On Scientific Method, Induction, Statistics, and Skepticism

This paper is part of an attempt to explain the distinctive status of modern science within human intellectual life, and in particular the way in which it seems to undermine the legitimacy of other, traditional modes of thought. I argue that the key feature of modern scientific methodology is its testing, not of our judgments, but of the concepts employed in them. Statistical techniques represent a systematic way of doing this. I show, however, that something similar happens on a larger scale in so-called scientific revolutions. (Hence I argue that there is no discontinuity between the methods of “normal” and “revolutionary” science.) My aim in this paper is to explain how universal statements, as they occur in scientific theories, are actually tested by observational evidence, and to draw certain conclusions, on that basis, about the way in which scientific theories are tested in general. 1 But I am pursuing that aim, ambitious enough in and of itself, in the service of even more ambitious projects, and in the first place: (a) to say what is distinctive about modern science, and especially modern physical science, as a human intellectual activity; and (b) to show how this distinctiveness explains the unique status of modern science in human intellectual life. So I will begin by saying a few words about that larger project. One might doubt, first, whether that project is legitimate. Although everything is different from everything else, the question “What is distinctive about X?” is not necessarily well put, because X may not be anything—that is, anything distinct. And there are indeed many philosophers, otherwise of the most diverse intellectual backgrounds and tendencies, who would deny, in various ways, that modern science is a distinct thing, or that its status is unique. It seems to me that they deny something terrifyingly obvious— a fact which confronts us far more urgently than the fact that, say, ravens are black. I will put off further remarks about this until the end of the paper, however, because the discussion of my more limited present aim will focus precisely on the ways to tell when a question is well put, and whether there is such a (distinct) thing as X . That one’s methodological problems are also the subject of one’s investigation is a sign that that investigation is philosophical, although (or because) also a threat to its coherence. Second, some points about methodology. I am not a sociologist, or even a historian, nor (unlike some recent philosophers of science) will I pretend to be $Id: scimethod.tex,v 5.6 2003/10/17 15:11:53 abestone Exp $ scimethod.tex; 17/10/2003; 10:53; p.1