Proposal talk:Verifiability policy considered harmful

From Strategic Planning

Philosophically compatible alternative

Would this be a better proposal as Develop systems for accuracy review? 18:09, 14 August 2009 (UTC)Reply[reply]

I think the two go hand in hand, and have expressed as such on the accuracy review talk page. Thanks! —Preceding unsigned comment added by (talkcontribs) --Fasten (Wikinews: Aktion Deutschland Hilft asks for donations after the earthquake in Indonesia) 13:10, 20 October 2009 (UTC)Reply[reply]


This policy proposal is attempting to resolve an incessant problem on Wikipedia. As noted many people think that just because they can go somewhere on the web and find a page that corroborates what they believe to be true, then the “fact” is verified. I have come up against this on such a regular basis, that I have substantially reduced my contributions just so I do not have to deal with this.

The verification policy is not the problem. It and every other Wiki core principle is very thorough, once you get past the title. The principles together form a solid basis for edits and they are not that different from those used to write scholarly works. The problem is the poor manner in which information is synthesized and the slip-shot manner the core principles are used to defend an editor’s POV, not the content of the information itself. Once you back that POV with some form of administrator power, even in a non-controversial issue, then you have problems.

I think we can arrive at the purpose behind this proposal without actually having to revise the existing verification policy, although it might be re-worded clearer. (Anyway, do you really believe changing/adding “truth” to the policy would really help? I do not.).

For example, the existing policy might have some instructional pages taken from actual incidents from Wikipedia showing what is appropriate and what is not. These pages can be very thorough, taking an incident and incorporating actual discussions, the logic used as proof, the items used as references, etc. We might also be able to use technology as a way of developing a checklist or means test for the inclusion of material. For example, I claim fact "A"; do I have several references not related to each other; are these references by an authority from the field of study from which I claim the fact; how many of these references are published materials; are they scholarly publications; does this claim constitute a claim from the "body of human knowledge"; etc.

I have so many "not so funny stories" of just how this issue can go wrong.GMJ 18:56, 15 August 2009 (UTC)Reply[reply]

I don't think that adding the word truth to the policy would help, but I do think that the policy is inadequate. I mean, who cares of "Jane Smith" got an article about "contentious topic" published in "respectable publication". Does she know anything about "contentious issue"? Did her article cite sources for her research? Does "respectable publication" specialize in material about "contentious topic". It doesn't seem to matter, as long as her remarks were published, they're in.
What is needed, I think, is a more clearly defined reliable source policy to back up V, and a process whereby neutral members of the community can vote yea or nay on a particular source (such as AFD, FA, FP, etc) wrote this unsigned post in 11:11, 18 August 2009, signature added by User:Rursus afterwards
I agree with GMJ. The verifiability policy is not causing the problem of some policy head-thumping wars between biasly minded hotheads. Those hotheads are a few persons that wasn't filtered away by the policy as intended. The troubles will be much worse if that policy is dropped, it will most likely be a full scale war followed by a general desertion of seriously minded editors. There are other policies about not disrupt wikipedia to prove a point, and about not using the policies as weapons, that are expected to filter away trolls using verifiability as a weapon. Rursus 14:14, 18 August 2009 (UTC)Reply[reply]

"Truth from Verifiability" instead of "Verifiability, not truth"

I think that the current verifiablility policy is now a mistake. As it states on the standard is "verifiability, not truth". I think that the standard should be "truth from verifiability". Which would you rather have, an encyclopedia that was always true, but lacked any sources and direct ways to verify it, or an encyclopedia that was had a source for everything, but was always wrong?

For example, in the first, it would have "The Earth orbits around the Sun." In the second, it would have "The Earth is motionless.[1]" When people come to wikipedia, they want the best truth that Humans know. The only reason that we care about verifiability is that we believe that if we can't verify the information, it will also sometimes not be true.

I figure there are three kinds of knowledge.

  1. Historical knowledge is what has happened. For this, you absolutely need reliable sources. Either the US declaration of independence was signed in 1776, or it wasn't. The current verifiability policy seems to think this is the only type of knowledge there is. The way to as close an approximation of truth for this is to try and find the best sources, and use them.
  2. Scientific knowledge, or experimental knowledge is of the form if I do this, what happens. It can be replicated and retested. This is the Galileo looking through the telescope part, don't just read the reliable sources of Ptolemy and Aristotle. For this, what you need is someone to have done the experiment. For verifiability in this, describe the method of doing the experiment accurately enough that it could be replicated, and describe the results of the experiment. If the experiment is not easy to do, then you need someone to have put up the equivalent of a lab notebook.
  3. Mathematical knowledge is the last form. For this, all you need for verifiability is the assumptions, and a correct proof. Reliable sources completely do not matter for this.

What is really needed for verifiability is that either the average reader of the article or the average 20 year old should be able verify the what is stated in the article. For a article on advanced mathematics, the proof should be at the level that someone reading the article would understand. For an article about World War II, it better have references. It is thru verifiability that we know that a article is as true as we can make it. But the purpose of verifiability is to obtain truth.

This is not saying that you can just put up a mathematical proof, and that is acceptable. The original research and notability requirements would still hold. But if someone wants to put up a implementation of Binary Search [1] they don't need to put a bleeping citation, it just needs to work, because someone who really needs the true algorithm can verify for themselves that it works.

Truth from verifiability. Jrincayc 03:43, 21 August 2009 (UTC)Reply[reply]

The purpose of Wikipedia is to collect and redistribute human knowledge, that is, what humans believe today that is true. This can be acheived in only one way: we look at what the experts of each field have published (as recently as possible) about what they believe is the truth. Generally we are not experts so we cannot search for the truth ourselves.
Your proposed policy is simply impossible to apply. How do you define truth? Sure thing, we all want to know the true real truth (...), but all we have is the published body of human knowledge. Most probably, a large part of what we believe today to be true, in a couple of centuries will be replaced by new and "more true" knowledge or at least massively amended. Truth is evasive.
Your example is wrong, and it's not surprising that you couldn't come up with a better one. Quoting Ptolemy for the movement of the Earth goes against the verifiability policy, because a 2000 year old (primary!) source can hardly be considered reliable for a scientific fact, in a world where even 100 years can easily make scientific information obsolete.
You're right about one thing: blind verifiability is bad. We need to strive for information that is not simply verifiable (that is, merely published as such), but that can be judged as true by making sure the source is reliable, that it is not a printing error, that it corroborates with other sources, etc. --AdiJapan 07:16, 26 August 2009 (UTC)Reply[reply]

It sounds like we may be in agreement that Wikipedia's current stance of "Verifiability, not truth" should be replaced or modified. As I see it, the goal of the verifiability policy is to help the reader verify that article is true. For example, I read through Stress article. It doesn't have inline citations, but that is okay since it is a physics type article, and I was able to follow the logical aspects of it. It does have inline derivations which are more useful. We don't have to be experts, what we have to be is able to explain why the experts believe what they do so that the reader can follow their reasoning and see if they believe it to. Maybe what I want should be called "Deep verifiabilitity". Wikipedia should make it as easy to verify as possible in a deep a level as humans currently understand the topic. In otherwords, I want the verifiability to include both "what" and "why". Jrincayc 02:47, 20 October 2009 (UTC)Reply[reply]


The purpose of Wikipedia is to represent and redistribute human knowledge, that is, the information known to humans. That is, information representing reality; and part of reality is "what humans believe to be true". With very few exceptions (matematics and logic being one) "what humans believe to be true" is only conditionally true, true within a certain context. There is a truth according to the Palestinians, a truth according to the Israeli government, a truth according to the Israeli settlers, etc. The purpose of Wikipedia is to keep out truth, except where it can be placed in its proper context. Show me any place where a user is entering "Truth" and you have shown a place where the purpose of Wikipedia is being thwarted. Truth is always PoV. - Brya 16:14, 26 August 2009 (UTC)Reply[reply]

Then by your own definition, Wikipedia ought to be restricted only to math and logic? No, the purpose of this proposal is to give editors the ability to say "that source is not an expert in this field, find something better" or "a self published encyclopaedia is not acceptable". WP:V permits any randomness to be included in Wikipedia as long as it's "published" elsewhere first.
No, truth should be restricted in Wikipedia to math and logic only. The rest of Wikipedia should give information, not truth. Basically, indeed anything ever published elsewhere could be included in Wikipedia, but in practice there is a hierarchy; with mainstream views being much more important and deserving of mention and space than crackpot views. However, there is a big difference between a mainstream view and the truth. "The Truth" will always be a lie. - Brya 18:21, 28 August 2009 (UTC)Reply[reply]
Ok, forget "truth", but wikipedia needs to be more than a blog for cross posting various bits of op-ed and ngo published fluff. There needs to be some mechanism for determining the legitimacy and accuracy of a source, and discarding that which, while relevant, is no more insightful then the random comments of a man on the street. -- 23:47, 28 August 2009 (UTC)Reply[reply]
Yes, and one would hope that this would be "good sense" by the contributor. If every contributor was self-aware enough and self-critical enough to limit himself to what he actually knew (at a professional level) there would be a lot less problems. Wikipedia is chock-full of opinions. - Brya 08:13, 29 August 2009 (UTC)Reply[reply]

I found this I just wish it was binding and indexed for quick reference later. -- 18:16, 8 September 2009 (UTC)Reply[reply]


  1. Ptolemy's Almagest


Some proposals will have massive impact on end-users, including non-editors. Some will have minimal impact. What will be the impact of this proposal on our end-users? -- Philippe 00:18, 3 September 2009 (UTC)Reply[reply]

Changing the verifiability policy will have no direct impact on non-editors. If we can improve the verifiability policy it will help new editors that sometimes find it bizarre. For example here is one discussion of it [2]. Jrincayc 14:53, 13 September 2009 (UTC)Reply[reply]

Wikipedia-specific proposal

This, from reading, seems a very English Wikipedia-centric proposal. It's submission may be well-intentioned, but alas smacks of forum-shopping on a controversial point. --Brian McNeil 03:31, 20 October 2009 (UTC)Reply[reply]