Was Lewis a Kantian?

I recently received my copy of Lewisian Themes. This is a very nice collection that includes an interesting article by Rae Langton entitled "Elusive Knowledge of Things in Themselves". In her article, Langton compares Lewis's argument for epistemic humility, i.e. the thesis that we do not have knowledge of things in themselves, that we do not know the intrinsic properties of objects, with Kant's argument for humility, as interpreted by Langton.

Lewis's argument is motivated by Ramseyan reasons, namely that if "T [the complete and final theory of the world] is multiply realizable, then some fundamental properties are hidden from us" (p. 132). More carefully, we cannot use observation to identify which realisation of T is actual. This is because no matter how T is realised, any evidence available to us could only provide us with support for the Ramsey sentence of T. Our knowledge of the fundamental properties is only knowledge of them as role-occupants and Lewis says that "to the extent that we know of the properties of things only as role-occupants, we have not yet identified those properties. No amount of knowledge about what roles are occupied will tell us which properties occupy which roles." Lewis believes that there are indeed multiple realisations of T on the basis of his permutation argument and replacement argument. According to the permutation argument, we can permute fundamental properties and the laws governing them that feature in T while leaving the rest fixed, thereby bringing about a scenario that differs at the fundamental level but is cognitively indistinguishable to us since all the original roles are occupied. This needs to be qualified in that the permutation argument relies on combinatorialism and quidditism (cf. p. 132 footnote). (For a criticism of quidditism cf. J. Schaffer's paper "Quiddistic Knowledge" also in Lewisian Themes.) The replacement argument invokes the idea of replacing the fundamental properties with 'idlers' or 'aliens', i.e. fundamental properties that play no active role and do thus not feature in T. Again, the fundamental level is different but in such a way that it does not affect the roles being occupied.

Though leading to a similar conclusion as Langton's Kant, the arguments are quite different since the "Kantian answer to the epistemological question depends not on the multiple realizability of realizers, but on a kind of receptivity of knowledge, and a kind of irreducibility of causal power" (p. 134). Langton then discusses to what extent a contextualist epistemology (as outlined in Lewis's paper "Elusive knowledge" (1996)) can help us overcome our ignorance of things in themselves. While such a strategy may be reasonably successful in dealing with Ramseyan humility insofar as we might be properly discounting alternatives, this does not apply to Kantian humility since it is not based on there being alternative realisations of the role properties. The Kantian argument "did not exploit the idea of possibilities that fail to be ruled out be our evidence" (p. 136) and can thus not be dealt with by means of a contextualist strategy according to which we can properly ignore certain possibilities when we are in the right contexts.

(A detailed discussion and criticism of some of the technicalities of Lewis's argument can be found in "Humean Humility?" by S. Leuenberger.)