Monday 25 February 2008

McDowell's Kant VI

We have come to my final post in this series on McDowell's Kant. Earlier I argued that the two-tier approach reduces Kant’s picture to the Lockean view, according to which the understanding is not active in experience but only in the judgements we make based on experience. Once the understanding has been removed from experience, we are left simply with raw sense-impressions that are now presumed to serve as a basis for judgement. But this means that sense-impressions must rationally constrain one’s empirical beliefs, which just is the “Myth of the Given” as Sellars understood it.

The alternative approaches I considered, {STRATEGY1} and {STRATEGY2}, are the Kantian analogues to the coherentist and bald naturalist positions respectively. Both conceive of the relationship between sense-impressions and beliefs as being merely causal. However, there are important differences between the two. {STRATEGY1} still preserves the notion of the understanding as the domain for rational inferences. But on {STRATEGY1}, sense-impressions can only exert causal but no rational (or justificatory) influence on the activities of the understanding. As Davidson’s famously (or is that infamously?) puts it, “only beliefs can justify beliefs.” On this view, the content of any one of our beliefs is determined solely by its relation to other beliefs, since the sense-impression (striped of all rational currency) can no longer determine their content. McDowell sees this view as ultimately self-defeating since it prevents us from “so much as making sense of the notion of a belief at all."(McDowell [1994]. p. 9.) {STRATEGY2} dispenses with the understanding altogether, reducing the entire chain leading from sense-impressions to empirical beliefs and judgements to a set causal relations. The price of {STRATEGY2} is that we are forced to give up the idea that rational relations are sui generis. In brief, according to {STRATEGY2}, saying that A justifies B is on a par with saying that A causes B.

According to McDowell, the problems articulated in (Q1) and (Q2) never arise if we conceive of sense-impressions as themselves having conceptual content (McDowell [1994], pp. 9-10). Thus understood, it at once becomes clear how sense-impressions can stand in justificatory or evidential relation to beliefs and endow our beliefs with empirical content. McDowell sees Kant as pointing towards (if not actually arriving at) a solution along these very lines:
The way to stop oscillating [between the myth of the given and coherentism] is to conceive empirical knowledge as a co-operation of sensibility and understang, as Kant does. To avoid making it unintelligible how the deliverances of sensibility can stand in grounding relations to paradigmatic exercises of the understanding such as judgements and beliefs, we must conceive this co-operation in a quite particular way: we must insist that the understanding is already inextricably implicated in the deliverances of sensibility themselves. Experiences are impressions made by the world on our senses, products of receptivity; but those impressions themselves already have conceptual content.(Ibid, p. 46)
By McDowell’s lights, when Kant says that the understanding is required for experience, he is not only saying that it is required for empirical knowledge or judgement—i.e., W-experience—but for sense-impressions themselves, that is for experience in a sense continuous with Lockean experience. Thus, McDowell straightforwardly rejects the two-tier approach described earlier. McDowell, as I understand him, can therefore be described as holding to a version of the three-tier account. Like Longuenesse and others, McDowell sees the role of the understanding within experience as different from its role in empirical judgement. However, he parts ways from most advocates of the three-tier approach by insisting that there is no pre-conceptual manifold of intuition upon which synthesis must operate. McDowell puts the point as follows:
The relevant conceptual capacities are drawn on in receptivity….It is not that they are exercised on an extra-conceptual deliverance of receptivity. We should understand what Kant calls “intuition”—experiential intake—not as a bare getting of an extra-conceptual Given, but as a kind of occurrence or state that already has conceptual content. (McDowell [1996], p. 9)
By McDowell’s lights, the only sensible impressions are those that are already endowed with conceptual content and can therefore stand in grounding relations to beliefs. Moreover, these sensible impressions are not in turn to be seen as arising from the exercise of the understanding on some more basic type of sensible impression. There is, as McDowell puts it, no “sheer receptivity”. The understanding is “already inextricably implicated in the deliverances of sensibility themselves”.(McDowell[1996], p. 46) On this account, the problem of how “blind” intuitions can determine the rules by which they are to be synthesized does not arise.

McDowell therefore agrees with the Kantian slogan: “intuitions without concepts are blind.” [A51/B75] But he draws from it a very different conclusion to that underlying the views described earlier. Previously, this phrase was taken to highlight the fact that intuitions, conceived in terms of the pre-synthesised sensory manifold, are unable to play any guiding role in activities of the understanding when selecting which rule to apply to said manifold. This is supposed to emphasise the importance of bringing intuitions under concepts in order to make them understandable. McDowell grants the point, but then takes the Kantian suggestion to be, not that the synthesis of the imagination needs to act upon the manifold before it can serve in judgement (i.e., that what was blind must be made to see), but that sensory intuitions should never be seen as blind to begin with.

The questions, (Q1) and (Q2), with which I have been concerned may be put as follows: how can that which was blind (i.e., unconceptualised intuition) be made to see (i.e., intuition brought under a concept)? Admitting that sensory impressions are blind seems to suggest that they cannot guide the understanding in the application of the rules under which the manifold is synthesised. But this means that impressions can exert no rational influence on the understanding. But if impressions can exert no rational influence on the understanding, then it remains unclear how experience can justify our empirical judgements. The way to avoid this conclusion, McDowell suggests, is to deny that impressions are blind. Assuming that the Kantian dictum, “intuitions without concepts are blind”, holds, the only way to deny that impressions are blind is to say that they involve concepts. Once we picture impressions as conceptual, we are (to put the matter crudely) picturing them as speaking the same language as the understanding. Sense impressions are now able to play the guiding role necessary vis-à-vis the understanding’s rule application to the manifold.


Longuenesse, B. [1998], Kant and the Capacity to Judge.

McDowell, J. [1994], Mind and World.

McDowell, J. [1996], “Woodbridge Lectures”.

*Although not directly cited, my reading of McDowell's Kant is significantly influenced by the work of Hannah Ginsborg

Monday 11 February 2008

McDowell's Kant (Part V)

We are only one post away from my summary of McDowell's own position (I'm sure the suspense has been just eating away at you by now!) But first, I want to consider two more alternative positions (for reasons that should become clear by the end of the post).

The first alternative response to (Q1) and (Q2), which I will henceforth refer to as {STRATEGY1}, would be to argue that unconceptualised intuition merely plays a causal, rather than normative, role in determining how the sensory manifold is synthesised. The difference between the causal and normative can be made vivid by contrasting the case in which A triggers B with the case in which A indicates B. On the present view, the intuition does not suggest or indicate which rule the imagination ought apply. Rather, it merely triggers the application of a certain rule. Take for example the case in which one’s vision is affected by a red cube. On the present account, the red cube causes one’s imagination to apply the red-rule and the cube-rule as opposed to the green-rule and sphere-rule. However, it does not indicate that the imagination ought to apply the red-rule and cube-rule.

One advantage of this view is that it allows us to preserve the Kantian doctrine that intuitions without concepts are blind. According to this picture, the intuitions do not tell the imagination to apply a particular rule, it merely causes the application of the rule. But its greatest strength is also its greatest weakness. For once we conceive of the intuition as merely exerting causal influence on the imagination, then it becomes unclear how the intuitions can, as Kant says, provide the “matter” of the experience. One upshot of this fact is that the intuition plays no constitutive role in determining the content of experience. That is, intuition may determine whether the understanding applies the red-rule or cube-rule, but it plays not part in making up the rules themselves. Thus, the repertoire of rules available in the understanding for the imagination to apply is determined solely by the understanding itself.

Significantly, this view seems to be in opposition to the kind of approach Kant himself would endorse since it essentially collapses the distinction between pure and empirical concepts. Thus, insofar as we wish to remain true to Kant, we have good textual grounds for rejecting this picture. But even if we have no qualms with departing from Kant on this score there still seems to be non-textual grounds for being sceptical about this approach. Specifically, {STRATEGY1} makes unclear how experience could be about a mind-independent world. McDowell expresses the worry as follows:
But if our freedom in empirical thinking is total, in particular if it is not constrained from outside the conceptual sphere, that can seem to threaten the very possibility that judegments of experience might be grounded in a way that relates them to a reality external to thought. And surely there must be such ground if experience is to be a source of knowledge, and more generally, if the bearing of empirical judgments on reality is to be intelligibly in place in our picture at all.McDowell [1996], p. 5.
The problem with {STRATEGY1} is that by making empirical concepts (in addition to the pure concepts) solely the product of the activity of the understanding it eliminates any role the manifold may play in determining the content of our judgements.

A second strategy, henceforth {STRATEGY2}, would be to give up on the idea of normativity altogether in favour of a more thoroughgoing causal account. On this view, it is not simply the case that the unconceptualised intuition causes the imagination to apply the rule that it does, but the rule applying activities of the imagination, under the direction of the understanding, is itself to be understood in purely causal terms. Whereas on {STRATEGY1} normative rules guide the interaction between the understanding and the imagination, and causal laws guide the opposite direction from the unconceptualised intuitions to the imagination, on {STRATEGY2}, both directions are to understood in purely causal terms. On this picture, what are taken to guide all of the imagination’s activities are no longer rules, understood as concepts, but rather brute causal laws. This view has the advantage of eliminating any problem concerning how the unconceptualised intuition, understood as blind, could direct the imagination’s application of rules since the rules the imagination applies are themselves now seen as causal laws of exactly the same kind that govern the unconceptualised intuition.

However, this view suffers from the same defect as {STRATEGY1} since it fails to explain how experience can provide the “matter” of judgement. Moreover, {STRATEGY2} removes the activity of the understanding from experience altogether. It therefore entails the denial of (K3). Thus, whatever the merits of such a proposal, it is not a Kantian one. Rather than offering a solution to (Q1) and (Q2), it actually represents an abandonment of the philosophical approach that gave rise to the problem in the first place. In this regard, {STRATEGY2} can be seen as employing a type of quietism, albeit at great cost to the Kantian view of experience.

It should be clear to the observant reader that the three approaches to (Q1) and (Q2) considered in this and a previous blog posts parallel the three options McDowell considers and rejects in setting up the problematic of Mind and World; the Myth of the Given, on one hand, and the Coherentist and Bald Naturalist positions, on the other. In my next post, I will outline McDowell's alternative.

Monday 4 February 2008

McDowell's Kant (Part IV)

In my previous post I adumbrated the three-tiered approach to (Q1):
(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?
If we adopt the three-tiered account, the problem raised by (Q1) begins to take a slightly different form. Commentators such as Strawson and Sellars have suggested that by Kant’s lights, we come to see some object α as φ by forming an image of α in accordance with the rule or concept φ. On this view, when one comes to see something as red, one is forming an image of that thing in accordance with the rule or concept red. However, the question quickly arises as to what determines which rule the imagination applies in its synthesis of the sensory manifold and what role (if any) does the manifold itself play in its own synthesis?

In the case of the “pure concepts of the understanding”, or the categories, the answer is relatively straightforward. Since all human experience is subject to the categories, then it is the understanding, and the understanding alone, that determines which rules the imagination applies to the sensory manifold. However, the above question becomes most relevant, not with regards to the pure concepts, but with respect to what Kant calls “empirical concepts”, such as red cube or green sphere. For example, we may ask: when confronted by a red cube, what directs the imagination to synthesise the sensory manifold according to the red-rule and cube-rule as opposed to, say, the green-rule and sphere-rule? If our perceptual experiences are going to be of or about mind-independent objects, then it would seem that the sensory manifold must play some role in determining which empirical concepts the imagination employs in its synthesis.

Clearly, if we are preserve the receptivity of experience, then the sensory manifold must play some part in determining which rule the imagination applies in its synthesis. This means that the manifold must, in some sense, be able to “tell” the understanding which rule the understanding should apply in synthesising the manifold. But therein lays the problem. In the “Transcendental Logic” Kant notes that “intuitions without concepts are blind.”(A51/B75) The blindness of the unsynthesised intuitions (i.e., before they have been brought under a concept) seems to preclude them being able to play any guiding role in which rule the imagination applies in its synthesis. But if unconceptualised intuition is able to guide the understanding in its rule application, then it would seem that unconceptualised intuition is not really blind after all. The question that now faces the three-tier approach may be put as follows:
(Q2): How are we to reconcile the idea that sensible intuition determines the content of experience (such that it serves as its matter) with Kant’s insistence that unconceptualised intuitions are blind?
I believe (Q1) and (Q2) are the questions that serve as the backdrop for McDowell’s appropriation of Kant in Mind and World. There, McDowell frames the present questions in terms of “the idea that experience must constitute a tribunal, mediating the way our thinking is answerable to how things are"(Mind and World, p. xii). According to McDowell’s Kant, we can only make sense of the sensory manifold playing a role in determining which rule the understanding applies in its synthesis by conceiving of the manifold as in some sense “justifying” the application of one rule as opposed to another. I will have more to say on McDowell’s conception in my next post.