Wikipedia talk:Verifiability

RfC: Do images need to be verifiable?

Prompted by a discussion at WP:RSN, in which several editors have asserted that images are not covered by this policy, I ask the question "Do images need to be verifiable"? This policy clearly says "This is strictly applied to all material in the mainspace". Should this be altered to specifically exempt images? Dlabtot (talk) 20:39, 6 July 2010 (UTC)

---

 

You may want to pose the question with more focus. Images have several "sourceable" elements: copyright and content. The latter, given the linked discussion, seems to be what you're really asking about. Or are you hoping to address the former as well? (For example, a Geocities site - i.e. self-published, non-expert - would likely not be acceptable at FA level to support prose of "1 000 tons of X were harvested in country Y in 1922". Would, however, a declaration on that site of "this image was published in 1922" be acceptable support of a PD-US license?) Эlcobbola _talk 21:04, 6 July 2010 (UTC)

This question is entirely unrelated to any copyright issue, which is why I didn't mention copyright. Dlabtot (talk) 21:10, 6 July 2010 (UTC)

• Comment. Traditionally, the mages have not been covered by WP:V, in the way textual information has been. Wikipedia specifically encourages the use of free images, particularly user-created images. To quote from WP:IUP: "Wikipedia encourages users to upload their own images". In fact, all of these user-created images are self-published, which applies to almost all images in the Commons. Their use on Wikipedia is extremely widespread and commonly accepted. Such images do not really fall under WP:SELFPUB rubric since that section of WP:V concerns the usage of self-published sources "as sources of information about themselves", which is not the case with user-created images (the creator of the image is usually not the subject of the article). When somebody uploads to the Commons the photo of a person or a building, we mostly rely on the good word of the uploader that the image is indeed of the person or the building that the uploader says it is. The same applies to the case where someone uploaded the photo they took of a plant or an animal. The de facto practice has been to allow the use of such user-created images and, for the most part, it works reasonably well. I think that in cases where there is a factual dispute regarding an image's content (rather than about it copyright), one has to go by the general WP:CONSENSUS rules rather than to invoke WP:V. Incidentally, the same principle applies to WP:OR, which, I believe, de facto does not cover user-created images (which, in the most literal sense, all constitute original research). Nsk92 (talk) 21:16, 6 July 2010 (UTC)

What the de facto practice is, seems irrelevant, as we are discussing what the policy should be. And we are actually talking about cases where the factual content of the image has been challenged. I don't understand what you mean by "one has to go by the general WP:CONSENSUS rules rather than to invoke WP:V" -- these two policies work together, and are not in any way in opposition. Everything on Wikipedia, without exception, is decided by consensus. Do you have a rationale as to why the verifiability policy should not apply to images? (If that is in fact your position - hard for me to tell.) Dlabtot (talk) 21:24, 6 July 2010 (UTC)

IMO the de facto practice is the correct one and reflects what the policy should be and in fact what it has been, in terms of how WP:V has been interpreted thus far. I have explained why requiring the traditional WP:V standards for images is counter-productive: this would disallow the use of essentially all user-created images (basically almost all of the stuff in the Commons). The benefit of allowing to use user-created images far outweighs the problems caused by the relatively small number of user-created images where there are disputes about their factual accuracy. Nsk92 (talk) 21:34, 6 July 2010 (UTC)

Well, no, applying the verifiability policy as written would not disallow the use of essentially all user-created images, unless essentially all user-created images were challenged or likely to be challenged. Dlabtot (talk) 21:37, 6 July 2010 (UTC)

The basic meaning of WP:V is to require that the information contained in Wikipedia articles be based on published reliable sources. User-created images are always self-published. Nsk92 (talk) 21:48, 6 July 2010 (UTC)

• Comment - This misunderstanding is caused entirely by the fact that WP:V and WP:NOR are separate policies. If they were merged into one, the policy on original images would be clear. Yaris678 (talk) 07:37, 11 July 2010 (UTC)

Long reply from WhatamIdoing

What does it mean for an image to be "verifiable"? Are you trying to verify that the contents usefully illustrate the article? (That's a determination you make with your very best editorial judgment, not a fact that you need to support with a reliable source.) Or that the caption in the article matches the description in the commons file? Or that the description in the commons file is accurate/not written by a liar or a hopelessly confused person? I don't really see how anyone can answer Dlabtot's question until it is clarified.

On 01 May 2010, I proposed an addition to WP:IUP to address this, and I recommend reading that. It developed out of several similar questions there, and one in which a denialist tried to exclude an image of an enveloped virus from an article (the story ran something like, "The virus does not exist. Therefore the image must be faked or mislabeled. I demand a high-quality scientific paper that specifically identifies this particular image as being this particular virus... even though an electron micrograph of one enveloped virus pretty much looks like another.").

There are two important, and separate, considerations: The first is whether we're willing to trust image uploaders, and the second is the value of illustrations.

In terms of trusting our image uploaders, the community has a long-standing practice of assuming good faith about image descriptions, and I think it's served us well. If I walk down to the city park, take a picture of the playground, and upload it with a basic description, nobody is going to say, "Where is your third-party, properly published, reliable source that proves this picture of a playground really, truly is a picture of a playground? Didn't you think to bring along a credentialed journalist to certify that this is a playground?" The image is, at some level, self-verifying (any non-blind person can determine that a picture of a playground is a picture of a playground). For anything more complex (e.g., that this image shows this particular playground, and not some other playground in some other town), the community's choice is ultimately between trusting me, and dumping the vast majority of free-use images. I think that most of us are generally satisfied with this trade-off.

(As a point of procedural interest, what the community's practice actually is is critically important, since actually practice is our primary source of policy. It is, in fact, the "real" policy, whereas this page is merely an attempt to describe the real policy. The most useful and most widely supported policy pages document actual practice, not what a couple of high-minded editors think the community ought to do.)

Injection of a clear fluid

The bigger problem is dealing with editors who mistakenly think that images do something other than illustrate the subject. We are not trying to provide documentary evidence that a thing exists or is done: We are trying to give readers some visual cues and help them better understand the subject matter.

For example, the image to the right could legitimately be used in several articles. We don't need to have an independent reliable source tell us whether it's insulin, saline, or something else in the syringe, because it doesn't matter. You could use this image just as easily to illustrate "Insulin injections are often IM" or "Morphine may be given by IM injection" or just about anything else (except, perhaps, cobalamin, which is intensely red). The point of including the image isn't to provide irrefutable proof that IM injections are done (that a virus exists, that quartz is a trigonal crystal, that the vehicle has hubcaps, or whatever the subject is); it's to give readers an idea of what an IM injection is. Even a completely "unverifiable" image can often serve this encyclopedic and educational purpose admirably well -- even if we later learn, for example, that the syringe in this picture was empty, or that there was no needle on the end of it.

On this point, something that is documented to be X, but does not actually look like X, is a poor illustration for an article about X. Something that may or may not be X, but looks exactly like it, is an excellent illustration for that article. We really do care about what the images look like, not whether they're "real".

I still believe that the best place to address this issue is at WP:IUP, not here. WhatamIdoing (talk) 22:14, 6 July 2010 (UTC)

WhatamIdoing has it right... images should never present information on their own or in a vacuum... they should, instead, illustrate the information that is presented in the text. That text is what needs to be cited. And, images that do not illustrate something in the text are useless. Blueboar (talk) 22:50, 6 July 2010 (UTC)

@WhatamIdoing: Thanks for that excellent outline, which I support. If there is ever an article consisting entirely of images, we will need to revisit this discussion. Meanwhile, articles present their information with text, and assertions in the article should be verifiable. An article should not say something like "because this picture shows a certain virus, it follows that the virus exists"; it should not do that because the point made is not verifiable. Images are used to illustrate articles, and disputed cases need to be resolved via consensus, as normal. Johnuniq (talk) 23:09, 6 July 2010 (UTC)

• About that image to the right. It's descriptive title is Intramuscular injection into the Vastus Lateralis. Combine that with the lack of a verifiable source. So now let's imagine the picture with the following caption: Intramuscular injection of a clear fluid into the Vastus Lateralis. Should the image be retained or should the caption be edited to reflect only its visual information? BruceSwanson (talk) 16:12, 7 July 2010 (UTC)

Your question presents a false dichotomy. The image should be retained, and a variety of captions are perfectly acceptable, including captions that go well beyond the visual information in the image.

We don't know that this image really is an injection of a clear fluid into the stated muscle: we see less than two square inches of skin (could be a different muscle, or even an arm instead of a leg). It could be an empty syringe, a different muscle, or a non-injection. But it looks like an intramuscular injection of a clear fluid into a leg, and so it could be used (as appropriate to the article's needs) with a wide variety of captions, including:

• "IM injection",
• "IM injection into the vastus lateralis muscle" (trusting that the uploader got it right),
• "IM injection into the upper leg" (not quite trusting the uploader's knowledge of anatomy),
• "In an IM injection, the needle is held at a 90 degree angle to the skin",
• "Insulin is commonly delivered with an IM injection",
• "IM injections of morphine can be used to control pain after surgery",
• "Inexpensive 'insulin syringes' are used to deliver a wide variety of medications, not merely insulin", and even
• "Images are not always what they purport to be. For example, this image might show an actual injection, or it might be an expensive theatrical prop, used to simulate an injection without actually puncturing the actor's skin."

Any of these (and dozens of other options) are acceptable, so long as the image is used to illustrate the article's contents. WhatamIdoing (talk) 17:28, 7 July 2010 (UTC)

Based on its visual information only, how many different captions could you come up with for this image? BruceSwanson (talk) 18:11, 7 July 2010 (UTC)

A fair few. But I'm not sure why that matters. Does the relevant article have text that says something like "the virus appears on an electron microscope as a black dot approx 10nm wide, within an annular feature of 50nm diameter and 10nm width", and can that statement be referenced? If so the image needn't be referenced. However there *are* cases where an image adds information independently of the text of an article, and which needs to be verified; for instance [1]. Trustworthy though the creator of that image is, someone quite properly raised a concern at FAC whether the process outlined in the image was verifiable. A source was easy to find, and was added to the description page. Compare also [2] and [3]; the former is a scan of an image from a reliable source and we can rely on it, while the latter is user-created with no source information, and there is every chance that the user who created it has introduced some inconspicuous minor error. The Land (talk) 20:04, 7 July 2010 (UTC)

The interpretation of some images, like The Blobs is entirely dependent on the context provided by their sources. If the source isn't verifiable, then such images have no meaning at all, do they?

From Wikipedia:IMAGE#Pertinence_and_encyclopedic_nature:

“ Their origin must be properly referenced. In the case of an image not directly attributed to its creator, (e.g., in the case of reproduction of ancient artwork or artifacts) it is not sufficient to merely indicate the image's immediate source, such as an URL, but the identity of the image's content (author, manuscript, museum id) must be given. ”

BruceSwanson (talk) 00:35, 8 July 2010 (UTC)

IMO that section of WP:IMAGES needs to be revisited. The next sentence says, "Images that are not properly identified, (e.g., images with descriptions such as "a cuneiform tablet", "a medieval manuscript", etc.) are unencyclopedic and hence, not useful for Wikipedia" -- which is simply wrong. An image about which we know nothing more than (the copyright status and) the fact that it is "a cuneiform tablet" is perfectly encyclopedic and entirely usable as an illustration (perhaps in the infobox) at Cuneiform script, with a caption like "Cuneiform tablets are made from clay" or "Cuneiform tablets were widely used for about 3,000 years". The same principle applies to a Medieval manuscript: we don't have to know any details to use it as an illustration of the generic concept.

Electron micrograph of HCV

As for what you call "The Blobs", here's a list of potentially acceptable captions:

• Transmission electron microscopy (TEM) can be used to produce images of very small objects, such as viruses.
• TEM produces flat, two-dimensional images, rather than the three-dimensional structures that are produced by scanning electron microscopy (SEM).
• TEM shows internal structures, rather than the external surfaces that SEM shows.
• Electron micrographs are black and white images.
• Micrographs of small objects are often fairly fuzzy.
• Including a linear scale in a micrograph helps the viewer understand the scale of the objects shown.
• An electron micrograph of a virus.
• An electron micrograph of an enveloped virus.
• An electron micrograph of an enveloped virus with a diameter of approximately 60 nm.
• These viral particles have a diameter of approximately 60 nm.
• The Hepatitis C virus has a diameter of approximately 60 nm.
• The Hepatitis C virus is an enveloped virus.
• A micrograph of one enveloped virus looks much like another; consequently, visual inspection is not used for identification.
• Enveloped viruses can be very small.
• The core of an enveloped viral particles is relatively small; most of the volume is due to the envelope.

Note that some of these statements require support (in the main article text), but all of them would be "legal" captions for this image. The choice of the best caption depends on the nature of the article and the aspect that editors are trying to illustrate. The best caption for use in Micrograph is not the best caption for use in Hepatitis C virus. WhatamIdoing (talk) 04:11, 8 July 2010 (UTC)

Support WhatamIdoing's reasoning and comment. Also note that policy is meant to document, not determine, consensus of community practice. If the community practice is different from the descriptions of the policy, the policy should change. Also note that, possibly as a run-around for this specific image, the Hep C EM has been nominated for deletion at commons [4]. What happens if an image uploaded to wikipedia is deleted from the commons? WLU (t) (c) Wikipedia's rules:^simple/_complex 10:45, 9 July 2010 (UTC)

I agree with BruceSwanson. Their origin must be properly referenced. There is not point in adding made up captions to images with dodgy provenance. If the image does not come with a useful or relevant description, making one up is little more than original research. --Gavin Collins (talk|contribs) 00:57, 11 July 2010 (UTC)

Is this a playground?

Gavin, can tell me how you'd do that? Imagine that you walked down the street and took a picture of a playground, much like the one on the right. You upload it to Commons, and you add it to as an illustration for an article about your hometown. What's your "proper reference" that says the picture of the playground is a picture of a playground? What's your "proper reference" that says this picture was taken in a given place? Do you really want editors to assume such bad faith that they declare you too be too stupid or too confused to know where you took the picture?

The community's policies are what we do; they are inherently practical. Can you give me a practical example of how you would provide a "proper reference" for an editor-created photograph like this snapshot of a playground? WhatamIdoing (talk) 19:47, 11 July 2010 (UTC)

I can think of one proper reference he would have already supplied: his name. BruceSwanson (talk) 07:40, 13 July 2010 (UTC)

Essjay controversy. Real names and alleged qualifications don't mean squat, particularly when advocating for a tenuous or fringe point of view. The fact that you allegedly use your real name doesn't make your contributions to AIDS-related articles any more credible, but your open advocacy for AIDS denialism does make them less. The fact that I use a pseudonym is irrelevant, since I can justify my edits using high-quality sources and appropriate policies. WLU (t) (c) Wikipedia's rules:^simple/_complex 14:55, 13 July 2010 (UTC)

Random break (verify images?)

The above discussion deals with photographs (or micrographs) only. Do the same principles apply to drawings? As well as necessarily being original research, a drawing necessarily has an element of creativity. Should the editor adding the image (not necessarily the same person as the uploader) be able to verify that the drawing is a reasonable representation of its subject? Take this image, which is the subject of a long and heated discussion on Talk:Pope Leo XIII. The coat of arms itself (the cypress, comet etc.) is heraldically accurate, but the surrounding bits, which I believe is called the shield, are a creative interpretation by the artist, using accepted heraldic elements. The artist says that the ornature is similar in style to coats of arms on the pope's tomb, in a vault in the vatican etc., but not that it is a faithful rendering of any of those. Now, per WhatamIdoing, the image would be fine in articles on heraldry, drawing or suchlike with a caption such as "a drawing of a coat or arms in neo-gothic style", but ought it to be used in an article about its subject without verification that it is a faithful representation of a specific, physical coat of arms? NB This is a genuine query about policy, not an attempt at canvassing. I am, of course, informing the participants in the discussion. Scolaire (talk) 20:13, 12 July 2010 (UTC)

A drawing is not necessarily original research. If, for example, I redraw a schematic diagram that appears in a book (because the copyright owner did not give permission to put an exact copy in Wikipedia) and my diagram is electrically equivalent to that in the book, there is no original research. If my diagram is laid out differently enough from the book, it isn't a copyright violation either.

In the case of heraldry, at least in England, what is granted by the sovereign is a blazon, and possibly other elements, expressed in words. The words are interpreted by various artists, and the result of each artist's effort is an achievement. There can be many different achievements for the same arms, and if no rules of heraldry were broken, they are all equally valid. Jc3s5h (talk) 20:51, 12 July 2010 (UTC)

I'm reluctant to add to the soup, but isn't your assertion of the electrical equivalence of the diagram unverifiable original research? —chaos5023 (talk) 21:17, 12 July 2010 (UTC)

Not at all. Of course, the description of the image on Commons should include a citation to the source. Anyone who understands schematics can compare the book to the version I draw and see they are electrically equivalent. Jc3s5h (talk) 21:44, 12 July 2010 (UTC)

I'll withdraw the word "necessarily" then, both for OR and creativity. An electrical diagram such as you describe would be neither OR nor creative. But if you then expand the circuit by adding extra resistors, switches or whatever, in order to illustrate a particular point in an article, it is no longer a faithful copy of the drawing in the book; should you not then be able to verify from a written source that that is how the circuit in question would actually work? Creativity would only arise if you used curved lines instead of straight lines to illustrate a resistor, or a heptagon instead of a circle for a voltage source. In the case of an "achievement" in heraldry (I was hoping to avoid technical terms) the scope for creativity is far greater. If no rules of heraldry were broken, they are all equally valid for heraldic purposes, but are they equally valid for the purpose of illustrating a Wikipedia article, or is there a point at which the creative element reduces the image's usefulness in providing information, in the absence of specific attributability (if that's a word)?

There is an analogous situation for photographs. In this picture, for instance, the creative element is not the photographer's, but the costume designer's. Again per WhatamIdoing, a caption such as "Man in Tombstone, Arizona, wearing a stagecoach driver's costume" should not need to be backed up by sources, but could you really caption it "Typical stagecoach driver's costume of the 1880s" without a reliable source? Scolaire (talk) 09:16, 13 July 2010 (UTC)

It seems a written description is not enough for you because you want style cited also, and that linking to images is not useful unless the illustration exactly matches the sourced image (this is an issue of copyright, though). It seems by your reasoning, an editor could always request to source something or anything further back just a little more. Sure, so and so used it but heraldry is hereditary, prove his father used it, then his grandfather and so on. And how would one source a style to heraldic displays? How can we know if a painting shows Jesus being crucified and not just some bearded man? Because some museum says so? And where is their source? And even if their source can be traced to the original patron, how can we know that is what the artist really had Jesus in mind? There is a requirement that an editor provide sourcing, yes, but it is also incumbent upon the challenging editor to read and understand what is presented and the sources given. An argument of, "I don't know therefore I should challenge it" is simply a waste of time and resources. After all, how can you prove something to someone that can't understand what he is looking at? [tk] XANDERLIPTAK 11:55, 13 July 2010 (UTC)

In the case of electrical diagrams, the diagram should either illustrate the article text (which is properly sourced), or be an electrically (but not necessarily artistically) faithful copy of a diagram in a reliable source. Since we do not have the same standard of verifiability for images that we do for text, we can require that any editor challenging the image to have the skill to prepare a schematic diagram from an unambiguous text description, just as we can require that any editor who challenges an x-y plot of some data in the text have the skill to make plots.

In the case of heraldry, one would expect a person who possesses arms to have several different achievements which would be used for different purposes. Ordinarily one achievement is no more valid than another. A challenge to a particular achievement in Wikipedia could be a verifiability challenge (i.e. the achievement does not correspond to the blazon, which is sourced) or could be on artistic grounds (e.g. it is ugly, the elements are drawn in such an unusual way that many people with knowledge of heraldry would not recognize them, etc.) Jc3s5h (talk) 14:44, 13 July 2010 (UTC)

I would indeed expect a person who possesses arms to have several different achievements which would be used for different purposes. For any given drawing, then, I would simply ask, which of the documented achievements does the drawing reproduce? Verifiability, see? Again, the analogy with the electrical drawing is not a good one. If my expertise is in history, and my concern is with historical accuracy, you cannot "require" me to have the skill to prepare a heraldically accurate image, or to argue my case in heraldic terms.

Xanderliptak's crucifixion analogy, on the other hand, is a very good one. If I were to upload a visually stunning drawing of the crucifixion, in which Christ is depicted as blonde, short-haired and beardless, this would not, as far as I know, contradict anything in scripture. But could I hope to replace existing images in a dozen articles including Jesus Christ, Christianity, Catholicism and Protestantism without being challenged on verifiability? Or could I "require" any editor challenging the image to have the theological skills to prepare a depiction from an unambiguous text description? Scolaire (talk) 18:40, 13 July 2010 (UTC)

What must be verified depends on what is claimed. If the text claims that a picture represents the arms of Duke X, it is merely necessary that the picture conforms to the blazon. If the text claims the picture represents the arms that appear on the set of silverware Duke X purchased in 2008, then there should be some kind of citation that demonstrates a close artistic correspondence between the silverware and the picture.

As for pictures of the crucifixion of Jesus, they are unverifiable in the normal meaning of the word. The picture should be accompanied by some verifiable statement that makes it clear why that particular image should be included in the article. Jc3s5h (talk) 19:00, 13 July 2010 (UTC)

I think Jc3s5h is on the right track. Images need to usefully illustrate the article contents. An image that doesn't usefully illustrate the article (e.g., when, through a quirk of styling, lighting, or other feature of the image, the image doesn't look recognizably like the subject of the article) should not be used.

In the case of what Bruce calls "The Blobs", anybody who's seen a handful of electron micrographs for enveloped viruses -- or who is willing to ask his or her favorite web search engine for more information -- will look at this one and conclude that the image is basically what the Hepatitis C virus should look like, assuming, of course, that you agree that HCV exists, and that the well-sourced physical description of the virus in the text is accurate. (You might also decide that it's not an especially high-quality micrograph by recent standards.) This image looks like HCV, and therefore is a potentially useful illustration in the HCV article.

By contrast, an image that didn't look like HCV, even if the authenticity was attested to by a dozen stellar sources, would be far less useful. It doesn't matter how many sources say that File:APEC Police Helicopter, Opera House, 2 Sept 2007.JPG shows a particular model of police helicopter: it's not a useful illustration for police helicopter. Images primarily need to be useful, not verifiable.

Scolaire, if you want to include images of a blond Jesus (an artistic conception that Albrecht Dürer was famous for), you wouldn't get a verifiability challenge (unless your caption said something patently unverifiable, like "This 20th century painting proves that the historical Jesus had blond hair and blue eyes"). Jesus Christ currently has two blond-haired images in it. I doubt that anyone has ever challenged the images as "not being verifiably accurate depictions of the historical Jesus". WhatamIdoing (talk) 19:01, 13 July 2010 (UTC)

Blond, short-haired and beardless? Where? Scolaire (talk) 19:08, 13 July 2010 (UTC)

File:Adoration of the shepherds reni.JPG: blond, short-haired, beardless. (You didn't specify "adult". ;-) WhatamIdoing (talk) 18:39, 14 July 2010 (UTC)

(edit conflict)

See the archived discussions.LeadSongDog come howl! 19:21, 13 July 2010 (UTC)

There seems to be a confusion over what verifiability means. It looks like Scolaire is not asking for sources that shows something is true, but is asking for sources that explain why and how something is true. Were an artist like Albrecht Duerer or Michelangelo Buonarroti to have their images of Jesus placed in articles, Scolaire seems to believe that because their depictions do not offer up the brown haired and bearded man, that they somehow fail or need additional sourcing other images do not. We do not need a letter form the artists stating that is is Jesus and this or that is why hey made him as they did. If someone wishes to exclude the images on other grounds, fine, but not on verifiability. [tk] XANDERLIPTAK 03:36, 14 July 2010 (UTC)

Fact is, I asked a straightforward question regarding creativity and verifiability. I have tried to avoid getting bogged down in detail over a single example by repeatedly giving new and very different examples, but each time the response focusses on the fine details of my examples instead of on my question. I am also detecting a certain amount of hostility which I don't really understand (except in the case of Xanderliptak whom I specifically asked elsewhere not to personalise the discussion). Now, my question was:

• Is there a point at which the creative element reduces the image's usefulness in providing information, in the absence of specific attributability?

Here is a selection of answers:

• we can require that any editor challenging the image to have the skill to prepare a schematic diagram from an unambiguous text description
• The picture should be accompanied by some verifiable statement that makes it clear why that particular image should be included in the article.
• Images primarily need to be useful, not verifiable.
• See the archived discussions (a link to eight talk page archives of ~100kb each about what Jesus looked like).

These answers are all very much at variance with each other, and I can't really see where any of them is policy-based. I really would like to see a more generic and more coherent (and less dismissive) answer to what I think is a pertinent question to the current discussion on verifiability of images. Scolaire (talk) 06:25, 14 July 2010 (UTC)

Unless the image is a copy, or what one might loosely call a "visual paraphrase" of a reliable source, I believe it should be apparent to a person with a reasonable amount of skill in the subject matter of the article to see that the image correctly illustrates the text of the article and the caption of the image. Jc3s5h (talk) 18:21, 14 July 2010 (UTC)

I think the reason so few people can answer your question is that the question itself is premised upon a misconception of why we add images in the first place. Images should not be added to "provide information"... they should be added to "illustrate information". In my opinion, an image should not be added to an article unless it illustrates something that is stated in the text of the article. It is that text that needs to be verifiable, not the image that illustrates it. Now, we can argue that a particular image does not accurately illustrate what is stated in the article and should be removed or replaced with a better image, but that is a different issue. Blueboar (talk) 18:39, 14 July 2010 (UTC)

Okay, that's a good jumping-off point. So, the examples so far:

• The coat of arms: this illustrates the blazon (have I got that right?) correctly, but the extraneous detail surrounding it distracts attention from the thing being illustrated (the blazon). If the "achievement" was verifiably that of a specific coat of arms on a wall in a specified room of a specified building, only then would it usefully illustrate that additional fact.

Injection: A coat of arms is the colloquial term for a heraldic achievement, and both refer to the pictorial representations of the blazon but not to any specific depiction. A specific pictorial representation is referred to as an 'emblazon'. However, with an heraldic achievement (coat of arms), any pictorial representation is fine unless one is trying to reference a specific emblazon. The same say any image of a Pieta is fine to represent the Pieta, unless you wish to depict Michelangelo Buonarroti's famous example in particular. [tk] XANDERLIPTAK 01:08, 15 July 2010 (UTC)

• The stagecoach driver: This illustrates the paraphernalia of a stagecoach driver of the 1880s, but the brightly coloured shirt (if we ignore the shades) distracts attention from the thing being illustrated. If the shirt was verifiably copied from a shirt in a Wild West museum, only then would it usefully illustrate that additional fact.
• The hypothetical self-drawn crucifixion picture: That would illustrate the crucifixion as well as any other image, but the unconventional (short-haired, clean-shaven) Jesus figure would distract attention from the thing being illustrated. If the figure was verifiably typical of a particular school of artists, or religious sect, only then would it usefully illustrate that additional fact.
• Jc3s5h's hypothetical electrical diagram does not have any extraneous detail or "creative" element, so verifiability does not arise.
• WhatamIdoing's playground or "injection of a clear fluid" likewise have no extraneous element, but if there were, say, a Samurai warrior in the playground or a pink ribbon around the syringe, we might reasonably ask what those extra elements were illustrating and whether they were verifiable.

Would you agree? Scolaire (talk) 20:26, 14 July 2010 (UTC)

No, I don't think that we'd ask whether these elements were "verifiable". We'd ask whether -- in our own best editorial judgment -- they were appropriate illustrations of the already-sourced text. If the particular achievement is described in the text, and the image looks like what the text describes, then there are no "additional facts" to be verified. If the text says that stage coach drivers wore colored, printed shirts (or, more plausibly, leather cuffs), then the image correctly illustrates the text. A self-drawn crucifixion picture might be challenged as less appropriate, interesting, valuable, or encyclopedic than a famous artwork, but if you've got a body dying on a cross, nobody's likely to say, 'Yeah, well, can you prove that being hung on a cross is really crucifixion?' As with the playground, extraneous elements might make the image less appropriate -- but a playground's still a playground, even if a Samurai warrior is standing in the middle of it. We don't need to find a reliable source to "verify" that the Samurai warrior was present at the playground; we need to ask whether that's the best choice for illustrating the concept or existence of the playground.

Note that the locus of the decision is what we're debating: The community's actual practice (and therefore real policy) is to empower you, the individual editor, to decide whether to choose an image of a playground with or without the Samurai warrior. You get to do this with your very own best editorial judgment, rather than relying solely on whether you can find a source that says there's an annual Japan Day festival in Central Park. WhatamIdoing (talk) 21:18, 14 July 2010 (UTC)

I agree with WhatamIdoing, these are not questions of verifiability, but are editorial decisions. If you have two images, say a playground with a samurai and one without, and there is not a specific playground being described in the article, then it is an editorial decision on which image to use. Now, if this samurai is a child dressed up in a costume, that image may best be used to show both what a playground is and what they are used for. If it is just to illustrate what a playground is, then go with the samurai-less one. If you have an image of a flu vaccine being given, but the nurse is in the way and all you really see is the nurse's back, then it might be better to use an image of another vaccine being given to better illustrate the process. Simple diagrams or images are best suited for engineering, medical, chemical or other scientific subjects, where artistic license could confuse viewers with excessive embellishments. If the article is about an historical figure before the invention of photography or is about something artistic, like the coat of arms, then it is likely best to use the most aesthetic and artistic illustration, or a contemporary one if possible for historical weight. [tk] XANDERLIPTAK 04:38, 15 July 2010 (UTC)

"Editors think"

There seems to be some confusion about recent changes to the policy. In this edit, the policy was altered to remove the words "editors think". Wikiblame tells me the term "editors think" has had that position in the policy since February 2010. Before that, it said, "we think". I believe that people are perhaps confused and thinking that User:Jwy is trying to implement a change to the policy; in fact, he's challenging a change. I've put it back given that evident confusion, but it seems that the phrase is contentious; perhaps a discussion is in order? --Moonriddengirl ^(talk) 20:05, 7 July 2010 (UTC)

I agree with Moonriddengirl's edit/revert or what ever it is. Whether the specific words are a change or a return to old language does not matter to me... what matters is that the edit accurately states the policy. It does. Blueboar (talk) 22:56, 7 July 2010 (UTC)

Self-published sources

This section appears to create a loophole when it states, "Self-published material may in some circumstances be acceptable when produced by an established expert on the topic of the article whose work in the relevant field has previously been published by reliable third-party publications". Many leading political commentators have academic credentials, but use blogs and magazines to write opinions that would not be accepted by literary journals. For example, if a political leader is a scholar on government (and there have been many), we should not treat his comments on his party website on the government of the day to be of equal weight with an article he subjected to peer review. TFD (talk) 21:08, 7 July 2010 (UTC)

You are correct that we should not give such self-published material equal weight with peer reviewed material... but giving something less weight does not necessarily mean giving it no weight at all. The key to including such material is to phrase it as being an opinion, and not presenting it as accepted fact. Blueboar (talk) 12:19, 8 July 2010 (UTC)

There's a category problem here: Things like political (also cultural, economic...) opinion, identification of trends, estimates are not likely to be peer-reviewed because they are not factual but personal. Weight is given to them based upon the position of the person or organization making them. Verifiability here is making sure that what was attributed to them is accurately presented here, and often that's an easy if not trivial task. A common dispute among editors is whether the person has some standing (i.e. self-evident authority) or audience (i.e. self-evident readership/listenership) that makes their opinion, analysis, etc. appropriate for the article and that's not a verifiability dispute. patsw (talk) 13:00, 13 July 2010 (UTC)

Bolding

I think it is poor form to bold the second half of this sentence:

Academic and peer-reviewed publications are usually the most reliable sources where available, such as in history, medicine, and science, but material from reliable non-academic sources may also be used in these areas, particularly if it appears in respected mainstream publications.

Bolding the sentence alters its emphasis. As written, it emphasizes the reliability of academic sources (while qualifying that other sources may also be acceptable). By bolding the second half of the sentence, we've changed its emphasis in a way that conflicts with its structure. More to the point, the sentence is clear enough without heavy-handed markup.

I removed the bolding here. It was restored 2 minutes later by SlimVirgin, with an edit summary stating that "we need the bold because the first part of the sentence is being adhered to, the second part ignored." I have not seen systemic problems of this nature which demand changes to core policy. What specific instances is this intended to address? Do other editors besides SlimVirgin see this as a systemic problem requiring us to mark up core policies? We should probably resist the temptation to go and bold the particular subclauses of policy which support our position in specific content disputes. MastCell ^Talk 23:57, 8 July 2010 (UTC)

I agree with Mastcell. I see no reason to bold one part of a sentence, implying that other parts of the policy are somehow subservient to this one part. If emphasis on that one part is needed in particular circumstances, the editor involved can bold it him/herself when quoting the policy. There's no need to bold the phrase on the main policy page. LK (talk)

Agreed. I think that bolding should be reserved for brief encapsulations of the policy, like verifiability, not truth or who has the burden of providing references. We should use ordinary language to avoid misunderstandings of policy, not shouting. RJC ^Talk_Contribs 00:19, 9 July 2010 (UTC)

There have been recent statements on the CC workshop that show some editors believe high-quality reliable sources can be excluded from articles just because they're not scholarly sources. This is quite wrong, and there's nothing in the policy that implies it. Indeed, the sentence you're focusing on was written after considerable discussion to make clear that non-scholarly sources had to be included. Yet the second part of the sentence—after "but"—is being ignored by certain editors when it suits them:

Academic and peer-reviewed publications are usually the most reliable sources where available, such as in history, medicine, and science, but material from reliable non-academic sources may also be used in these areas, particularly if it appears in respected mainstream publications (emphasis added).

Something therefore needs to be done to correct the misinterpretation, either bolding or new wording, but bolding is obviously simpler. SlimVirgin ^{talk|contribs} 00:56, 9 July 2010 (UTC)

Do they maintain this even after being pointed to this sentence in the policy? At some point, you cannot reword or highlight a statement to counteract illiteracy, since "but" can function only one way in that sentence. I'd be interested to see what was being said. Was it a matter of banishing information supported only by a non-academic source, or of different information's being present in academic and non-academic sources? RJC ^Talk_Contribs 01:25, 9 July 2010 (UTC)

RJ, yes, they know what this policy says, and they're ignoring the second part of the sentence in order to prioritize scholarly views. The particular issue on the climate-change arbitration case was, of course, global warming. There is a big dispute with scientists on one side (for the most part), and other commentators on the other. By misreading this sentence, there was an effort to exclude the non-scientific perspective, including material in The New York Times. But it's not happening only on climate-change articles. Religious editors want only biblical scholars to be used, engineering editors want only engineering sources etc etc. There's an intellectual apartheid developing as each group seeks to promote the sources he agrees with and leave out dissenting voices.

I don't want to bold the sentence for the benefit of those editors. The people I have in mind are the less experienced editors who find their reliable sources being rejected because they're not specialized enough. I would like them to be able to come to the policy and see at a glance, without having to search for it, that if their source is The Times of London, it's welcomed by this policy. SlimVirgin ^{talk|contribs} 01:48, 9 July 2010 (UTC)

That sounds like trying to use WP:RS to sidestep the debates that would arise regarding WP:NPOV, WP:UNDUE, and WP:FRINGE. They are reliable sources, but competing reliable sources get us into the problem of point-of-view. Perhaps we could add some language to that effect. Regarding the biblical-scholars-only problem, most biblical scholars aren't going to support what religious editors want. Is the problem one of theologians refusing to accept "religious studies" sources, or are they rejecting pop-atheism sources? RJC ^Talk_Contribs 03:34, 9 July 2010 (UTC)

I'd rather not focus on any particular issue as an example, RJ, because this is something I keep noticing. There is a tendency for editors to want to exclude material that doesn't fit within the specialism they focus on, even if the material has been published by high-quality sources. The policy makes clear that all high-quality reliable sources are welcome, but people are reading one part of the key sentence and managing to ignore the other. :) SlimVirgin ^{talk|contribs} 15:43, 9 July 2010 (UTC)

First of all, I don't see people at the climate-change Arbitration case saying that the New York Times can't be used as a source, so the stated basis for this change seems either hyperbolic or non-existent. Either way, a diff might help. More importantly, if the concern is that editors in an Arbitration case are making erroneous statements about policy, then please address it at the Arbitration case, with the editors in question and with the Arbs - not by going to a policy page and artificially emphasizing the half-sentence that you believe supports your argument. MastCell ^Talk 05:33, 9 July 2010 (UTC)

The issue is not just CC editors — that's just an example. There are many other examples, in many areas, which I have seen myself over the years, where editors become confused and assume that scholarly sources always trump news media or other popular sources, to the point of excluding the latter kind. Addressing this in one specific arb case, or otherwise in 20 different places, makes no sense — this is exactly why we have policies, so we don't have to repeat ourselves and can have centralized discussions and standard rules we can point to. Crum375 (talk) 05:54, 9 July 2010 (UTC)

Stylistically, this bolding is far from optimal; I almost reverted on those grounds alone before remembering to check the talkpage myself before requesting justification. As far as usefulness to the policy, I am really not seeing the necessity. It may be perfectly appropriate in a particular situation to add emphasis when quoting policy at another editor, but I think that the point for the general case is better served by the unformatted text. - 2/0 (cont.) 06:54, 9 July 2010 (UTC)

I just discovered SlimVirgin's statement at the Arbcom case on Climate Change. It appears that his/her edit to this policy page is a continuation of the same dispute, and an attempt to change policy to make a WP:POINT or help his/her case before Arbcom. I find this unacceptable, and will oppose this change on principle alone. LK (talk) 10:07, 9 July 2010 (UTC)

That is nonsense, LK, and I'd appreciate some AGF. I made clear when I first posted above that I'm responding to statements during that case by others, where they state clearly that they're interpreting V to mean that scholarly sources are prioritized, which is a false interpretation. If people can interpret the wording of the policy that way, it has failed. SlimVirgin ^{talk|contribs} 15:48, 9 July 2010 (UTC)

The bolding, if any, should be limited to the words may be. Even unbolded that clause is of dubious value. The choice of publication alone does not establish the reliability of a source. A first tier general news publisher such as The Times still has content that escapes thorough fact checking, particularly its reader letters and advertising. That is what the peer review process and the publication of review papers are for. When we resort to general news media on scientific topics we should do so only with considerable caution. The underlying principle must remain that we use the best sources available in the subject area. LeadSongDog come howl! 12:57, 9 July 2010 (UTC)

Academic sources should be given priority, but high-quality non-academic sources are also acceptable. But they are second tier, second choice, second fiddle, second to appear. They should appear as attributed statements, but should still appear. I see this as quite basic, the kind of interpretation any experienced editor should be able to make, understand, point out and accept. The bolding seems unnecessary; the point is clearly made and anyone who attempts to use it to stifle the appearance of pop sources can, as MastCell says, be rebutted by emphasis in the quotation. WLU (t) (c) Wikipedia's rules:^simple/_complex 13:20, 9 July 2010 (UTC)

There is nothing in the policy that says non-academic sources are second choice. Sometimes they will be, and sometimes not. To imagine that we'd prioritize some minor paper by some minor academic over coverage by multiple high-quality news organizations is clearly false. SlimVirgin ^{talk|contribs} 15:48, 9 July 2010 (UTC)

I agree with SlimVigin... both academic and non-academic sources of equal quality should be given equal weight. Blueboar (talk) 16:04, 9 July 2010 (UTC)

That sounds like saying that equals should be treated equally: fine enough, but it doesn't tell us what makes things equal. In general, peer-reviewed sources are more reliable than those with only editorial oversight. "Academic and peer-reviewed publications are usually the most reliable sources where available," with the caveat that non-academic sources can be reliable as well (not that they are just as reliable). There may be exceptions to that rule, but that does not make it any less the norm. RJC ^Talk_Contribs 16:10, 9 July 2010 (UTC)

I agree with Slim. There are several metrics regarding reliability and suitability of a source in a particular situation, (all of the general wp criteria, plus all of those in the context of the particular statement that they are used to support.) We have to reduce any attempts by someone to use one policy sentence to pass judgement on or exclude a source based on just ONE (e.g. academic vs. non-academic) of the MANY metrics. While bolding is an unusual way to do this, it shouldn't be taken out unless immediately replaced by another method. (e.g. wording) North8000 (talk) 16:19, 9 July 2010 (UTC)

There may be a broader issue here that the bolding attempts to correct, but in an unclear manner. Just what is the disparity between academic and nonacademic sources? I agree with the flat statement without bolding - academic sources tend to be better, but nonacademic are not excluded. If we need to clarify academic versus nonacademic, we should do so through wording, not through emphasis - but first we need to clarify if there is support for academic over non.

And I still believe the bold is unnecessary, distracting and undercuts the very intent of the sentence - reliability is paramount and academic sources tend to be more reliable. But that is tend to be not an absolute. WLU (t) (c) Wikipedia's rules:^simple/_complex 16:47, 9 July 2010 (UTC)

It seems there are a couple of ways to further emphasize the acceptability of non-academic sources. I don't know that the relative worth of them is a WP:V/WP:RS issue so much as a WP:NPOV/WP:UNDUE issue, since I thought of verifiability as a minimum threshold test rather than as a scalar quantity. One way to make the statement easier to find would be to have it as its own paragraph rather than buried in a statement about context.

Academic and peer-reviewed publications are usually the most reliable sources where available, such as in history, medicine, and science, but they are not the only reliable sources in such areas. Material from reliable non-academic sources may also be used, particularly if it appears in respected mainstream publications. Other reliable sources include university-level textbooks, books published by respected publishing houses, magazines, journals, and mainstream newspapers. Electronic media may also be used, subject to the same criteria.

The concern over using WP:V and WP:RS to exclude material that someone argues is "less reliable" could be addressed by adding the following to the end of the final lede paragraph (which begins, "Verifiability is one of Wikipedia's core content policies ...")

The other content policies govern the weight that should be given to various views, not verifiability. That information should be verifiable is the minimum threshold for inclusion in the encyclopedia.

Or something along these lines. RJC ^Talk_Contribs 16:55, 9 July 2010 (UTC)

I'd be fine with having it in its own paragraph, RJ, if that would highlight it. There also used to be a sentence there about the importance of making sure all majority and significant minority positions are included, but it may have got lost during a copy edit. It might be worth restoring that too. I'll take a look for it. SlimVirgin ^{talk|contribs} 16:59, 9 July 2010 (UTC)

The proportionate representation of various viewpoints is covered under WP:NPOV; I'm not sure it needs to be duplicated here, particularly as they'll inevitably get out of sync. I don't really have a strong feeling about whether or not to use a separate paragraph. I guess I think the previous, non-bolded wording is amply clear - it categorically states that non-academic sources may be used - and it seems to have been relatively stable, so I don't see the need to change anything. I guess I could be convinced by actual examples of problems that would be solved by these proposed edits, but so far I've only seen testimonials. MastCell ^Talk 17:09, 9 July 2010 (UTC)

I agree, NPOV covers significant minority opinions in WP:UNDUE. What might be worth including in general is something that highlights the fact that all of these policies need to be read together. I see RJC's point - WP:V gives the first threshold - did someone say X? Then RS tells gives the next - is their opinion worth including? Finally, NPOV (the most finickity of the three) is about interpreting that source with an eye to proportionality, context and relative merit. Easily the hardest to do and the one requiring the most judgement. The types of sources will in large part depend on the topic - you'll prefer textbooks over newspapers for a drug/history/physics article, but that's an absurd requirement for actors/singers/TV shows. NPOV is far more about context than V. WLU (t) (c) Wikipedia's rules:^simple/_complex 18:22, 9 July 2010 (UTC)

Perhaps I'm rigid and overly hierarchically-minded, but I treat RS as an explanation of V, namely has V been met (since unreliable sources can't be used to verify something). V: is there a view that we can verify? NPOV (with its explanatory UNDUE and FRINGE): how much should that view figure in the article? RJC ^Talk_Contribs 18:39, 9 July 2010 (UTC)

I agree with RJC, but in practice many views/sources get blocked from inclusion into articles based on a position that could not possibly be justified based on WP:UNDUE. This sort of gets into one of the fundamental problems in Wikipedia: when there's mixed support for including a prima facie reliable source (say, a review published in a journal indexed in its field like a medical journal in PubMed), how is the issue resolved? In standard Wikipedia practice, you might start with a third opinion, maybe hit up a noticeboard or even move onto a RfC, but very often there is still mixed support. Thus, the person(s) who are the loudest wins. That's how it often it is in "consensus" based systems, particularly ones where the consensus is not governed by some authority who can put pressure on the other people. Thus you see the trend towards removing people who disagree with you permanently from the discussion, since that's the only way consensus-based systems can deal with these situations. Compromise is nice, but it's hard to compromise when the more aggressive side won't even allow any inclusion. I recently encountered this over Presidency of Barack Obama where a couple editors refused to allow a GAO report on ARRA's transparency initiative in a paragraph which is only about ARRA's supposedly unprecedented transparency. A couple other editors support it, but given the extreme diligence and borderline ownership of the editors opposed to inclusion, there's probably no hope. I also ran into this at Efficient markets hypothesis when removing what was obvious WP:SYNTH. My perspective is that Wikipedia aims to be comprehensive and should defer to including prima facie reliable sources and their perspectives to the extent that these perspectives are not duplicated, and the WP:UNDUE should almost never be a question when the reliable source hurdle has been passed. II | (t - c) 00:50, 17 July 2010 (UTC)

Repeating the same citation in different sections of an article

Having come across citations of the same source in different sections of an article recently, I just wanted to make sure that this is annoying rather than necessary. It is true, the various occasions the source was used were relevant to the sections concerned, however surely there is no need to repeat the practice time and time again and one citation would suffice?

What exercises me is, that the editor concerned has really done some sterling work in their recent edits and I wouldn't wish to discourage them continuing to do so. But it is just this one point I wanted to get the opinion of others on this.

Is there a particular way to cite a subsection of the same source without repeating the whole source every time? Dieter Simon (talk) 00:20, 10 July 2010 (UTC)

I'm not really sure if I understand the question, and would have to see the article itself. If you are talking about the same source being used multiple times in the same article, this is encouraged - the <ref name = > tags are expressly for this purpose. It seems obvious to me that you would do this, if you have a source that is useful in multiple sections. For instance, if I have a well-received, critically acclaimed book X, written by renowned scholar Y, written about notable topic Z, and the article is about Z, it is quite reasonable to expect article Z to cite the book repeatedly. See, for instance, Rosalind Franklin, which cites The Dark Lady of DNA by Brenda Maddox dozens of times and quite appropriately. Note that in this case, it cites the book as a reference and uses an Author, Page notation throughout the footnotes. There are other ways of doing it, but if I understand your question correctly, this is a common practice. Depending on how controversial and/or disputed the material is, and how thorough and complicated the topic, you may end up using several citations in even the same sentence (but then WP:OR can be a problem). WLU (t) (c) Wikipedia's rules:^simple/_complex 00:46, 10 July 2010 (UTC)

Yes, thank you, WLU, it is what I wanted to know. Yes, of course, the article is Ragged School, but what you said anwers my query. Dieter Simon (talk) 00:58, 10 July 2010 (UTC)

In my opinion, that page is pretty OK. Many of the citations could be collapsed by the use of <ref name = >, personally I would use the {{cite web}} and {{cite book}} templates, many of the references to books need page numbers, and there are a couple unsourced (or unclearly sourced) sentences and sections (Legacy in particular lacks any source) but it's certainly far from terrible. None are world-breakers, most are simply stylistic. I've made some of the changes I mentioned. WLU (t) (c) Wikipedia's rules:^simple/_complex 02:12, 10 July 2010 (UTC)

I agree with you, it isnt bad at all. Thanks for the useful pointers. I must take another look at the guidelines. Dieter Simon (talk) 16:48, 10 July 2010 (UTC)

Your opinion required

Need your interpretation of what I believe to be an exceptional claim. Keep or remove? Please see Talk:Chinese armies (pre-1911)#What to do with this claim?. Gun Powder Ma (talk) 20:11, 21 July 2010 (UTC)