Stephen P. Stich, The Fragmentation of Reason

Abstract

This is a lively, engaging, and provocative treatment of some central issues in meta-epistemology. The issues I wish to address begin in chapter 4, where Stich asks how normative standards for cognitive processes might be discovered and defended. Let me turn, in particular, to his critique of analytic epistemology: the attempt to identify epistemic value by analyzing normative epistemic terms found in everyday language. Stich's critique is largely based on the specter of cultural diversity in epistemic concepts. Why should we choose among competing cognitive practices on the basis of the standards implicit in our own language and culture, once we recognize that other languages and cultures may have different epistemic concepts and standards? (One wishes there were more definite evidence that there are cross-cultural differences in epistemic concepts. The one example Stich discusses, relegated to a footnote, is pretty inconclusive, a mere 'hint' as he himself puts it.) I agree that unless one is inclined toward chauvinism in matters epistemic, one should not value conformity with an evaluative notion simply because that notion is prevalent in one's society. But I doubt that anyone endorses epistemic standards on the grounds that they are one's own; presumably they are invoked because they mark out something valuable, either intrinsically or extrinsically. The question is whether one should cease to value them given the hypothetical discovery that other languages and cultures have different (though not necessarily conflicting) concepts. Not so, I submit. Our very own language already has a multiplicity of distinct epistemic standards. As I argue elsewhere, 'justified' is not quite equivalent to 'rational' or 'intelligent'; cognitive processes can be evaluated in terms of the distinct dimensions of reliability, problem solving power, or problem