ABSTRACT
Consider first the anti-monistic movement, represented most notably perhaps by the work of Bachelard, Koyré, Popper, Lakatos, Feyerabend and Kuhn. Both Bachelard and Kuhn come very close to the position, whose roots lie in Vico, and which I shall characterise as super-idealism, that we create and change the world along with our theories.1 Neither Kuhn nor Feyerabend have managed to sustain the intelligibility of the concept of a clash between incommensurable descriptions, or to say over what such descriptions clash. Popper has not shown how the falsification of a conjecture could be rational, unless nature were uniform. And he has not furnished any ground for assuming that it is, in the face of Humean and Goodmanesque possibilities. Nor has Lakatos shown how unless nature were uniform, it would be rational to work on progressive rather than degenerating programmes; or, for that matter, pay any attention to the history of science. More generally, the theorists of scientific change have found it difficult to reconcile the phenomenon of discontinuity with the seemingly progressive, cumulative character of scientific development, in which there is growth as well as change.
