Twenty years have passed since Paul Feyerabend and I first used in print a term we had borrowed from mathematics to describe the relaationship between successive scientific theories. ‘Incommensurability’ was the term; each of us was led to it by problems we had encountered in interpreting scientific texts (Feyerabend 1962; Kuhn 1962). My use of the term was broader than his; his claims for the phenomenon were more sweeping than mine; but our overlap at that time was substantial. Each of us was centrally concerned to show that the meanings of scientific terms and concepts — ‘force’ and ‘mass’, for example, or ‘element’ and ‘compound’ — often changed with the theory in which they were deployed. And each of us claimed that when such changes occurred, it was impossible to define all the terms of one theory in the vocabulary of the other.

In 1962, Thomas Kuhn offered an account of the difficulties which attend the reception of revolutionary scientific ideas: “Communication across the revolutionary divide,” he wrote, “is inevitably partial.” (Kuhn 1962, p. 149). Eight years later, in the postscript to the second edition of The Structure of Scientific Revolutions, he sounded the same theme, characterizing the participants in revolutionary debates as people who attach their terms to nature differently, so that their communication is “inevitably only partial” (Kuhn 1970, p. 198). In his most recent discussion of these issues (Kuhn 1983), Kuhn has explained more clearly than before the nature of the differences in language which he takes to be present in revolutionary debates.

Here is the example that L. J. Savage (1954, p. 21) uses to adumbrate the principle: