Wikipedia talk:Fruit of the poisonous tree

disagree in major part

Although the essay's thesis is logical and the legal point is correctly stated for U.S. law, a very large part of Wikipedia, perhaps a majority, would have to disappear. Importantly, an expert may determine that someone, although an unreliable source in Wikipedia's terms, was right, and may create a reliable source stating that other person's information. That is a benefit to society of their expertise. Examples:

• Linguists have been turning to fiction for dialogues attributed to time periods from which no one has survived for evidence of actual language. Dictionaries use this evidence to compose many entries and Mark Twain has been such a source.
• Michelson and Morley tried to establish the existence of aether as a carrier of electromagnetism and failed, but Einstein decided they were wrong, reinterpreted their work, and developed the special theory of relativity.
• Also, Einstein's work established that Newton was wrong about gravity.
• An attorney who believed that a certain known individual other than Lee Harvey Oswald had been involved in murdering President Kennedy defended a media article with that view against a libel suit brought by the person allegedly so involved. The outcome, according to the attorney, who wrote a book about it, was that the jury went further than needed to decide the case and concluded not only that libel had not been proven but that the CIA had been part of the murder. This can be disagreed with from several angles, but whichever relevant view might be wrong is still reportable in Wikipedia if a reliable source (in Wikipedia's sense) can be found to support each view even if tracing all the way to the bottom would set one or more views and the media reporting them as unreliable or would set the original statements as original research.
• Theology allows unending disputes about the reliability of sources, disputes followed by the chopping of heads, etc.

Original research is, of course, permitted, so long as it is not original in Wikipedia. Most scholarly conclusions are published as original research in scholarly journals. Some are wrong, although it may take some time to prove that they are wrong. Any of these conclusions may be reported in news media, school textbooks, and other secondary and tertiary sources and what those state about the scholarship may be reported in Wikipedia, even though some of them are ultimately found to be wrong, and sometimes even after they are found to be wrong. An example is with cold fusion, where it is legitimate to state in Wikipedia the original researchers' findings and conclusions even though they were, for our purposes, ultimately disproven (the disproof also being reportable in Wikipedia).

Editors should not be expected to dig into the grounds of a reliable source to further validate it. Nor should an editor merely delete content that is cited to such a source. Instead, if one editor reports from a reliable source and another does the digging and concludes that it is not reliable in the sense discussed in this essay, the more reasonable outcome is that the second editor has found a dispute that may be reportable as such in the Wikipedia article, by reporting contradictions between sources, probably as a criticism.

An editor also has an option not to add content because of a disagreement with the sources. That is what most of us do on most subjects.

And an editor might want to refrain from reporting only from low-quality sources even though they (barely) meet Wikipedia's criteria. Some time back, there was a disagreement on a talk page about whether children's books and other children's sources were reliable enough for Wikipedia on subjects addressed by college textbooks. I find it helpful to limit my substantive editing to subjects about which I have at least some working knowledge, so that I add sourced content that is not mysterious to me within the scope of my knowledge. It prevents writing stupid content. It doesn't prevent disagreement or uncertainty, but it does raise the bar or standard for refutation.

Nick Levinson (talk) 04:53, 9 November 2015 (UTC)

Expanded to cover Wikipedia processes

I've just expanded this essay to cover failures of Wikipedia process, in addition to sourcing. This does create a rather weird multi-pronged essay structure, but I think that's better than having to turn WP:FRUIT into a disambiguation page. {{u|Sdkb}} ^talk 05:40, 20 September 2020 (UTC)

Focusing on fact-checking and editorial controls

I'm trying to think of a way to rewrite this that would make it work better. As it is it has problems; the entire purpose of an RS, after all, is to take things that we could not cover directly, then filter them through a process of fact-checking and editorial controls, while providing interpretation and analysis; when an RS is working properly, this converts stuff from an unreliable source into a reliable one. If they weren't allowed to do that, we could never cite anything ever - anyone could take an academic paper and say "aha, but the experiments and surveys it uses as sources are not WP:RSes, and the papers it relies on ultimately use those things, so it's not an WP:RS." Or they could say "aha, we cannot cite this New York Times article, because it quotes random people on the street, who are not RSes." In practice this essay would allow anyone to disregard any source they disagreed with at any time. Even if we somehow limited it to sources we have "formally" declared unreliable or even deprecated (which would be a terrible idea because those designations are not supposed to have that weight), it would make it impossible to ever cover the activities of anything we have designated as unreliable. If an unreliable news source says something so important that it's covered everywhere, are we unable to ever report the bare fact that they said it, even if it gets massively covered everywhere and has a dramatic impact visible through numerous RSes? Nonsense. On top of this, the "process" that this poorly-conceived essay effectively suggests would invite editors to perform WP:OR by second-guessing a RSes sourcing; and OR is, in fact, policy. --Aquillion (talk) 19:07, 30 May 2022 (UTC)

After reflecting a bit, I think I found a fix for the main issue. The key here is whether the reliable source applied its fact-checking and editorial controls to the material it reprinted; in situations where there is reason to think it didn't, FRUIT might make some sense. But that point has to be emphasized prominently - FRUIT makes no sense if it forbids reliable sources from reporting on unreliable sources; it makes some sense if we interpret it as forbidding eg. RSes that uncritically republish entire articles from unreliable sources. (Though I would wonder how long something can continue to be considered an RS if it does that.) --Aquillion (talk) 19:29, 30 May 2022 (UTC)