Monday 28 January 2008

McDowell's Kant (Part III)

The Space of Reasons is Back! First off, a special word of thanks to all the commenters who kept The Space of Reasons active while I was away. They’re a number of great insights in your responses, many of which foreshadow much of what I was planning to say. To recap, we began looking at the following question:
(Q1): How are we to reconcile the Kantian claim that experience involves making judgements with the Lockean conception of experience as the means by which objects are given to us?
As promised, I will like to sketch a response to (Q1) based on some thoughts gleaned from the work of BĂ©atrice Longuenesse who limns what may be described as a three-tier approach (for reasons that will eventually become clear). On Longuenesse’s reading of Kant, the understanding is active not only in W-experience, but in N-experience as well (though she of course does not put the matter in exactly those terms). However, she distinguishes between two “aspects” of the activity of the understanding—namely, the “sensible” and the “discursive”.

According to the sensible aspect, “the understanding is a rule giver for the synthesis of imagination…In this first aspect the activity of the understanding, or actualizing of its rules, is nothing else than the productive synthesis of imagination.” [Longuenesse 1998, p. 63] The discursive aspect of the activity of the understanding, by contrast, “reflects sensible synthesis under concepts, whether empirical or pure.” [Ibid]

The first step towards solving the problem posed by (Q1), is to recognise that the activity of the understanding, as it factors in W-experience, is different from that which factors in N-experience. In W-experience the understanding is responsible for the “combination of concepts” in empirical judgements, while in N-experience, the understanding is responsible for the “combination of the manifold of intuition” in sensation.

The second step towards responding to (Q1) is to recognise that N-experience itself may be further subdivided in terms of pre- and post-synthesis manifolds. The pre-synthesis manifold represents simply that which is given by the senses while the post-synthesis manifold represents the end result of the synthesis of imagination, now understood as the sensible aspect of the understanding. Thus emerges the three-tier structure of the Longuenesse reading of Kant. The first tier is that of the un-synthesised sensory manifold, the second tier is the synthesis of imagination carried out by the sensible aspect of the understanding, and the third tier refers to the synthesis of concepts carried out by the discursive aspect of the understanding.
The three-tier account avoids the trivialising consequences of the two-tier account by preserving some of the anti-Lockean oomph in the Kantian account of experience. On this approach, judgement is involved in N-experience, though it now serves an intuitive rather than discursive function. In its intuitive capacity, the understanding results in one’s experience of the objective world being a certain way, without itself committing one to the claim that it is that way. This brings us back to the two ways in which I suggested that experience contrasts with judgement on the Lockean account. First, experience is seen as passive while judgement is seen as active. Second, experience involves merely a registering that things are a certain way while judgement involves a commitment to them being that way. What the three-tier account effectively does is sacrifice the empiricist idea that N-experience is purely passive, while preserving the empiricist idea that N-experience itself involves no commitment to things being a certain way. Thus, it makes good on one empiricist intuition while sacrificing another. This allows us to preserve one of the central insights of Kant’s Copernican revolution—namely, that experience itself consists of judgements—without sacrificing the empiricist contrast between experience as registering the way things are and empirical judgement as commitment to things being that way.

However, one may well question whether the three-tier account represents a genuine alternative to the two-tier account. Recall, on the three-tier model, N-experience is itself subdivided into pre- and post-synthesis stages. Moreover, the pre-synthesis stage (i.e., the first tier of the three-tier account) is now identified with the raw deliverances of the senses, before the understanding has had a chance to operate on them. The question may therefore be asked, what prevents us from simply identifying the first tier with Lockean experience and the second tier with empirical judgement? If this question goes unanswered, then the trivialising objection simply re-emerges.

I do not believe the attempt to reduce the three-tier account to the two-tier account in this way is successful. Notice, the reduction only gets off the ground if the first tier is identified with Lockean experience. But this identification is not right. Recall, that on the Lockean view, experience provides us with data which can be used in empirical judgment. However, the unsynthesised manifold that characterises the first tier could not possibly fulfil this role. It is only after the synthesis of imagination, which characterises the second tier, that the subject is in a position to exercise empirical judgements in the manner the empiricist envisions. It is this fact that gives teeth to Longuenesse’s observation that “the psychological data empiricist assume depend themselves on operations empiricists cannot account for.” [Longuenesse 1998, p. 38]

By denying (K3), the empiricist is removing all right she may have for treating experience as structured enough to serve as a basis for perceptual judgement, or at least so the Kantian thought goes. Thus, by Kant’s lights, it is first and second tiers combined that is to be identified with Lockean experience.


Longuenesse, B. [1998], Kant and the Capacity to Judge. Princton, NJ.: Princeton University Press.