Scientific writing, style, and the trolley problem

Image: Trolley by McGeddon CC BY-SA 4.0 via

Our scientific literature has a reputation for being not much fun to read: colourless, tedious, and turgid.  By and large, it deserves that reputation (and I would include my own papers in that assessment).  There are exceptions, of course, but they’re few and far between.  I’ve speculated before about some of the reasons for this.  But there’s a possibility I think I’ve been missing, and I’m going to use this post to think through it.

One thing I see fairly often is early-career writers struggling because they think there’s a single best way to write a given piece of text. Although this is wrong, it’s an easy thing to think*, and it’s easy to become paralyzed at the keyboard by the belief that one hasn’t found that single best way yet. I wonder if there’s a connection between this problem and style that I haven’t appreciated before.  I wonder if writers think that there’s a single phrasing that will be the clearest to, and most appreciated by, everyone – or to turn that around, a single phrasing that isn’t unclear to anyone.

What does style have to do with this?  Well, think for a bit about style in fiction writing.  I promise, I’ll get back to scientific writing, but bear with me. In fiction (and literary nonfiction), style and “voice” are enormously variable.  Think of the no-nonsense terseness of Ernest Hemingway, the dizzying busyness of James Joyce, the sardonic choppiness of Kurt Vonnegut, or the conversational yet lyrical sweep of Toni Morrison – and that’s just four that occurred to me while writing this sentence.  And each of those styles has its fans and its detractors.  Those who love Hemingway may loathe Joyce, or Morrison, or Mario Vargas Llosa, or Barbara Kingsolver, or for that matter Barbara Cartland (sorry**).

But if every stylistically marked piece of writing turns some readers off, then a superficially attractive conclusion is that one can turn nobody off by rinsing away from one’s writing all vestiges of style.  The problem, though: what’s left will look a lot like our scientific literature – bland, colourless, and uninteresting.  I don’t think this is too strong.  I read a lot of science fiction, and I know that I can identify (blind) a passage by Ursula LeGuin, or by Sherri Tepper, or by Clifford Simak just by its rhythm, its tone, and other elements of style.  I’m quite sure I can’t do that with scientific papers.  We’ve converged, I think, on a common scientific style that’s essentially an absence of style.

And that’s where the trolley problem comes in.  The trolley problem is a superficially simple, but endlessly knotty, problem in ethics and philosophy.  You probably know it (especially if you’ve watched The Good Place, and if you haven’t, why not?).  Here’s a simple version.  Imagine you’re at the controls of a trolley.  Ahead on the tracks are five people, and you can’t stop the trolley in time to avoid hitting them.  You can, however, switch the trolley to another track with only one person on it.  You’ll hit them instead – so should you actively decide to kill one person, to save five?  And if that one seems easy, there are elaborations and variants enough to keep you up night after night.

What does this have to do with writing?  Well, the attempt to expunge style from our writing is an attempt to steer the trolley down a track nobody is standing on (although rather than counting up people’s deaths, we’re less disturbingly counting up people who are confused or turned off by a paper we’re writing).  But:  there is no track that nobody is standing on.  We write stylelessly because we think that won’t turn anybody off – but we simultaneously complain about reading dull and turgid literature.  Have you ever put aside a boring paper in favour of reading, or doing, something else?  I have.  There are people on the styleless track.

If we accept that lack of style has readership costs just as style does, we come to what I’ll call the trolley problem of writing.  Imagine that I write a paper with an unusual bit of writing style – perhaps a colourful metaphor.  Image that doing so leaves 10 readers confused or repelled; but imagine that the style gets my paper a little social-media play or word-of-mouth that recruits 11 new readers.  I’ve saved 11 readers, but run down 10: writing with some style has increased my net readership, even if it’s turned some people off.  This is a trolley-problem win, isn’t it?

Thinking of writing style as a trolley problem converts our decisions about style to a quantitative optimization.  Or at least, it would if we could quantify the readers standing on each stylistic track.  I can’t do that, unfortunately: the question of how scientific readership responds to style is rarely and poorly studied.  But there are anecdotes suggesting that style both repels and attracts readers:

  • That style can repel is evident from the occasional rant (often on Twitter) by someone from the just-the-facts-ma’am school of scientific writing. Such folk will insist that we write without adverbs, or insist that “the data speak for themselves”.  They’ll explain the importance of exclusive adoption of the passive voice, or of the banishment of contractions, on the grounds that anything else “sounds unprofessional”.  You’ve met these people; you may have even been taught by them.
  • That style can attract can be seen very occasionally when a paper gets passed around with the suggestion that people read it, not because the results are staggeringly and broadly important, but because it’s a great read***. And consider Hurlbert 1984 (Pseudoreplication and the design of ecological field experiments). This is a 25-page, fairly difficult paper about a statistical issue that’s important but not exciting.  In fact, this paper could easily have been the statistical equivalent of watching very complicated paint dry.  And yet it’s been cited over 4,000 times, and after 30+ years is still widely read.  Why?  I believe it’s at least in part because of Hurlbert’s discussion of “demonic intrusion” and “non-demonic intrusion” in experiments.  I know that when I was a grad student, that’s what made me stick with the paper, and years later, it drew me back to re-read it.  I don’t think I’m the only one.

So, the big trolley-problem question is this: what kind of stylistic choices might attract more readers than they repel (trolley-problem wins), and what kind of choices might repel more than they attract (trolley-problem losses)?  I don’t know the answer to that – nobody does.  But at least framing the question this way makes clear something that I think is rarely said out loud.  I’d like to think it’s something that matters – something that offers us a path to making our scientific writing less tedious.

© Stephen Heard  June 26, 2018

My assessment of Hurlbert (1984) is borrowed from The Scientist’s Guide to Writing, my guidebook for scientific writers.

*^Partly because as high-school students and undergraduates, we often learn to write in courses that involve our writing being graded right or wrong against a rubric – with the idea that, as on an exam, a perfect grade comes by knowing each of the correct answers.

**^I know, I have this weird obsession with Barbara Cartland as a writer.  She appears in every one of my writing talks and in my writing book.  There’s a reason – and here it is.

***^This happened to me quite recently with this paper.  It’s about ecoevolutionary dynamics in butterflies, but even if you don’t care about that (or understand what it means), you should read the paper.  It’s that good.


20 thoughts on “Scientific writing, style, and the trolley problem

  1. Marco Mello

    Very nice post! Thank you for writing it. In my opinion, a good first step towards a more enjoyable scientific literature would be to allow different writing styles in journals. Instead of pushing everybody to the colorless, dry IMRAD style, why not allow each person to write their manuscript as they please? Of course, in science, we need to keep values such as clarity and technical correction, but papers could become a lot more fun to read. For instance, by using some storytelling techniques (


  2. John Pastor

    I recall reading with pleasure, as a graduate student, the papers of Robert Whittaker, Peter Marks, Bill Reiners, Herb Bormann and Gene Likens, Margaret Davis, Robert MacArthur (especially his warbler paper), and others of that generation. These people wrote with individual and enjoyable style. They did not always follow the introduction-methods-etc format. Natural history motivated the problems they worked on. Because of that, I could see those very same problems on a walk through my local woods. Good style makes people naturally feel that the author’s scientific problems are their very own as well. Maybe we should have our students do close readings of these classic papers for style as well as scientific content.

    Liked by 1 person

    1. ScientistSeesSquirrel Post author

      John, I’m curious – do you think you could have identified a paper as being by Whittaker (for example) if the author line were blinded? (I guess also without subject-matter cues, which are hard to blind.) Were their styles individually recognizable? Or is that an extreme and I’m painting an unfair picture with my literary examples?


      1. John Pastor

        Hi Steve,
        I think I could identify a Whittaker paper from a Reiners paper, for example. But certainly I could distinguish one of these scientific classics from a more recent paper even given just a paragraph or two without knowing the title or authors. There is a big difference between the way they wrote then and we write now. I don’t think the range of styles in these earlier scientific papers is as extreme as, say, the difference between Hemingway and Faulkner. Perhaps the more extreme differences in the styles of literary authors arose because they were trying intentionally to create different styles, whereas we are just trying for clarity (in theory, at least)..
        In my view, the best scientific styles came out of Oxford and Cambridge in the 1930s and 1940s. These writers include Charles Elton, Alister Hardy, David Lack, and Paul Dirac.It continues to be a real pleasure to read what they wrote. Must have been something in the air back then.

        Liked by 1 person

    2. Brian McGill

      Great point. The two books I’m working on right now having me diving back into the 1950s and 1960s MacArthur, Hutchinson, May, Pielou, etc and even the 1980s Jim Brown, Mike Rosenzweig, etc. The 80s are not as distinct from today as the 50s of course, but they all had much more freedom to structure their papers as they saw fit and to tell a story. I was walking in the words and saw this. This suggest a model. This fits this data. This makes me speculate …

      Like Steve said I think we have squeezed a lot of good variation out. I imagine I would need a good chunk of a paper to distinguish but between writing style and mathematical style I bet I could distinguish well-known authors with a pretty high degree of confidence.

      Which raises an interesting point, I don’t think we’ve squashed individual style out of mathematical expression as much as writing.


      1. John Pastor

        For those who have wanted to try a more story-telling style that takes motivation from natural history observations, please consider sending 1500-2000 word papers to The Scientific Naturalist, a new series in Ecology ( I think Steve mentioned this in one of his posts, but I can’t find where. I am the editor for the series – go to any recent issue of Ecology to see some examples. (Thanks, Steve, for letting me use your blog for some shameful promotion!)


  3. Jeremy Fox

    I want to see a follow-up post based on variants of the original trolley problem. 🙂 The trolley problem is infamous because slight changes to the problem set-up shift many people’s intuitions about the right action to take. Is the same true for intuitions about scientific writing style? What’s the scientific writing style equivalent of being reluctant to push one person off a bridge onto the track to stop a trolley that’s barrelling towards five people tied to the track? 🙂

    Liked by 1 person

  4. Jeremy Fox

    An undergrad philosophy professor of mine thought that the morally correct solution to the original trolley problem would be to somehow use a weighted die (or series of weighted die rolls) to choose where to direct the trolley, in such a way as to give every person an equal pre-roll chance of dying.

    I suggest that scientists should choose their writing style in the same way. Use a random number generator to give every potential reader an equal chance of being pleased with the style. 🙂


  5. janig717

    I feel part of the problem is that scientists sometimes lack an emotional connection with the work they choose to do. In medical research, there is so much emphasis on the numbers of publications that people forget to ask questions that really move them. They do what works. Difficult to write stories that way.

    Liked by 1 person

  6. Elizabeth Moon

    While working on my second degree and then in grad school in the 1970s, I was lucky enough to read excellent scientific writing both as essays (e.g. in NEJM and The New Yorker) and books (Lewis Thomas’s _The Lives of a Cell: Notes of a Biology Watcher_, among others) as well as in the papers I read for various classes. Later, I read more work by scientists who were also good writers (Oliver Sacks, for one) and nonscientists writing about science clearly and well (John McPhee, especially his series on geology.)

    One of the things that becomes obvious to a professional writer (as I am now) is how much clearer writing becomes when it is not forced. Trying to avoid any personality is a forced style. Trying to infuse a personal, unique style is a forced style. When a writer concentrates on style–when the style becomes the message–then communication suffers.

    When communication really is the intent, regardless of style, both the communication and the writer’s inherent voice come through. (This is not a popular position among many writers, especially early in their apprenticeship, because intentional style and intentional difficulty have been, at times, held up as necessary elements of “serious” writing.) But as a grad student in microbial ecology, I began to realize that uninteresting writing frequently bored the reader past errors in logic of the paper…errors that the author might have noticed if they had written more lucidly. (“Gee, did I really say that? Do that? Ooops.”)

    Every practiced writer has a writing voice, a style made up of that writer’s background experience of reading and writing. A unique vocabulary, if it could be charted, of cliches to avoid, of fresh figures of speech that pop into the head but are derived from their appreciation of others’ figures of speech, both in writing and overheard while listening to people. It takes most writers a lot of output for that voice to rise out of the mishmash of recently acquired reading experience–we all start out sounding like the last several books we read.That would be true for the great scientist-writers as well. (The only cure for derivative writing is more writing, more varied reading, and a lot of listening to real people talking about things that interest them. If you can find a group of much older people in which a discussion of sports doesn’t sound like sound bites from a sports channel…that’s a great resource.)

    So a scientist who wants to communicate better (and who wouldn’t?) does not need to study style, or try to force any particular literary habits other than clarity. A writing voice will emerge from using it *to communicate.* To my ear, Stephen Heard’s voice has become more distinct in the few years I’ve been reading his blog, more easily distinguished from that of other scientists and science journalists I read. Communication is always about more than the meaning of individual words, though that’s important. Emotion (scary thought?) is always part of communication. Yes, the scientist who is bored by their research will likely produce a boring paper. Either change the research or figure out how to be excited about it. It’s just about impossible to write well about something that bores you. The boredom shows through and the bored writer worries about whether the reviewer will like the paper, rather than focusing on the paper itself. (If you really have to write about something that bores you, it helps a lot to write a paragraph that fully expresses the boredom–and then delete it. No guarantee but at least your mind won’t be stuck on how bored you are once you’ve expressed it. Most of us have no trouble writing about what we *don’t* like.)

    Liked by 1 person

  7. Jonny Velasco

    I still don’t know how to feel about passive voice. I’ve been taught that academic (especially technical writing) focuses more on the action more than who’s doing it, which make sense in the field, but it is *really* grating to read, especially when grading 20ish lab reports a week…


  8. Pingback: Friday links: break your methods, scientific writing vs. the trolley problem, and more | Dynamic Ecology

  9. Pingback: The Disco Era of scientific writing | Scientist Sees Squirrel

  10. Pingback: “Scientist Sees Squirrel” is nominated for a People’s Choice Award! | Scientist Sees Squirrel

Comment on this post:

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.