ABSTRACT
Before starting, let me motivate details of the forthcoming construction. The undercutting argument (hereafter UA), raised in Part I, shows that certain aspects of our web of belief are required by the machinery of conceptual change; and because whatever is required for the evidential justification of a conceptual change cannot itself be flushed out with the debris, those aspects are unrevisable. I first explored this strategy for unrevisability while investigating one strand in the traditional notion of the a priori: to see by means of it whether logic, in whole or part, must be held rigid during conceptual change (see Azzouni 1994).1
