From logical to probabilistic empiricism: arguments for pluralism

Starting from the 1960s, the conception of the philosophy of science forged by logical empiricism was gradually superseded by a new way of doing philosophy of science characterized by a progressive opening to the pragmatic components of science. On the one hand, the so-called post-positivist movement urged the need to consider the historical and sociological aspects underlying the formation of scientific knowledge, while on the other, the semantical view of theories prompted a shift of interest from theories to models and called attention to all aspects of experimentation. A decisive step in this direction was taken by Patrick Suppes, who labeled his own perspective “probabilistic empiricism” to emphasize the crucial role of probability and statistical methods in science. The turn of the 1960s brought about a pluralistic attitude regarding both scientific methodology and the spectrum of disciplines of interest for philosophers of science. Pluralism is currently a widespread position that different authors develop in different ways. A peculiarity of Suppes’ probabilistic empiricism is the strict link he establishes between pluralism and the statistical methodology for experimentation and forming hypotheses. After an outline of the major features of the logical and probabilistic versions of empiricism, followed by some remarks on pluralism, this article draws attention to the desirability of adopting a bottom-up approach that assigns crucial importance to the context, in tune with Suppes’ constructivist and pragmatical attitude. With an eye to the foundations of statistics, it is argued that pluralism and the adoption of a bottom-up approach are necessary choices in the light of two examples: cluster analysis, and the use of probability and statistics in courtrooms, which is discussed in the light of a controversial legal case.