Thursday, March 19, 2009

What Does Everyone Know?

Yesterday I ranted a bit about a lame way that some reviewers try to sink a manuscript by making unsupported/unreferenced claims that there is nothing new in the manuscript. This reminded me of an issue that is related to what I called we-already-knew-thatism. And that is:

When does something become Common Knowledge that (almost) everyone really does know, and therefore no citations are necessary, and when does an 'old' concept still require a citation?

I talked about some aspects of this a while ago (Onset of Conventional Wisdom), but, as happens from time to time, I seem to have more to say on the topic.

There are certain statements that I think everyone would agree need no citation. If, for example, it is somehow relevant to your research that the Earth is not flat, you likely don't need to provide references to support the concept of a non-flat Earth.

Other statements would have required a citation during an earlier time, but at some point most people stopped providing citations because the concept became Common Knowledge. That's the situation that interests me today.

We interrupt this blog to say that as I was typing that last paragraph, a grad student wandered into my office to complain about a paper in which the author didn't cite other people's work in his discussion, as if he (the author) was taking credit for everything or as if the other people's work was Common Knowledge, even though (in my student's opinion), it isn't.

Anyway..

What is the typical time range for the evolution of a concept from citation-required to everyone-knows-that/no citation required? Is there a typical time range?

Probably not, but if I had to guess a typical time range, I would say about 12-15 years, corresponding approximately to the time from an academic's grad school days to the point at which they become wise and all-knowing. A more realistic answer, however, would be that the time range also varies with the fame of the person who published the original papers on the concept of interest. In this case, the concept of fame includes other elements that contribute to or hinder one's ascent to galactic academic fame (nationality, gender, academic pedigree) and therefore neverending citation nirvana.

Another factor is related indirectly to what I discussed yesterday: If someone manages to publish a paper on something that is instantly obvious -- as in, something that makes so much sense but no one had said it that clearly before or no one had bothered to publish it before or there was a breakthrough that explained something rather fundamental but perhaps not surprising -- I think those types of concepts very quickly become Common Knowledge because it is difficult for us to imagine a time when the issue was in doubt. Then, although the author may have done a great service to the community by explaining or clarifying an important puzzle or issue, the citations might trail off pretty quickly after publication.

Such is (academic) life, I suppose. I certainly can't say that the papers I write never contain uncited statements that should be cited, but one thing I do when editing my students' manuscripts is to indicate uncited statements that I think need a reference. It's good to get into the habit of being aware of this issue early-on in your writing career. It's the right (fair) thing to do, and also, as my student's comment today showed, people might think you are a jerk if you are too sparse with the citations. It's probably best if you start your career without a reputation for being a jerk; you can always acquire that reputation later if you want to.

14 comments:

  1. Relatedly, when one lab publishes an exciting paper that challenges a piece of Common Knowledge, everyone cites it. Later, when another lab publishes a smackdown paper that makes it clear that the first lab's paper was incorrect, people think, "Oh, so Common Knowledge was correct after all." Thus because the Smackdown paper instantly enters the collective accepted wisdom, it may not get cited much.

    ReplyDelete
  2. As a grad student, I definitely have a hard time with when I should cite things. I also tend to have a hard time with citing the correct things. How am I supposed to know who really came up with some idea or concept first? I think my default is simply to cite EVERYTHING with whatever articles I can find, although this is probably not the best course of action. (Just the one that might get me in the least trouble.) After all just because something is common knowledge in some subfield of physics, does that really make it common knowledge for everyone? At what point should the condition for citing/not citing be is this common knowledge for people who are going to read this paper? Often I think I could probably get away with citing a lot less, as things that I didn't previously know seem to be common knowledge for most people in my field already. Or maybe that's just my imagination.

    ReplyDelete
  3. So true. Folklore is a real problem in mathematics because there's a strong vested interest in keeping it that way. Less competition.

    ReplyDelete
  4. The citation bar is, I think, set at different heights in different [sub-]disciplines. I believe in some fields one is essentially required to cite the seminal paper, even if that paper is 50 years old: the world is not flat [Eratosthenes, 240BCE].

    ReplyDelete
  5. I've come across this a lot more lately than I did in the past but ... a lot of journals are now putting a citation limit on submitted manuscripts. One journal I've sent manuscripts to set a limit of 40 citations for new research manuscripts, 50 if you're submitting a review. I'm not sure that I agree with limits because they may force submitters to cut important citations in favor of bumping up cite numbers for their own work or work of friends.

    ReplyDelete
  6. How am I supposed to know who really came up with some idea or concept first?

    That should come with time. Hopefully you're reading a lot of manuscripts in your field. Tracking their citations will lead you to track those citations, which will lead you track those ... so on and so forth. Eventually, you'll get to the original concept.

    ReplyDelete
  7. I think once a result has made it into the textbooks it no longer needs to be cited. Until then, it probably should be cited. This may differ from field to field.

    ReplyDelete
  8. Citations serve two purposes: 1. Give credit where credit is due, and 2. Help the reader understand the arguments.

    Whenever I look for citations to my own work, I concentrate on purpose 1, but purpose 2 is the real deal. If I discovered the Sun, I'd like to get cited every time it is mentioned, but it would not add clarity for readers and would be a waste of space because once I'd pointed it out, only blind people could not see it up in sky all day long.

    I try to only cite references that add clarity to my discussions (or, necessarily, represent potential reviewers), although I've rarely seen a reference to my work that I thought spurious.

    ReplyDelete
  9. I've spent today reading some of the 'Commmon Knowledge' papers in a line of research that I am trying to develop. I was amused to see how uncertain the authors were at the start and how they became more certain over time *even when there seems to be no really good scientific reason for the increase in certainty* (other than the fact that no one got round to publishing the 'challenge' paper because the science was a bit difficult. Very educational! For me it certainly proved the worth of reading some of the original papers (which date back to the late 1960s in this case).

    ReplyDelete
  10. Here's where I get lost, similar to what Nicole said:

    Whose "common knowledge" do I need to be concerned with?

    Because apparently, the majority of people in this country are unaware that there are 3 branches of government. Do I use what they would "know" as the standard of common knowledge, or do I look at the people in my field? For the sake of argument, would I need to cite a definition of the field of psychology in an academic paper? I think it's unnecessary, but maybe I'm wrong.

    And another aspect of this that gets me all confused sometimes - do I cite the source where I learned of something (eg., a book or whatever), or do I need to go look up where that person got their information from and cite that instead? And then what if the info I'm using actually was a citation from ANOTHER source that the book author's source cited in THEIR paper? And if what I'm trying to use is the original source's spin on that information, do I cite both or all of the relevant sources? *head explodes*

    It's endless. And yes, it leads to less mucking up of concepts and theories when you trace back to the original source, but at the same time it can be a truly daunting task. In my earlier days of schooling, we were taught to cite the source where WE got our information from, not where the source got their information from.

    The timeframe in psych seems to be very different than other sciences too. For example, although everyone in the field of psych is familiar to some extent with Milgram's authority experiments, you can't mention them in a paper without citing them, and they're really, really, really old experiments and Milgram is a household famous name.

    ReplyDelete
  11. Yes, the "common knowledge" problem is very tricky in interdisciplinary work. Many things are obvious to one part of your audience, and totally news to another. You have to take a rhetorical attitude towards it. I think John has the right idea (comment above).

    Increasingly, journals in my field are discouraging heavy citation. Editors complain that it contributes to poor writing and many journals have a weak rule that is something like: 1 manuscript page of citations per 4 manuscript pages of text.

    ReplyDelete
  12. And another aspect of this that gets me all confused sometimes - do I cite the source where I learned of something (eg., a book or whatever), or do I need to go look up where that person got their information from and cite that instead?

    At least in my branch (biology) you should cite the original source. Citing someone who cited the original source only in passing, doesn't really deserve the credit you're giving them.

    Of course, in some instances ... citing a review which covers "Common Knowledge" sort of takes care of everything ... you're citing it in case it really isn't "Common Knowledge" and you're giving someone credit for doing a bunch of legwork (i.e., compiling the "Common Knowledge" into a single repository).

    ReplyDelete
  13. I'm an undergraduate (a senior, 34 years old), so take all of this as coming from that perspective. I tend to treat citations similar to how I treat proofs (in math) and comments (in software): I write them relative to my level.

    That is, if I know something and think my peers likely already know it, or if it's in a textbook, I don't cite it. If I came across something in my research, I do cite it, even though it might be something that a more advanced person wouldn't cite. I am sort of considering my peers as my audience (nobody reads undergrad papers anyway), so if they would think, "Where does that come from?" then I cite it. (Of course, I also cite the sources for my ideas, etc., in a spirit of academic fairness. But I'm talking more about whether the readers need a citation.)

    Similarly, in math proofs, I try the write the correct proof such that it satisfies me, which means I may make remarks that would be too obvious for someone more advanced.

    It seems to me that this would work on a more advanced level as well. If nobody reading your paper is going to wonder who came up with this idea that the earth revolves around the sun, then you needn't cite it.

    ReplyDelete
  14. Working on my first thesis right now, and this is one of the problems I've encountered. How much do I have to cite background information? As a thesis my background info starts out more general (there is evidence that the universe began in this fashion and evolved in some fashion, but we can't see it, so a study of foregrounds is necessary) and becomes more specific (detailed references to specific foreground studies). But do I need to go in and reference where I came up with the theoretical history of the universe?...

    I guess to me it seems like "audience" is another factor to consider when making references. If you are writing for more people in your own specific field, it seems like many more things can be "common knowledge" than if you were writing for intelligent non-experts. It's an interesting border!

    ReplyDelete