War II that may or may not have been written by people who had only read secondary sources. (Emperor Jimmu is covered in exactly two primary sources from the 8th century, and so for a book to be considered a secondary source its author would need to have specifically consulted one of these two. If they give a bibliography and don't include either of them then they are a tertiary source.) Such a source can't possibly be taken as providing reliable information on Emperor Jimmu, unless it just happens to accidentally agree with reliable secondary or tertiary sources. Most people who write books about World War II in English do not speak Japanese and have no direct exposure to any of the context surrounding Emperor Jimmu, and so they're apt to misinterpret whatever sources they do use.
And then occasionally we find sources that appear to have taken at least some of their information from Wikipedia. This essay probably meets the requirements of WP:RS for most of its statements, and its author meets WP:GNG so his opinions might technically be noteworthy as well. And Wikipedia is not directly mentioned anywhere in the essay (it doesn't contain any bibliography). But careful analysis indicates that its author probably consulted Wikipedia for at least some of his information, since he duplicates Wikipedia's somewhat idiosyncratic style guidelines for writing Japanese people's names. To the best of my knowledge, Wikipedia and Wikipedia-dependant (i.e., quaternary) sources are the only ones that would refer to "Takahashi Korekiyo" (familyname givenname) and "Junnosuke Inoue" (givenname familyname) in the same paragraph. Of course, that kind of discussion probably belongs on WP:WINARS more than here... (talk) 17:08, 7 June 2014 (UTC)
There's no such thing as a quaternary source. Indeed, some fields (e.g., law) do not even recognize the existence of tertiary sources. You should read WP:LINKSINACHAIN. A translation of a primary source is still a primary source. A quotation or close paraphrase of a primary source is still a primary source. And a lousy source is a lousy source, no matter whether it's primary, secondary, or tertiary (see the WP:NOTGOODSOURCE section of that same page). WhatamIdoing (talk) 19:14, 7 June 2014 (UTC)
Indeed I do not believe quaternary sources exist, but of course if reliable sources unambiguously claim the existence of quaternary sources I am willing to reconsider. Otherwise in my view the introduction of the term is original research. Arnoutf (talk) 20:22, 7 June 2014 (UTC)
Part of the trouble we have with terms like "tertiary sources" is simplistic statements about which kind of source is what. A translation is usually a primary source but this isn't necessarily so. I have on my bookshelves an annotated translation of the Domesday Book 1086, and with all due respect for WAID, it's very much a secondary source. A transcript of an interview is a primary source, but if a professional journalist has conducted the interview, edited and summarised what was said and given it a framing narrative and some context, then that's... well, we could quibble about whether it's technically a primary source or not, but the point is that it's a much better source than a transcript.

Evaluating sources is the encyclopaedist's primary job and it takes careful and reasoned judgment. I've always felt that this business of saying "this is primary, this is secondary, that's tertiary" is suboptimal and I'd prefer more nuanced language.—S Marshall T/C 11:05, 23 July 2014 (UTC)

I agree that (most) annotations are secondary sources. The translated text itself, however, is a primary source. It's possible (even common) for sources to contain both. WhatamIdoing (talk) 15:50, 3 August 2014 (UTC)
Re: "I've always felt that this business of saying "this is primary, this is secondary, that's tertiary" is suboptimal"... Amen to that, brother. This policy focus too much on defining the nature of sources, to the neglect of explaining the appropriateness of sources for specific types of statements. Very often, it's the nature of the text we write that is more important to NOR than the nature of the source.
On a similar note... I have noted this before, but it keeps falling on deaf ears - in the early renditions of this policy, the focus was not on using primary, secondary, and tertiary sources... instead the focus was on not turning Wiikipedia into a primary source (which is what happens when you add Original research). That was an important point (one that is directly related to the entire concept of NOR)... I wish we could return it to the policy. Blueboar (talk) 16:10, 3 August 2014 (UTC)

Cannibalistic tertiary sources[edit]

We don't seem to address the problem of "cannibalistic" tertiary sources, which are most commonly magazines and newspapers with low editorial standards. The two most common patterns of this are:

  1. The repetition of and elaboration upon sensationalistic stories from other similar publishers, without adequate fact-checking by anyone at any of these publishers, going back to the original (incorrect) article.
  2. Gathering and summarizing "facts" from biased or other questionably reliable sources, usually also tertiary, and re-publishing falsehoods as if they were common knowledge and well-established, simply because they've been restated frequently enough to become accepted, at least among a particular audience.

See the history of Negativland for an example of how the first tendency can actually be intentionally socially engineered to spread false stories. Most often it's just a blind process of rushing to publish salacious material. Sometimes it's entirely innocent and a factor of AP and other newswires leading to the same story being seeded with minor variations in thousands of venues more or less simultaneously, but retractions/corrections seeing far less print.

The second sort of case is typified by animal breeder and pet owner publications, e.g. Cat Fancy magazine and similar publications. In these cases, articles, which never cite any sources at all, are rehashed in slightly varied form every few years, about every major breed or type of common interest. They're mostly written by breeders or others with financial incentive to play up positive qualities (real or not) of their subject, and editorial control at such publications supports outlandish claims of breeders (their primary source of income in the form of paid advertising), and they're generally based on previous articles in the same or similar publications, going back decades. To make matters worse, when they paraphrase interesting (often incorrect) factoids about this breed or that, they often alter the wording a bit to avoid an appearance of plagiarism, but in the course of doing so, change the meaning of the original claims to something even less plausible. The three most common problematic results of this publication pattern are nonsensical claims about breeds and their collective "personalities", promotion of controversial breed origin claims as if proven, and acceptance of folkloric stories about the breed as if verified fact.

The latter pattern is also, of course, used by pseudoscience publications, but we already cover those at WP:FRINGE. The broader problem we're not addressing at all other than with the vague statement that some tertiary sources are more reliable than others. WP:RS is almost as a vague: "Tertiary sources such as compendia, encyclopedias, textbooks, obituaries, and other summarizing sources". [And note that obituaries are questionably reliable not because they're summarizing in nature but because they're usually written by or rely upon material provided by family and friends of the deceased and have the editorial purpose of promoting only good things about the subject, but that's another matter.]

Template:Tertiary can be used to ID problematic citations to tertiary sources in articles, but absent NOR and/or RS being clearer that such sources as discussed here are tertiary, other editors are apt to remove such tags as wrongly applied to sources they feel must be categorically considered secondary simply by virtue of the fact that they're in magazines and newspapers not "compendia, encyclopedias, textbooks". NOR's own wording perpetuates the logic error that a publication rather than what it publishes is or is not tertiary: "within any given tertiary source, some articles may be more reliable than others". While, yes, we can consider an encyclopedia as a whole to be tertiary, it's not true that a magazine or newspaper is necessarily or always secondary; any given article may be tertiary or (in the cases of investigative journalism, interviews, and editorials), primary.
 — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  15:53, 17 July 2014 (UTC)

A few quick thoughts:
  • Are you familiar with the distinction between an obituary and a paid memorial advertisement? It's important.
  • How exactly do you know that these magazine stories about pet breeds are wrong? WhatamIdoing (talk) 16:48, 17 July 2014 (UTC)
  1. Yes, though I may have been blurring the distinction here unnecessarily. I have made (cautious) use of journalistic obituaries in the course of working on bio articles, and in some cases they repeat known falsehoods because their writers will trust what the subject or the subject's family and friends have to say about the deceased rather than more independent sources. I clarified by the above with a small insertion
  2. Because I do serious breed research, and catch such publications in errors and distortions, the evolution of which can be traced over time. A fact or claim from, say, two or five generations ago may be repeated in somewhat distorted form in perhaps 1997. It then gets warped further and recycled, even further from the original statement (which may have been false to begin with or is obsolete) again in 2001, and further in 2005, and so on, until it looks like "lots of sources agree" on some particular factoid, which is actually nonsense. An example is the common claim that many Manx kittens with partial tails have their tails docked for medical or fraud purposes, and that lots of "Manx" kittens are actually just random cats with their tails cut off. This may have been true, on the Isle of Man itself in the late Victorian and early Edwardian eras when tourism to the Isle of Man produced a sudden and then-unfulfillable demand for Manx kittens; there's no evidence this is true any longer, and because the breed has other defining characteristics, and pets are routinely X-rayed by veterinarians, such a ruse wouldn't really fool any one who knew the breed, nor anyone at all for long. Yet cat magazines still repeat this and many other false stories about Manx cats (and many other breeds; I'm not singling that one out). An even worse case is when breed articles are written by breeders and full of all sorts of romantic, unprovable nonsense about the breed in question and "its temperament", as a form of thinly-veiled promotional material. There's an overwhelming amount of this crap about virtually every breed (of cat, dog, you name it). A third case is when dubious claims made by specific breeders are promoted by these publications, presumably due to advertising dollars or personal relationships in some cases, but more often probably just because the publications' budgets are small and their fact-checking abilities very limited. An example of this is the claim by the breed's originator, disproven by genetic tests, that the Pixie-bob breed is hybrid of domestic cats and the North American wild bobcat, a completely different genus. One of our own cat breed articles repeated this nonsense this very month, and the article on the breed did so only a year ago; it got the idea, as I recall, from an article in Cat Fancy. It's not that such magazines are always unreliable. Many of their veterinary articles are written by actual vets based on real research (sometimes even cited). It's that a certain class of article in them, the breed profile, are fundamentally tertiary in nature, and simply uncritically repeat whatever has previously been published about the breed, with an eye to especially re-presenting anything "interesting" or "fascinating" or "unique" or "appealing", i.e. that which is most likely to actually be distortion and half-truth and supposition at best, or outright lies promoted by breeders for monetary gain at worst.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  01:50, 30 July 2014 (UTC)
Please take a look at the Primary, secondary and tertiary sources section of the Wikipedia:No original research page. Though this discussion has merit; it has nothing to do with tertiary sources. — Robert Greer (talk) 20:53, 17 July 2014 (UTC)
I've already read that, and do not see what you mean. I'm addressing a particular type of tertiary source, one that is not of the compendium nature we're usually thinking of when we talk about tertiary sources (e.g. other encyclopedias). That's, well, the entire point of my post. There are tertiary sources, in rather large quantities, that we're misclassifying as secondary simply because of their medium (especially magazines and magazine-like websites), that are not actually, at least in some of their articles, doing any real journalism or research, but simply consuming and regurgitating previously-published material with insufficient regard for its veracity.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  02:50, 30 July 2014 (UTC)
I completely agree with SMcCandlish that there is a serious problem with poor quality tertiary sources (and sometimes ones we might classify as secondary as well) that endlessly repeat incorrect scraps of information. For an example in a different area, search for "lysichiton melt snow" in google. You'll get thousands of hits in which it's said that one or more species of Lysichiton have flowers that produce heat and so melt snow, which is quite false (only Symplocarpus species are thermogenic). Far too many articles are supported by poor quality web-based sources. Peter coxhead (talk) 07:44, 31 July 2014 (UTC)

There is a reason why this policy page does not address the issue... Poor source quality isn't an OR issue... its more of a reliability issue. You are not engaging in Original research when you read something in a poor quality source and repeat it on Wikipedia. You might be engaging in bad (or at least incomplete) research... but not Original research. Blueboar (talk) 11:46, 3 August 2014 (UTC)

I think that Blueboar's answer hints at the solution: you need to play the "my gold-plated academic source trumps your dozen silly little magazine articles" game. It can be tedious, but it's ultimately effective. If you can find an indisputably good source that talks about this problem in detail, then so much the better (and I hope that you'll use it to expand Dog breed and similar articles). WhatamIdoing (talk) 15:56, 3 August 2014 (UTC)
That's the rub, though. There is virtually no source at all for most breed-related information other than these cannibalistic tertiary sources mired in breeder-sourced bias and "telephone game"-caused decreasing veracity in successive generations of near-plagiarism of older sources. I'm unaware of anything like a Journal of Breed Studies. For some farm livestock breeds of especial importance there are, here and there, some academic articles about their origins and such, but these are rare. There are also veterinary science journal articles with reliable material on all sorts of animals, but veterinary facts are usually not the sort of thing in dispute in our animal breed articles. It's almost always breed origins and alleged behavioral traits ("good with children", "unsually dog-like for a cat", "likes water", "very intelligent and playful", etc., etc., etc.). Much of it is no more accurate than astrology. If we take it as a given that these sources are tertiary, biased and poorly fact checked (yes, these are WP:RS and WP:V issues), then having WP make factual claims based on such sources would seem to me to be original research, because we're cobbling together a falsely synthesized portrait of breed, and calling it factual, when the underlying components are often nonsense. I'm at a loss for what to do about it (other than, of course, leading by example in sourcing well). At this point, though, I keep getting reverted even on identifying tertiary sources of this sort with {{Tertiary}} fairly often, with disgruntled edit summaries that their dog or cat or rabbit or whatever magazine is just as reliable as any other publication, which is often simply not true.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  10:21, 13 August 2014 (UTC)
Oh! There's a name for what's going on in these sources and what it's doing to our articles here: woozle effect. It's a RS / V problem inasmuch as it's happening already in sources, but it becomes an NOR problem when we extend it from those sources into our articles, and this is precisely what's happening with animal breeds. And many other topics; the more fannish they are the more it's happening.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  12:08, 13 August 2014 (UTC)
So I asked above how you know that these are false. You said that you did "serious breed research". I assumed, perhaps incorrectly, that this "serious research" involved things like research publications, which would then be reliable sources. Apparently not? Apparently your source of knowledge is your own, unpublished, personal experience? WhatamIdoing (talk) 15:44, 11 September 2014 (UTC)
What are you talking about? "Apparently" you need to go read WP:AGF for starters. When you say "Apparently your source of knowledge is your own, unpublished, personal experience", you clearly are not "assum[ing]...that this 'serious research' involved ...reliable sources", i.e. assuming good faith. So, how about you stick to the issue raised here, instead of following me from forum to forum personalizing title and style disputes? (That's an issue I'll raise on your talk page with a WP:ARBATC {{Ds/alert}}.) I cite my sources in the articles I write and improve; I'm not going to spend hours laying out a catalog of them for you just because you want to play personally antagonistic games on policy talk pages. Sources vary from topic to topic anyway, naturally. If you think I've violated some core policy like WP:NOR, there are noticeboards for that. Perhaps you're not using them because you know you can't sustain such an accusation, and would likely be WP:BOOMERANG sanctioned for making it. The short answer to the question that could have been worded with much less personal investment in hostility and aspersion-casting, is that I invest a great deal of time, effort, and often money (over US$2,000 so far) obtaining (or obtaining access to) reliable source material for WP articles, ranging from journal articles behind paywalls, to antiquarian sources not yet available from Gutenberg Project and other archives, to brand new books one cannot even find used yet. To take cat breeds for example, I have a huge stockpile of cat-related publications, boxes and boxes full of them, and can easily compare and contrast articles, the claims they make, who is writing them, and what the writers' disclosed backgrounds are (e.g. breeders of the breed in question vs. those with no obvious conflict of interest). But this isn't about domestic animal breeds. I was using that as a simple example, fresh in my mind. Not every observation about sourcing related to animal breeding needs to be hunted down by you across Wikipedia and responded to with vitriolic character assassination, thanks.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  11:05, 14 September 2014 (UTC)
I'm confused. Following you from forum to forum? Maybe you should look through the talk page history; you'll find that I've contributed to this page since 2008. Article titles? Since when does "This breed of dog has this kind of temperament" have anything to do with article titles?
Here's the conversation we've had, from my perspective:
Me: "How exactly do you know that these magazine stories about pet breeds are wrong?"
You: "Because I do serious breed research, and catch such publications in errors and distortions"
Me: "you need to play the "my gold-plated academic source trumps your dozen silly little magazine articles" game. It can be tedious, but it's ultimately effective."
You: "That's the rub, though. There is virtually no source at all for most breed-related information other than" silly magazine articles.
Me: "You said that you did "serious breed research". I assumed, perhaps incorrectly, that this "serious research" involved things like research publications, which would then be reliable sources. Apparently not? Apparently your source of knowledge is your own, unpublished, personal experience?"
So I'm going to have to ask you to pick one position or the other here: either there are sources for breed-related information (e.g., about their temperaments), or there aren't. It is not possible for the answer to this question to be both yes and no; it is not possible for the answer to be "I invest a great deal of time, effort, and often money [on] reliable source material" and "There is virtually no source at all".
  • If there are such published, reliable sources, then cite them.
  • If there aren't any such published, reliable sources, then I am back to the basic NOR assumption, which is that there are only two types of knowledge in the world: that which can be sourced (e.g., something published in a research journal) and that which cannot (most importantly, something learned through personal experience). We cannot cite anything in the latter category, full stop.
And if anyone's got any clue how this relates to other forums or to article titles, which weren't even mentioned in this discussion, I'd be happy to hear it. WhatamIdoing (talk) 02:18, 15 September 2014 (UTC)

"advance a position" should be added back[edit]

This was recently wrongly removed. The editors that need to be dealt with with a SYNTH policy are the editors who are trying to advance some dubious contention. The editors who passively "reach or imply a conclusion not clearly stated by the sources" includes pretty much EVERY editor who is simply building an article since there are obvious, uncontroversial assumptions about chronology, etc that have to be made if editing is to be possible.

To be sure, people can still abuse WP:SYNTH even if "advance a position" is included. For example, RS #1 says something about Lars von Eisenach doing something in Leipzig in July 1888. RS #2 says something about Lars von Eisenach doing another thing in Leipzig in August 1888. I have the one sentence follow the other when writing the Lars von Eisenach bio. Some editor who doesn't like what is implied about Lars accuses me of SYNTH. Why? Because I "reached or implied the conclusion" that we have the same Lars von Eisenach here. Apparently RS #1 has to also refer to what Lars did the next month and/or RS #2 has to also refer to what Lars did the previous month. Otherwise we've got a chronology in the article "not clearly stated by the sources." I could still be accused of "advancing the position" that there is one Lars von Eisenanch instead of two, but the WP:SYNTH criers would at least be a little more circumspect about accusing me when the charge requires accusing me of some deliberation (advancing a "position") as opposed to accusing me of passively "implying" the sort of commonsensical "conclusion" we have to make everyday if we are to EDIT Wikipedia.

The number of "conclusions not clearly stated by the sources" is near infinite. If a 2014 article talks about Barack Obama and doesn't clearly state that he is the U.S. President, that he's President is a conclusion "not clearly stated", no? So a Wikipedian's brain can't put the two together? Requiring someone alleging SYNTH to accuse the transgressor of "advancing the position" that Obama is President would mean fewer spurious SYNTH accusations because editors aren't likely to accuse people of advancing utterly non-controversial positions that no reasonable person would challenge. They are, however, likely to accuse people of "implying conclusions not clearly stated" if that's all it takes to make out a SYNTH charge that justifies the removal of content they don't like.--Brian Dell (talk) 17:08, 26 July 2014 (UTC)

I agree that it's necessary to consider what the synthesis does, but the phrase "advancing a position" seems to me to be too strong, because it implies some undesirable purpose. I edit plant articles mainly, so let me give a plant example. Source 1 says that bee-pollinated flowers are blue to violet coloured with protruding lower "lips". Source 2 says that species X has blue to violet coloured flowers with a protruding lower lip. I write that species X is bee-pollinated, citing sources 1 and 2. I wouldn't want to describe this as "advancing a position" and certainly not "advancing a dubious contention" (with a bit more detail added, it's almost certainly true), but it's definitely SYNTH. The point is that if I were writing as myself elsewhere I could quite legitimately write something like "It is likely that species X is bee-pollinated, based on its floral characteristics", citing sources 1 and 2. But here, in Wikipedia, I shouldn't.
The problem with all attempts to define SYNTH is the fuzziness of the words needed to do so. "Conclusion" in relation to SYNTH is the part after the "therefore": A and B therefore C. SYNTH is where there are sources for A and sources for B but not for the entire proposition and where C is not indisputable (as, say, a conclusion reached by a simple mathematical operation would be). In your example about Obama, it's not a "conclusion" that he's the US President; it's a separate fact than can be sourced but doesn't need to be since no sensible person would challenge it. Juxtaposing information in a biography based on different reliable sources isn't SYNTH. Inferring reasons or intermediate behaviour based on that juxtaposition would be. Peter coxhead (talk) 06:24, 2 August 2014 (UTC)
SYNTH can be complicated... we really can't assume that two sources that mention someone named Lars von Eisenach are talking about the same person. They might be... but then again they might not. To know whether they are talking about the same person, we need both sources to include some sort of over-laping information. For example, if both sources say that the Eisenach they are talking about lectured at Cambridge University in 1885, and traveled to Berlin for an audience with the Kaiser in 1890, then we know that they are both discussing the same Eisenach... and once that is established, it isn't SYNTH to place the information mentioned only in one source in juxtaposition to the information mentioned only in the other. On the other hand... if the two sources don't overlap in any way, then it is SYNTH for us to make the assumption that they are talking about the same person. The assumption is not directly supported by the sources. Blueboar (talk) 12:44, 3 August 2014 (UTC)
Re "we really can't assume that two sources that mention someone named Lars von Eisenach are talking about the same person." — I think we can, unless there is reason to question that they are not the same person, or there are WP:BLP considerations. Seems like assuming they are not the same person unless proven that they are, in cases that do not involve WP:BLP, is an impractical and disruptive approach to editing Wikipedia. --Bob K31416 (talk) 13:54, 3 August 2014 (UTC)
Actually, Peter is "advancing a position": he is advancing the position that this particular flower is bee-pollinated.
This was discussed #Synthesis of published material that advances a position above; the general idea was to try it out and see whether it produced actual disputes (i.e., something more serious than people not being used to the wording). I believe this is the first question about it so far. WhatamIdoing (talk) 16:09, 3 August 2014 (UTC)
Here's a dictionary definition of position:
"6. A point of view or attitude on a certain question: the mayor's position on taxes.[3]
The example that Peter gave may have been clearer if he left out his parenthetical comment "(with a bit more detail added, it's almost certainly true)". Maybe this was what you were referring to when you thought he was advancing a position? Anyhow, here's the example again without that part.
Source 1 says that bee-pollinated flowers are blue to violet coloured with protruding lower "lips". Source 2 says that species X has blue to violet coloured flowers with a protruding lower lip. I write that species X is bee-pollinated, citing sources 1 and 2.
This is coming to a conclusion using original research, specifically synth. There is no indication that this is advancing a position, using the above definition. Whether right or wrong, it looks like an attempt to reach an objective conclusion. --Bob K31416 (talk) 20:22, 3 August 2014 (UTC)
Precisely, and very clearly put (sorry I wasn't as clear!). It's wrong in Wikipedia regardless of its objectivity (or even truth), which is why it's better to remove "advance a position". Peter coxhead (talk) 13:15, 4 August 2014 (UTC)
Agreed. Advancing a position is WP:NPOV not (intrinsically) WP:NOR matter. Also, the fault in the above is worse than just WP:SYNTH, it's a classic case of the propositional fallacy of affirming the consequent; i.e., it's an invalid way to reach a conclusion even before concsider whether there is novel synthesis. I wouldn't use an example like this, because it may be taken to imply that synthesis is okay as long as it doesn't suffer that "If A, then B; B, therefore A" fallacious reasoning pattern, which most cases of SYNTH do not (or fewer people would engage in it).  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  06:25, 5 August 2014 (UTC)
Off the main point, but it's a fallacy only if you assume that "bee-pollinated flowers are blue to violet coloured with protruding lower lips" means "if a flower is bee-pollinated then it is blue to violet coloured with protruding lower lips". However, in evolutionary terms, pollinator preference and floral morphology co-evolve, so there's not a unidirectional cause and effect. So "bee-pollinated flowers are blue to violet coloured with protruding lower lips" should be read as equivalence not implication. Peter coxhead (talk) 07:09, 5 August 2014 (UTC)
The flawed conclusion lies in assuming the equivalence and ignoring the possibility of sub-set... it may be true that all X (bee-pollinated flowers) are Y (blue to violet colored with protruding lower lips)... but that does not necessarily mean that all Y (flowers that are blue to violet coloured with protruding lower lips) are X (bee-pollinated). X could be a sub-set of Y, and there could be Ys that are not Xs. Blueboar (talk) 16:07, 5 August 2014 (UTC)
I'll respond on your talk page to avoid hijacking this page any more! Peter coxhead (talk) 17:02, 5 August 2014 (UTC)


I'd like to see this blurb promoted into this Wikipedia:No original research policy page. I believe too many editors use "Primary" as a synonym for "bad" or "unreliable" or "unusable". People are often looking for rules of thumb that make their lives easier, but that often means a lot of disruption when those rules of thumb are being applied to cases where they do not apply. Editors need more encouragement to consider the particulars of the case at hand.--Brian Dell (talk) 03:31, 30 July 2014 (UTC)

I don't mind linking to that essay... but were you aware that this policy already address the issue of Primary sources fairly directly? To quote: Unless restricted by another policy, reliable primary sources may be used in Wikipedia; but only with care, because it is easy to misuse them. (It then goes on to discuss what constitutes a misuse). Blueboar (talk) 12:54, 3 August 2014 (UTC)
Perhaps that phrase would be the ideal place to link to PRIMARYNOTBAD. WhatamIdoing (talk) 15:59, 3 August 2014 (UTC)

Question about routine calculations[edit]

I'd like to get some clarification about the intended scope of WP:CALC. Some confusion has arisen here Talk:Bon_Secours_Sisters#.22Average.22_death_rate about how best to describe the historic mortality rate (over 36 years) in an Irish orphanage. The various data sources, which seem mostly reliable, suggest that short-term epidemics and the like account for large chunks of the deaths. Thus, using the term average is not appropriate, since the information is coming from different sources, the data is so skewed, and many of the incidents are not independent events. I'm sure there are many similar cases in other articles, which surely must go beyond the intended policy defined for routine calculations. Any thoughts? jxm (talk) 16:05, 30 August 2014 (UTC)

If the problem is the term average, then you could try non-mathematical substitutes like "overall" or "typical". WhatamIdoing (talk) 16:36, 11 September 2014 (UTC)

Insider's view[edit]

I would like to suggest a change of wording in the Primary, secondary and tertiary sources section. We say

Primary sources are original materials that are close to an event, and are often accounts written by people who are directly involved. They offer an insider's view of an event, a period of history, a work of art, a political decision, and so on. Primary sources may or may not be independent or third-party sources.

I would like to suggest that "an insider's view" be changed to "a first-hand view", because "an insider's view" can be read to mean that the source is connected to or affiliated with the subject. In other words, that the source is not independent. I realize that we point out the difference between "primary" and "not independent" in the very next sentence. But it seems that the message is not getting through.
On Talk:Soka Gakkai, there is a group of editors who insist that any book from the publishing arm of that organization is a "primary source" and hence can't be used in the article. To back up that claim, they cite this section. So I'd like to suggest this change of wording. And is there any other way to rewrite this section so that it can't be misunderstood? --Margin1522 (talk) 21:20, 10 September 2014 (UTC)

I agree that "insiders view" is not right... but I don't like "a first-hand view". I know it does not really work for history related articles... in part because not every primary sources in the field of history actually does offer a "first-hand view". Some primary sources start off as originally being secondary but, over time, evolve into being primary sources. For example, Livy's history of the Punic Wars is not a "first-hand view" (he was not there and did not actually see the events he is describing)... It would have originally been considered a secondary source... but today it is considered a primary source never the less, due to the fact that it is the closest we come to a contemporary source.
Also, I don't know if "first-hand view" accurately describes what the definition of Primary source for the Sciences. I know we have discussed before the fact that different fields of academia have different definitions of primary and secondary ... but we have never actually resolved the issue.
I think it might be time to completely rewrite the PSTS section entirely... or perhaps to hive it off into its own guideline. As written it causes more confusion than clarity. Despite the fact that we clearly say that Primary sources MAY be used, but must be used with caution... we still get too many editors who think the policy says Primary sources are never allowed. We need to correct that and better explain when and how they can be used.
I will also remind everyone why the terms "Primary", "Secondary" and "Tertiary" were originally added to this policy... they were added to make the point that WIKIPEDIA should not be made into a primary source ... which is what happens when you add unpublished information, analysis, observations, conclusions, synthesis, etc. To put it in simple words: the original intent was not to say "using primary sources causes OR" (although I agree that sometimes it can)... the original intent was to say "adding OR causes Wikipedia to be a Primary Source... and it should be a Tertiary source". Blueboar (talk) 22:30, 10 September 2014 (UTC)
Picking up Blueboar's point, neither phrasing is really right for scientific sources. The issue in scientific articles is mainly notability: something published only in an original journal article may not turn out to be accepted by the relevant scientific community in the longer term – for example the research reported may not be replicable. (Nanobacterium and Cold fusion offer cautionary tales; I'm concerned over the speed with which Wikipedia jumped on the Dendrogramma bandwagon – I don't see secondary sources there.) So the general approach is to rely on secondary or tertiary sources for notability and then, cautiously, use original sources such as journal articles if these are the only ones that contain details needed in the article. This is particularly true of medical articles, where WP:MEDRS has attempted (rightly but not always successfully in my view) to set out the relevant distinctions. Peter coxhead (talk) 10:48, 11 September 2014 (UTC)

Illustrations violating NOR[edit]

I just flagged two images at Talk:Paraceratherium#P._transouralicum_with_calf_-_source.3F that could not be sourced. There is an editor defending their inclusion, claiming that there is consensus that NOR does not apply to illustrations. I am challenging this position and wonder if anyone has any input on whether his claimed consensus to suspend NOR does, in fact, exist in the broader community. Samsara (FA  FP) 16:17, 11 September 2014 (UTC)

How about linking to the original discussion where I explain what the consensus is? And conclusion from last community discussion about this:[4] And why do we need this same discussion on five different talk pages? FunkMonk (talk) 16:44, 11 September 2014 (UTC)
(edit conflict) Samsara, does the image actually look like what it purports to be? If so, then it's okay. We do not require people to provide sources to prove the identity of the subject; we only require that images look like the subject. See WP:PERTINENCE. WhatamIdoing (talk) 16:45, 11 September 2014 (UTC)
There is clearly a consensus that NOR as applied to text does not apply to images; it's clearly stated at WP:OI. Peter coxhead (talk) 16:59, 11 September 2014 (UTC)
I read that, and it says, so long as they do not illustrate or introduce unpublished ideas or arguments. And that is the core of my request - to please clarify that the representation is not introducing unpublished ideas. Samsara (FA  FP) 17:06, 11 September 2014 (UTC)
As I've already explained, the article text describes how scientists think the animal looked like, and the illustration follows that, without introducing new ideas. FunkMonk (talk) 17:08, 11 September 2014 (UTC)
If the text is based on sources, and the illustrations, according to what you just said, based on the same sources, why are you refusing to use the same sources to support the illustrations? Why are you being evasive about this? Samsara (FA  FP) 17:20, 11 September 2014 (UTC)
I'm not saying the illustration is using the same exact source as the article (Prothero's book). I'm saying the sources confirm the accuracy of the illustration. The two statements are not the same. FunkMonk (talk) 17:22, 11 September 2014 (UTC)
Explain to me why the source(s) can't be used to support the illustrations. Samsara (FA  FP) 17:25, 11 September 2014 (UTC)
There are several different sources the illustrator could have used that say the same. I cannot know which exact source it was, and therefore cannot add it to the image description, but that is irrelevant, as long as I know it is confirmed by the sources I do know. FunkMonk (talk) 17:27, 11 September 2014 (UTC)
You still have not named a single specific source that supports either of the contested illustrations. Samsara (FA  FP) 17:32, 11 September 2014 (UTC)
I've already mentioned Prothero's book. Do I have to quote myself? See also practically every source listed in the article. The point is, I can not point out which exact one the illustrator followed, and it is irrelevant anyway, because they agree with each other. There is a controversial idea that it had large, elephant like ears, but that is not shown in the illustration, so it is not a problem. FunkMonk (talk) 17:38, 11 September 2014 (UTC)
You cited Prothero in response to my other query, not this one. Here's the thing I find fishy about your evasive behaviour. When we find a text source that is lacking a citation, we look in the literature to confirm or reject the stated fact, and if, on balance, we find mostly or exclusively literature supporting the fact (or if the support is more recent and a revision of the earlier rejection), then, assuming our finding in the literature matches the claimed fact exactly, we supply the citation without changing the text. So my question to you is still, why can't you provide ANY SOURCES and insert them into the appropriate image description page, that support the faithfulness of the reconstruction? It would be as simple as saying, "see this literature for confirming the claim". But I have never seen anyone on Wikipedia as stubbornly refusing as you, to do a small simple thing that would be an improvement of the encyclopedia. And that makes me suspicious. It's almost like you don't want to be the one inserting a source because you know there's something wrong with it. Samsara (FA  FP) 17:50, 11 September 2014 (UTC)
Short answer, because you don't ask nicely, and because I don't have to, per the policies cited above. Constantly assuming bad faith doesn't exactly make me less "evasive". If you weren't so confrontational, then yes, I'd happily do it, I'm quite cooperative when people are nice. But I don't take orders, especially if they are not founded in any existing Wiki policies. FunkMonk (talk) 17:53, 11 September 2014 (UTC)

I'm reading this and I'm slightly baffled. The whole point of WP is collect the knowledge of our species. That's why articles require sources. Reconstructions are fine, and someone needs to make them, however, without sources directly referencing the reconstructions, there is no way to tell the difference between something based on the fossil record (or someone else's possibly copyrighted reconstruction), and something someone drew randomly. Often, the articles and written descriptions are not enough for other editors to check if the reconstruction is sensible, perhaps a vandal with a sense of humor has made a drawing and posted it, regardless, things need to be verifiable. If reconstructions can't be verified, they are useless. For an example, see Arandaspis. It has a lovely reconstruction, but it is unsourced. The written text is nothing like specific enough to glean the actual shape of the animal, nor should it be. The reconstruction can't be checked easily, and there is no indication of what was used to make it. In no other realm do we "trust" editors to create material without sources, and things as speculative as reconstructions need to be even more scrutinized, due to their unknown nature. This is also a large difference between photographs and reconstructions. Photographs are rarely speculative in nature, and indeed are capturing an actual state of existence at the time of their creation. Therefore, "own work" is good enough. This simply does not apply to reconstructions.

It isn't a large burden to ask for sources to be listed with reconstructions when they are uploaded, and it would bolster the credibility of the reconstructions greatly, which in the end is a good thing for WP, as anything that makes WP more credible is a good thing. Everything on WP needs a source eventually, so why not start collecting them now for the good of the project? pschemp | talk 03:08, 12 September 2014 (UTC)

I actually do not think that illustrations are exceptional in this case. All your arguments apply equally to photographs, particularly if the subject of the photograph cannot itself be easily verified. While the photograph itself has captured an actual state of existence— barring any manipulation, of course— you still don't know, without verifiable sources, that the photograph is a true and accurate depiction of what it purports to be. Just as with the cases you give for reconstruction, maybe the photographer mischievously uploaded an incorrect image or didn't correctly identify the subject in the first place. While images of famous buildings, landscapes or people may not have this problem, many other images do. Regards, Orange Suede Sofa (talk) 04:08, 12 September 2014 (UTC)
Photogrpahs and speculative illustrations are fundamentally different objects. Yes photogrpahs can be manipulated. However this is not relevent to the restoration debate, because restorations are based on a guess. A good one but a guess nonetheless. It isn't difficult to cite the source the restoration was based on. In most cases, for a photograph it is impossible to cite a source other than the creator of the image because there isn't one. Again I say, why not take steps to improve the credibility of WP when they are so simple?pschemp | talk 15:41, 12 September 2014 (UTC)
Pschemp, no one is saying rationales for the restorations aren't useful. But images should not be removed or deleted if one isn't provided. They can be removed/modified when inaccuracies are pointed out, not before, that's how it goes for all Wikipedia images. FunkMonk (talk) 09:35, 12 September 2014 (UTC)
I never called for removal or deletion, instead I said that everything needs to have a source. Very few of these restorations have a source. Tell me again what the harm is in requiring restorations to have source material? Sourcing is a pillar of WP. All restorations are based on something, so it needs to be cited. It can't that difficult to add sources to the existing pictures. Time consuming, yes, but that's what editors do. pschemp | talk 15:41, 12 September 2014 (UTC)

The purpose of images is to illustrate what is said in the text of the article. For example, if the article text describes what a group of scientists think a prehistoric creature looked like, then it is perfectly appropriate to include a drawing based on that description. It can even be created by a Wikipedia user (we explicitly allow such user created images). The key is to have an appropriate caption that links the image to the text. The caption should not bluntly imply "this is what a Blueboarodon looked like"... but instead should make it clear that "this is how Prof. Jones, et al. describe what a Blueboarodon looked like" And... if a different group of scientists disagree with Prof. Jones... and describe "Blueboarodon" differently... then a similar drawing should be included to illustrate their view of what the creature looked like (with an appropriate caption). In other words... for situations where POV comes into play... the caption should make it clear that the image is illustrating an opinion... and not necessarily illustrating a fact. Blueboar (talk) 10:20, 12 September 2014 (UTC)

Doesn't "restoration of" suggest pretty clearly that it is, well, just a restoration, a hypothesis? Also, we usually keep restorations out of the taxobox and use images of skeletons, so that this first image gives a less speculative impression. FunkMonk (talk) 10:22, 12 September 2014 (UTC)
All good hypotheses are based on facts. No is saying that speculation isn't allowed on WP, but speculations should be able to cite their sources. A restoration with no sources cited is indistinguishable from OR, and violates WP:VERIFY. "(Wikipedia) content is determined by previously published information rather than the beliefs or experiences of its editors. Even if you're sure something is true, it must be verifiable before you can add it." Everyone knows that restorations are speculative, but they should be as accurate as possible, and there has to be a way to check this using previously published info. With sources, a restoration is confirmed to be based on something, and it becomes acceptable. pschemp | talk 15:51, 12 September 2014 (UTC)
As pointed out above, there are different standards for image verifiability than to text verifiability. Please resolve the underlying issue before applying non-existent policies to specific cases. FunkMonk (talk) 17:11, 12 September 2014 (UTC)
To be quite honest, I think this is a convenient misinterpretation of "policy" (one of the two "policies" cited being in actual fact a guideline) that you and a few others have come up with for your own convenience. WP:V is quite clear on there being no difference at all (see cited passages elsewhere in this discussion or read the page itself). And you all know that dreadful moment when a friend or acquaintance approaches you with the, "say, I heard such and such about Wikipedia, is it true?" I highly doubt that anyone would want to be the person explaining why illustrations on Wikipedia escape intellectual scrutiny. There simply is no logical explanation or valid excuse for it. Samsara (FA  FP) 21:12, 13 September 2014 (UTC)

I wish people would stop periodically trying to shit on the collective entirety of our paleoartists. They're saints for dealing with this garbage. I want to give a loud THANK YOU to our many intelligent and talented artists who make paleontology accessible to lay readers with their illustrations. Abyssal (talk) 16:04, 12 September 2014 (UTC)

We're simply asking for verification that will enhance the value of the works that paleoartists are contributing. Samsara (FA  FP) 16:34, 12 September 2014 (UTC)

Proposing changes to Wikipedia:WikiProject Palaeontology/Paleoart review[edit]

I think the above discussion has demonstrated that different standards are being applied to illustrations in different areas of Wikipedia. This is generally undesirable. All information on Wikipedia should be equally verifiable. Wikipedia:WikiProject Palaeontology/Paleoart review has set itself up as an internal peer review process that, according to its own understanding, awards a sort of seal of approval to illustrations that fall into the "paleoart" category. Unfortunately, the requirement of verifiable, reliable sources supporting each illustration is not currently one of its principles. I propose that such a principle be explicitly added and observed. I am deliberately keeping this discussion here as it is a direct corollary of the above. Stakeholders will be notified. Samsara (FA  FP) 16:46, 12 September 2014 (UTC)

Time of RfC formation: 18:28, 12 September 2014 (UTC)

Yes, different standards are applied to images than to text. See WP:OI and WP:PERTINENCE, as pointed out above. Many do add sources for their restorations, but we should not remove images without it. I find it a bit funny you would want to change the policy of a single Wikiproject, instead of proposing to change the established Wikipedia policies that you are actually disagreeing with. The former will not be changed before the latter is. Please resolve the underlying issue before applying non-existent policies to specific cases. FunkMonk (talk) 17:08, 12 September 2014 (UTC)
Note to others: FunkMonk is the most frequent contributor to paleoart review, and by reverting requests for additional verification at Paraceratherium drew my attention to the problem outlined above. Samsara (FA  FP) 17:49, 12 September 2014 (UTC)


  • Support as nominator. Samsara (FA  FP) 22:19, 12 September 2014 (UTC)
  • Support requiring sources Don't care how that happens. Doesn't necessarily need to go into Paleoart, because Paleoart people should be following overall WP guidelines.pschemp | talk 16:02, 15 September 2014 (UTC)


  • Oppose as these are not requirements that are imposed on any other biology illustration topics, or for that matter any other illustrations at all.--Kevmin § 01:16, 13 September 2014 (UTC)
    • The requirement should be imposed on all topics, yes. However, other than the parallel dinosaur project, I'm not aware of any others that have set themselves up with the pretense that internal peer review can substitute for reliable sources. I believe the outcome of this discussion should be treated as a precedent, for all topics and all similar "review" projects. Samsara (FA  FP) 04:08, 13 September 2014 (UTC)
  • Oppose - this is a very arbitrary place to start. Propose to change the underlying issues, which are the WP:OI and WP:PERTINENCE policies, before you try to change this. FunkMonk (talk) 10:13, 13 September 2014 (UTC)
    WP:OI says, "Original images created by a Wikipedian are not considered original research, so long as they do not illustrate or introduce unpublished ideas or arguments" (emphasis in original). That is, it requires that the ideas be published and verifiable, i.e. explicitly sourced. Samsara (FA  FP) 13:08, 13 September 2014 (UTC)
Nope, that is your own interpretation. If you actually read the sourced article, you would yourself be able to see whether the images introduce controversial ideas or not. I have read the literature used in the text, and they don't. We do remove restorations once they turn out to be inaccurate, as happened at Spinosaurus yesterday[5][6], not before.FunkMonk (talk) 13:10, 13 September 2014 (UTC)
  • Oppose – this is just instruction creep. WP:OI is clear and all that we need. Let me declare an interest, since I have contributed a number of illustrations of reconstructions of extinct plants. We simply could not provide such illustrations in Wikipedia articles without the freedom to draw them ourselves, obviously always based as closely as possible on reliable sources. If an illustration is inaccurate, then the proper course is to remove it and discuss at the article's talk page if the removal is opposed. Peter coxhead (talk) 10:41, 13 September 2014 (UTC)
    • No, it's trying to bring a WikiProject's operation back into line with policy by changing the WikiProject's instructions to participants (there's already text there, so there's no creep - just replace the text that misinterprets WP:OI with a more appropriate line.) Samsara (FA  FP) 13:08, 13 September 2014 (UTC)
  • Oppose - The quote you keep putting up is going to destroy your points. What I, as well as many other paleoartist have to deal with is speculation. Not every single animal illustrated is known from a complete specimen, with all the exact integument and colour preserved, not even animals like Anchiornis, Archaeopteryx, and Microraptor, who are known from many specimens that preserved large amounts of colour and integument. If unpublished speculation was not allowed, than we would be forced to remove every single reconstruction of every single prehistoric article! My main point is, all paleoartists have to speculate, even if they are illustrating animals that everything preservable is known. IJReid (talk) 14:59, 13 September 2014 (UTC)
    • That's not a sound argument. Some things are outside the scope of Wikipedia, and those that are purely speculative are exactly that. Samsara (FA  FP) 15:41, 13 September 2014 (UTC)
      • Yes, but if they are completely speculative, they wouldn't pass the image review in the first place! As long as it matches all the known bones, even if only one bone is known, then it is not completely speculative. For example, see the illustration for Dromaeosauroides. It is only known from two teeth, but the image is still usable because the teeth of the illustration match those that are known. My argument is more sound than arguing that we should completely remove every image with some speculation (which is all of them!). IJReid (talk) 16:20, 13 September 2014 (UTC)
        • You did raise my concern mentioning the illustration for Dromaeosauroides - only known from two teeth. When I checked I was glad to see it clearly labelled Hypothetical restoration, based on related genera. I presume that the people working this part of Wikipedia are well familiar with the related genera. I presume they did a reasonable job reconstructing it based on available information. While this case stretches pretty far, I consider the Encyclopedia better with this approximate image than no image. It's informative. Alsee (talk) 16:53, 13 September 2014 (UTC)
Just as a sidenote, I made that illustration, and I did send it to the original describer and finder of the fossils for approval, and they liked it. I made the tail straighter due to their remarks, for example. FunkMonk (talk) 17:40, 13 September 2014 (UTC)
  • Oppose - In a nutshell: Don't pointlessly nuke good images. I am not involved in Palaeontology or Paleoart. I oppose unless very concrete problems can be demonstrated. I certainly agree that an image with sources to support it is more valuable and more resistant to challenge. WP:OI does not currently require images to meet the proposed standard, and I believe the proposed standard would be very disruptive if it were generally applied. WP:PERTINENCE provides extensive guidance on selecting appropriate images, and it provides abundant guidance on rejecting images. An image with speculative elements, which otherwise fits known information, is often the best available image and well satisfies WP:PERTINENCE. Based on a brief examination of wp:WikiProject_Palaeontology/Paleoart_review they seem to have put in place an excellent elevated standard of review, listing many criteria for easily and firmly excluding images that would otherwise be acceptable. The best available image for illustrating an article should not be excluded on the sole basis that does not come with the proposed form of sourcing. An unsourced image can be challenged on the basis of any reasonable source indicating it fails WP:PERTINENCE, or an image can be replaced on general argument or sources indicating some other image is superior. Alsee (talk) 15:58, 13 September 2014 (UTC)
  • Oppose as instruction creep and also as trying to change a policy in the middle of a content dispute. WhatamIdoing (talk) 21:25, 13 September 2014 (UTC)
    Paleoart review is not policy. I personally think NOR, RS and V cover what we need in terms of policy. Others are trying to make a case that policy needs to be changed, but that's not what this proposal is for. This proposal is about changing the operating principles of paleoart review, which currently claims that user created images are exempt from NOR: User created images are not considered original research, it says near the top on the project page. That needs to be changed, as it's very obvious that user created images could be used to advance original thoughts. Regards, Samsara (FA  FP) 22:10, 13 September 2014 (UTC)
Agreed: Paleoart review is not policy. I see it as an excellent subject-area essay providing a heightened standard, to assist editors in quickly reaching consensus for removal of problem images. Agreed: user created images could be used to advance original thoughts, and images can be rejected on WP:OI if a reasonable concern is raised about an image. BTW, it's a really Bad Idea to start an RFC and then use the voting area to argue with every single person who don't support you. Take it to the comments section. (Oh wait, you argued with everyone except me, I'm not sure whether to be flattered that my reasoning was impervious to attack, or insulted that you didn't bother with me) Alsee (talk) 04:18, 14 September 2014 (UTC)


This really needs to be addressed as a site-wide thing, not targeted at a single Project. I will say I do support the concept that if a WP-editor's construction of a figure from otherwise text descriptions that require more than just understanding basic knowledge of that field, eg requiring interpretation or creativity beyond what is the norm, that is original research and should be removed. That is, a couple specific ideas that we would allow is the creation of a coat of arms from the text-based heraldry statement, drawing out a chemical molecule based on it's IUPAC name, or creating a map based on a list of locations that are named by a source - there is no "interpretation" in doing any of these. But I can see for the case here for paleontology that a wikipedia's attempt to draw out a creature based on a high-level description (eg "Suchandsuch is a 10 foot long 4-legged creature with a dorsal fin") is not appropriate. I would assume that there is a "language" in paleontology that describes common body parts and if the drawing was created from that type of description, that would be reasonably okay, but if one has that description to that level of detail, I would also think one could find the image provided by the researchers that determined that. This doesn't just apply to paleontology, obvious, but it is a good case of where this can go wrong. And note this only applies to images created by WPians from sourced materials, not images created by researchers themselves. --MASEM (t) 16:10, 13 September 2014 (UTC)
Did you take a look at wp:WikiProject_Palaeontology/Paleoart_review? They lay out extensive criteria to reject an image if anyone can identify a conflict with known evidence. They seem to be taking good care to prevent anything from going wrong. Alsee (talk) 16:37, 13 September 2014 (UTC)
No other part of OR policy works that way: you don't write article text that is otherwise kept unless someone can prove it invalid; we require all material included to be verifyable from the start. So no, that review process is bad. (But as I said, this should be set at policy at the Wikipedia level, not specifically against one project's policy). --MASEM (t) 17:14, 13 September 2014 (UTC)
Please consider your argument carefully - by that interpretation I believe a large majority of existing images would have to be stripped from articles. We simply can't use copyrighted images published in sources. wp:Verifiability says: In Wikipedia, verifiability means that people reading and editing the encyclopedia can check that the information comes from a reliable source. Wikipedia does not publish original research.. WP:Original images says Original images created by a Wikipedian are not considered original research, so long as they do not illustrate or introduce unpublished ideas or arguments. This goes directly to the interaction between images and verifiability. The proposal here is that a wikipedia photo of a Ford Mustang is "not verifiable" if the photo doesn't come with some sort of sourcing to verify that it is a Ford Mustang. The only sourcing on the file says "Source: Picture was taken by me in Arizona". Are you saying we exclude that image? The only way to actually Verify an image is by visual inspection. Someone looks at the image and says "that's not a Mustang because [detail is wrong]". In this case we're talking about images that claim to be "reconstructions of extinct animals". The only way to verify that is by visual inspection. Any reader seeing an image labeled as a reconstruction of an extinct animal is clearly on notice that the image is a best-effort reconstruction with speculative elements. --The images are what they claim to be. Alsee (talk) 05:39, 14 September 2014 (UTC)
That is absolutely not what is being said here, because you are talking about things that exist and thus others that are reasonable experts in that field can verify that: a photo of a Ford Mustang can be confirmed by a car enthusiast, while a photo claimed to be taken in Arizona can be confirmed by someone local should that been necessary. The issue is when there is no way to verify the image by any expert. Keeping in mind that the image above with the parent and its child has been shown tied to a source, take the case where there was no source at all, and the artist - not a paleontologist - guessed what the child looked like. There is zero way anyone can validate that would be correct. With a source that shows what an expert might think it looks like, then we can rest easy on the reliance of this expert source that it is correct. But if we can't, then it is unverifiable original thought and should not be used. It's the nature of the education guess that has no sourcing to back it up that is a problem. --MASEM (t) 06:13, 14 September 2014 (UTC)
I agree on implementing the same standards site-wide. Simpler, less evasive language on policy pages might be a step towards that. However, my concern remains that there are currently two WikiProjects known to me (Palaeontology and Dinosaurs) that see it as legitimate to regard their own internal peer review as an adequate substitute to reliable sources. That's exactly the kind of thing that NOR was originally designed to prevent. And that's also the fulcrum of the debate: if the paleoart peer review were structured to ensure compliance with material published in reliable sources and then made that information accessible in file descriptions, it would be a great addition to Wikipedia. Operating as it is now, it runs counter to our purpose. Samsara (FA  FP) 17:09, 13 September 2014 (UTC)
I can go browse images on wiki commons, I can find an image labeled qqqqqq.jpg, that image can be sourced as "Made by me", and I can use that image to illustrate any relevant article in the encyclopedia. The only way to do that is based on a visual inspection of the image which leads me to believe it would would be an informative illustration for that section. We have abundant guidance on image selection, but ultimately it all comes down to visual inspection. Does the image depict something useful and appropriate to the article? Alsee (talk) 06:00, 14 September 2014 (UTC)
WP:OI would exclude speculative elements that have the intent or significant effect of advancing original ideas, but not to the extent of prohibiting reasonable good faith reconstructions where filling in details is unavoidable. Alsee (talk) 16:37, 13 September 2014 (UTC)
There's a line where that can be crossed however. Take for example coats of arms. There is room for interpretation in drawing out the symbols from the text heraldry description, such that if you asked 10 people to create a coat of arms from a text blurb, you'll get 10 very different images. But coats of arms have a very strong element definition that it's not so much the exact style that is needed but the clear understanding of what it is to represent; those 10 version may have a different version of a eagle for example, but it's the existence of the eagle on that that identifies the coat, not exactly how it looks. When people are drawing otherwise unrepresented extinct animals, there is not that same 1-to-1 in all cases between what is identified by paleontologists and the illustration. Take the case of the image in the section above, where the original drawing (which seems fine) is photomanip'd to include a child of the species. There's no source to explain what that child would have looked like, the extrapolation appears to be coming from the fact the species is compared to the hippo. There's no "obvious" way to draw the child relative to the parent. As such, we as WP editors should not be making that step. If a paleontologist made that and published it as their claim, sure, we'd be okay. --MASEM (t) 17:21, 13 September 2014 (UTC)
Some features, like for example colour, are unknown, so an educated guess and decision must be made, based on available sources. Same goes for the calf. I wouldn't have drawn one myself for the reasons you mention, but I don't think the image should be removed just because it appears in it. Incidentally, it does not look much different from earlier depictions of such a calf.[7] FunkMonk (talk) 17:46, 13 September 2014 (UTC)
Educated guesses fall under "unpublished ideas or arguments" of OI, so would be invalid. --MASEM (t) 18:43, 13 September 2014 (UTC)
I repeat: "based on available sources". In a sense, the documentary I linked to[8] shows a "published idea" of how a calf would look like, and it matches the one shown in the image. FunkMonk (talk) 18:45, 13 September 2014 (UTC)
In this specific case, then, that image should be fine as long as the documentary is sourced, going along w/ the fact that I would assume the producers of that show are reliable. (as BBC, I'm expecting that not to be an issue). But we're talking the very general case here; without that documentary, the drawing of the calf would be an OR educated guess and inappropriate. --MASEM (t) 19:53, 13 September 2014 (UTC)
That's why we take it on a case by case basis over at the review. Many images have been weeded out over the years[9], and many have been corrected, based on various sources. It is pretty effective, especially since most illustrators won't look at some obscure policy when they upload an image. Much of the paleoart we have has been uploaded by people we have no direct contact with. Most of it is useful and accurate, so we only remove them once something is demonstrably wrong with them. FunkMonk (talk) 20:10, 13 September 2014 (UTC)
WP:V does not work on a case by case basis. It is applied consistently to all material used in article space. There is no exception for paleoart. Reference it or risk losing it: any material challenged or likely to be challenged must be attributed to a reliable, published source using an inline citation. I challenged your material. Policy says you must provide inline citations now. Samsara (FA  FP) 20:43, 13 September 2014 (UTC)
Risk losing good content for no apparent reason? That is simply a stupid idea. FunkMonk (talk) 10:10, 14 September 2014 (UTC)
If we cannot verify it, including the necessary steps to get to the educated guess, it is not good content for WP. --MASEM (t) 02:26, 15 September 2014 (UTC)
And I guess that's why Wikipedia has different sourcing policies for image and text verifiability. So we don't lose useful images for no good reason. I wish we would stick to the actual Wikipedia image policies in this discussion, and not apply non-existent ones/ones that specifically only apply to text. FunkMonk (talk) 10:28, 15 September 2014 (UTC)
What's missing from your argument is how a contributor several years down the line would find out that there was once a link made between some documentary, and that illustration. How are you giving that person a chance to evaluate the accuracy of the illustration? You're not, is the simple answer. Requiring the readers' blind trust is not an acceptable option if Wikipedia wants to maintain or improve its reputation. We have the means to inform them about the sources that back up our articles. I don't see how it's a good idea to withhold this. Samsara (FA  FP) 20:33, 13 September 2014 (UTC)
As I said, an image with the sort of source information you're asking for is more valuable and more resistant to challenge than one with out it. So I agree it's not a good idea to deliberately withhold this. The disagreement is that you are asking to change policy to categorically exclude useful informative images when there is no identifiable problem with them. Alsee (talk) 03:43, 14 September 2014 (UTC)
Lack of sourcing is a problem in itself. It is preferable to provide sources for images rather than remove them. However, supplementing sources that an illustration is actually demonstrably consistent with also seems to not be "allowed" in the scheme of the Palaeontology WikiProject. I've suggested a solution along the lines of "this previously unreferenced illustration was reviewed by X and determined to be consistent with the views given in Authorityman, T. H. E. 2014. Big book of knowledge. Egghead University Press." I'm still waiting for a logically coherent explanation of why this would be a bad idea. Samsara (FA  FP) 04:33, 14 September 2014 (UTC)
There is a transparency principle encapsulated in WP:V, which is that the reader should be able to verify the accuracy of individual statements without reading ALL of the sources given at the bottom of an article. Illustrations are not intended to be a loophole, so should similarly indicate the evidence from reliable sources that they are based on.
Here are some relevant passages from WP:V:
[Wikipedia's] content is determined by previously published information rather than the beliefs or experiences of its editors. Even if you're sure something is true, it must be verifiable before you can add it.
All material in Wikipedia mainspace, including everything in articles, lists and captions, must be verifiable. All quotations, and any material whose verifiability has been challenged or is likely to be challenged, must include an inline citation that directly supports the material. Any material that needs a source but does not have one may be removed.
Samsara (FA  FP) 17:36, 13 September 2014 (UTC)
However, at present, WP:OI clearly says that images are not the same as text. If you want to change this, start an RfC specifically to this end. At present you're effectively quoting selectively. Peter coxhead (talk) 18:02, 13 September 2014 (UTC)
That's incorrect. I've quoted the WP:OI passage previously. I understand that you think it means something different than what I think it means. WP:V is rather clear on what is intended and imo resolves the issue over the interpretation of WP:OI. Samsara (FA  FP) 19:16, 13 September 2014 (UTC)
@Samsara: well, we'll have to disagree as to whether WP:OI needs "interpretation" via WP:V. I don't think it does. Peter coxhead (talk) 09:32, 14 September 2014 (UTC)

I'm a little concerned about the overall direction of this proposal. Let's leave aside the area of paleoart, since some people seem to have their heels dug in a bit there. Let's ask about normal pictures:

  1. I take a picture of something in a city I'm visiting. I upload the image to the article, and someone "challenges" it because he wants a published source that proves this picture really was taken in that location, or really does show what it appears to be a picture of. He's named no specific problem with it and has no knowledge of the subject. There are no published, reliable sources that say that I took this picture in that city. What do you think should be done?
  2. I walk over to my neighbor's house and take a picture of a plant. She tells me the name of the plant. I upload the image to Commons and add it to an article about the genus. Someone "challenges" it, saying that there's no good reason to assume my neighbor has any idea what she bought for her garden. Everyone agrees that the photograph looks remarkably like the plant it is alleged to be, but there are no published, reliable sources that tell us what she bought. What do you think should be done?
  3. I'm working in a medical lab, and I take a picture of a cell under a microscope. I upload the image to the article, and someone "challenges" it, because he wants a published source that proves this cell really does depict what I say it does. He's named no specific problem with the image, and it looks just like all the other (non-free) images of this kind of cell that anyone else has ever published. There is, naturally, no published, reliable source that says "on this day, this editor really did have this kind of cell under her microscope". What do you think should be done?
  4. It's the same cell, only this time I've drawn it instead of photographed it. It still looks like pretty much every other diagram in a textbook of this cell, but I didn't actually use any of them. If I'm going to say where I got it, then naming those other sources would basically be a lie: they weren't my sources. What do you think should be done? WhatamIdoing (talk) 21:42, 13 September 2014 (UTC)
In all those cases, that's a combination of common sense and AGF. Importantly, these are all things that others can verify because the item (the city, the plant, the cell) all exist today, compared with the hypothetical look of a creature that died out years ago. Now, arguably, the one case that would be one to look for better info is the plant one, because by no means is "your neighbor" an expert in botany, but again, there are likely editors that are reasonable experts in biology to be able to judge if the naming was correct by the photograph alone. There's no novel interpretation going on in any of these cases. --MASEM (t) 22:21, 13 September 2014 (UTC)
So long as the photo looks like the plant that my neighbor says that it is, does it even matter if she's right? Some species can't be differentiated by looking at the plant, or require looking at parts of the plant that aren't visible in my photo. WhatamIdoing (talk)
Probably not, though if someone who is an expect comes in and challenges that the photo may not be the species shown, I would expect efforts to be made to find a confirmed photo of the species (one named by a botanist or the like). But this is a different issue from the point above. --MASEM (t) 06:06, 14 September 2014 (UTC)
Yep, as my OPPOSE noted the proposal would be very disruptive if generally applied. Good images are valuable, and our policies give both good guidance and considerable case-by-case leeway in determining what constitutes a "best image" for a given article-use. Paleoart seems to have developed an excellent subject-area list of criteria. Alsee (talk) 04:41, 14 September 2014 (UTC)

Assuming this did pass, it would mean having [ ref ] links hanging off of images all across Wikipedia. This doesn't bear directly on the argument itself, but it would be an ugly-as-hell consequence. Alsee (talk) 12:26, 14 September 2014 (UTC)

That's well-established practice, actually. If you look at articles such as evolution and DNA, you'll find that about half the captions already include a citation. Note, though, that the primary issue here is not about including sources in captions. The primary issue is providing sources at all, whether just in the file description or on article captions as well (which WP:V already says you should provide when challenged to do so). Samsara (FA  FP) 12:48, 14 September 2014 (UTC)
We would not be required to have the ref in the caption; the file page absolutely must have the source(s) to avoid the original thought, and if editors felt the caption on the image could use it, that's fine. --MASEM (t) 13:42, 14 September 2014 (UTC)
Much of this can be resolved by careful and appropriate wording of a caption. For example, if an editor has drawn a picture of animal X, then an appropriate caption would read "Artist's depiction of X". Now... the artist may have drawn a poor or inaccurate depiction... in which case, the solution is to substitute a different (better or more accurate) artist's depiction.
However, sometimes the issue isn't over the quality or accuracy of the depiction itself... instead the issue is that there is a dispute in the sources over what the animal looked like. This is where we have to have captions that directly link the image to the specific source that was used to create it. This is best done through source attribution (as opposed to citation). If source A says the animal had/has webbed feet, and source B says it didn't/doesn't... we can include two images (side by side)... the one showing webbed feet captioned: "Artist's depiction of X, according to A", and the other captioned: "Artist's depiction of X, according to B". Then both POV's on the issue of webbed feet are presented in accordance with our WP:NPOV policy. Blueboar (talk) 14:04, 14 September 2014 (UTC)
The disputed images seem to say "Restoration of some dinosaur", which I eventually concluded was probably that field's jargon for "artist's depiction of some dinosaur". My initial thought was that it was a restoration of an old or damaged artwork, i.e., painting restoration. WhatamIdoing (talk) 02:23, 15 September 2014 (UTC)
I had a similar first impression. I think reconstruction might be a better term. Is restoration the standard term taken from professional/RS usage? Alsee (talk) 09:47, 15 September 2014 (UTC)
"Restoration" is most frequently used for palaoart. "Reconstruction" seems a bit more directed at man-made objects. But they're pretty much interchangeable for this purpose, and could both be used. And both make it clear that these images are hypothetical to some extend. FunkMonk (talk) 10:28, 15 September 2014 (UTC)

Arbitrary Section Break[edit]

It's time for a history lesson folks. When WP was first created, images and text were added all over the place, without citations and sources. WP:NOR and WP:VERIFY came out of this time as editors realized that in order to be a credible project, WP had to cite and source, and make sure copvio wasn't happening. That means a lot of text was gone through and cited, and a lot of pictures had to be re-found and sourced. Was this fun? No. Was it necessary? Yes.

Thus, regardless of who or how or even when, everything on WP needs a cited, documented source. Artists' renderings are not exempt from WP:NOR and WP:VERIFY. This information needs to be found and inserted into the restorations image descriptions, just like we did with images way back when. If any of you were around then, there were not immediate mass deletions, but a reasonable period of time set for this to be done. That is what needs to happen again. No whining about who or how, or when, or that it would be too hard is allowed. We aren't children throwing a tantrum. Those weren't acceptable excuses and behaviour for WP then, and they aren't now. Set a time period, (1 year is reasonable) and fix it. pschemp | talk 15:58, 15 September 2014 (UTC)

Yet again, before that can be enforced, you need to get the wording at WP:OI and WP:PERTINENCE changed, since it makes a clear distinction between article text and images when it comes to sourcing. Nothing will happen until then. We follow Wikipedia's actual policies, not the arbitrary preferences of various editors. FunkMonk (talk) 16:06, 15 September 2014 (UTC)
Yet again, WP:VERIFY and NOR take precendence. I've read OI over and over and it still requires that drawing are not just made up. Verify requires that things can be verified. To do that you need sources listed. pschemp | talk 17:48, 15 September 2014 (UTC)
pschemp, I'd really appreciate it if you would tell me how you would handle each of the four situations I listed above. WhatamIdoing (talk) 16:09, 15 September 2014 (UTC)
@Pschemp: you're mixing up two very different issues here. Images do not need sources in the sense of citations; they need sources to ensure that they are not in copyright, but these sources aren't subject to WP:RS. Thus I can't write a description of a species of plant based on my own observations, but must paraphrase a published source and cite it. But I can (and do) upload photographs of species of plant taken by me. If we could only use images taken from reliable sources, we'd have very few in Wikipedia, as they will mostly be copyright. Reconstructions of extinct species certainly need to make clear that it is only a reconstruction, and be accompanied by some reference to the origin of the information used to make the reconstruction, but this can't cover every last detail of a complex image. Peter coxhead (talk) 16:40, 15 September 2014 (UTC)
And by conflating photographs with drawings, you are avoiding the issue because, photos cannot cite peer reviewed sources, drawings can, and are required to by VERIFY. They are two different things. pschemp | talk 17:48, 15 September 2014 (UTC)
Of course photos could cite reliable sources; we could decide to use only photos previously published in such sources (and now copyright free of course). We don't because this would remove most available photos. The same applies to drawings. Peter coxhead (talk) 18:05, 15 September 2014 (UTC)
So where is such a distinction made? WP:PERTINENCE mentions both illustrations and photos. FunkMonk (talk) 17:56, 15 September 2014 (UTC)

Citing sources is a fundamental principle of Wikipedia. Everything that has a source, needs to cite it. This is not a difficult concept. pschemp | talk 17:50, 15 September 2014 (UTC)

And yet again, that applies mainly to the text, images (of any kind) follow different regulations, per Wikipedia's own policies. Repeating the same faulty argument over and over again does not help your case. FunkMonk (talk) 18:06, 15 September 2014 (UTC)
"Everything that has a source, needs to cite it" is not actually the rule. There are only four kinds of material for which sources are required. There's a quick summary of these four kinds at WP:MINREF. WhatamIdoing (talk) 01:13, 16 September 2014 (UTC)
Actually, that's exactly what WP:VERIFY says. "Readers must be able to check that Wikipedia articles are not just made up. This means that all quotations and any material challenged or likely to be challenged must be attributed to a reliable, published source using an inline citation." And, restorations are material that is so likely to be challenged, that they already has been. pschemp | talk 02:35, 22 September 2014 (UTC)
And again, that applies to text. You fail to acknowledge that there are different standards for text and images again and again. FunkMonk (talk) 13:40, 22 September 2014 (UTC)
There are not. You're simply citing the same non-policy document over and over, and it doesn't add up. You cannot pull a fast one by using an illustration instead of text to convey information. There is no circumvention intended, or permissible. That should be blindingly obvious to any observer. Samsara (FA  FP) 14:07, 22 September 2014 (UTC)
"You're simply citing the same non-policy document over and over" Because it is an actual policy, and not the arbitrary opinion of a random editor. You keep arguing in circles. Again, change the Wiki policies in question before proposing to change specific projects that follow those policies. FunkMonk (talk) 14:20, 22 September 2014 (UTC)

Simple summation across sources[edit]

I've been accused of SYNTH. What I think is ok, according to "Best practice is to research the most reliable sources on the topic and summarize what they say in your own words", is to sum up number of devices sold across sources. You can find say, Android devices sold for a year (or quarter) in a source but I find it really hard to find A source with a single number for longer periods. Similarily, I can find numbers for smartphones and tablets separately but as they run the same operating system (and have compatible apps) a combined total is useful information.

What get's murky is saying the aggregated numbers are "installed base". Those numbers (or estimates) are very hard to find in one number/source. And comparing installed base of Android vs. PCs. If one number is clearly higher than the other, would it be ok to say that one is more popular than the other? If both numbers come from the same source or different. comp.arch (talk) 10:43, 16 September 2014 (UTC)

@Comp.arch: I think this belongs at WP:NORN, not here as this page is for discussing changes to the policy page. Dougweller (talk) 11:45, 16 September 2014 (UTC)
Thanks, didn't know about the other place as it seems may be right place to ask questions. I however, think my understanding doesn not violate the guildelines and maybe the need to be clarified where the line is drawn. comp.arch (talk) 14:17, 16 September 2014 (UTC)


I've just written Wikipedia:Interviews, an essay to try and clarify the how interviews relate to Wikipedia's sourcing policies. I've tried to make it reflect policy and current practice as best I can, but there are bound to be things that need to be clarified further, or perhaps even flat-out errors, so I'm hoping to get others to look over it so that it can be as accurate as possible. If anyone here could take some time out to check it and/or improve it, I would be very grateful. — Mr. Stradivarius ♪ talk ♪ 11:59, 18 September 2014 (UTC)