Wikipedia talk:Notability (science)/Archive 1

This is the original talk archive of the pre-merged proposal on Wikipedia:Notability (scientific contributions) [1]

Some thoughts

"Textbook" might be better defined, I'd like criterion #1 to be taken to mean used in a course at an accredited university. This could eliminate NN kook theories qualifying via vanity press "textbooks" and the like.

I like criterion #2 paper cited a hundred times, but see some wiggle room here. Two papers that mention the theory each cited some number of times summing to 100 (but if 75 of the citations are the same article...), a hundred papers each cited once...? I thinnk this is a criterion that should be tested out on some exemplar theories (see below) and see how it works.

Criterion #4 is going to be the tough one, operationalizing "popular media" and "politcal circles". Blogs and the like, drawing a line between mainstream media etc. We should look into how this debate is progressing in other notability guidelines, they must have the same problem...

I was wondering if some test cases of notable, non-notable and borderline theories would be appropriate to help in discussing the severity of each of the criteria (my inital reaction is that the bar is too high, and the best way to test this is to advance some cases that I think merit inclusion). This could get into an excessivly lively debate with supporters of some theories that are currently over-represented on WP, and so these test cases might be best kept in the back pocket for now, but I think anyone contributing to this effort would do well to collect a few pet examples for guidance, future debate.

Cheers, Pete.Hurd 17:26, 1 September 2006 (UTC)

On textbook, I thought about explaining "general" and "specialized" a bit more, but I agree some qualifier about usage should be added. I'm not sure if "widely used" is helpful. Pindyck and Rubinfeld clocks in at Amazon rank #600 right now, although I can see that drop precipitously in two or three weeks. Osborne and Rubinstein is currently at #29,000, so gauging usage from Amazon ranks seems a questionable business. "a course at an accredited university" leaves a lot of wiggle room for crank profs who assign their own sketchy textbooks to their classes, so I would at least go with "multiple", but it's a criterion that shouldn't be too hard to verify. Lots of class syllabi are online.

I would guess that for the "hundred papers each cited once" case the textbook criterion applies. Or at least a "handbook of {field X} criterion (which I still need to add). I agree the 100x1 could be stretched out to 75x2, etc.

"Political circles" could be restricted to "legislative hearings" or somesuch. I'm clearly targeting intelligent design here.

I'll retrieve some AfD's I consider "test cases" later. ~ trialsanderrors 17:57, 1 September 2006 (UTC)

We don't need this for science

Notability's primary purpose is to make sure that Wikipedia does not turn into a directory, in accordance with what Wikipedia is not. There are directories in the fields of people, businesses, products, services, and web sites. There are no such equivalent science directories for an encyclopaedia to avoid being. Sciences don't have dates of birth, addresses, sale prices, or telephone numbers. ☺ In the absence of directories which can act as potential sources, notability simply devolves into a general prohibition on original research, i.e. on ideas that have not been written about by people other than their creators. Uncle G 11:58, 2 September 2006 (UTC)

I don't follow this argument at all. First, notability serves principally as a filter for "WP:NOT an indiscriminate collection of information": That something is 100% true does not mean it is suitable for inclusion in Wikipedia, not as a filter for "WP:NOT a directory". Second, there are of course directories for scientific output. They're called abstract databases and citation indices. Third, writing an article about a topic for which there exists only one published source is not original research, but it will most likely fall under WP:NPOV#Undue weight. Fourth, scientific output is a massive exercise in trying to create notable subjects, but only very few succeed in being noticed by the scientific or general public. Most attempts to create new theories or nomenclature simply go nowhere. Which is the very reason we should look at applicable mechanisms on how to separate the includable from the unincludable research. ~ trialsanderrors 22:05, 3 September 2006 (UTC)

I think the best argument for a notability guideline would be an example of a clearly NN theory, which doesn't qualify for deletion as OR. There are some areas of WP, eg schools, where failiure to implement notability guidelnes has produced useless catalogues in contravention of WP:NOT. Can such a fate be imagined here? Pete.Hurd 22:27, 3 September 2006 (UTC)

• Currently ongoing: Wikipedia:Articles for deletion/Radical Induction Theory of Ulcerative Colitis
• Deleted per closing admin, stood up in DRV: Wikipedia:Articles for deletion/Research into health benefits of Falun Gong
• Closed as "no consensus", which IMHO was a horribly wrong call, esp. given that the article is still only a Wikiquote: Wikipedia:Articles for deletion/Economic totalitarianism (4th nomination)

~ trialsanderrors 23:02, 3 September 2006 (UTC)

• Another: Wikipedia:Articles for deletion/Mikhail Lebedev -- Samir धर्म 01:31, 6 September 2006 (UTC)

The AfD seems to have been deleted. ~ trialsanderrors 10:08, 29 September 2006 (UTC)

• A case of a "fringe" theory: Wikipedia:Articles for deletion/Heim theory
• Possibly encyclopedic topic with rather confused content: Wikipedia:Articles for deletion/Money in the utility function

We absolutely need this for science

We need absoutely need this for science, which, with the sheer volume of extant literature, is very much at risk for the directory phenomenon. Here's an example. In 2000, three goofballs in Ottawa (I'm allowed to call them goofballs) published [2] on the computer modeling of red blood cells in an optical trapping system. Is it verifiable? Of course! There's the peer reviewed journal link right there, and it's been cited 19 times by other papers. It's a great study, and paved the way to a cell sorter that they made. Is is notable? If I made an article called Computer modelling of red blood cells in optical trapping systems should it stay, or should it go? I'd be in favour of a notability guideline to help decide that. Otherwise we'd be debating between WP:V and "Wikipedia is not an indiscriminate..." to determine whether or not to keep that article.

It's even worse in medicine, where Wikipedia can easily turn into another PubMed. PubMed is a phenomenal directory. Everything in PubMed is verifiable. How does Wikipedia differ? I think it's in terms of notability primarily. The fact is that there's a lot of (1) verifiable minutia (that's important only in a small sphere) and (2) non-notable junk that's published in peer review journals and found in PubMed. I think a notability guideline is essential to help sort out what to keep. Otherwise, we'll become a repository for the "Chelation therapy and the elevated copper theory of ulcerative colitis" (PMID 531503) and what not (WP:BEANS, I know). -- Samir धर्म 09:40, 4 September 2006 (UTC)

Problematic

Okay, I get it - you have seriously high standards. But that aside: It has been published in a selective, peer-reviewed scientific journal and has received at least 100 citations. That seriously needs to be far better developed. Depending on your field, 100 citations can be a huge number, or small (but not really do to importance). Apart from which, what about 2 founding papers with 50 cites apiece? What if some are review papers? A more realistic scenario for occurances is something like

1. Theory A published (say 10 cites)
2. Covered in a review or two (5 cites)
3. Theory B published, essentially a modification of theory A (say it relativises or quantises it.) 10 cites
4. Some appendexes get covered (3 more cites)
5. Another review (3 cites)

And infinitely more complicated scenarios. This kind of standard (but almost all the ones there) are definitely square pegs proposed for round holes. Realistically, anything that's had two or more papers published on it in a top level journal (i.e. Nature, Science, The Astrophysical Journal, Astronomy & Astrophysics et al.) with two or more independant authors/groups working on it is verifiable, encyclopaedic, notable, whatever. Although who knows? The Alpher-Bethe-Gamow paper was essentially forgotten when it was published in '48 because nobody was really working in the field, and experiments couldn't get there, but it was still hugely important. WilyD 19:06, 26 September 2006 (UTC)

I agree. I'm not happy with the 100 criterion at all, and I'm perfectly fine with scrapping it completely. On the A-B-G paper comment: You're assuming that it is our job to evaluate what's important. It's not. Our job is to evaluate what others found important, and if those others match our criteria for reliable sources. So going back in time to the Wikipedia of 1948, and deciding that A-B-G deserves an entry because future researchers will find it important is just a misconception of what notability means. Notability isn't our reflection of the subject's merit, it's our reflection that enough outside sources gave reliable accounts on the subject that we can cover it objectively without engaging in our own research. ~ trialsanderrors 18:48, 27 September 2006 (UTC)

No, in 1948 it's not really cited widely, but it's the seminal paper of the field (the field is just not well explored at this point). I'm not talking really about future importance, but importance to it's field, which may be huge or tiny. I'm not sure exactly how to qualify this, but it's clear to me that any field in which work has been done and peer reviewed paper(s) published should have at least a little coverage, per Wikipedia, she ain't paper. Roughly speaking, the most important item or two per field should receive coverage, even if the field's obscure. Things like citations muddle up importance and popularity. The most important paper on Quark stars probly doesn't have a hundreth the citations of the most popular paper on String theory. Now, future hindsight will (of course) tell us that string theory is bunk, but that's not the point. The most prominent theory/observations on Quark Stars should be documented as well - it's verifiable and encyclopaedic. Vainity and Spam are a bit of a concern in all fields, of course, but we do need to be cautious that it doesn't result in the exclusion of legimiately encyclopaedic content of very narrow interest. WilyD 01:19, 30 September 2006 (UTC)

Just want to make sure here, but you seem to be arguing against a standard of notability in favor of verifyability alone. I'm not defending the 100 number, but I don't see how discarding some measure of "popularity" allows us to come up with a criterion of "importance" that isn't going to regress. If the argument "this paper/theory is encylopedic, despite the topic of TRIVIAL-BUT-NOT-TOTALLY-KOOKEY-THEORY being really obscure and unpopular, this paper/theory is a seminal contribution to the field". Any standard that cuts out Quark Stars is broken and should not be implemented, does that mean that a guideline cannot be crafted that includes them and excludes NN fluff? Maybe the question can't be answered without some NN fluff exemplars. Pete.Hurd 17:58, 30 September 2006 (UTC)

Verifiability is not the issue with scientific topics, objectivity is. That's covered in the introduction, where WP:V doesn't even make an appearance. The Falun Gong AfD above is a good test case where a group of editors tried to use verifiable but insignificant research and pseudo-research to push a "Falun Gong benefits your health" POV. Those are the cases we should be worried about. I don't see how Quark Star fails this test, even though it might not make the 100 cites criterion. That's the reason why we're trying to establish a number of OR criteria, to account for the differences in validation between research fields. The 100 cites criterion is simply geared towards the fields that mainly use peer reviewed journal articles as validation mechanism. The other issue is that from reading the Quark star article I'm unable to deduce its scientific significance. This is another purpose the Science notability guidelines can serve: as style guidelines on how to apprach science topics from an encyclopedic perspective. Far too many science articles are written as How-to's rather than encyclopedic accounts. ~ trialsanderrors 08:04, 1 October 2006 (UTC)

Well, that's why I rather like the issue of how in 1950, Alpher Bethe Gamow fails these criteria despite being significant research published in a highly respectable journal. It was just an obscure topic in the amount of research being done on it (mainly because there was no observations, so it was a tough topic) - but this could be clearly identified as important. What I would like to see is something that's more flexible to the field. So that smaller fields with less researchers can nonetheless have verifiable, encyclopaedic contributions included. I'm not talking about trivial stuff, but important stuff with few citations because few researchers are working in that area. A biologist may find 100 citations a small number (I have no idea) but it's a huge number in my field (astronomy). Topics that are busy may allow one to scoop up 100 cites across a score of papers (I've no idea if this is acceptable under this criterion. The textbook criterion may save a lot of stuff, though (if I know wikipedia) obscure or hard to find textbooks are likely to cause bellyaching (if the stuff can't be googled to get clear, concise, digestable results - I'm sure you've been down that road)

Beaming myself back in time to 1950 and being asked whether A-B-G should be included I would probably have said no (based on your comments, I don't know the actual circumstances), and that's a decision I could live with today. Because in 1950 the ultimate verification of the importance of A-B-G was a future event, and therefore falls under WP:NOT a crystal ball. In science (and the arts), for every iea that changes the world there are 99 that fail. And while some of the failures are quite notable and can be covered under NPOV, most of them are simply spurious. But that doesn't keep their creators from claiming that they're simply ahead of their time and ultimate recognition will undoubtedly follow... ~ trialsanderrors 01:24, 3 October 2006 (UTC)

No, in 1950 we know it's important it just isn't popular because of the area of science its in. From what I understand you're trying to exclude cranks, not Hans Bethe and George Gamow type scientists who just work in a field where there aren't a lot of researchers - but you're not doing very well in trying to do so. WilyD 00:47, 4 October 2006 (UTC)

The Alpher-Bethe-Gamow paper was essentially forgotten when it was published in '48 because nobody was really working in the field if this is an accurate description, then I can't see how a notability guideline can be crafted that would recognize it as notable without a crystal ball. I can't see how a notability guideline that would include it based on it's status then would be of any use now in excluding NN science. We ought just to concentrate on separating crank from mainstream and give up on notability in that case. Pete.Hurd 01:39, 4 October 2006 (UTC)

This is (essentially) my whole argument. That notability in science does not necessarily reflect encyclopaedic value. WilyD 02:31, 4 October 2006 (UTC)

I think I get what you're saying, but I'd stick with the equation of notability and encyclopedic value and re-phrase your view to be that metrics of popularity don't properly judge notability in the case of scientific theories, papers etc. Does that seem an honest representation? (comment added after indentation reset comment below) Pete.Hurd 17:12, 4 October 2006 (UTC)

[reset indentation] To recapitulate your argument, you're saying that we're committing a Type II error by excluding a theory that is 1. important but 2. unrecognized. No we're not, it is simply not our within our editorial discretion to make judgements on importance if recognition is not established. If A-B-G was not vetted by the scientific community in 1950 should not have been in the Wikipedia of 1950, no matter what future researchers did to establish its importance or validity. Note that truth is not an inclusion criterion for Wikipedia, only reliablity and comprehensiveness of outside sources. Your example isn't even a valid case of a false negative, and even if false negatives exist, the problem of false positives (POV pushing of spurious theories and researchers) seems to be the more pressing one. It's also important to make clear that exclusion is never a judgment on importance or validity, which seems to be your major concern. ~ trialsanderrors 09:18, 4 October 2006 (UTC)

No, that's not what I'm saying. What I[m saying is that recognition should be normalised to the field of study. A-B-G was vetted by the scientific community in 1950 - it just wasn't cited much because very few papers were being written on the subject of big bang nucelosynthesis. But for a paper on the subject (or a theory on the subject) it was very important. No editorial opinioning needed (except *maybe* a judgement on what field its in). I'm not talking about future researchers. What I'm against is deleting verifiable, encyclopaedic content with no real justification (Other content might be fraudulent isn't a good argument for deleting unrelated content). WilyD 16:20, 4 October 2006 (UTC)

So how did this vetting occur in 1950? ~ trialsanderrors 07:57, 5 October 2006 (UTC)

By publishing it in a highly respected journal under the names of two prominant sciences (and one underappreciated graduate student). WilyD 13:37, 10 October 2006 (UTC)

There are hundreds of articles in respected journals that get zero citations. ~ trialsanderrors 22:56, 10 October 2006 (UTC)

Aye, you & I would consider that evidence of non-notability, WilyD not necessarily. Non-resolvable debate. Pete.Hurd 02:33, 11 October 2006 (UTC)

Well the resolution would be to move this proposal into WP namespace and have a larger community find consensus. ~ trialsanderrors 09:25, 11 October 2006 (UTC)

It's hard to say that moving it to the WP space would find a concensus. There are certain topics where it's obvious a notability criterion is needed (essentially, to keep out spam/vanity). But even those have some flexible points to allow for the argument I'm trying to make. That Science is a rediculously vague topic, and because the number of Geneticists is much larger than the number of Cosmologists, the number of citations for a Cosmology paper to be "as notable" (whatever the fuck that means) as a Genetics paper is lower than the number of citations the Genetics paper has. While the textbook criterion tries to replace that, it does so in a way that hacks at one of Wikipedia's greatest strengths: Being current. WilyD 13:01, 17 October 2006 (UTC)

We're only current if the situation gets immediate attention. E.g. we require two albums for a band to pass WP:MUSIC, that's something that usually takes two to three years to accomplish. But it seems to me that you're looking for a criterion that bestows notability based on being published in a top journal in the field. I'm not all too happy with such a criterion because a lot of half-baked stuff gets published even in top journals, but I could live with it. What I don't want to sneak into Wikipedia is the Falun Gong stuff, where editors try to use a single study published in a fringe journal to push their POV. ~ trialsanderrors 05:23, 19 October 2006 (UTC)

WP:MUSIC is actually very flexible, and a good example of how you can set up a "notability" criteria list without excluding gobs of legitimate content. What I'm suggesting is more like the WP:MUSIC criterion Has become the most prominent representative of a notable style or the local scene of a city (or both, as in British hip hop); note that the subject must still meet all ordinary Wikipedia standards, including verifiability., for instance. WilyD 14:30, 19 October 2006 (UTC)

care to formulate it and put it into the proposal? Also, no criterion is there to exclude. Only the total of criteria can exclude (and of course only with a qualifier that this is an approximate guidline and nothing is set in stone). ~ trialsanderrors 18:38, 21 October 2006 (UTC)

Err, maybe - but I really don't like the citation count criterion, especially for something as vague as "science" - no notability guideline requires a number higher than two for any criterion. The point is that in a lot of fields, 100 is a very inappropriate guideline, for a variety of reasons. Citation counts are a very strong function of time [adsabs.harvard.edu]ADS offers a normalised citation count function, for example. If I look up Hubble's most important paper, I find - 324 cites[3] ADS's cite count may be incomplete (I'd be surprised if it weren't) - but it's clear to me that Hubble expansion of the universe is several orders of magnitude more notable than any reasonable threshold. Sure, it passes other criteri(a)(on), but in a lot of ways all the criteria here block very recent work - attacking one of the strengths of Wikipedia. WilyD 14:50, 31 October 2006 (UTC)

[time to reset the indentation again] Actually #4 covers recent work. You're trying to create some hypothetical of a groundbreaking scientific discovery that should get immediate encyclopedic attention but is somehow ignored in the scientific community and the informaed public that is really in my book some romantic notion of the lone scientist working alone against the conceived wisdom of his/her peers. How often do we actually have those cases? And compared to that, how often do we have cases of fringe scientists who claim this romantic notion for themselves and try to use Wikipedia for self-promotion? ~ trialsanderrors 07:07, 2 November 2006 (UTC)

But that's really not what I'm getting at at all. I'm talking about theories or what have you that are the leading edge of their field, but have little popular coverage/low citation counts because the field doesn't have a lot of researchers working in it. I'm not talking about work that goes against the wisdom of the scientific community - but work that *is* the wisdom of the scientific community, but because of the small number of researchers active in the area, doesn't meet the very high standards set by things like 100 citations. Self-promotion can be squashed with WP:SPAM and WP:VAIN - if that's all you're concerned about, there's no reason to make a guideline like this at all - it doesn't accomplish anything except to excise legitimate content in that case. WilyD 19:10, 2 November 2006 (UTC)

Here is the current version from WP:PROF: There is no objective criterion for establishing that a publication is "widely" cited. Wikipedia editors should consider not only the absolute number of citations (as provided by a citation index) but also the number relative to other publications in the same field which are generally acknowledged to be important. Which seems like a good answer to your concern. ~ trialsanderrors 18:58, 3 November 2006 (UTC)

Yeah, that seems like a much better way to gauge cites. WilyD 19:33, 7 November 2006 (UTC)

thoughts on...

NPA personality theory and it's author Anthony M. Benis up for AfD? Pete.Hurd 03:39, 2 November 2006 (UTC)

Seems like an obvious case of bandwagoneering to me. I also asked User:Psychonaut if he has an opinion on this essay. ~ trialsanderrors 06:55, 2 November 2006 (UTC)

Probability-based strategy? for those inclined to offer opinions, mine are on the article's talk page. Ummmm, please feel free to point me to a better place to solicit options on these things.

Lysenkoism isn't science

Re your proposed criteria for notability in science: The fourth criterion is a major no-no.

The scientific merit of the theory is disputed in, or rejected by, the scientific community, but it has received significant attention in political circles and ongoing coverage in the popular media. In this case the article should make note of this status.

Doesn't look at all sensible to me. It is a common misconception that the scientific community defines science; empirical evidence does that. However, if the scientific merit of a theory is rejected by the scientific community, you can be assured that this is for an excellent reason, i.e. the theory is not supported by evidence, directly contradicts verifiable evidence, or is non-testable. That the theory remains popular for political or populist reasons has no bearing on the facts. To mark a theorem as notable on these grounds would be a form of Lysenkoism, politics disguised as science.

As we've seen with the ID debate, one tactic used by proponents of non-scientific or pseudo-scientific (or just plain crackpot) theories is to gain credibility by debate rather than presentation of evidence. Your fourth criterion would play straight into the hands of these people and severely damage the reputation of wiki on scientific matters.

The theories that meet this fourth criterion may well be described as "notable" for some reason or other, but scientifically notable they are not! Steve 14:04, 6 November 2006 (UTC)

Emphatic disagreement. Science is not the sum total of what is empirically verified, it is the process of doing science. Lysenkoism is notable as science, scientists are familiar with the idea and it's history, it is part of the knowledge of any educated biologist as part of their scientific training. It is important to know of it because of what it says about how science is conducted. Newtonian mechanics are notable, Lysenkoism is notable, Lamarkian inheritance is notable, etc etc etc, these things are not airbrushed out of "science" once they are no longer held as *true*. Admittedly, Newtonian mechanics, Lysenkoism, and ID lie on a continuum, in that most scientists ascribed to Newtonian mechanics in it's day, while a smaller proportion ascibed to Lysenkoism and fewer still to ID, but that just kind of underlines your initial point. Pete.Hurd 16:33, 6 November 2006 (UTC)

I think you are confusing the body of knowledge with the 9-to-5 job. The process of "doing science" only is science if it is producing new knowledge empirically consistent with the extant body of (empirical) scientific knowledge. If it isn't then you have the form without the substance: pseudoscience (...even if it's done in labs by men wearing white coats!) To not only not do the science, but claim scientific validity anyhow is one step worse than pseudoscience. I'm afraid that is exactly what definition 4 would allow. Lysenkoism is, as you rightly say, notable. What it isn't is science. So, why include it in the scientific body of knowledge when all that inclusiuon can do is unjustifiably promote one and unjustifiably corrupt the other? Moreover, it is fundamentally incorrect to see Lysonkoism, ID and Newtonian mechanics as part of the same continuum. They are chalk and cheese. One of them is a well-formed scientific theory that, within well-understood limitations, is experimentally verifiable. The others are political or religious movements. Why try to shoehorn them together? If there is a point to be made about these subjects, it is not sensibly achieved by mislabelling any of them. If there is a point to be made about valid but fringe science then yout shouldn't put it in the same pot as pseudoscience. If it is valid science, it will be testable and it will pass those tests and scientists everywhere will be happy.

If you want to put changes in the scientific body of knowledge over time into perspective then you have to differentiate between those that have been superseded by better theories but are still valid within known limits (e.g. Newtonian Mechanics) and those that should never have been termed science at all (e.g. Lysenkoism). Even if it is with 20-20 hindsight, we should not be calling those ideas science. We should call them for what they really are. That's not airbrushing them out. It's just being accurate.

Now, it seems to me that what Trialsanderror is trying to achieve with definition 4 is to include controversial but otherwise valid theories as science. Well, if they really are valid (i.e. empirically consistent) then he doesn't need to as they already are science. What makes such science notable is the non-scientific handling (objection or promotion) of it. The Bush administration's policy of denying climate change on economic grounds would be a good example. So, play up the politics when the science is good, but please don't do it when the science is bad. It only encourages them. Steve 23:21, 6 November 2006 (UTC)

It's been a few years since I last took a class on epistemology, but it was my understanding that the philosophers were pretty much in agreement that there was no workable criterion which separated "science" from "non-science". The idea that the quality of empirical fit, or verifyability, is the hallmark of science doesn't work well when applied retrospectively to cases such as the copernican revolution, when novel heliocentric theories made poorer empirical predictions than epicycle enhanced earth centric ones. I'm not aware that there is an off the shelf bullshit detector test that will readily distuinguish "valid but fringe science" from "pseudoscience" as one distinguishes cheese from chalk. I wish there were, but I don't think there is. I would not want for instance to use the empirical verifyability test to say whether string theory is psudoscience or not. I don't think Trialsanderror is making a strong epistemological claim about whether something is or is not *scientifically true* with rule 4, just attempting to determine whether something belongs in an NPOV encyclopedia. Pete.Hurd 23:54, 6 November 2006 (UTC)

Pete asked me to chime in, although I have only a little to add. Pete is right about the state of the demarcation problem (a.k.a. the problem of saying what is and is not science) in the philosophy of science. It is important to distinguish two kinds of problems: (1) whether a theory is a scientific theory and (2) whether the prominent advocates of a theory are using proper scientific methodologies in supporting their theory. A good example to distinguish these two is young earth creationism. The theory is absolutely scientific; it makes empirical claims which are subject to confirmation or disconfirmation. Its proponents on the other hand, are probably not engaging in scientific practice, because the hypothesis has largely been disconfirmed. With respect to (1), most of the popular solutions have been abandoned. Falsifiability was a popular criteria 50 years ago, but it has many problems. "Subject to confirmation" replaced it, but confirmation holism (including weaker versions) provide a problem for this view. Other views have been proposed, and if you have a favorite one, we could talk about it. Either way, though, I think Pete's right that this is tangential to the purposes of this essay. --best, kevin [kzollman][talk] 04:03, 7 November 2006 (UTC)

I think we should quite simply not take on the role of deciding what is science, not-science, pseudoscience, bad science, valid science, or whatever else one might think the notability guideline tries to do (hence the disclaimer at the top of the draft). Our role as encyclopedists is not to examine the quality of the original expression, but to detect if it created an echo. If the echo is strong and resoundingly negative, it is still an echo and within the scope of this notability guideline. If it only occurs outside the channels of standard scientific discourse, it is not, but might still be relevant and notable as a current event. The only scientific contributions we should not include are the ones that create no echo, and the discussion between WilyD and me is mostly about where to draw the line for this, i.e. is publication in a peer-reviewed/top-tier journal enough or do we require more for inclusion? ~ trialsanderrors 06:20, 7 November 2006 (UTC)

I think that immediately calls NPOV into question on a massive scale. In an empirically-based subject area, the only true measure of NPOV is agreement with empirical facts, whether those facts are distateful or not. Any article that apparently legitimises a theory in the face of empirical fact is immediately non-NPOV. Now, I believe you are correct that wiki should not decide what is science, but it needs to at least take a reasoned and supportable stance on what is science or it is immediately non-NPOV. (IMHO, deciding that stance is not possible within the existing wiki model. It's a great and worrying flaw. However, that is not a debate for here.)

One of the reasons that I have belatedly become involved in wiki is related to this: the manipulation of wiki by fringe and partisan interests seemingly under the guise of NPOV. Many contributors believe that to say a subject is misguided or just plain wrong is not NPOV. That is a fallacy. Facts are not points of view. If something is demonstrably incorrect then the only NPOV is to say so. To grant even a vestige of respectability to a subject on the basis that it is claimed to be science even though it cannot be shown to be science is to play right into the hands of the proponents of what is, by definition, a falsehood. In my mind, definition 4 allows precisely this. So, if something is pseudoscience, even notable pseudoscience, it needs to be labelled as such or it is not NPOV. Steve 10:57, 7 November 2006 (UTC)

"if something is pseudoscience, even notable pseudoscience, it needs to be labelled as such" there is no rule for doing this. A great deal of effort by very clever people has failed to generate a rule for doing this. Entire academic disciplines have been built around attempts to do this, no success. "the only true measure of NPOV is agreement with empirical facts" agreement with empirical facts is not a simple metric, Philosophy of science#Theory-dependence of observation, Philosophy of science#Indeterminacy of theory under empirical testing. Attempts to impose such rules on science, e.g. Behaviorism's attempts to impose a rigid logical positivism criterion for "science" on psychology and subsequent witch-hunt against all other methodologies as "witchcraft" was a miserable failiure, the behaviorist thought police greatly stifled scientific progress until they were de-throned. There are such things as bad science and pseudoscience, but you're not going to be able to classify them cleanly, and not without a bunch of debated points. That's not a bad thing, that's what belongs in the articles. A simple "theory true / theory not true" flag in the upper right corner of the article is just not possible, and not a constructive aim for WP articles. Pete.Hurd 16:04, 7 November 2006 (UTC)

I think you are exaggerating the scale of the problem, particularly by using a subjective, disputed and non-rigorous philosophical counter argument :). In practice, there are very few cases where someone would not be able to make the judgement call. Often, that someone can be anyone. Occasionally you need a co-ordinated body with delegated responsibility to make the decision. Unfortunately, that last is not the wiki way. As I've hinted, this is a major flaw in the way wiki approaches scientific entries. Steve 18:44, 7 November 2006 (UTC)

Your certainly right that people would definitely be able to make the judgment call, the problem is different people would make different calls. Equally qualified scientists are at odds over whether or not String theory is an appropriate science. Similar claims have been leveled against Bohmian mechanics (or any interpretation of quantum mechanics). Equally qualified scientists disagree over the prospects of sociobiology and evolutionary psychology. Some scientist accuse the use of generalized utility functions in economics as non-falsifiable and thus non-scientific. Steven J. Gould and Richard Lewontin accused the adaptationist program as being either a non-science or bad science. Many theories in sociology and social psychology are criticized as being so general that they could fit any fact. Shall I go on? --best, kevin [kzollman][talk] 19:23, 7 November 2006 (UTC)

[Edit conflict, response to Steve] First, as for every objection I always ask back how you would phrase it. And second, I don't see how we can implement a criterion that tags even the flimsiest scientific proposition as "pseudoscience" without taking on the role of evaluating the facts, a role we don't have per WP:NOR. If a proposition that claims to be scientific is soundly rejected by the scientific community we should report it as such based on sources: "X is a proposed explanation for Y that has been rejected in scientific circles" or "is controversial in scientific circles". Clearly we have an article on Intelligent design which more or less (and I'm sure after lengthy wrangling) follows that pattern. ~ trialsanderrors 16:25, 7 November 2006 (UTC)

I wouldn't phrase it at all. I believe definition 4 is entirely unneccessary. To qualify something as you describe, by saying "has been rejected in scientific circles" is, to me, quite acceptable. You (quite correctly) state the "X" as what its proponents actually maintain and you then (quite correctly) state that science rejects it. In fact, there is one particular article that needs rewriting that I would like to rewrite that way. Evaluating facts, by the way, is not OR its [[WP:V]. Interpretation of them might be OR, but that's a different matter in my book. However, I do stick to the position that notable bad science is not notable science; it's notable something else. Steve 18:44, 7 November 2006 (UTC)

Even if we could find a metric for fitting empirical facts and label articles which fail to correspond, I think it would be a bad idea to do so. Pete already mentioned that the first heliocentric models where worse than the previous models which utilized epicycles. Similarly, General relativity had little speaking for it when it was first published, although it has since been tested. Several interpretation of quantum mechanics are not confirmed (in the sense that you would prefer one over the other). This is a general phenomenon known as Kuhn loss, and there are more examples. On the other side, much of intelligent design is very well confirmed, remember the hypotheses are specifically designed in order to fit the facts. "God made it so that it would be like it is." Similarly, conspiracy theories are designed so that they could never fail to fit the facts. The criterion of "fitting the facts" is a very poor guide to what we think of as legitimate science. --best, kevin [kzollman][talk] 18:01, 7 November 2006 (UTC)

Some thoughts

As it stands, I want to bring up tw articles:

1. MOND
2. Caloric Theory

It seems to be obvious to me that both these topics "merit" articles. It's not clear to me whether the first one passes by any criterion here (though it probably does get past the good citations/reasonable journal criterion. The second one passes the textbook criterion I'd guess (at least, I have to imagine such a textbook exists). While you're trying to ask how to exclude pseudoscience from being presented as science, or to stop violates of WP:SPAM and WP:VAIN, I think it's wise to keep in mind the kind of stuff that needs to be accounted for the in the mix. WilyD 15:41, 8 November 2006 (UTC)

I don't know what the first one is. Is this supposed to be MOND (Modified Newtownian Dynamics)? If you want you can add a criterion for superseded scientific theories. I think it's pretty clear from the categorization alone that the problem with vanity doesn't go for obsolete theories if they're labeled as such. ~ trialsanderrors 19:10, 8 November 2006 (UTC)

OK, it might've helped to cite the 1983 papers in the references to make the search easier, but the citation count for the first paper (ASTROPHYSICAL JOURNAL 270 (2): 365-370 1983) is 341, for the second (ASTROPHYSICAL JOURNAL 270 (2): 371-383 1983) 167, and for the last (ASTROPHYSICAL JOURNAL 270 (2): 384-389 1983) 93. That's clearly within #2. But re-reading #2 now we should also account for contributions that did not start out in peer-reviewed journals but were then widely discussed in the field. ~ trialsanderrors 20:54, 8 November 2006 (UTC)

I havn't been following closely, but if the citation counts peer-reviewed articles citing and that meets #2, then it might not really matter whether or not the target article was in a peer reviewed journal? Pete.Hurd 21:01, 8 November 2006 (UTC)

Like, conference proceedings that are then cited in peer reviewed journals?As for MOND, I'm surprised the cite counts are that high (I'll admit I didn't check) but that's probably still very low compared to the leading alternative (CDM) - the first year WMAP paper has 10x that many cites (3584 by ADS) for instance. Or do we just mean to imply (well cited relative to an average paper in the field? WilyD 22:52, 8 November 2006 (UTC)

[Squeeze in comment] I don't think a comparison to the cite count of the leading alternative is necessary. Our criterion for WP:NPOV#Undue weight is that it is being held by at least a significant minority (and prominent adherents). ~ trialsanderrors 05:56, 9 November 2006 (UTC)

I suppose conference proceedings, or perhaps a book like A New Kind of Science makes for an interesting test case. There are certainly plenty of non-peer reviewed things that crop up in the science citation index. I'm remaining mute on the actual thresholds, but I think rule #2 could be relaxed in terms of the original publication is peer reviewed if it gets cited as if it were peer reviewed, then that's just as scientifically notable, know what I mean? Pete.Hurd 23:18, 8 November 2006 (UTC)

ANKoS is a book by a notable author, so it really doesn't fall under this guideline (I didn't read it, although I liked a review that called it something like "one infinitely smart monkey typing away on a typewriter"). Clearly cellular automata are notable, but I don't know what contributions came from it that have been introduced into the normal pipeline of scientific ideas that we could consider here. I changed #2 to allow contributions written on the backs of beer coasters. ~ trialsanderrors 05:23, 9 November 2006 (UTC)

I used the ANKoS example because (IIRC) Wolfram self-published the book because 1) he could afford to, and 2) it avoided the hassle of submitting his work for peer-reviewed publication (a hassle he didn't want to deal with). Wolfram presents the work as amounting to a very important scientific theory, if it is (debate elided here) it's not peer reviewed but may be (at home where citation index access is painfully slow, so won't check) cited out the wazoo by other authors peer-reviewed work, blah blah blah...

"allow contributions written on the backs of beer coasters" excellent! now if you could just convince NSERC to similarly recognize my long-underappreciated Full Moon Pale Rye Ale magnum opus in six coasters^H^H^H^H^H parts (not only insightful, but strangely refreshing). Pete.Hurd 06:27, 9 November 2006 (UTC)

Oh, ok. Since I'm equipped with an Anderson Valley Winter Solstice Ale and the Townes Van Zandt bittorrent isn't eating bandwidth I can quickly confirm that ANKoS got about 360 cites. About your magnum opus, I doubt it meets our notability standard unless you can provide at least six coasters for notable brews that cite it. ~ trialsanderrors 06:49, 9 November 2006 (UTC)

Test case: Winning strategy

Here's an interesting one: Pete, Kevin and I think this article's got nothing to do with game theory, Trovatore says it's got nothing to do with combinatorial game theory. The term itself has quite a bit of a footprint in the usual search engines, but I haven't seen anything yet that might amount to a feasible definition or a trajectory of common use. It might be interesting to see how it fares at an AfD, where "90 hits on Google scholar" is often an argument in favor of keeping. ~ trialsanderrors 20:31, 14 November 2006 (UTC)

It's definitely a real concept, used in relation to the Axiom of determinacy in set theory. There is a small group of articles that are sort of strange birds, they aren't really part of the cannon called game theory, although they're not really combinatorial game theory exactly either. I have talked with a few set theorists around here and can verify the legitimacy of the article, but I don't really no where to put them. --best, kevin [kzollman][talk] 20:45, 14 November 2006 (UTC)

It seems to me that it started out as something akin to what's described better in Determinacy#Strategies and then morphed into something that reads like half-remembered ideas from a combinatorics class. The question is is there anything worth salvaging or should we redirect it to Determinacy? And if we salvage it, what are the sources to rewrite it from? Certainly if we want to put it into a game theory context things like "winning", "rule" and "lookup" table need to be explained. ~ trialsanderrors 21:26, 14 November 2006 (UTC)

Quantum ecosophy

Up for deletion. Enjoy. ~ trialsanderrors 22:37, 16 November 2006 (UTC)

And a bit more serious; Emergent philosophy (not up for AfD yet). ~ trialsanderrors 22:46, 16 November 2006 (UTC)

Up on RFC, Eutherian fetoembryonic defense system (eu-FEDS) hypothesis. Pete.Hurd 00:22, 19 November 2006 (UTC)

Six cites? That could be interesting to follow. Didn't read through it yet, but I noticed this. ~ trialsanderrors 02:43, 19 November 2006 (UTC)

Yeah, I noticed that too, threshold seemed much lower... More to say later... Pete.Hurd 03:59, 19 November 2006 (UTC)

FWIW, [4] Pete.Hurd 06:27, 20 November 2006 (UTC)

I'd say the article is perfectly AfD-able, especially with the author's comments which pretty much amount to WP:OWN. WP:RFC doesn't seem to create too much traffic. ~ trialsanderrors 03:24, 21 November 2006 (UTC)