Philosophy of Science: normative or descriptive? Lessons from IARC’s Methodology

In any introductory course in philosophy of science we ask the question whether philosophy of science should be normative or descriptive. This question has a history of its own, which also influenced what has become the ‘mainstream’ in philosophy of science, and that marked a difference between philosophy of science on the one hand and the social studies of science on the other hand.

For one thing, even the terms ‘normative’ and ‘descriptive’ have no univocal meaning.  On the one hand, the normative component may, for instance, concern the conceptual / theoretical framework of the sciences or instead the way in which science uses / applies such framework.  On the other hand, the description of science may happen at different levels: the level of well established theories, systematised in handbooks, the level of the practice of research (whether contemporary or historical), the level of the social dynamics of science, etc.


This is all very conceptual so let’s reason on a concrete example. Take for instance the methodology used by the International Agency for Research on Cancer (IARC) to write the ‘monographs’. The monographs issue an evaluation carried out by a group of experts and they concern the carcinogenicity of given substances, chemicals, etc. The IARC monographs attracted quite some attention from the public and the press because of the decisions issued about glyphosate (a herbicide massively used world-wide) and about consumption of red and processed meat.



The methodology of the IARC has been thoroughly examined in philosophy of science, and specifically in the philosophy of causality and of medicine. It has been used as an illustration of the so-called RWT (the Russo-Williamson Thesis), then generalised in the position called ‘evidential pluralism’. According to this view, to establish causal claims (for instance whether and to what extent exposure to glyphosate or red meat consumption cause cancer) one need to establish that the putative cause makes a difference to the effect (correlation) and that there is some mechanism linking the putative cause to the effect. This is of course a very simple formulation that doesn’t make justice to all the nuances that several scholars developed over the years. In fact, much has been said – and still is to be said – about what evidence of difference-making and of mechanisms exactly consist of, how they work in practice, how they ought to be combined, etc.


But I mention the IARC methodology not to give yet another argument in favour of evidential pluralism, but rather in favour of a certain philosophical methodology and of certain scientific practice.


To begin with, philosophers engaging with the IARC methodology use it as an illustration of a philosophical idea, but also use it to sharpen their ideas. We aren’t passive observers of a scientific practice (a purely descriptive task) and we don’t ‘glue’ our ideas onto the scientific practice (a simple-minded normative task). A good philosophy of science ought to look for appropriate interactions between ideas and science-in-practice and remain open to change. At the same time, IARC scientists engaging with EBM+ try to implement the ideas behind evidential pluralism to improve on their methodological set up. So such collaboration between EBM+ philosophers and IARC scientists go well beyond a scholarly debate.


This is brief and certainly not exhaustive of the various interactions between these two worlds, but hopefully this is enough to draw some conclusions about scientific method and about philosophical methodology. These remarks may sound obvious. Yet, in the aftermath of several media attacks to the IARC or the public distrust in vaccines, they are instead worth restating.


Scientific method is not written on the stone nor is it infallible. And yet, there is a lot that scientific method achieved and most procedures are reliable and solid. At present, the IARC methodology represent one of the best examples of a solid methodology, one that attempts to integrate the results from largely successful scientific enterprises: epidemiology, biology, molecular medicine, etc. This does not imply that the method has nothing to improve. And in fact the ‘Preamble’, the text fleshing out the IARC methodology, is continuously discussed and revised.


What can philosophy of science contribute to this? Quite a lot. For one thing, philosophy of science can help disentangle important conceptual distinctions. For instance: What does it mean that a substance is cancerogenic at the population level? And what does it mean for an individual person or patient? How much do we need to know about biological mechanisms or bio-chemical pathways in order to establish that a substance is cancerogentic? Why does a community disagree? And what scientists disagree about? How can we disentangle genuine epistemic factors for disagreement from socio-economic-political factors (e.g., vested interests)? Once we ask question like these, it is easier to see that the normative vs descriptive characterization of philosophy of science misses what is interesting about philosophy, namely its relevance for thinking about science and for contributing to make it every day better.


By Federica Russo

Assistant Professor,

University of Amsterdam


Leave a Reply

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