Monday, 17 December 2007

McDowell's Kant (Part II)

In my previous post, I concluded that the question that motivates McDowell’s Kant may be put as follows:
(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?
One straightforward response to (Q1) is to distinguish between two meanings of the word “experience” as used by Kant in the Critique. The first, which I will refer to as narrow experience (or N-experience), refers to raw “sensory impressions”. N-experience corresponds with the Lockean conception of experience. The second, which I will refer to as wide experience (or W-experience), refers to “empirical knowledge”.

On the present proposal, when Kant talks about experience involving the understanding, it is W-experience that he has in mind. Thus, on this picture, the categories only apply to experience understood in terms of empirical knowledge or W-experience. However, N-experience, which corresponds to the Lockean conception, does not involve the activity of the categories or the understanding. The upshot of this distinction is that it is only for W-experience that the understanding is required, thus avoiding the problem articulated in (Q1). I will refer to the reply to (Q1) that trades on the two meanings of experience adumbrated above as the two-tier approach.

However, many would consider two-tier approach problematic since it seems to undermine the efficacy of Kant’s argument in the Critique as a reply to Hume. Recall, according to Kant, the purpose of the Transcendental Deduction is to show that the categories are a priori conditions of the possibility of experience (See: A 84/B 116—A 95/B 129).[2] Now, according to Hume, we do not perceive objective, publicly observable states of affairs, but only subjective “impressions”. The only way to gain access to the objective world is via an inference to physical objects as the cause of our impressions. However, Hume impugns any such inference as unwarranted. Thus, we are left only with the flux of unconnected subjective experience.

Kant’s strategy in the Deduction is to argue that Hume’s view of experience is mistaken since it omits certain structural features without which we could not have the sorts of experience we unquestionable do have. Kant begins with an assumption which Hume himself accepts:
(K4): Experience is possible
He then sets out to show in the Deduction that:
(K5): If experience is possible, the categories have objective validity
If successful, Kant’s argument is supposed to establish the conclusion:
(K6): The categories have objective validity.
Given (K6), Hume is no longer entitled to his denial of the reality of causation and his sceptical argument is thwarted.

But suppose we interpret the word “experience” as it appears in the Deduction to mean “empirical knowledge”. Then we must build into (K4) everything Kant means by empirical knowledge as set forth in the Critique. Now empirical knowledge, for Kant, is knowledge of an objective world governed by causal laws. However, if we take experience as it appears in (K4) to include all of this, then Kant’s argument would beg the question against Hume. Thus, if we interpret “experience” as empirical knowledge, we would no longer be entitled to see the Deduction as a reply to Hume. The upshot of this is that if we wish to maintain that the relevant arguments in the Critique are effective vis-à-vis Hume, we must also renounce the two-tier account.

In my next post, I will look at a more sophisticated reply to (Q1) that avoids the above criticism, and which is loosely based on the reading of Kant advanced by Béatrice Longuenesse.


5 comments:

Anonymous said...

It doesn't seem to me that Kant thinks too much about this "two-tiered" notion of experience. His point would be that N-experience is, strictly speaking, impossible (i.e. "intuitions without concepts are blind"), and therefore any concept of N-experience collapses into W-experience. To get out of your dilemma then, what would need to be shown is that Kant's strategy in the Aesthetic and Analytic is to show the necessity of that collapse (i.e. show why our normal, noncritical conception of experience can only be explained with reference to a deeper conception of experience which is dependent on certain commitments).

The Minking Than said...

Man I thought I could understand philosophy when I see it. I guess I have to read more. :). Again not quite surprising that you don't have too many comments. My high school mathematics teacher always used to say this - people don't ask anything when they, either did not understand anything or understood everything. I have to read a lot more philosophy before I can start to make meaningful comments to your posts. But looks like you are well versed in your knowledge on philosophy. Good Luck.

j. m. c. dow said...

..."when Kant talks about experience involving the understanding, it is W-experience that he has in mind." this suggests that there is a type of experience that does not involve the categories and the understanding. i'm with the anonymous commentator that suggests that N-exp. collapses into W-exp. experience just is empirical cognition all the way down, involving more and more simple petite judgments or concepts as you get closer to basic indexicals and demonstrations. the way you characterize the dialogue between kant and locke is also puzzling, because you portray kant as in some sense ignoring Locke's notion of N-experience, rather than showing that all experience involves the pure concepts. the problem with the two-tiered approach is not that it doesn't allow a proper answer to Hume, but that it doesn't capture Kant's view as cast in what McDowell calls "the Kantian dictum" (A51/B75). also, i don't recognize Kant's view in the following: "the only way to gain access to the objective world is via an inference to physical objects as the cause of our impressions." i'm not sure this is the only way, since one might read the argument in the first analogy or the refutation as suggesting that material objects are necessary as such, not merely as causes of impressions. your interpretation casts Kant's view of intentionality as a sideways-on view, with causes on one side and impressions on the other. while there is room for such a view at the scientific image level (and idea that McDowell tends to ignore, esp. in the Woodbridge lectures), such a view does not capture the manifest image level. e.g., why shouldn't it be possible for me to have non-inferential access to a lecturn after having kicked it? i don't have infer that it is there; it simply is there in the experience, though this experience is suitably conceptualized. i am also puzzled by what you take Kant to be showing in arguing against Hume for "the reality of causation". this needs to be clarified in order to understand the dialogue between them. it seems like the charge that Kant begs the question largely comes from your assumption that Kant's notion of experience is by definition W-exp, rather than N-exp (suitably altered to capture Hume's view). but why should we think that Kant bases his notion of experience solely on terminological reasons, rather than that he provides arguments in either Locke's or Hume's terms why N-exp is not sufficient to capture our manifest image representations of subject-predicate relations, cause-effect relations, or simultaneity relations. i suppose i am suggesting that it should not be supposed that kant's deduction on its own is sufficient to answer Locke or Hume's accounts of the intentionality of experience, but instead, that the analogies are necessary to show the flaws in N-exp (whether in the Lockean or Humean guise). what do you think?

cunninghamjj said...

It seems to me that with the distinction between apriori concepts (the categories) and empirical concepts (our familiar stock of sortal, demonstrative, and characterising concepts) which Kant clearly makes, we can avoid the problem which you have identified.

With this distinction in mind we can read Kant as claiming that the categories apply to both N-experience and W-experience. But that empirical concepts only apply to W-experience.

This allows us to ascribe the correct anti-Humean result to the deduction since Kant is there concerned to show that the categories necessarily apply to N-experience. But it also allows a role for the 'giveness' of sensory experience in something like the Lockean sense you discuss. Although n-experience does require the capacity to judge, since it is constitutive of one's possession of pure concepts that one have such a capacity, the content of such an experience which is correctly characterised by empirical concepts does take the form of a roughly Lockean 'impression'.

There is, moreover, quite a bit of textual evidence to suggest that empirical concepts do not necessarily apply to N-experience. First of all there is Kant's empiricist sounding theory of empirical concept acquisition (they are derived from n-experience and so the having of an n-experience cannot already imply that we possess them). Second of all there is the constant talk in the axioms of intuition and anticipations of perception of the content of our n-experiences (empirical intuitions) being 'indeterminate'. If empirical concepts were to necessarily apply to experience then such content would be determinate.

En Passant: All of this seems to me to sit pretty well with what Longuenesse says about Kant's theory of empirical concept acquisition. On her view, it is the capacity to make hypothetical, categorical, modal ... judgements about the content of n-experience which enables us to derive empirical concepts from that experience. Such a capacity entials that the categories apply to n-experience. Perhaps you will cover the Longuenesse stuff in your next post.

Joe said...

It seems to me that with the distinction between apriori concepts (the categories) and empirical concepts (our familiar stock of sortal, demonstrative, and characterising concepts) which Kant clearly makes, we can avoid the problem which you have identified.

With this distinction in mind we can read Kant as claiming that the categories apply to both N-experience and W-experience. But that empirical concepts only apply to W-experience.

This allows us to ascribe the correct anti-Humean result to the deduction since Kant is there concerned to show that the categories necessarily apply to N-experience. But it also allows a role for the 'giveness' of sensory experience in something like the Lockean sense you discuss. Although n-experience does require the capacity to judge, since it is constitutive of one's possession of pure concepts that one have such a capacity, the content of such an experience which is correctly characterised by empirical concepts does take the form of a roughly Lockean 'impression'.

There is, moreover, quite a bit of textual evidence to suggest that empirical concepts do not necessarily apply to N-experience. First of all there is Kant's empiricist sounding theory of empirical concept acquisition (they are derived from n-experience and so the having of an n-experience cannot already imply that we possess them). Second of all there is the constant talk in the axioms of intuition and anticipations of perception of the content of our n-experiences (empirical intuitions) being 'indeterminate'. If empirical concepts were to necessarily apply to experience then such content would be determinate.

En Passant: All of this seems to me to sit pretty well with what Longuenesse says about Kant's theory of empirical concept acquisition. On her view, it is the capacity to make hypothetical, categorical, modal ... judgements about the content of n-experience which enables us to derive empirical concepts from that experience. Such a capacity entials that the categories apply to n-experience. Perhaps you will cover the Longuenesse stuff in your next post.