The twilight of the journal articles

Scientific journal articles serve two purposes: as a permanent record of an experimental result (which in theory is repeatable), and as a token of achievement to be exchanged at some point for more money and a nicer office. Publishing makes or breaks your career. Since the stakes are very high, people feel a need to make their papers as impressive as possible. This leads to obfuscated prose.

In fairness, journal articles are sometimes opaque because of the need for precision. But most of the jargon is unnecessary, particularly in the introductions and conclusions. Just as a person may build a house they don’t want to live in because they’re worried about the resale value to someone else, a scientist may write a paper not because they think it’s clear and persuasive, but because it uses the coded language that they imagine their all-powerful overlords want to read.

It’s just like a tenth-grade chemistry lab report really. The apparatus was old and dirty. The scale wasn’t calibrated properly, and the results you got look like crap. What to do? Dress it up in some fancy-pants language and hope for mercy. I may not be pious, but at least I can genuflect!

I hate lab-report writing. It’s a great pity that high school students are rewarded for the tortured passive voice constructions that pass for sentences, because by the time they get to college, they think that’s what smart is. Far better if more lab reports read like this masterpiece, one of my all time favorite pieces of scientific writing: Electron Band Structure In Germanium, My Ass.

Abstract: The exponential dependence of resistivity on temperature in germanium is found to be a great big lie. My careful theoretical modeling and painstaking experimentation reveal 1) that my equipment is crap, as are all the available texts on the subject and 2) that this whole exercise was a complete waste of my time.

All of this leads me to the topic of science blogs. I love science blogs because they are conversational vehicles in which working scientists try to sort out what’s really relevant and important. They’re like the conversations you have at a science conference between sessions, but I don’t get to go to many of those. So I read science blogs, and what fun!

I got started thinking about this because I was reading Neil Saunders’ blog and came across this item: Did someone just admit that journal articles don’t communicate science effectively? In it, he reflects that perhaps “the traditional journal article is increasingly ineffective as a communication tool.”


Coincidentally, I also stumbled across this entertaining presentation on How to write a great research paper (PDF) by Simon Peyton Jones at Microsoft Research. Straightforward and compelling, it’s definitely worth a read.

3 thoughts on “The twilight of the journal articles”

  1. I was doing some surfing on your MATLAB work mentioned in the book “Crowdsourcing” and happened across your blog. Good stuff. I am actually writing my next blog post about your MATLAB contests and the concept of brilliant tweaks to improve a product.

  2. In support of those excellent slides about how to write a good research paper, you might be interested in Medawar’s 1964 paper, “Is the scientific paper a fraud?” I welcomed this challenge to the accepted academic edifice, having waded through hundreds of papers myself.

    I got some great perspective from Prof. David Jensen’s “Research Methods for Empirical Computer Science” course at UMass Amherst. It was the only course I’d ever taken in grad school that asked the meta-questions and challenged a lot of the crippling assumption that young grad students have. The reading list at that link has got some other interesting gems in it as well. It’s important to know that the purpose of the readings isn’t to teach you something about the papers’ topics, but rather to get you thinking about the style and quality of the papers themselves. The course taught critical thinking skills, infinitely more useful.

Comments are closed.

%d bloggers like this: