Fast Science and Philosophy of Science

Jacob Stegenga

We are living through one of the strangest and most anxiety-provoking times that most of us can remember—in lockdown, separated from friends, lovers, colleagues, work, extended family, and in some cases immediate family, to slow the spread of this new virus, SARS-CoV-2. The threat of this virus, and the effectiveness and harms of the social policies meant to mitigate this threat, have become the most important scientific issues of a generation. So it is worth asking: what is the role, if any, of philosophy of science during this pandemic and global lockdown? Should we be trying to get in on the dispute between, say, Neil Ferguson (the most prominent epidemiologist whose models predicted dire consequences of the pandemic and who encouraged strict lockdown policies) and John Ioannidis (the most prominent epidemiologist who has criticized the dire model forecasts and lockdown policies)? I recently posed this question to colleagues on social media. The responses were insightful, and suggested that the discussion could benefit from broader engagement with our discipline. Thus, here I reproduce some of the motivation for the question and summarize several themes from the responses.

So far in the lockdown—May whatever-it-is-today—I am not aware of a systematic piece written by a philosopher of science on the pandemic or the policy response (if you know of one, I would be grateful if anyone reading this would direct me to it). Of course, I do not mean a journal article—our journals move too slowly for this—but I mean something deeper and more impactful than what normally appears in social media or blogs. Eric Winsberg has amassed an impressive collection of resources about the pandemic on his Facebook feed, and commentaries there have been insightful.^[1] Alex Broadbent has written a couple of short articles and produced a powerful video about the impact of the lockdown on African communities. Jonathan Fuller published a discussion of epidemiological models in the context of evidence-based medicine. Erik Angner has reminded us of the importance of epistemic humility during a pandemic of overconfidence.^[2] What’s keeping me up at night, though, is this: should philosophers of science be trying to assess the merits of the various scientific arguments pertaining to SARS-CoV-2 that are now having such profound implications on policy, in a rigorous yet publicly visible manner, at a pace which accords with that of the relevant scientific work? This has been a moment of what we might call ‘fast science’. Ought philosophy of science attempt to engage with, contribute to, and criticize fast science, as it unfolds?

Both the harms (broadly construed) of SARS-CoV-2 and the effectiveness and unintended consequences of the social policies meant to mitigate those harms have been underdetermined by available evidence. At issue are the plurality of epidemiological models generating discordant predictions, the reliability of the various models, the empirical substantiation of model parameters, the evidential basis of the effectiveness of many lockdown strategies, and a complicated intertwining of social values and science. Moreover, stances on the pandemic are, obviously, politicized. The pandemic and global policy response seems, therefore, apt for analysis from the perspective of philosophy of science.

On the other hand, the relevant scientific issues require technical expertise in epidemiology and economics, and up-to-date knowledge of a rapidly shifting epistemic landscape. A meta-level operative principle for any commentator on the pandemic ought to be epistemic humility, especially given the secondary pandemic of over-confidence among policymakers, armchair epidemiologists, and the scientists themselves.  Moreover, because positions on the pandemic are often politicized, a philosopher of science who defends the Swedish approach (for example) might be assumed to have values and motivations aligned with, say, Trump supporters.

That said, our colleagues in philosophy of science include experts on scientific models, on methodological problems in medical science, on causal inference, on the relationship between science and values, on inductive risk, and on many other issues pertinent to the pandemic and policy response. Obviously, the stakes are high: some models are telling us that millions of people might die from the virus, while other models are telling us that millions of children might starve as a result of the global lockdowns. Should we be using our expertise to assess these scientific arguments as they unfold? There is a temporal dimension to my question. I am not asking whether or not philosophy of science should, in the comfortable future and in obscure journals, assess the merits of epidemiological models, or the relationship between scientific expertise and policy. I am asking if we should be assessing the particular details and predictions of these models, now, and how they are being deployed in policy, today, in a manner that, to the extent possible, reflects the manner of presentation of the pertinent scientific arguments (namely, the speed of their articulation, their visibility, and their impact). Should philosophy of science engage with fast science?

This question generated a range of interesting answers when I posed it on Facebook. My interlocutors agreed to let me draw on their responses for this post. I have tried to summarize their views accurately, though I apologize for any misattribution.

Two distinctive features of fast science that are especially important are the pace at which it is produced and disseminated, and the impact it has on popular thinking and public policy. The Imperial College model predictions of Covid-19 cases, hospital bed use, and deaths were developed rapidly and posted online without peer review, and this almost immediately influenced policy in the United Kingdom and the United States. Some of most prominent empirical findings about the virus, such as the various seroprevalence studies, have moved almost as rapidly and were also posted on platforms with no peer review, and were then widely publicized. A strong criticism of the Santa Clara seroprevalence study was posted on a personal blog of a statistician (Andrew Gelman), and these criticisms were soon quoted by popular press. Richard Horton, the chief editor of The Lancet, says that he has hired ‘surge capacity’ staffing to deal with the dozens of papers submitted to his journal every day about Covid-19. The scientific work on the virus is growing exponentially: by early April the number of scientific papers on the new virus was doubling every fourteen days. The number of page views at medRxiv have increased by a factor of one hundred since December. The UnHerd interview with Swedish epidemiologist Johan Gieseke was viewed more than one million times on YouTube after just two weeks of its posting.

So much science having so much impact, yet philosophers of science have been relatively quiet.

One argument that was raised in favour of reticence, or at least to explain our discipline’s degree of silence about the relevant science and policy response thus far, is that philosophers tend to be good at debating issues that require a long time to clarify. We are less accustomed to engaging with rapidly developing scientific debates that have immediate policy impact. Anya Plutynski noted that it can be difficult to write something that is both general and timely, given how quickly the scientific landscape changes during an event like the coronavirus pandemic. The deep questions we tend to address require cautious and often slow scholarship. Our usual mode of thinking, speaking, and especially writing might end up muddying the public debates, suggested Marshall Abrams, and this could have unforeseen negative consequences. Mauricio Suárez concurred by noting that philosophers will be more useful in the future, retrospectively articulating views about what happened, and possibly what went wrong. We are typically good at dealing with settled science rather than with developing science, and this, suggested Craig Callender, perhaps exposes a fault with our discipline.

The insight of Plutynski, Abrams, and Suárez appealed to the fact that philosophers of science are not accustomed to such fast science. The comments by Callender and Plutynski suggested that this might be an aspect of our discipline worth remedying. Perhaps we ought to have more competence and willingness to engage with fast science.

But how? Some of the responses noted practical difficulties that many of our colleagues are facing. Quayshawn Spencer and Boaz Miller argued that while there may be a role for our discipline in such a situation, many of us are simply too busy, because on top of our regular duties we have been thrust into new personal and professional roles, such as caring for children at home while teaching classes online. This interacts with the point above: a disciplinary norm, noted Miller, is to learn the relevant science well, including its history and up-to-date empirical evidence, before writing about a topic. This is difficult to do when our personal and professional lives are turned upside down by the very scientific and policy issue we could be assessing.

Another practical difficulty for philosophy of science to engage with fast science is to have venues in which such engagement can occur. Our pre-print platforms are meant for future articles that exemplify slow, careful scholarship, and anyway, unlike in medicine, what happens in a philosophy pre-print archive stays in a philosophy pre-print archive. A prominent and senior member of our discipline wrote to me saying that a short article he had written, relevant to understanding the pandemic, had been rejected by a number of visible sites devoted to popular science essays. In any case, most of us do not have established publishing relationships with, say, The Atlantic, or Aeon, though Fuller noted that venues such as Nautilus and The Conversation have been open to some slightly deeper analysis than is afforded by many venues.

For philosophy of science to have some impact when engaging fast science, the kinds of expertise in our discipline must be recognized. Eran Tal mentioned the ‘silencing of non-epidemiologists who speak about the science’. I have heard fellow academics say about the various questions pertaining to the pandemic and policy response that they would sooner listen to epidemiologists than philosophers of science. On its face, there might be some grounds to this, though the wild disagreements between epidemiologists, the controversies surrounding their models and empirical methods, the many causal inferences in question, and the value-laden nature of the policy responses all suggest that some careful philosophical engagement would be worthwhile.

An interesting idea was suggested by Jeremy Howick: given how much disagreement we see between epidemiologists today, medicine now looks a lot like philosophy. This could be taken as another reason to think that philosophers of science have a role to play in the current crisis. Yet, the conclusion that Howick drew was that because there are so many unknowns, the precautionary principle applies. Responding to this, Julian Reiss noted that the precautionary principle is ambiguous—we might be especially precautious about harms of the virus, or we might be especially precautious about the harms of the lockdowns, but the precautionary principle itself is silent on how to balance the competing relevant concerns. A related worry was raised by Eric Schliesser, who reminded us that precautionary principles have been used to justify illiberal and inhumane acts, such as occurred during the war on terror.^[3]

Juliette Ferry-Danini noted that in France, philosophy of science is being instrumentalized in public discourse to defend preposterous positions. Although it is not exactly ‘silencing’, insofar as the public instrumentalization of philosophy of science requires the discipline to have at least some visibility, it is a sort of intellectual silencing when a medical researcher, Didier Raoult, cites Feyerbend to disparage the ‘dictatorship of the methodologists’ when defending his non-randomized and non-blinded trial of hydroxychloroquine as a treatment for Covid-19. Ferry-Danini’s article criticizing Raoult has been viewed more than fifty thousand times—an inspiring example of philosophy of science engaging fast science.

Perhaps the straightforward reason for the silencing of non-epidemiologists in this crisis is based on the notion that one needs discipline-specific technical skills to engage with the details of epidemiological models. The debate about the infection fatality rate of Covid-19, or how to predict its impact on mortality four weeks from now, involves technical issues requiring expertise to fully understand, let alone criticize. Christopher Clarke suggested that unless one has expertise about the particular models in question, then philosophers of science should limit themselves to urging some general lessons about the importance of epistemic humility, about the reliance of models on assumptions, and about the entanglement of facts and values. Philosophers of science can offer general lessons about risk, uncertainty, and clarify details about models and methodology—Rachael Brown, Thomas Sturm, Fuller, and Jonathan Sholl seemed to agree that this kind of work can be valuable, even if it does not involve a dive into technicalia or a defence of a partisan position. André Ariew pointed us to an example in Germany of a committee composed mostly of humanities scholars who are advising on the pandemic policies.

Of course, many of our colleagues are experts about models, causal inference, and the relationship between science and policy. Moreover, in the public discourse surrounding the pandemic and policy response, Eugene Earnshaw noted that ‘there are very few contributions that come from a position of unassailable expertise in any case, and the sort of meta-scientific issues that philosophers of science are familiar with are particularly relevant right now’. This statement summarized what I took to be the most common thread in the responses to my question. Making a similar point in an email, Elliott Sober wrote that ‘philosophers of science who are scientifically well informed and analytically acute can make substantive contributions here’. This point was emphasized by many contributors to the discussion, including Miller, Spencer, and Plutynski.

A related point was mentioned by Miller: Our discipline has a mixed attitude toward scientific expertise. Many philosophers of science are concerned with the demise in public trust of science, evident in debates about teaching intelligent design in classrooms or in mistrust of climate science. Conversely, many philosophers of science are concerned about excessive trust in particular domains of science, exemplified by, say, unreliable research on pharmaceuticals. Of course, it is perfectly consistent to hold that science is the best means to learn about our world, while also criticizing sloppy science. Tal noted that one future task for philosophy of science could be to mitigate the backlash against science (and especially, I reckon, model-based and policy-oriented science) that might occur when people start to question the gross exaggerations of the various Covid-19 model predictions and the unprecedented policy responses.^[4]

To engage with fast science, Ariew noted that not only is technical expertise required, but so is thick skin. Moreover, the incentive and reward structures in academic philosophy might discourage such engagement, at least among pre-tenure philosophers of science. Developing this idea, Sholl argued that to publicly engage with fast science is a risk, and thus requires the luxury of not worrying about what one’s peers or future employers might think. (At this point in the discussion my sister got in on the conversation, speaking from her personal experience that philosophers tend not to shy away from disagreements.)

In an email, Nancy Cartwright agreed that philosophers can and should weigh in on serious scientific and policy issues, though she noted that her own skills are at a meta-level rather than assessing particular data and models—I reckon that this is a diagnosis for many of us.  Jay Odenbaugh raised a number of such meta-level questions about models that are especially salient in the present context and that are worth quoting in full:

When should policymakers ignore models and when should they use them for policy guidance? What are good rules of thumb for critically thinking about models for laypeople? Do we need models for policy or can we simply get by with observational studies? How should we determine who is a good or bad modeller? How do epidemiological models compare to other types of models in other domains?

Although persuasive answers to these questions might require the sort of slow, meta-level scholarship that our discipline is accustomed to, such answers might better prepare our discipline for more rapid and impactful engagement in future episodes of fast science, which could emerge from all sorts of phenomena, such as a climate crisis, advances in artificial intelligence, or indeed, another virus.

In an expression of gratitude to our colleagues for their thoughtful contributions, the final response to the discussion on my social media post came, naturally, from my mother. This discussion was limited primarily to my circle of interlocutors on social media. I would like to hear what the readers of this blog and the broader discipline think about the role philosophy of science should play during the sort of fast science that we are witnessing today during the pandemic and lockdown. What do you think?

Jacob Stegenga
University of Cambridge
jms303@cam.ac.uk

^[1] See also Eric Schliesser and Eric Winsberg writing in The New Statesman here.
^[2] See also this good blog post by Liam Kofi Bright and Richard Bradley, ‘The Masque of Rona’.
^[3] See also this post on Digressions & Impressions.
^[4] Schliesser and Winsberg’s The New Statesman article is a good example of this.