A blog about the implications of the statistical properties of language
Bibliography management: A lesson in casual written American English
Research-funding grants are the lifeblood of most academic research, and nowhere more so than in a medical school. I’ve been in the same lab at the same medical school for 15 years, and we’ve been a pretty stable group for some years now; we have an excellent funding record, so you’d think that we’d have the grant-writing process functioning like a well-oiled machine. In a lot of ways, we do–at this point, everyone knows their role in the process, and we can write a good grant fairly quickly. Seulement voilà, technology paradoxically slows us down. Here’s an email that I wrote addressing our technology problem, with some possible solutions. What do you care? You probably don’t–unless you’re either studying or teaching English, as many of the readers of this blog are. For you, here’s my email, written in a high-casual register of American English, with lots of typical lexical features of that kind of writing. I’ll give you an annotated version that explains some of the weirder quirks of the language, and then the full text of the email.
Annotated version: English notes
Over the past few years, we’ve struggled with the citations and bibliographies of our grant proposals. From a technical perspective, the issues have been that (1)
we write our grants in Word, (2) we used EndNote for years, but it eventually pretty much stopped working, for reasons that we were never really able to discover/fix, and (3) we tried Mendeley, which worked OK for individuals, but we were never able to get the group sharing functionality to work well, which is a problem for us, since the writing of our grant proposals has generally been a group effort.
Word, EndNote, and Mendeley are commercial products that are widely used either for writing (Microsoft Word) or for managing citations (EndNote and Mendeley).
Our recent solution has been to have one person do all of the citation management for everyone.
The interesting thing here is the noun phrase all of the citation management. Normally, in informal writing (as on this blog), I would write that as all the citation management, without the preposition of. Both are acceptable, at least to some speakers; if a speaker only accepts one of them, it will be the version with the preposition. Presumably I used the version with of here, despite the fact that I tend to use the version without it, because I was writing at the high end of the casual register (since it’s a work-related email), and because I often adjust my writing style somewhat when writing stuff that I know will be read by non-native speakers, as is the case in our lab.
This has not been hyper-efficient, as the citations come from (potentially lots of) different people, and you have to repeatedly exchange the document containing the research proposal back and forth between whoever currently has the conch and the person who’s managing the citations.
Having the conch is a reference to the important/horrifying movie The Lord of the Flies, from Sir William Golding’s important/horrifying novel of the same name. It’s the story of a group of British schoolboys who find themselves marooned on an island and quickly revert to a state of what one might call “savagery” if it weren’t for the case that “savages” typically behave a hell of a lot better than the aforementioned British schoolboys. In the story, the boys find a conch shell, and whoever is holding it gets to speak without interruption. One of my professors in grad school used the expression when managing data elicitation sessions in our courses on linguistic field work, and it stuck with me. Incidentally: the French word for conch is le bulot, which is interesting because (a) it’s a minimal pair for the [u/y] vowel contrast that Americans can typically neither hear nor produce (and I include myself in that, despite having gotten a master’s degree for a thesis on phonetics and subsequently living on and off in France for the past three years), and (b) if you know the word, you can order conch in French restaurants, and it is really good.
It’s hard enough to manage the conch-holding, and when you add the problem of passing the document back and forth between that changing person and the person who is managing the citations, it all becomes a bigger hassle than it seems like it ought to be.
The interesting point here: is a hassle big, or large? In theory, the two words are synonyms, but in practice, they have very different patterns of usage. These differences are a matter of statistical likelihood, versus invariant patterns–you could see either big hassleor large hassle, and I actually initially wrote this as large hassle, but it didn’t feel quite right, and I changed it. You can read about a nice study on how the big/large thing works in English here; briefly, big tends to be used for objects, while large tends to be used for quantities. A “hassle” is not an object, but neither is it a quantity; I guess that the not-a-quantity thing won here.
I suggest here some options, and will state up front that I advocate option (2).
1) Continue the system where we have a single person who manages the citations.
2) Bite the bullet and move from Word to LaTeX.
“I believe in traditional western medicine. Bite this bullet.” Picture source: https://goo.gl/3GE7Ea
To bite the bullet: to do something that you know is going to be unpleasant and that you have been, and/or would like to, put off until a later time. As far as I know, the way that it’s typically used is the way that I used it here: you follow it with the word and, and then follow that with a verb phrase that refers to the thing that you don’t want to do. The expression comes from the old cowboy movie stereotype where the hero gets operated on without anesthetic, and to keep himself from screaming in pain (wouldn’t be manly), bites down on a bullet during the surgery. I’ll just point out that this would be hyper-stupid—if you wanted to break a tooth, biting down hard on a bullet would be an excellent way to do it.
Costs/benefits of the two options:
1 (continue what we’re doing): I’ve laid out the costs out above. The benefit is that we already know how to do this.
2 (move to LaTeX): The cost: not all of us are comfortable with LaTeX. Crucially, Mandelbrot has historically not been thrilled with it, which is important not only because he’s the boss, but because as the PI, he is the one who is doing the last-minute editing, which makes his efficiency crucial at the moment when efficiency matters the most–that is to say: at the last minute. The benefits: (1) the BibTex citation management system for LaTeX is hyper-good, and (2) we now have a mechanism for letting multiple people work simultaneously on a LaTeX document, as we recently bought a lab license for ShareLatex.
Mandelbrot is a reference to Benoît Mandelbrot, a French/American mathematician who showed that Zipf’s own explanations of Zipf’s Law were super-unlikely to be correct. I’ve used his name here to anonymize my boss’s name. LaTeX is a typesetting language that is often used by scientists to write their articles; back in the days when Google was new, one of our natural language processing profs used it as an example of how word sense ambiguity could lead to getting very NSFW (Not Safe For Work) results for an innocent search.
Cordially,
Zipf
The entire email
Hello, folks,
Over the past few years, we’ve struggled with the citations and bibliographies of our grant proposals. From a technical perspective, the issues have been that (1)
we write our grants in Word, (2) we used EndNote for years, but it eventually pretty much stopped working, for reasons that we were never really able to discover/fix, and (3) we tried Mendeley, which worked OK for individuals, but we were never able to get the group sharing functionality to work well, which is a problem for us, since the writing of our grant proposals has generally been a group effort.
Our recent solution has been to have one person do all of the citation management for everyone. This has not been hyper-efficient, as the citations come from (potentially lots of) different people, and you have to repeatedly exchange the document containing the research proposal back and forth between whoever currently has the conch and the person who’s managing the citations. It’s hard enough to manage the conch-holding, and when you add the problem of passing the document back and forth between that changing person and the person who is managing the citations, it all becomes a bigger hassle than it seems like it ought to be.
I suggest here some options, and will state up front that I advocate option (2).
1) Continue the system where we have a single person who manages the citations.
2) Bite the bullet and move from Word to LaTeX.
Costs/benefits of the two options:
1 (continue what we’re doing): I’ve laid out the costs out above. The benefit is that we already know how to do this.
2 (move to LaTeX): The cost: not all of us are comfortable with LaTeX. Crucially, Larry has historically not been thrilled with it, which is important not only because he’s the boss, but because as the PI, he is the one who is doing the last-minute editing, which makes his efficiency crucial at the moment when efficiency matters the most–that is to say: at the last minute. The benefits: (1) the BibTex citation management system for LaTeX is hyper-good, and (2) we now have a mechanism for letting multiple people work simultaneously on a LaTeX document, as we recently bought a lab license for ShareLatex.
Zipf's Law 