One of the classic “methods of proof” that is part of a classic joke provides the “Proof by reference to inaccessible literature”, which goes a bit like this:
“The author cites a simple corollary of a theorem
to be found in a privately circulated memoir of the
Slovenian Philological Society, 1883.”
Assume, for a moment, that you are an academic writer (anywhere from an undergrad student to a PhD researcher with 10 grants and 100 publications).
Assume that the source actually exists, and is not just made up. Of course, if it is truly inaccessible (or virtually so), then the reader, professor, or reviewer is going to have a difficult time verifying it and convincing himself that you didn’t just make it up…
Could this actually happen in real life in academic writing? Has it? Are there any academic writing guidelines/policies that specify how “available” a source has to be before you can cite it? Can you cite it as long as you keep a personal copy in your office or something and make it available for review, or know where a copy can be found? What if you take notes and document the source, and then the only known copy “disappears”?
Does it matter if the referenced document is a primary source (e.g. a never-published firsthand account of the First Battle of Bull Run in a love letter that your grandfather said that his grandmother got from her first boyfriend), or a secondary or tertiary source (e.g. a set of handwritten notes dated 1820 that you found in the ruins of the old schoolhouse)?
Reference to actual academic policies, or actual incidents would be interesting!
If it is something like a mathematical theorem (as per your example), or indeed an argument of any sort, you ought to be able to just reproduce it (write it out again). However, if we are talking about inaccessible evidence then this sort of thing is bound to happen all the time. Lots of crucial historical documents, for example, can be very hard to get at, but, come to that, the evidence for many scientific findings is pretty much inaccessible in practice: you might have to spend years re-doing the experiments (the worst part usually being learning for yourself how to actually carry them out).
In practice, researchers have to rely on one another being honest about their sources, and a good part of academic training is concerned with (informally) instilling this ethic of academic honesty. (That is why actual academics are horrified by things like plagiarism, when many students don’t see what the big deal is.)
But actually most areas of life are like that: You have to rely on most people being honest most of the time. If people mostly lied and cheated indiscriminately, society could not function. Fortunately (or, rather, because we go to great lengths to socialize out kids into being mostly honest) most of the time they do not.
I’ve certainly run into this while reference-chasing. The thing is, though, if the result is any good, or of any significance, it’s probably going to show up elsewhere, in more reputable (i.e., accessible) sources.
You also see private communication cited as references sometimes. Usually, these are for side notes to what the paper citing them is about: You wouldn’t want to rest an entire paper on them. And generally, the person who privately communicated to you (likely someone at the same institution, or an old advisor/student) also has a paper in the pipeline on that subject, and just isn’t quite ready for publication yet. So if you find a reference to “private communication” in a paper that’s more than a couple months old, you can usually poke around for works by that person since then and find it.
The most extreme case I ever saw, my old advisor once wrote a paper with a cite to “M. Scott, private communication”. I still haven’t figured out whether the reviewer let it through because he got the joke, or because he didn’t.
I was recently reading an article in which an historian cited a document believed to have been destroyed in the collapse of the Cologne Historisches Archiv in 2009. Now, it is I suppose possible that another historian might have obtained images of that document or made detailed notes which could confirm the information being cited. But the chances must be that the notes made by the historian writing the article are the only surviving record of that document’s contents. So what’s he supposed to do? At least by using it in print, he is ensuring that something of its contents is not completely lost.
Another type of case is when an archive is closed except to a single researcher, such as an authorised historian or biographer. Any cites to those records must be taken on trust. Except that the historian with the privileged access must work on the assumption that the archive in question will eventually be made available to others. That might not happen until long after he or she is dead, but even then it would still be an obvious move for future historians to go back and check on their honesty.
Of course, there have been cases where writers have cited obscure publications which have turned out never to have existed. The recent fuss over the supposed meeting between Dickens and Dostoevsky is a good example of that. Claire Tomalin now accepts that she was duped.
One assumes that the cite was in reference to the invariance of some constant or other definition of relationship between metrics.
There is a whole range of things that go on with references in papers. Some of it depends upon the field and nature of the research presented. Mathematical papers for which a core thread of a derivation depend upon a difficult to substantiate proof are going to be called on it in the review process.
On the other hand, many citations are not of this form. One of the important aspects of citations is that the researcher show absolute honesty in attribution. If an idea or contribution comes from someone you spoke to about the issue in the corridor, citing a private communication is perfectly reasonable, indeed expected. My favourite such acknowledgement is Claude Shannon acknowledging J. W. Tukey as the inventor of the word “bit” to define binary digit.[1] (Although this wasn’t actually a citation with formal reference, the spirit is there.) But the flow of a paper must provide the source of input to the new contribution to the state of understanding that is being made. If you got an idea, or some other contribution to the work, from a source, it should be cited. Otherwise you are implicitly claiming the idea as your own, which is simply not done. Citations are also an academic lifeblood. It isn’t enough to be published, to be cited is worth many many times as much. It proves that your work is valued. So proper citations of other works is paying your dues. It is generally agreed that the primary paper in an area is the one that is cited, and other follow up papers only if they contributed additional work that is directly pertinent to the paper being written. Thus allowing the pioneers in an area to gather the maximum credit for their work via citation counts. Eventually a result ceases to be cited when it becomes impicitly part of the background knowledge.
So, citation of obscure publications, private communications, and in the modern time, blogs, and web pages, isn’t always evidence of shoddy reasoning, it is evidence of correct attribution of ideas. Thus it is important it be done, and modern publishing embraces all of these sources.
[1] C. E. SHANNON, A Mathematical Theory of Communication, The Bell System Technical Journal, Vol. 27, pp. 379–423, 623–656, July, October, 1948.
I know of one physicist who cited an imaginary person e made up, thus making it even more ephemeral than Chronoc’ advisor’s cite – at least someone else made up Montogemoery Scott, and lots of people knew who the character was.
He cited this imaginary person on two occasions. I thought that he’d included him as a co-author on one paper, but I haven’t been able to find that one.
Then, of course, there’s the case of the chemist who included his cat as a co-author, not unlike Rita Mae Brown on her mysteries.
Has it ever happened at the undergrad level, where things are more, shall we say, “mechanical”? E.g.
robert_columbia
Advanced Discrete Mathematics 443
UCLA
Prof. Jones
Spring 2013
"In this paper, I will establish that finding all the roots of a polynomial equation in polynomial time is not possible.
…<evidence that finding the roots of a polynomial equation in poly time is equivalent to solving the Traveling Salesman problem in poly time>…
According to Robinson (1), solving the Traveling Salesman problem in poly time is equivalent to the Halting Problem. Every computer scientist knows that the Halting Problem is not solvable in a finite number of steps. QED.
Notes:
Robinson, J. Initiatory Documents of the Second Degree of Loyal Engineers of the Prairies, Billings Lodge, Montana, 1940.
"
The Professor contacts the lodge:
Prof:“Excuse me, my name is Professor Jones, at UCLA. One of my students cited your “Initiatory Documents of the Second Degree of Loyal Engineers” in a paper, and I wanted to see how to get a copy to verify the source.”
Lodge Rep: “I’m sorry, that document is available to initiates only. In order to be initiated, you must work as a Civil Engineer in Montana for at least 5 years and be approved by a majority vote of the current membership.”
Prof: “But the semester ends in three weeks! I have to get this graded now.”
Lodge Rep: “Bummer, man! Sorry, can’t help you there.”
Are there any schools that have rules against students doing this sort of thing, or is it the professor’s problem?
I doubt very much there are explicit rules prohibiting such a thing. But whether or not to accept a reference in grading a paper is going to be at the professor’s discretion, so it’s the students problem, not the professor’s. Using an unpublished, unrefereed reference, even if publicly available, as a crucial reference in a paper would not be acceptable in the first place. If I were grading such a paper, I would insist on a better reference.
Even more trivial than that. The paper was on the Alcubierre metric (a theoretical metric describing something which might be considered a warp bubble), specifically the view of the exterior space from the bridge of such a ship. In the introduction, there’s a bit about “It might be argued [citation] that the center of the bubble would actually be in the engineering section of the ship, but we use “bridge” here to mean the center of the bubble, for simplicity”.
I’m curious, though, about the imaginary citation CalMeacham mentioned.
If it was mathematics and in a generally inaccessible place (or even difficult to access–see explanation below) what I actually do is include the argument but say something like, “This result is due to X and cite X’s paper in the inaccessible place. We include the argument here to make this paper self-contained.”
Nowadays, many math journals cost several thousand dollars a year and most libraries have stopped subscribing. So checking out the journal would involve inter-library loans and take a long time. I am aware of at least journal that is subscribed to by exactly one library in all of North America. Reduce that number by one and it becomes inaccessible except by extraordinary effort.
It is odd that while the costs of producing a journal have fallen drastically, the price has risen precipitously. There are a number of online journals that are absolutely free because, the major costs, typesetting, printing, and mailing, are gone.
Ah. So you would say something like (as in my previous example) “…According to Robinson (1), solving the Traveling Salesman problem in poly time is equivalent to the Halting Problem. Robinson’s work is difficult to obtain a copy of and a copy of the relevant proof is reprinted in Appendix A. Every computer scientist knows that the Halting Problem is not solvable in a finite number of steps. QED”
So then, the reader can read the proof, and you have acknowledged that it’s isn’t your proof so there’s no question of you being seen as plagiarizing?
Or, I suppose you could just indicate the note in the footnote (or entry on the Works Cited page) itself, directing the reader to Appendix A.
Then, the question arises, “What if copying enough of the work is so substantial that it isn’t ‘fair use’ under copyright? Though, that might be attenuated if the work is hard to get…”
I’d love to tell, but, even after all this time, I wouldn’t want to get anyone in trouble. I can’t even tell you the imaginary name, since you can locate it with that and a reasonable search engine.
The Chemistry guy, though, made it into the news. I just can’t recall his name (or the cat’s)
I had this odd fantasy of going back to school and intentionally taking a cross-country or international journey over Spring Break to find a (legit) citation for a paper that can’t be found at any local library, just to mess with the professor.
I don’t think anyone has mentioned so far the simple fact that languages are various, and many relatively inaccessible to non-specialists – analogous to the notion that many sources indeed are held in unique sites.
Here’s a two-fer: the Husserl Archives at Louvain. Only a handful of people (that I know of) can read Husserl’s unique shorthand. There are many thousands of pages written by him that will likely never be transcribed, and only exist because of fortuitous opportunity to smuggle them in empty fishbarrels, thanks to those filthy German pigs and that German pig Heidegger and the rest of those Nazi scum.
Also, consider a not-inconsiderable amount of work still left untranslated to German from Polish by Ingarden. I don’t know the extent of such untranslated work (Ingarden often translated his own work into German, or else wrote directly in German language), but it’s safe to say your average academic doesn’t read Polish.
And there’s the trivial matter of, say, an undergraduate citing and translating himself or herself a quotation from a standard language like German or Italian or Spanish for an English paper (to be read by an English Professor). English Literature Ph.D.s are not particularly known for being good with languages, as a group, so a cite from a very dense, very long source in a core language like German could likely be scrubbed over by a would-be wit by providing a false page number reference or slightly mistranslating a passage in the favor of his or her thesis. This could certainly happen in a published paper on an obscure topic in a not-very-high-stakes field like literary criticism, unlike, say, philosophy or literary theory (iff the editors of the journals of practice of the latter fields have held strong against continentalism and twaddle).
Prof: “Uhh, <student>, did you have to use all these Moroccan sources on your paper about foreshadowing in Macbeth? There’s plenty here right in the school library. I’m having an awful time finding someone who can read the language to help me verify that you translated your sources correctly. The only Arabic speakers we have teaching here speak Egyptian Arabic, and they tell me that the Moroccan dialect is incomprehensible to them.”
<student>: “Hey, don’t blame me, my father got me those on his trip last month. Seemed a waste to not use them considering I already know the language…”
Very simple: you paraphrase it, rather than copying it verbatim. In fact I would invariably do that anyway. For one thing, I do not feel I understand it unless I am capable of paraphrasing it and it relieves me of any worries about copyright. I recall from a year or two ago reading a claim that took about three lines. By the time I sorted it out, it occupied a half page of close reasoning. Here is a direct quote of my wording: “The forward implication is based on the proof of \cite[Theorem 1]{Hoc}.” In this case it wasn’t the publication that was obscure (Transactions of the Amer. Math. Soc.) but the argument itself. But that doesn’t change the principle.
The question does arise, what if I am just quoting the theorem without understanding the proof. I don’t like to do that; you can propagate errors that way. But I have done it, if the argument was based on methods I wasn’t familiar with. I had something lack a few years ago and it was crucial to what I was doing. I still feel uncomfortable about that.
On rare occasions I’ve had to refer to an ancient article (early 1800s) that I was unable to obtain a copy of. In such a cases, I’ve resorted to a second-hand cite, that is, citing a later article that quoted or referenced the article. (That is, Smith 1811, as quoted in Jones 1978.) But I hate doing it, because sometimes when you do get the original reference you find they said something different or it was taken out of context.
I am reminded about an anecdote about anthropologist Louis Leakey, who grew up in Kenya and spoke fluent Kikuyu. When he was at a university in England, he wanted to use Kikuyu as his foreign language requirement, but the university couldn’t find any examiners who knew the language. The sent out inquiries to linguists find out if they could recommend any experts in Kikuyu who were living in England, without explaining why they needed one. The name that came back was Leakey himself!