From Duhem to van Fraassen: the empiricist tradition of ‘saving the phenomena’ against Galileo’s mathematization of nature
Michela Massimi
University College London

The debate on scientific realism has raged among philosophers of science for decades. The scientific realist’s claim that science aims at giving us a literally true description of the way things are, has come under severe scrutiny and attack by antirealists, in particular by Bas van Fraassen’s constructive empiricism. All science aims at is to save the observable phenomena, according to van Fraassen.

If we want to understand the philosophical origins of van Fraassen’s view on “saving the phenomena”, we have to go back to the empiricist tradition behind it, namely to Pierre Duhem’s illuminating series of historical essays called To Save the Phenomena. Duhem traces the philosophical tradition of “saving the phenomena” back to Plato. Indeed, Plato introduced what Duhem calls the “method of the astronomer”, which influenced the development of Greek astronomy for many centuries: the astronomer should be fully satisfied when the hypotheses he has introduced succeed in saving the phenomena, without aiming at giving a true description of nature. Duhem placed himself within this philosophical tradition and opposed it to the new “method of the physicist” inaugurated by Galileo. It is only with Galileo that the philosophical trend of “saving the phenomena” came to an end. Galileo wanted the foundations of astronomy to conform to reality; he believed that the Copernican system was not just a system of “mere contrivances for the calculation of astronomical tables but propositions that conform to the nature of things”.

In this paper, I explore the philosophical relevance of Galileo’s method to re-assess some widespread assumptions behind the notion of “phenomena” in current philosophy of science. I contend that the current impasse between scientific realists and antirealists depends in part on a naïve conception of phenomena that they surprisingly share; namely, phenomena as empirically accessible images of otherwise unobservable entities. By contrast with this naïve conception, I claim that Galileo’s mathematization of nature introduced a brand new conception of “phenomenon” that marked the real beginning of the scientific revolution, but that has also been largely misunderstood by modern philosophers of science following Duhem’s empiricist tradition. For instance, van Fraassen has recently portrayed Galileo’s mathematization of nature as follows: “Galileo’s discipline was to determine beforehand a small set of properties and restrict scientific descriptions to those. Not coincidentally, of course, they were the properties representable by geometry and arithmetic: number, size, shape…(…) Modern science began with Galileo’s and Descartes’ evangelical reification of the scientific image of the world”. Interestingly, van Fraassen traces back to Galileo’s mathematization of nature the scientific realist’s tendency to ‘reify’ nature.

By contrast with this empiricist reading of Galileo, I defend Kant’s reading. In the Preface to the II edition of the Critique of Pure Reason, Kant presented Galileo’s experiment with the inclined plane as a crucial episode in the history of science that marked the beginning of modern physics. Most importantly, Kant saw in Galileo’s experiment and, more in general, in his mathematization of nature, the emergence of a brand new conception of “phenomena” which he regarded as paradigmatic and upon which he patterned his own view about scientific knowledge and its being confined to phenomena. Kant neither endorsed phenomenalism nor embraced the empiricist view that all there is to science is to save the phenomena. Instead, he suggested a radically new conception of “phenomena”, and hence of scientific knowledge, which in my opinion not only does justice to Galileo’s method but can also shed light on the current impasse between scientific realists and constructive empiricists.

In this paper, 1) I spell out what Kant’s new conception of “phenomena” is; 2) I clarify in what respect it is patterned upon Galileo’s own method; 3) why it differs from the Duhem-van Frassen empiricist conception of phenomena; 4) why it can be all the more relevant to overcome the current standoff between scientific realists and constructive empiricists.