Kant's Theory of A Priori Knowledge (review)

Robert Greenberg. Kant's Theory of A Priori Knowledge. University Park: Penn State University Press, 2001. Pp. ix + 278. Cloth, $45.00.

This is one of the deepest and most carefully reasoned books on Kant I have read. It is a book for the scholar of the first Critique, not the "educated layman," but it very much needs to be read by the former. Even for Kant scholars, it is not easy. Apart from demands imposed by the tightness of the reasoning, one may simply be uneasy with the fact that, as instanced in a relentless and detailed critique of the main species of the "customary" reading--represented primarily by Allison and Guyer--Greenberg is defending a radically new view of Kant's reasoning in the Critique.

Typically, one either supposes Kant really has no theory regarding the possibility of specifically a priori knowledge in the first place, or--either for excessive subjectivism, reference to "things in themselves" (in more than an Allisonian "methodological" spirit), or both--it is a disaster. (Or, as in Guyer and before him Strawson, one distinguishes between a Kant of each stripe.) On the no-theory option Kant is primarily concerned with the conditions of empirical knowledge and the relation of experience to objects. So occupied, he in effect redefines what it is to be concerned with the (non-analytic) a priori in the first place: it just is to be concerned with such conditions. In turn, he takes it for granted that, insofar as certain representations are indeed such conditions, they are sufficiently related to experience and its objects to meet the demands of knowledge. To the contrary, according to Greenberg, it is just this which Kant sees cannot be taken for granted.

Regarding the ontology of Kant's theory, Greenberg argues (18-26 and chapter 2passim) that it is and ought to be distinct from that of the knowledge that is its target, that is, from the world of "appearances." But this is not to say that what is in question is a world of "things in themselves": things that, as Greenberg takes that notion, would have to be viewed "according to properties or laws that are intrinsic to them" (101), to make sense of the theory in the first place. Rather, the ontology comprises "things simpliciter": things that might (though not by us) be viewed as "things in themselves." What the theory tells us, first, is thata priori knowledge is possible only insofar as it relates to things precisely as appearances. Second, this is possible only insofar as experience itself relates to things as appearances. Thus to the contrary of standard readings, the primary focus is on the possibility of experience, and its relation to objects, as a condition of a priori knowledge, not vice versa. This is of course not to deny the importance of space, time, and the categories as conditions of experience. It is just that their relevance to the possibility of a priori knowledge cannot consist wholly in the fact of such conditioning.

Greenberg concedes that the book falls short of defending the objective validity of Kant's theory, as opposed to its internal coherence and--as exemplified mainly by way of a detailed analysis of the second-edition Transcendental Deduction (chapters 11-3), a convincing account (chapter 10) as to what was needed but lacking in the original version, and (chapter 9) a major independent contribution to the discussion of Kant's views regarding "logical forms" and "functions"--the validity of Kant's reasoning. That just illustrates the general point: just as the ontology of the theory needs to be distinct from that of the knowledge that is its target, the point from which the objective validity of the theory is assessed needs to be part of some further theory. This may be dissatisfying. "But is that not how things always stand in philosophy?" (269).

There is no way that...