Engineering Gone Astray: A Short History of Carnap's Unsuccessful Explication Projects

Christopher French

UBC, Vancouver

Abstract: 
Rudolf Carnap’s brand of scientific philosophy consisted in the use of technical frameworks, like formal syntax and semantics, to reconstruct (or better, to explicate) vague or inexact concepts. The aim was usually to make philosophical questions meaningful, especially questions about meaning itself, analyticity and necessity. Carnap also applied this method to the reconstruction of scientific concepts in order to resolve controversies in the empirical and mathematical sciences -- controversies typically arising from confusions about the meaning of scientific concepts (e.g. in Der Raum and The Logical Syntax of Language). Carnap’s method matured in his most productive years, in the late 1940s and 1950s, when he explicitly stated a method of explication which makes this process of scientific concept construction explicit. Working on these topics within a community of like-minded philosophers, like John G. Kemeny and Yehoshua Bar-Hillel, Carnap used his previous work in semantics from the later 1930s and early 1940s to provide explications for the logical concepts of probability, information and entropy. Interestingly, Carnap also proposed ways in which these concepts could be used to intervene in scientific controversies. First, Carnap argued for a sort of concept pluralism by suggesting that both the logical and frequency concepts of probability have a legitimate role in science. He then suggested that a semantic concept of estimation -- defined in terms of both the frequentist and logical concepts -- could provide a general theory of estimation functions and so a general framework for the foundations of theoretical statistics. Carnap’s semantic concept of information was meant to provide exactly what the statistical concept lacked: a notion of information based on the semantic content in a signal. Lastly, Carnap proposed that a logical concept of entropy, also defined as a semantic concept, proposed as a way to resolve certain interpretative worries about the statistical concept of entropy in statistical mechanics. As it turned out, however, all three of these explications projects, albeit in different senses, failed. The question that my talk raises is: why did these explication projects fail? Typical reasons, usually having to do with Quine and Kuhn, are philosophical or methodological: reconstruction projects just don’t answer philosophical or methodological questions in an adequate way. Taking for granted that these are serious worries about explication projects, this paper suggests there are other, historical and sociological, reasons for why Carnap’s later explication projects failed. If explication projects are like engineering projects, Carnap was just a bad engineer. I use archival and published information to articulate the various ways in which Carnap’s work in inductive logic was poorly circulated, delayed and misunderstood. The thesis is that the fact that Carnap’s explication projects did fail does not provide sufficient reason to reject Carnap’s method tout court. By turning to the historical context of Carnap’s technical projects, the paper aims to spark a new conversation about how to disentangle the technical, rhetorical and historical reasons from the methodological and philosophical reasons for Carnap’s failed projects.