Every paper tells a story – or at least, it should

Image: Storytelling chair © BeyondTimelines, CC-0 via pixabay.com

There are plenty of strong opinions and controversies about scientific writing: the active voice vs. the passive, how to describe a result with P = 0.06, even – and perhaps most spectacular for ratio of strength of feeling to importance – one space or two after a full stop.  But one controversy that astonishes me is that over whether or not scientific writing is “storytelling”.  Spoiler alert: of course it is.

It may seem odd to use the word “storytelling” for scientific writing, because we tend to think of that word as associated with fiction. But in fact the match between conventional notions of “storytelling” and scientific writing is very close. Here’s how I put it in The Scientist’s Guide to Writing:

What does it mean for your paper to tell a story?  Successful fiction…sets up and then resolves some interesting question in a reader’s mind, by exposing compelling characters to a well-defined plot.  A scientific paper does the same.  It has characters: the rocks, chemicals, equations, or other entities that you studied.  It has plot: the methods you applied to your characters and the results you obtained from them.  Most importantly, it raises and answers an interesting question.

Once you start thinking about scientific writing this way, it completely revolutionizes the writing process.  At least, it did for me.  Writing a paper is no longer writing a diary (first I did this experiment, and got this result; then I did that experiment, and got that result; then…).  Instead, writing a paper means deciding on the paper’s important question and its answer.  Then: what plot elements (results, and methods to support them) does the reader need to understand the answer?  Which other plot elements should be excised as red herrings or distractions?  (There’s always an experiment that you ran, but that in the end didn’t shed light on the central question).  Which characters advance the plot, and which distract from it?  Storytelling a paper means writing backwards: start with the answer you want readers to understand; then choose the results needed to support that conclusion; then describe the methods it took to reach those results.  That’s how you tell an effective story.  (It may even help to have a baby, a werewolf, and a silver bullet.  Andrew Hendry explains.)

If you don’t think carefully about this, you could get the idea that something awfully shady is going on.  And something shady certainly could go on (and has) at the hands of a bad actor.  So let me be crystal clear: the idea that scientific writing should be approached as storytelling does not mean that when we’re writing science, we should write fiction.  We choose results to present not to shape the conclusion, but to make the conclusion as clear as possible.  We exclude not experiments and results that point against our hypothesis, but experiments and results that, in hindsight, don’t point in any useful direction at all.  We’re storytelling to make our nonfiction clear; but to do that, we can have some of the same concerns about our readers, and use some of the same techniques, that fiction storytellers do.

Every time someone talks about scientific writing as storytelling, there’s horrified pushback* – hence, my inclusion of this on the list of scientific-writing controversies.  Part of the pushback arises from the confusion with fiction storytelling.  But there are also people who object that any selection or interpretation of data is subjective, and thus not scientific.  The data speak for themselves, they argue – but of course data don’t speak for themselves.  If data spoke for themselves, we’d simply post our raw data files on the web and be done with it – no need for an Introduction or a Discussion section, or a statistical analysis, or even (I suppose) any text at all.  Of course, just posting our data wouldn’t make our science objective anyway.  We make lots of choices in choosing questions to ask, designing methods to address those questions, and gathering and processing data with those methods, and those choices inevitably have a subjective component.  Science doesn’t succeed because it’s done objectively; science succeeds because it gives us a way to gain knowledge despite human subjectivity**.  Science is not a big pile of facts, and scientific writing is not the production of a big pile of uninterpreted data.

Each time I write a paper, I make a lot of decisions about what goes in, and what comes out – not just experiments, but topics, paragraphs, sentences.  All those decisions work the same way: does the material I’m trying to decide about help me tell the scientific story I want to tell, or distract from it?  I may not be writing Moby Dick (fortunately, given its length)*** but I’m still a storyteller.

© Stephen Heard  July 16, 2019

Because I’ve thrown shade at Rod Stewart elsewhere for the wretchedness of “Maggie May”, I should acknowledge that the title of this post borrows from his considerably-less-wretched “Every Picture Tells a Story”.

---

*^Someone – almost certainly someone who didn’t read past this post’s title – will post with righteous outrage on Twitter about how I’m encouraging people to pollute our sacred and objective literature with invention.  Sigh.

**^This does not, of course, mean that anything goes!  We can and should reduce subjectivity; it’s just that it’s naïve to think we can eliminate it, and we should acknowledge that.

***^OK, I don’t work on whales, so I’m definitely not writing Moby Dick.  I work on plants and insects… so, maybe what I’m not writing is Flight Behaviour, or Angels and Insects, or – you knew I’d get here – The Very Hungry Caterpillar.

 

Advertisement