Scientific communication around the recent pandemic has been much debated. Some scientific public figures proved to be wrong and safety measures such as mask-wearing and surface contamination had to be revised as coronavirus-related research studies were increasing. Compared to the standard timescale at which scientific results are endorsed and published, this pandemic left no time between the bench and communication to the general public. Science appeared to be erratic when it was – wrongly – assumed to be linear. Conspiracy theories emerged because it was easier to believe that these inconsistencies were justified by malevolent organizations operating in the shadows, rather than to correct the way science production is perceived. How scientific research is presented within and outside of research communities has been heavily formatted and codified. It might be time to revise scientific storytelling.
First established as a cornerstone of journalism, storytelling has become an official “soft skill” required to communicate effectively. Scientists have followed the trend and adopted its principles to present their research in a clear and, ideally, objective way. A structuration process is certainly in agreement with scientific ethics, but still questionable if it distorts reality and tries too hard to hide exceptions or contradictory elements.
In a research paper, the conventional pattern used to deliver the scientific message is as formatted as the incipit/development/conclusion of a novel1. Introduction/Methods/Results/Analysis/Discussion is in itself, storytelling already!
On top of this first conventional structure used in experimental sciences, a strong and compelling narrative can layer additional ingredients. The text would typically spotlight a prominent scientific issue that has been left aside until then, argue that the outlined methodology is much more efficient and portable to other institutions than previously published results, and eventually promise a potential cure for a problem affecting a large part of our population.
Then, elements are usually re-ordered according to a deductive logic, by opposition to an inductive proof. In the deductive model, the hypothesis is expressed then the experiments are planned accordingly. It is the famous prelude to each research proposal, requiring to define “a scientific question” immediately followed by the no less daring requirement of “the relevance of the scientific question”. The inductive model looks in contrast much more like random sampling of the phase space. An inductive paper might start like this:
When on board H.M.S. Beagle, as naturalist, I was much struck with certain facts in the distribution of the inhabitants of South America, and in the geological relations of the present to the past inhabitants of that continent.
Darwin, On the Origin Of Species By Means of Natural Selection, Oxford University Press, 1905
It turns out that collecting data on the beak size of the Galapagos finches led to the development of the theory of Evolution. Modern research is still partially done through inference, as exemplified by screening approaches. In this type of studies, if one of the resulting phenotypes happens to be unexpected, therefore promising, it might be investigated further on. As all molecular pathways eventually lead to the biological Rome, it is not unlikely to find in the process a candidate gene involved in cancer. The whole story can then be re-written in a way that makes more sense from the deductive point of view. Science by inference might be the way to truly explore uncharted territory but it is usually not the way the final story is pitched or the preliminary data sold. The logics used to tell stories are thus different from the ones that are actually used to produce new scientific facts and theories.
Presenting the facts in the right order is not just a way to build a strictly deductive framework but also to trap the reader on a mind route. Storytelling should not be used to legitimize the scientific arguments presented to a skeptical mind. These ideas are not new and were exposed by Bruno Latour in the 80’s: in a research paper, storytelling is about layering out the text “so that wherever the reader is there is only one way to go”2. A few years ago, this debate was re-initiated by Krzywinsky and Cairo, who were presenting some storytelling guidelines to publish a scientific paper. According to them, the text should follow a plot, streamline the ideas in a newspaper fashion and avoid excursions outside of the scientific argumentation if it does not support the published hypothesis. Later in the same year, Katz heavily criticized this framework, rejecting any storytelling that would distort the truth. He went further on, suggesting that scientific communication should get inspiration from legal language to avoid persuasive tactics.
Besides introducing persuasion where there should only be demonstration, reorganizing the narrative in a way that makes sense actually deprives trainees from understanding how a research path is progressively elaborated. Research papers are not just about presenting results to an experienced community of researchers. They are learning material for early career scientists or people changing research fields, and are supposed to lay out a methodology specifically developed to address a difficult and subtle problem. However, storytelling polishes all the discourse, removes the hurting angles, and turns precious material into a description that is of no use to the younger generation of scientists. All of this to lift and rejuvenate research in an absurd competition for the most bulletproof peacock among competing investigations.
Should we revise the way we brand research and actually stop branding elegant deductions, acknowledging instead bumpy and complex roads? Admen have understood a long time ago that storytelling helps to capture the brain attention and causes distraction. Don Draper once said “What you called love was invented by guys like me … to sell nylons”. Scientific storytelling might be just a way to sell more papers at the expense of objectivity.
1 Randy Olson, Houston we have a narrative (2015), reviewed by Luna in Science. 2 Bruno Latour, Science in action: how to follow scientists and engineers through society (1987)