"The Knowledge Machine": some brief impressions

Over the past few days, I have read Michael Strevens’s The Knowledge Machine: How Irrationality Created Modern Science . This book caught my attention because it creates a clash between two biases of mine: while, on the one hand, I am sceptical about method-centric accounts of science, I am also very interested in seeing treatments of the irrational elements of scientific practice.1 So, after seeing some enthusiastic reviews on Twitter, I decided to give this book a try.

This book presents an interesting overview of two giants of 20th Century Philosophy of Science — Karl Popper and Thomas Kuhn — and of the shortcomings in their overarching views of science. Like those two, Strevens subscribes to the idea that science is characterised by a unique method of looking at the world. However, he eschews Popper and Kuhn’s definitions of the method in favour of a timeless image of science as a practice rooted in the iron rule: the only proper way to resolve a scientific controversy is through empirical claims. Modern science distinguishes itself from other knowledge-seeking practices, such as the various philosophical traditions, by rejecting appeals to metaphysical, aesthetical or various other forms of rational considerations; instead, the choice between two theories must be grounded on an experiment that supports one hypothesis but not the other.

The iron rule is targeted not at the behaviour of individual scientists, but at the scientific community, as it defines the rules of the game of scientific argumentation.2 This fixed game plays the role that Kuhn ascribes to paradigms in normal science: providing the stability that individual scientists need to provide incremental gains to the accumulated body of knowledge. However, individual participation in the game of science is irrational, because not only one has to forsake the use of all forms of non-empirical — yet rational — argumentation, but also the experiments often require painstaking effort to resolve minute differences between theories. What draws scientists to play this strange game, then? As Strevens suggests, one might point towards a mix of individual and social factors, but they are ultimately related to the effectiveness of science as a social practice that generates knowledge about and intervenes in the world.3

I am not entirely convinced that Strevens’s account provides entirely robust responses to the various studies on the relevance of non-epistemic values in science. As his discussion of the Pasteur example suggests, a cynic could point out that empiricism only takes us so far, and some disputes are resolved through the able wielding of institutional instruments to induce an impression of success. This acknowledgement of the role of non-epistemic values is often associated with “radical subjectivists” such as Bruno Latour, but it does not need to collapse into a full-blown denial of science having any objective content. Here, I am reminded of Hans Kelsen’s approach to interpretation:4 empirical arguments, much like norms, can frame interpretation, but the latter is underdetermined by the former. While Strevens does present a valuable distinction between value-free science and claiming that uses of science — in particular in public policy — are not value-laden, even scientific practice itself is not as hermetically sealed as his image suggests.5

This treatment of values, as well as the relative lack of engagement with contemporary exemplars of science, soured me a bit on the book as a whole. Nevertheless, Strevens gives a fair presentation of the shortcomings in Popperian and Kuhnian images of the philosophy of science, which might be of interest both for scientists and for those like me who work in science-informed fields. In addition, the author presents a fascinating ideal for science: as an irrational practice that, unlike some of its most ardent enthusiasts state, has no claims to the entirety of human thought, but is propelled by this rationality toward growing knowledge about the world. I therefore recommend this book as a thought-provoking read about science, but policy-minded readers should feel comfortable to skim large sections of it, especially given how some points are frequently repeated.

  1. As Charlie Stross’s Laundry reminds us, computer science in particular often seems to be closer to an applied branch of theology than to a properly scientific discipline. ↩︎

  2. Here I am reminded of Frederick Schauer’s defence of procedural rules regarding the acceptance of evidence. Against philosophers of science such as Larry Laudan, Schauer claims those rules are not just a legal anachronism, or an instrument to enforce the non-epistemic values at stake in judicial trials, but a protection against the fallibility of judges and juries. See, e. g., Frederick Schauer, ‘The Role of Rules in the Law of Evidence’ in Christian Dahlman, Alex Stein and Giovanni Tuzet (eds), Philosophical Foundations of Evidence Law (Oxford University Press 2021). ↩︎

  3. Strevens also seeks to explain why we only had one Scientific Revolution — the one that began in the 16th century in Europe — as opposed to the various revolutions identified by Kuhn or multi-cultural scientific revolutions. He makes good use of historical arguments to show why Kuhn’s model actually overemphasised the frequency and the role of revolutions, and he shows the specific historical factors that lead to the European Scientific Revolution. However, the book is less successful in showing why historical drivers in other contexts could not have led to the adoption of an iron rule, something that would require a deep dive into other candidates. In my uninformed impression, this seems the companion question to Yuk Hui’s investigation of technology in China (Yuk Hui, The Question Concerning Technology in China: An Essay in Cosmotechnics (Urbanomic 2016)). ↩︎

  4. Hans Kelsen, Pure Theory of Law (Max Knight tr, The Lawbook Exchange, Ltd 2009), ch. VIII. ↩︎

  5. On the relation between the ideal of a value-free science and democratic politics, see Liam Kofi Bright, ‘Du Bois’ Democratic Defence of the Value Free Ideal’ (2018) 195 Synthese 2227. ↩︎

Researcher, Law and Artificial Intelligence

Currently researching the regulation of artificial intelligence at the European University Institute.