User talk:Phenylalanine/How to apply WP:NOR's "Directly related" principle

Well done

Nice to see this essay, well done. Jayen466 04:51, 11 January 2009 (UTC)

Thanks a lot! Phenylalanine (talk) 13:06, 11 January 2009 (UTC)

SYN example I spotted today

This one is a beauty: [1] It looks okay when you read it, until you realise that neither of the sources cited even mentions the organization, let alone the article subject. It is all smoke and mirrors. Jayen466 04:51, 11 January 2009 (UTC)

Very interesting. The FairPlay source does say that "the big sales blip during the campaign week which saw Game, Europe's biggest videogame retailer, lose a massive 80% of its share value overnight in response to disappointing sales. When the campaign was over, sales shot up again, just as we said they would." So the editor could have simply written something like: "FairPlay claimed dramatic results for the boycott, including disappointing sales for Europe's biggest videogame retailer and an 80% loss of the company's share price". Instead, the editor introduced a personal analysis "while direct and/or sole causation could not be proved" (not in the sources) and added much detail from off-topic sources. Does this advance an implicit position not found in any of the sources? I think the FairPlay source partially supports the suggested conclusion, but the editor went overboard to strengthen the case: "The pre-boycott price was not recovered until over four years later, in mid-January 2007." I think the example could be included in the essay, what do you think? The examples in the essay illustrate cases where the implicit conclusion amounts to a critical or contrary view about a topic related to the article subject. What we have here is synthesis which advances a supportive or favorable implicit conclusion.

An interesting question is the following: Is it original synthesis to add information from off-topic sources that describe in greater detail facts that are already mentioned in on-topic sources and that are used in those sources to support or criticize a given claim. This reminds me of a discussion I had about the raw foodism article [2]. Cheers, Phenylalanine (talk) 13:06, 11 January 2009 (UTC)

This question of "greater detail" on facts already mentioned can be vexing. I've agonised over my choice to include the sources numbered 26 and 27 in the article on Idries Shah, to source the sentence "The scholarly consensus today is that the "Jan Fishan Khan" manuscript was a hoax, and that the Graves/Shah translation was in fact based on a Victorian amateur scholar's analysis of the sources used by previous Rubaiyat translator Edward FitzGerald." (in this section). While other sources, notably Graves' biography and Moore's paper, mentioned the article subject prominently with respect to this controversy, citing personal correspondence, these two sources (Aminrazavi, Mehdi (2005). The Wine of Wisdom. Oxford, UK: Oneworld. pp. 155. ISBN 1851683550 and Irwin, Robert. "Omar Khayyam's Bible for drunkards". The Times Literary Supplement. Retrieved on 2008-10-05) did not mention him by name at all. In the end I came to think the inclusion is correct, because there is no new conclusion involved, and it's nice for the reader to know how the debate ended.

As for the FairPlay example, the editor re-added this section many times, with different wording; there may be earlier edits where the SYN attempt was even more apparent. Note that citing the FairPlay website of course had the additional problem of falling foul of Wikipedia:BLP#Using_the_subject_as_a_self-published_source (self-serving, and claims about third parties). The article is nominated for deletion; perhaps we should have a look through the article's history to look at other wordings the section had; it's potentially a useful example and, like you say, one that advances a favourable rather than oppositional view.

As for your raw foodism example, I am inclined to agree with you that the debate on the health effects of cooked/raw vegetables should be sourced to literature that specifically discusses raw foodism in the narrow sense. I can't imagine that there should be no such sources available. If these in turn refer their readers to other, more general sources that look at the health effects of cooking without mentioning raw foodism specifically, then I could see a successful argument being made to include details from these sources, because then we'd have a source saying that they are relevant to the topic. In my view, it just helps to keep an article focused on the topic. Jayen466 19:38, 11 January 2009 (UTC)

Your analysis is spot on. In this case, its harder to make the case for excluding content on synthesis grounds. I think we should still discuss this issue in the essay. Cheers,Phenylalanine (talk) 23:40, 11 January 2009 (UTC)

Here is an earlier version of the FairPlay edit, without the SPS. The wording always seems to have been pretty much the same. Given that you've shortened the two existing examples in the essay, I think there is room for one or two (or three) more. Cheers, Jayen466 22:11, 11 January 2009 (UTC)

This example is simpler and more straightforward that the other example. Great addition. --Phenylalanine (talk) 22:54, 11 January 2009 (UTC)

I've popped a summary of it in, pls review. Jayen466 00:20, 12 January 2009 (UTC)

Nice work. I also agree with the title "Speculative conclusions" since the conclusion reached appears to be more speculative than the conclusions in the other examples. --Phenylalanine (talk) 02:56, 12 January 2009 (UTC)

If you wanted to create a shortcut, WP:ORIGINALSYN has a ring to it. :-) Jayen466 19:47, 11 January 2009 (UTC)

Done! --Phenylalanine (talk) 22:49, 11 January 2009 (UTC)

I really like that shortcut. :) Jayen466 00:20, 12 January 2009 (UTC)

Proof that original synthesis applies to implicit conclusions

There is one nice point that Blueboar made on the NOR talk page: B may well agree with A that X is false, but for a different reason. It might be fun to add that. Cheers, Jayen466 19:52, 11 January 2009 (UTC)

Good idea. I added it. --Phenylalanine (talk) 22:46, 11 January 2009 (UTC)

Suggested tweak to "Original synthesis also applies to implicit conclusions"

What do you think of something like this in the first para of the new section:

The original research policy sates that "you must cite reliable sources that are directly related to the topic of the article". Editors who cite sources that do not even mention the article topic will sometimes defend their edit by telling you, "But it is directly related" – because in their mind, it genuinely is related. It can be very hard for them to see, let alone acknowledge, that the cited source does not refer to the article topic at all. If you find yourself in this situation, it may help to remind the other editor that the original research policy can basically be summed up in two clear rules:

instead of:

The original research policy states that "you must cite reliable sources that are directly related to the topic of the article". If you feel that "directly related" is too vague to be of any real use as a guideline, you're not alone. Happily, we don't need to worry about that since the original research policy can basically be summed up in two clear rules:

? Jayen466 23:09, 11 January 2009 (UTC)

I like your proposal. Could we mention the fact that the editor is trying to push a POV (so there is no possible confusion with acceptable background info)? --Phenylalanine (talk) 23:24, 11 January 2009 (UTC)

If we do, can we find a way of saying it in as unconfrontational a way as possible? Ideally, we'd want the "guilty" editor to be able to read this essay and realise – in their better self – that that is what they are doing. We may disagree with them, but from their point of view, they'd just be wanting to describe things the way they see them. Which is fine, but they must find sources that see it their way if it's to go in the article, and if there aren't any – Jayen466 23:44, 11 January 2009 (UTC)

Looking at the text again, I think the situation described suggests that the information is contentious, so POV pushing is probably involved. --Phenylalanine (talk) 23:56, 11 January 2009 (UTC)

Solutions

We might want to add solutions at the bottom of each example, following a pattern such as:

Solution: Find reliable, published sources that comment on ... [the topic] and ... imply/suggest/speculate about [conclusion C]. If no such source can be found that links [B] to the article topic, do not introduce [B] into the article. Jayen466 00:37, 12 January 2009 (UTC)

Yes, I agree. Phenylalanine (talk) 03:00, 12 January 2009 (UTC)

Another example

The Inflation article includes a discussion of the Gold standard, as the Gold standard has been proposed as a means of controlling inflation. However, most mainstream economists oppose a return to the gold standard. Reasons include: inability to use monetary policy, vagaries of gold mining arbitrarily changes the money supply and inflation rate, and more technical arguments about the effect on exchange rates and Balance of payments during trade surpluses or trade deficits. These criticisms are all sourced and discussed on the page about the Gold Standard. However, most of these sources discussing these problems refer directly to the gold standard, not to inflation, i.e. they discuss why a gold standard is a bad idea, not why its a bad idea to use a gold standard specifically to control inflation. They do not belong in the article about inflation; the only appropriate sources are those that talk about how the gold standard is good at controlling inflation. --Phenylalanine (talk) 03:44, 12 January 2009 (UTC)

I seem to remember looking into this at the time; there were a number of sources that made similar points in direct connection with inflation (I think I posted them at NOR talk). But without citing such sources, the relevance of the discussion to the topic is not readily verifiable to the reader.

A similar example is Japanese_war_crimes#Official_apologies, where last time I looked a number of the sources cited (e.g. the one on Brandt) made no reference to Japan whatsoever. Brandt's action may have been referred to in multiple secondary sources talking about Japanese war crimes – there is one at least in google books – but as it stands, that bit is WP:SYN (with some weasel words to boot). Jayen466 02:57, 15 January 2009 (UTC)

It gets even worse: the part that says "some point to an act by" is not even in the source cited. --Phenylalanine (talk) 03:31, 15 January 2009 (UTC)

Exactly. The "some" is the editor who put that in, and perhaps, at a stretch, the author of the linked book, although "pointing" is probably too strong a word for what that author did there. "Mention" is more like it. Jayen466 03:34, 15 January 2009 (UTC)

Deletion of material

Hi Jayden, regarding this edit, the material served to explain how the synthesis rule necessarily supersedes the directly related test, as some could argue that syn doesn't cover all cases. --Phenylalanine (talk) 03:19, 15 January 2009 (UTC)

Okay, let me look at it. Jayen466 03:22, 15 January 2009 (UTC)

What we had was,

"Also, given that adding uncontentious background information in an article, by definition, does not involve any new conclusion, it should be clear that adding such background information is not forbidden by the SYN policy.

The job of the original research policy is to restrict sources, not content per se (which falls under the jurisdiction of Wikipedia:Relevance). Consequently, because information that advances an implicit position not found in any of the sources is forbidden in principle, and because relevant neutral background information that does not advance an implicit position is permitted in practice, irrespective of the source cited (provided that it is reliable), we can conclude that WP:SYN has all the bases covered and, therefore, necessarily supersedes the "directly related" test."

I thought the first para was covered in the lede. I found it didn't follow logically from what preceded it. As for the second paragraph, I admit I don't really understand it, notably the clause "The job of the original research policy is to restrict sources". WP:REL seemed a very stub-like and not very well-known page to link to. And the last sentence "we can conclude that WP:SYN has all the bases covered and, therefore, necessarily supersedes the "directly related" test." again I have trouble understanding. I think we were trying to say that the presence or absence of a novel conclusion is what matters, not whether the source is directly related or not. But I thought we had made that clear already. Thoughts?

Lastly, it is clear that this is your essay. So if there is anything I have done you don't agree with, please feel free to revert it. Anything I offer is offered on an "If you like it, good, if not, I'll desist or try better" basis! Cheers, Jayen466 03:30, 15 January 2009 (UTC)

I appreciate your feedback. I would like the essay to explain the points made in the lead per WP:LEAD, so I'm not too worried about moderate repetition. Some might argue that, in some cases, if the "syn test" is applied and the material passes the test, that doesn't mean it's not original research because the material might still not be "directly related" to the article topic. I'm tying to prove that the "presence or absence of a novel conclusion is what matters," as some might question that assertion. I'm trying to convince these potential readers that once you apply the syn test, the "directly related" test no longer applies. --Phenylalanine (talk) 03:57, 15 January 2009 (UTC)

You say, "Some might argue that, in some cases, if the "syn test" is applied and the material passes the test, that doesn't mean it's not original research because the material might still not be "directly related" to the article topic." — I think that could not happen, because the "directly related" test is part of WP:SYN. So if the sources were not "directly related" to the topic, the passage would fail WP:SYN. Perhaps I misunderstand what you mean.

As for the lead content, we could take it out of the lead and stick it in later. I like the wording in the lead. Jayen466 04:17, 15 January 2009 (UTC)

Okay, I think I may have got you now. Is it that you want to enable addition of circumstantial detail that does not involve any conclusion? I.e. you are trying to strengthen the inclusionary rather than the exclusionary position with this passage? Jayen466 04:23, 15 January 2009 (UTC)

I like the wording in the lead also and I'm not suggesting we should remove it, I just want the body to explain in greater detail and illustrate what it says. In my opinion, if "directly related" is part of syn, then we are defining "directly related" in a very bizarre way, i.e. whether or not material is "directly related" depends on whether the material advances an implicit position or not. In my mind, both tests are incompatible, which is why some might argue that there are really two separate tests: the syn test and the directly related test. My argument is that the syn test necessarily applies to implicit conclusions and, therefore, necessarily rules out the directly related test. --Phenylalanine (talk) 04:36, 15 January 2009 (UTC)

If the sources cited do not explicitly reach the same conclusion, or if the sources cited are not directly related to the article subject, then the editor is engaged in original research.

Notice the disjunctive "or" here. Basically, we have two separate tests: what I would call the synthesis test and the "directly related" test. It seems to me that if synthesis also applies to implicit conclusions and we all agree that background information from generic sources is perfectly acceptable, then we have no more use for the "directly related" test. --Phenylalanine (talk) 00:43, 26 January 2009 (UTC)

I see that I went too far with this line of thought (i.e. the "directly related" test is still relevant as WP:PCR does not address the proper use of sources), and so I revised the essay accordingly. Comments? --Phenylalanine (talk) 03:09, 28 January 2009 (UTC)

Title

Should we perhaps rename the essay, "Synthesis of published material to advance novel conclusions" or something like that? The "implicit position" is a term that actually does not occur in the essay. Jayen466 03:22, 15 January 2009 (UTC)

Well it occurred before it was deleted ;-). I think the focus of the essay is on an aspect of the syn policy that is largely unknown (implicit conclusions), so I though it was appropriate to stress that aspect in the title. --Phenylalanine (talk) 03:26, 15 January 2009 (UTC)

I agree the "implicit" is important, it's just that the word "position" is not in the text (... much, it actually occurs once, I see now). Jayen466 03:32, 15 January 2009 (UTC)

I use position and conclusion interchangeably in the text. Do you think that's problematic? --Phenylalanine (talk) 03:41, 15 January 2009 (UTC)

I am not sure, and I am no longer at my best (10 to 4 a.m. here). At any rate, I note that the box I added saying ""Synthesis occurs when an editor puts together multiple sources to reach a novel conclusion that is not in any of the sources."" clearly duplicates the second bullet point above it, though the box text has the advantage of reflecting the verbatim WP:SYN wording. Jayen466 03:48, 15 January 2009 (UTC)

I can see the virtues of consistency here and "conclusion" does have the advantage of being more informative than "position". How about "Synthesis of published material to advance implicit conclusions"? Cheers, Phenylalanine (talk) 04:13, 15 January 2009 (UTC)

I think that would be good. We should perhaps standardise on either "implicit" or "implied" as well; otherwise the reader may wonder if there is supposed to be a difference in meaning. Jayen466 04:25, 15 January 2009 (UTC)

I like "implicit" because the word doesn't imply a logical connection, it may be that a passage simply brings to mind a speculative conclusion without a clear logical deduction. What matters for our purposes isn't the logical connection, but the end result: what would a reasonable, educated person conclude based on statements "A and B". I use "imply" to mean "entail", but perhaps that should be clarified. --Phenylalanine (talk) 04:47, 15 January 2009 (UTC)

Recent changes, comments?

Any comments on the recent revisions to the essay? --Phenylalanine (talk) 17:01, 25 January 2009 (UTC)

The move from the lead to the main essay I think works well. I still have a problem with this passage: "The job of the original research policy is to restrict sources, not content per se (WP:ROC). Because information that advances an implicit position not found in any of the sources is forbidden in principle, and because relevant neutral background information that does not advance an implicit position is permitted in practice (WP:PCR), irrespective of the sources cited (provided that they are reliable), we can conclude that the synthesis test has all the bases covered and, therefore, necessarily supersedes the "directly related" test."

I think most people will read this and say, "The job of the original research policy is not to restrict sources. Its job is to make sure that there are sources cited." (So rather than diminishing the number of sources cited in an article, as "restrict" could be meant to imply, articles will end up having more sources because of the NOR policy.) Could you clarify that sentence? Jayen466 01:10, 26 January 2009 (UTC)

"The goal of the original research policy is to ensure that sources are used correctly in Wikipedia articles. The policy does not address the relevance of content within articles (WP:ROC)." How about something like this? --Phenylalanine (talk) 01:33, 26 January 2009 (UTC)

I replaced the text with this: "The goal of the original research policy is to ensure that information is properly referenced with reliable sources. The policy does not address the relevance of content within individual articles (see WP:ROC)." Any better? Cheers, Phenylalanine (talk) 02:37, 26 January 2009 (UTC)

Where do we go from here?

Thoughts? --Phenylalanine (talk) 03:01, 28 January 2009 (UTC)

Just having had a quick look through it, I think it has become a lot more readable. Well done. I suppose eventually it could be taken out of your user space and put up as a Wikipedia essay, if that is what you want; see Wikipedia:Policies_and_guidelines#Essays and Wikipedia:Wikipedia_essays. Cheers, Jayen466 03:43, 29 January 2009 (UTC)

Thanks. That's a great idea. I'd also like to get some feedback from the broader community. I can maybe post a request for comment on several user talk pages, or at Wikipedia talk:NOR. --Phenylalanine (talk) 11:57, 29 January 2009 (UTC)

Yes. I mentioned the essay in a SYN-related discussion here just now, and will continue to drop references to it where appropriate. It will be good to get more eyes to go over it. Cheers, Jayen466 19:53, 29 January 2009 (UTC)

Thanks, I will drop some requests for comment on several user talk pages. --Phenylalanine (talk) 11:30, 30 January 2009 (UTC)

Other examples of implicit conclusions

I moved these two examples here from the essay to keep it as short and concise as possible. I think the two examples in the essay do a good job of illustrating the idea of implicit conclusions. --Phenylalanine (talk) 01:28, 31 January 2009 (UTC)

Example 1

The Wikipedia article's topic is self-evidently closely related to "X", but not to "Y".

• A = James believes that "X" is false because "Y" is true. [properly referenced]
• B = However, Paul believes that "Y" is false. [source comments on the proposition that "Y" is true, but not on the proposition that "X" is false or on James' belief that "X" is false.]

— Suggested conclusion:

• C = If Paul is correct, James may be wrong. [not in any of the sources]

Explanation:

Stating "A and B" in the article is original synthesis since it suggests "A and B, therefore C". The example suggests that the adversative argument B pertains to the validity of A.

Note: Paul may well agree with James that X is false, but for a different reason.

Solution: Find a source that says Paul disagrees with James over X because of his different assessment of Y. If none can be found, delete "B".

Example 2

Paleolithic diet article

• A = The Paleolithic diet excludes grains, including whole grains, on the grounds that they promote ill-health.[properly referenced]
• B = Consumption of whole grains has been associated with a decreased risk of several health problems.[source does not refer to the Paleolithic diet, nor does it comment on key related topics such as the Paleolithic, human evolution, hunter-gatherers or evolutionary medicine]

— Suggested conclusion:

• C = Exclusion of whole grains from the Paleolithic diet may not be justified. [not in any of the sources]

Explanation:

Stating "A and B" in the article is also original synthesis because "A and B" suggests "A and B, therefore C".

Solution: Check if there are sources on the Paleolithic diet that discuss the merits or demerits of eliminating whole grains from one's diet. If none can be found, delete "B".

• I agree. The essay is better for having less complexity in it. Cheers, Jayen466 14:15, 31 January 2009 (UTC)