Does science hide in narrative?

Are our most emotive and personal forms of description — the story, confession, history, narrative — neglected by scientific thought? Or does science hide in narrative? We don’t expect scientists to tell us a tale in the meticulous detail of their publications; narrative is reserved for the scandals of memoirs, the courtroom appeal, and the patient’s history. Narrative may even be powerful in science education, but that doesn’t mean it’s essential to the scientific process. In last week’s debatable, Boaz Miller asked whether scientific knowledge is anything special. This week, I want to dig deeper to ask if narrative is essential to all forms of knowledge, including scientific knowledge. Tell me why, within or without a story…

by

Michael Cournoyea
Michael Cournoyea is a doctoral student at the IHPST at the University of Toronto. He received his BSc at McGill University in Biology and Philosophy and has worked at the intersection of these disciplines for the last five years. He currently works as a don at Victoria College and is active in student life on campus. His work examines the pluralism and politics of causal explanations in medicine -- whether biomedical, evolutionary, phenomenological, or sociological. The implications of his work are pragmatic, engaging issues in racialized medicine, the sovereignty of patient health, and how we should live the healthy life.

23 Comments

  • Boaz Miller
    Boaz Miller Reply

    Scientific publications today try to avoid narrative, which is part of the view that the context of discovery does not matter. The only context that matters is the context of justification, which has only one narrative – the narrative of the scientific method. This is why science students are trained to de-personalize their lab reports and write that “results were obtained” (by whom?). Compare that to Kepler’s detailed personal and tormented reports of his repeating failures in the lab.

  • Curtis Forbes
    Curtis Forbes Reply

    That’s a good point, Boaz, and I think it’s probably informed by the “context-stripping” methods of the exact sciences, which are regularly deployed in laboratory settings. Rather than accounting for the social and physical context that the objects of their study (e.g. elementary particles, human beings, etc.) are usually found within, natural scientists are generally in the business of removing such context. They do this by, for example, implementing various shielding conditions in the laboratory. The idea is that by isolating their objects of study from all the confounding variables that usually surround them outside the laboratory in the messy, complicated real world, scientists can gain a better understanding of their objects of study.

    What I’m not sure of is whether this “hiding” of the narrative, in publications and theorizing, is a bad thing so far as scientific practice goes. These “isolating” laboratory methods are pretty effective in allowing us to control those objects of study, and to gain a better understanding of them apart from their usual surroundings.

    Where nuanced narratives about scientific practice are needed, to my mind, is when we try to understand science itself, i.e. when we are doing history, philosophy, sociology, and anthropology of science. The self-image that exact scientists often have of their discipline is, I think, a result of their applying the “context-stripping” approach to the study of their own craft. They strip away the entirety of the context of discovery, which includes all the funding initiatives that inform their work, the political climate outside the laboratory, their own personal motivations, etc. This leaves them with a picture of their activities as “pure hypothesis-testing,” or some other suitably abstracted, noble, and disinterested search for natural knowledge.

    But if we want to understand what science really is, stripping away all this social context is the wrong method. Physicists are actually the wrong people to tell us what physics is, for example, because their methods of context-stripping are ill-equipped to parse the complicated narratives that make up physical scientific practice, which is a multi-faceted social phenomenon with extremely variegated features. Physicists often claim priority over those in the humanities studying scientific practice in defining the nature of their own practice, simply because they are the ones doing physics. But people have good reason to be suspicious of the pictures physicists paint of their own practice, and not only because it sometimes takes an outsider to see certain things about a culture that remain opaque to “insiders.”

    And yes, I stole that position from Sandra Harding (Chapter 4 of “Whose Science? Whose Knowledge?”), where she claims that “Physics” (physics’s disciplinary self-image) is a bad model for physics (a bad picture of what physics actually is). Just to give credit where it is due.

  • Mike Thicke
    Mike Thicke Reply

    Very interesting questions! A few things come to mind.

    (1) Science does seem to value one sort of narrative – the causal narrative. If you can’t give a causal story for why some particular effect you infer is occurring, it will be very hard to convince other scientists that the effect is real. I forget who talks about this, but the story that comes to mind is that many scientists were reluctant to believe in continental drift because they thought continental drift was impossible. Despite there being good empirical evidence for the effect, there was no causal story, and thus scientists were very reluctant to accept drift.

    (2) The narrative of the active sperm attacking the passive egg is a classic in feminist philosophy of science. Seeing the (male) sperm as active reinforced ideas about the roles of men and women in biological and social contexts. And various “just so” stories are closely related to this idea — back to Curtis’s first debatable!

    (3) Robert Boyle was a proponent of reporting both successful and unsuccessful experiments (whether or not he practiced what he preached…). We know that in modern science unsuccessful experiments, and even experiments challenging previous results, are very rarely published. It might be a really valuable rule of publishing that (linked on the web or whatever) a full narrative of the experiment and previous failed attempts was made available with all publications. This is obviously really hard to enforce, but it would be really epistemologically valuable, as it tells you something when the experimenter failed 100 times, then found the right combination of parameters, and then succeeded. We might think it is pretty likely in that case that the experimenter just got lucky this time where he didn’t the previous 100 times.

  • alice Reply

    Have you read any of these?

    Sorry for bibliographic list! If you’re interested, let me know and I’ll re-purpose some of my thesis’ lit review on this into a blogpost.

    Beer, Gillian (2000) Darwin’s Plots: Evolutionary Narrative in Darwin, George Eliot, and Nineteenth-century Fiction, 2nd edition (Cambridge: Cambridge University Press).
    Curtis, Ron (1994) ‘Narrative Form and Normative Force: Baconian Story-Telling in Popular Science’, Social Studies of Science, vol. 24(3): 419-61.
    Kubli, Fritz (2001) ‘Can the Theory of Narratives Help Teachers be Better Storytellers?’, Science & Education, vol. 10: 595-599.
    Leane, Elizabeth (2007) Reading Popular Physics: Disciplinary Skirmishes and Textual Strategies (Hampshire: Ashgate).
    Ogborn, Jon, Gunter Kress, Isabel Martins & Kieran McGillicuddy (1996) Explaining Science in the Classroom, (Buckingham: Open University Press).
    O’Hara, Robert J (1992) ‘Telling the Tree: Narrative Representation and the Study of Evolutionary History’, Biology and Philosophy, vol. 7(2): 135–60.
    Mellor, Felicity (2003) ‘Between Fact and Fiction: demarcating science from non-science in popular physics books’, Social Studies of Science, vol. 33(4): 509-538.
    Mellor, Felicity (2007) ‘Colliding Worlds: Asteroid Research and the Legitimatization of War in Space’, Social Studies of Science, vol. 37(4): 499-531.

    • Luis Reply

      You might also add to this list the following article by the chemist Roald Hoffman. I use this article to encourage my undergraduate chemistry students to think more deeply about what is included in scientific articles.

      Hoffman, Roald (1988) ‘Under the Surface of a Chemical Article’, Angewandte Chemie International Edition in English, Vol. 27 (12): 1593-1764.

  • Eleanor Louson
    Ellie Louson Reply

    Great debatable and responses! One kind of narrative that I thought of after reading Michael’s post this morning is the kind that makes its way into the first section of many scientific papers: the introduction, containing a history and context of the problem being worked on by the scientists involved. Decisions about what to include in the previous work on the topic, what to emphasize or gloss over, who to describe as the founding person on that topic, etc. all work towards crafting a kind of narrative within which the most recent paper is meant to fit – and often it turns into a progressive story, fitting in well with what Curtis (crediting Harding) described as the subject’s disciplinary self-image.

    In addition, many branches of the natural and social sciences employ working narratives about their foundation/founder(s) and can tell stories about the development of their field that emphasize the giants upon whom they now stand – for example, experimental psychology sees its roots in Wilhelm Wundt’s Leipzig laboratory, dismissing most of his actual work to the extent that he’s pretty much a figurehead for the “experimental” approach.

  • Marci Reply

    Add to the reading list Carol Cohn’s “Sex and Death in the Rational World of Defense Intellectuals.” Feminist scholars of science studies (see Harding, Keller) have contributed to both “revealing” the hidden narratives in science as well as explicitly situating themselves as scientists with complex and sometimes conflicting personal narratives. Great discussion!

  • W. Dean Reply

    It looks to me like you guys (and gals) have been drinking too much of the sociology of science Kool-Aid. I can see why a catalogue of failed experiments would be useful to someone writing the history of a scientific discovery (it would undoubtedly add dramatic tension), but I fail to see what epistemic value it adds to a scientific argument. So I’d like someone to explain to me, nay, just give me one example of how adding a narrative into, say, a lab report would count as value added.

    Don’t get me wrong. Galileo’s “Dialogue Concerning the Two Chief World Systems” is solid gold. But he had the public, the theologians and his fellow scientists to convince, so he had to pull out all the stops. If he was writing within the modern academy to a group of peers working on the same subject and conducting the same experiments within an established field, he too would probably prefer economy of expression over a good read.

    • Curtis Forbes
      Curtis Forbes Reply

      Mike gave a great example in his #3 – when an experimenter got the reported positive result only after several trials and several tweaks to the experimental apparatus. Not sure why you don’t think that’s relevant – if parameters are set and re-set so as to achieve a desired result, that makes a difference to the credibility of the one trial where the ‘desired’ result actually obtained.

      If I remember correctly, W. Dean, you’re skeptical about climate models for reasons that take into account the historical course of climate science, and the ad hoc tweaking of these models to preserve anthropogenic warming. It seems to me that you’re committed to the epistemological relevance of narrative as a result …

      Another example may be when funding initiatives influence the agenda of scientific research – perhaps the idealized model discussed in a paper results from the need to find a model that serves external funding desires, and therein reflects the needs and desires of funding bodies rather than the whole truth about the object thereby described. Being up front about that might be epistemologically virtuous and valuable.

      I’m sure there are tonnes of other examples.

      Out of curiosity, why the antagonism towards sociology of science, or the very idea that telling a narrative about why one is doing scientific research might be epistemically valuable?

      • W. Dean Reply

        Curtis,

        Mike’s example is ambiguous. I could be wrong, but I can’t imagine too many situations where one could perform an experiment 100 times without success and not publish the fact that the experiments only worked in 1 of 101 trials. Maybe if a scientist made a calculation error in dosing, for example, and ruined 99 out of 100 experiments, we could say that we don’t have the whole story, because he only reported the one that worked. But botching a bunch of trials is less a compelling narrative of real science in action, than a pedestrian anecdote about a guy who confused grams and ounces or some such.

        The other case I can imagine where the “whole story” adds to our knowledge of thing is the following. Suppose a scientist administers 2 grams of a toxic substance to 100 rats and he reports that all 100 rats died. The full story, however, is that he also hit all 100 rats with a hammer before administering the toxic substance. Does the narrative add something? Yes: it explains why the rats really died. But it’s not obvious to me why, say, a professional obligation to provide a narrative would make any difference, since the scientist already threw ethics out the window.

        All tweaking and parameter adjustments are not the same thing. Finding optimal dosing or the fracture point of a material involves tweaking parameters of all kinds as part of finding out as much as possible about effects in use. Age, weight, sex and so forth affect drug dosing and composition, direction of force and weather and temperature affect the performance of building materials. Sound knowledge about optimal dosage and the limitations of materials are the result of thousands of combinations of various parameters.

        Ad hoc adjustments in climate models are birds of different feather. Where parameter adjustments were part of the process of knowledge gathering in the first case, they point to the incompleteness of the explanation in this one. Climate scientists must arbitrarily modify known variables or add invented variables that keep the model predicting past climate. There’s no malfeasance in this, of course, and under ordinary circumstances you might argue that the model is close enough. But when the stakes are this high, close doesn’t cut it.

        Your second hypothetical example about funding issues may have merit and you may be on to something. But, as I said in my original remarks on this topic, I’d like a concrete case of where narrative (and not simply a one-line disclosure) actually adds anything or how not including narrative leaves things out.

        Finally, I’m not hostile to sociology of science, but like bourbon you can have too much of it. What’s more, the sociology of science should also be looked at from its own sociological perspective. In a place like this where everyone belongs to the same subfield and shares much the same view about science, it seems to me that it’s easy to fall back on taking the conventional wisdom of the field as obvious. In other words, it’s important to be conscious of the fact that the Bubble Chamber can become an echo chamber.

        • Mike Thicke
          Mike Thicke Reply

          I think econometrics is an easy example here. See Leamer’s “Let’s Take the Con Out of Econometrics”. Or in medical research, consider John Ioannidis’s “Why Most Published Research Findings Are False”. There is also a classic example from Harry Collins’ gravitational waves book (title?), where Weber managed to find, and publish, a several standard deviation correlation between observations that turned out to be from data not even collected at the same time. It turned out that Weber was just running so many different tests for correlation that it was nearly inevitable that he’d eventually find a spurious correlation that was extremely suggestive. If we can’t know that Weber was running so many tests, and you wouldn’t just by reading his paper, then you wouldn’t be able to properly judge his conclusions.

          • W. Dean Reply

            Mike,

            As I said to Curtis, I’m not denying the importance of the “whole story” for philosophers and historians. I’m disputing its epistemic value for science.

            Now, the only one I’m already familiar with is Ioannidis. But the problem he found isn’t a lack of narrative (just poor research and biased methodology) and, more importantly for the case being made here, he didn’t discover it or argue it by constructing a narrative. Ioannidis examined the research methodologies used in studies and researchers’ disregard for contradictory results. The same goes for the example you offered: it just looks to me like poor methodology, partisanship or confirmation bias.

            Let me put my case another way: you have to show that these errors resulted from “omission” of the narrative not “commission” of bad or biased research. So far, however, all the examples offered here are really commissions told in narrative form and then claimed as paradigm examples of why “including the narrative” is important.

            • Curtis Forbes
              Curtis Forbes Reply

              I think one of the points implicitly being made is that inclusion of the narrative story might actually expose the commission of bad or biased research. By including a narrative about how research was conducted and motivated, other scientists might be able to make a better assessment of the validity of published research by making it easier to uncover poor methodology, partisansp, or confirmation bias. That’s an epistemic good, for scientific researchers, no?

              • W. Dean

                Curtis,

                The problem with that solution is its impracticality. There’s no reason to suppose a bad scientist (whether on account of bias, incompetence or both) will be any better at writing narratives.

              • Curtis Forbes
                Curtis Forbes

                Well, at least we’ve established that narrative has epistemic value for science, which was initially disputed, even if we have also established that it is problematic in practice …

              • W. Dean

                Curtis,

                That a bad scientist could be expected to do no better at narrative writing than at science hardly speaks to the usefulness of narratives.

              • Curtis Forbes
                Curtis Forbes

                But if the problem is only, as you say, the impracticality of including good narratives in scientific publications/reports because we cannot count on bad scientists (or even any scientist not trained to write narratives) to write good narratives, this implies that, should the inclusion of such narratives be made more practical (e.g. by having historians and sociologists go into the lab and produce a “good” narrative about potentially “bad” science, rather than requiring the scientist themselves to develop and deploy such skills for crafting good narratives), the context provided by a good narrative account of a scientific finding could serve as an epistemic good for science, given that it could (as I previously suggested) expose the commission of bad or biased research.

            • Mike Thicke
              Mike Thicke Reply

              If only there was a handy “good science” checklist that everyone agreed upon and scientists could attach to their submissions!

  • Nedra Weinreich Reply

    While I can see how narrative can be important for contextualizing the research process, as previous commenters have noted, I think the most important use of narrative actually comes after the research has been completed. What is the meaning of these results? Why is this research important, and what are its implications in the real world? How should it change how we do things or understand causal relationships? Without this type of information, the research may be pointless.

  • Rick Austin Reply

    Nedra makes a good point about context and meaning. We have a podcast with Eric Berger and Laurie Johnson, a print and broadcast journalist, respectively, at our site, http://www.ktexchange.org. It expands on the importance of narrative, especially to the lay reader.

  • Richard Oosterhoff Reply

    Nedra’s point about the timing of narratives might be developed further. Perhaps we can see Galileo’s work as doing the work of two modern genres: both grant-writing, and dissemination of results through journals. It only takes reading a few biologists’ CVs to realize that good grant-writing skills determine an enormous proportion of funding! And if you look at, say, the funding of US marine biology projects in the 1960s and ’70s, it quickly becomes clear that those who could tie the narrative of their project in with the Cold-War concerns of the US military were the projects that got funding. For instance, I know people who wrote and received grants to study deep-sea marine life, putatively to improve submarine technology . . . I’d be very much surprised if similar ecological, cancer, and military concerns have not shifted 21st-century grant-writing narratives. Grant writing is enormously dependent on narratives, implicit and explicit.

    On the other end, and in response to Dean’s observation, which certainly merits attention: One doesn’t have to be Popperian to realize the basic epistemic power of the asymmetry between verification and falsification. Implicitly or explicitly, in a scientific field “everyone knows” that certain methods and certain conclusions just don’t work. (As Dean pointed out regarding the sociology of science, what “everyone knows” can be true or false. I’d add that Dean is drawing on a narrative about sociology of science to explain this.) In biology, where I have a very little bit of experience, that kind of knowledge is often based on a narrative that is internal to the research group, and rarely made explicit in journal articles. It’s a narrative nonetheless. One research group will often misunderstand–or disagree with–another research group on the basis of different lab narratives (who trained whom, what methods are more efficient or trustworthy, etc.). Of course, this is taking the notion of “narrative” in a much larger sense than Dean intended. But it seems pretty hard to take, for example, the narrative of the Modern Synthesis out of modern evolutionary biology textbooks–even though major figures connected with that (W. Provine) are now spending all their time claiming there is no such thing as a Modern Synthesis. Narratives true and false at that level, I would think, enormously impact the direction of research, and also are the context without which journal articles cannot be understood.

Leave a Reply to Marci Cancel reply

Your email address will not be published. Required fields are marked *