Wikipedia talk:How to apply WP:NOR's "Directly related" principle

I would like to thank User:Jayen466, User:Blueboar, User:Bob K31416 and User:Gregbard for helping me with this essay. (I wish to acknowledge their contributions, without implying approval or disapproval of the essay). The original essay can be found here. Any comments and suggestions are highly welcome! Of course, feel free to edit the essay directly. Thanks! --Phenylalanine (talk) 21:12, 15 February 2009 (UTC)

Comment

 * All As(particular claim 1) are Cs(claims supported by source 1)
 * All Bs(particular claim 2) are Ds(claims supported by source 2)
 * No Cs(claims supported by source 1) are Ds(claims supported by source 2)
 * No Ds(claims supported by source 2) are Cs(claims supported by source 1)
 * Therefore
 * No Bs(particular claim 2) are Cs(claims supported by source 1) and no As(particular claim 1) are Ds(claims supported by source 2)


 * Really all this argument says is that claim 1 is not supported by source 2 and claim 2 is not supported by source 1. However, if it really is the case that claim 1 is supported by source 1, then that is a perfectly good reason to include the source in Wikipedia as supporting the claim. The same is true for claim 2.


 * If it turns out to be a controversy of language, I can easily see this policy being used to delete perfectly well supported material. Sometimes people use different formulations of the same truth which are perfectly acceptable, but some deletionist will claim has to be verbatim or close to verbatim.


 * In cases where it is a straight up logical necessity, I really can't see using a policy like this to delete such a claim. What if reliable source A makes the claim that "All 'widgets' are 'colresk'.", and furthermore that "All 'floobs', 'gackgacks' and 'blablas' are 'widgets'."


 * If reliable source B says that "All 'knakknacks' are 'gackgacks'." Then the claim:


 * "All knakknacks are widgets." would be perfectly justified in a Wikipedia article with the supporting sources A and B together. There is no way it could be untrue if the sources are reliable (which we are presuming they are).


 * In the course of natural language we will inevitably find imperfect formulations, both in the sources and the claims citing them. Some of them will be perfectly fine and appropriate, and we should gear the policy toward including them for sure. If we did not, we would be deleting perfectly well supported material. There will be other cases which will be clearly inappropriate. These will be the product of people's different interpretations of things.


 * But those truths which are true under every interpretation of two reliable and consistent sources, should be permitted. "All knakknacks are widgets." being one example. This still requires that the relationships which are put forth as true in the sources are relevant (or 'directly related) to the conclusion. So, I can't really support the formulation given in the box. You have to be able to put things together to some degree. Pontiff Greg Bard (talk) 22:33, 15 February 2009 (UTC)

Hi Gregbard, thanks for your comment. You said "I can't really support the formulation given in the box". Are you referring to the quote from WP:SYN: "Editors should not make the mistake of..."? I would agree with you that trivial deductions should be allowed (as is explained in greater detail here and here). WP:SYN is concerned with "Synthesis of published material which advances a position". This essay uses the term "significant" conclusion in a similar way (the word "position" is problematic, IMO, as it is often used in connection with WP:NPOV disputes and so can be misinterpreted in NPOV terms). --Phenylalanine (talk) 01:25, 16 February 2009 (UTC)

The Hawking example
Hi again, some useful developments since I last saw this essay. One quibble: Above the Hawking example, we say, "Below are several examples of implicit conclusions resulting from synthesis of published material. All of them represent violations of WP:SYN; they should be resourced, or emended." Now the Hawking example does not fit that pattern, and goes against the expectation set up in the reader. We could either delete or rephrase that sentence, or present the Hawking example in terms of the article on Black holes (and then say that while it is SYN in that article and should be removed, in line with the reader's expectation, it wouldn't be SYN in the Space-time article). Jayen 466 00:22, 16 February 2009 (UTC)


 * Yes, good point. I fixed the Hawking example as per your last suggestion. --Phenylalanine (talk) 00:59, 16 February 2009 (UTC)

Can implicit OR be mitigated by caveat language?
This is a great essay with some good examples. I was thinking however about the proper response to dealing with situations like this. Take the boycott example; the essay concludes "[w]e need a source linking the retailer's disappointing sales and share price drop to Organisation X's boycott. If none can be found, the second statement should be removed from the article." But I imagine an alternate solution that looks like this:


 * Organisation X, with which A. N. Other was associated, called for a pre-Christmas product boycott. In the run-up to Christmas that year, Retailer Z suffered disappointing sales, its share price dropped, and Organization X claimed credit for the losses on their website; however, a clear connection to the boycott has not been established.

In one of the later examples, I can imagine rewriting as:


 * The majority of scientists do not believe that the LHC will create black holes that are dangerous, primarily because of Hawking radiation. Theorist A states that Hawking radiation rests on dubious assumptions, but has not advanced an opinion on the safety of the LHC.

And in the next example:


 * Steven Hawking wrote of black holes evaporating, and used math that assumed that space-time was continuous. While this assumption is challenged by some theorists, these theorists have not challenged the conclusion of black hole evaporation.

In short, the question is whether implicit OR can be mitigated by caveat language that emphasizes that the implicit connection one might think is being made, is in fact not being made. Dcoetzee 00:12, 28 February 2009 (UTC)


 * Thanks Dcoetzee, the caveat language that you introduce effectively eliminates any false implications that may be inferred from the examples, though we can still infer the true logical implication, i.e. that "B" would not agree with the statement that "X is true because Y is true". I think it would be useful to further clarify this in the essay using your alternate solutions. Cheers, Phenylalanine (talk) 03:13, 28 February 2009 (UTC)