Kant on the imagination

A number of people who wind up at this blog do so by way of some form of Google search on Kant. I’m not sure how helpful those individuals have found the blog, but in case someone is looking for good reading on Kant’s theory of the imagination, and since it’s been on my mind lately, I thought I would post what is really no more than a selected bibliography on the subject. Please feel free to add to the list by way of comment. I would be interested to know what others have found helpful.

For reading in book form, what comes to my mind is Béatrice Longuenesse’s Kant and the Capacity to Judge (1998), especially Part III. This material is difficult, however. Henry Allison devotes a subsection to the imagination in Kant’s Transcendental Idealism (2^nd edition, 2004), 189-93 and discusses some of the sources that I mention below. One day soon I hope to return to Rudolph Makkreel’s Imagination and Interpretation in Kant (1990), but I confess that I have only begun to read it.

I have found the following articles quite useful and interesting—especially Wilfred Sellars’ essay. His illustrations, which as far as I can tell he drew himself, are worth a look by themselves. But his essay also paints a very rich picture (in rather little space) of what the imagination is and how it functions in perception.

Anyway, here are the articles:

P. F. Strawson, “Imagination and Perception” in Kant on Pure Reason, edited by Ralph C. S. Walker (Oxford: Oxford University Press, 1982), 82-99 (first published in 1971).

Wilfred Sellars, “The Role of the Imagination in Kant’s Theory of Experience” in Categories: A Colloquium, edited by Henry W. Johnstone, Jr. (University Park, PA: The Pennsylvania State University, 1978), 231-45.

J. Michael Young, “Kant’s View of Imagination,” Kant-Studien 79, No. 2 (1988): 140-164.

Hannah Ginsborg, “Lawfulness without a Law: Kant on the Free Play of Imagination and Understanding,” Philosophical Topics 25, No. 1 (Spring 1997): 37-81.

A bad criticism of Henrich’s analysis of B-Deduction’s proof structure?

Dieter Henrich has the distinction of having written two classic essays on Kant’s Transcendental Deduction. The first was published in 1969 and is entitled, “The Proof-Structure of Kant’s Transcendental Deduction.”¹ The second was published in a 1989 collection and is entitled “Kant’s Notion of a Deduction and the Methodological Background of the first Critique.”² I suspect that the reason these essays have their canonical status is that the main points they set out to establish have more or less become accepted as fact. In the former, Henrich claims that the two halves of the B-Deduction (§§15-20, §§21-26) constitute one argument where these halves serve as its main steps; they are not distinct arguments for the same conclusion. In the second essay, Henrich argues that the word “deduction” in “Transcendental Deduction” has a sense borrowed from the name of a kind of legal document widely used during the Holy Roman Empire and that this is relevant for understanding Kant’s argument.

But if the main lines of these papers have largely been accepted, the details have not. And here I want to focus on one way in which Henrich’s “Proof-Structure” paper has been challenged. Whatever problems Henrich’s proposal faces, it seems to me that this objection rests on a misunderstanding—due in part, however, to the way Henrich states his case.

Henrich presents his sketch of the B-Deduction’s argument in fewer than three pages (645-47). Consequently, one can imagine that a lot of important detail is missing. He argues that the result in §20 of the B-Deduction “contains a restriction: [Kant] established that intuitions are subject to the categories insofar as they, as intuitions, already possess unity (B143).” This conclusion leaves open, however, “the range within which unitary intuitions can be found” (emphasis Henrich’s). Henrich maintains that the “restriction” of §20 will be “overcome” by §26, employing the following strategy:

[W]herever we find unity, this unity is itself made possible by the categories and determined in relation to them. In our representations of space and time, however, we have intuitions which contain unity and which at the same time include everything that can be present to our senses. For indeed the representations of space and time have their origin in the forms of our sensibility, outside of which no representations can be given to us. We can therefore be sure that every given manifold without exception is subject to the categories.

Now, I wonder whether Henrich does himself a disservice by saying that §20 contains a restriction and that this restriction is overcome by §26. This most naturally suggests that Henrich thinks (that Kant thinks) that there are intuitions that (1) are “unitary” and those that (2) are not and that by §26, Kant concludes that not only the unitary intuitions, but also the non-unitary ones, are “subject to the categories” (by dint of having spatio-temporal form).

Unfortunately, I see little in Henrich’s essay to correct this impression, except some remarks made a few passages after the above citation. Picking up on A90-1/B123, Henrich seems to claim that the possibility that appearances and the understanding might be out of synch—that there might be a “disproportion between consciousness and givenness”—is not real. These remarks suggest that the above summary of Henrich’s analysis of the proof structure of the B-Deduction is wrong. It is not that Kant wants to show that those intuitions that lack unity are also subject to the categories; it is rather that such intuitions do not exist—at least, not for, or in, us. Thus, since only unitary intuitions exist, and since these are subject to the categories, the categories are valid “without restriction.”

This strikes me as ultimately the most natural way to read Henrich’s proposal. The results in §20 are conditional; it is misleading to say that they are restricted, unless one qualifies what one means by this.

With this in mind, I turn to a complaint that has been leveled against Henrich’s analysis by Hoke Robinson³ and Henry Allison.⁴ Here is Robinson:

To hold that the first step [of the B-Deduction] introduces a restriction which is removed by the second step is to hold that the a priori applicability of the categories is in principle narrower after the first step and broader after the second. In the text, however, it is just the reverse.

For in §20 Kant concludes that his results apply to intuitions in general, whereas in §26 Kant claims that the categories apply to intuitions of which we are capable, given our forms of intuition.

I have difficulty understanding this objection unless I take it to presuppose what I have suggested is a wrong understanding of Henrich. On this reading, Kant claims that unitary intuitions are subject to the categories in §20; by §26, he concludes that all intuitions, including the non-unitary ones, are. Thus the demonstrated applicability of the categories in §20 would be narrower than it is by §26. Hence the objection.

But it is hard to know what to make of this objection on what I have suggested is the correct understanding of Henrich’s proposal. For if the result in §20 is taken to be conditional, rather than restricted, then the applicability of the categories in §20 is neither narrower nor broader than it is in §26. The range of their applicability remains constant: to unitary intuitions. But Kant must show that there are no non-unitary intuitions to worry about in the second half of the B-Deduction.

This is not to say that Robinson and Allison have not raised important issues surrounding Henrich’s proposal (and indeed, I think that each has). It is to say only that I do not believe that this is one of them.

¹ Review of Metaphysics 22 (1969): 640-59.

² In Kant’s Transcendental Deductions: the Three “Critiques” and the “Opus postumum,” edited by Eckart Förster (Stanford: Stanford University Press, 1989), 29-46.

³ “Intuition and Manifold in the Transcendental Deduction,” The Southern Journal of Philosophy 22 (1984): 404.

⁴ Kant’s Transcendental Idealism, 2^nd edition, 161; see also 1^st edition, 351-52 n6.

Adding categories to Kant Blog

Over the next few days, I am going to attempt to add faux categories to this blog. Blogger does not currently support categories, but there are ways to get around this. I’ll be trying to employ a method using del.icio.us, which is described at phydeaux3. This may affect the site for a time (e.g., links that go nowhere).

UPDATE (4/18/06): I should note my gratitude to phydeaux3 for helping me get it to work!

Kant, the unity of consciousness, and the genitive

A number of months ago, one of my committee members (I won’t mention who, since he or she may not care to be mentioned here) urged me to be careful about my use of the expression, “representation of” and “idea of.” And a few years back, I was criticized by a professor—rightly, I see in hindsight—for not understanding how Hume used the genitive in the expressions “idea of” and “impression of.” So, for instance, if I say that I have a

(1) (mental) representation of a tree,

I may well mean by this that I have a mental state and that it picks out something distinct from it, viz., some tree, or perhaps trees in general. If it is some particular tree, it does not matter whether this tree exists. The point is that the representation is directed at something other than itself. But a sense datum theorist—like Hume—might employ a similar expression, e.g.,

(2) idea of a tree.

And by this Hume does not mean that the idea represents something else (as above). Rather, he means that one has a particular tree-ish idea, or tree-idea. The expression, “idea of x” functions as “feeling of shame” does. One does not mean by “feeling of shame” that one feels some shame; rather, one means that one has a shame-feeling (that one is in a state of shame). A sense datum theorist might infer from his tree-idea that he is in the presence of something else, viz., a tree. Such a sense datum theorist subscribes to what Hume calls the theory of “double existence” (e.g., there is a tree-idea and a tree corresponding to it). But that doesn’t make the idea of a tree about some tree any more than the puddles of water, which are evidence of rain, are about rain.¹

This is a long-winded way of saying that one must be careful about the use of the genitive case. Said committee member directed me to an essay by Thomas Lennon (on Locke’s theory of ideas) in which, building on the Port-Royal Grammar (1660), he distinguishes more than seven kinds of genitive case.² Wikipedia also has a list of different kinds of genitives.

This leads me to ask: what exactly is the nature of the genitive in Kant’s “unity of consciousness” (often: “unity of apperception”) [Einheit des Bewußtseins/der Apperzeption]? Here are some possibilities, none of which strike me as absurd (I concocted some of the genitive names):

(a) Possessive genitive: the unity belonging to (self-) consciousness.

(b) Objective genitive: the unity that is the object of (self-) consciousness.

(c) Compositional/material genitive: the unity constituted by or made up out of (self-) consciousness.

(d) Sufficiency genitive: the unity brought about by (self-) consciousness (analogy: “damage of the hurricane”).

(e) Make-possible genitive: the unity made possible by (self-) consciousness (analogy: “the grant of the NIH” (one made possible by funds from the NIH, but not created by or fully funded by the NIH, thus distinguishing this genitive from the sufficiency genitive)).

There may be some more that I haven’t thought of. And these options are not necessarily exclusive. I suppose that I have always assumed that (a) is correct. Note, though, that if one takes seriously that at issue is self-consciousness, then if (a) this unity belongs to self-consciousness it might also, in part at least, (b) be the object of self-consciousness. At the end of the day, I think that this is roughly what Kant thought.

But it is also possible that the foregoing distinctions are beside the point. One might argue that the expression, “unity of consciousness,” in Kant’s hands, is idiomatic or elliptical, referring simply to a single subject who is often, or perhaps essentially, conscious. Either way, the expression would then seem to pick out the unity possessed by a subject (or mind) that is conscious. Accordingly, the relation between unity and consciousness is mediated by a subject, in contrast to (a)-(e), where the relationship is in each case immediate. In short, one might contend that (a)-(e) take the expression “unity of consciousness” too literally: in employing it, Kant is simply indicating a conscious subject.

One might also suggest that “unity of consciousness” is elliptical for “unity of conscious representations.” This actually differs from what I had in mind with (a)-(e). But now I am getting away from talking about the genitive….³

Those interested in the unity of consciousness might look at Andrew Brook’s article, “The Unity of Consciousness,” on the Stanford Encyclopedia of Philosophy. He talks a little about Kant, there, too.

¹ I think that the distinction between (1) and (2) is typically said to be the distinction between the objective and subjective genitive, respectively.

² “Locke and the Logic of Ideas,” History of Philosophy Quarterly 18, No. 2 (April 2001), 158-59.

³ But I think this distinction is relevant to Kant’s distinction between analytic and synthetic unities of consciousness. I discuss this in my dissertation.

Brief break from bloggin‘

I’ll be out of town from Tuesday until Sunday night (4/4-4/9), so I won’t be updating this blog during that period. I don’t advertise this under the illusion that I have a large audience that will be disappointed by my absence; I advertise this on the remote chance that someone will comment on one of these posts and be disappointed that it does not appear soon thereafter. I’ll post comments, should there be any, when I return.

Kitcher on Kant and apperception

Lately I have been reading a lot about Kant and apperception. What apperception is for Kant isn’t transparent, and commentators have diverse opinions on this subject, as I remarked in a previous post. One distinction to which one ought to be sensitive is that between what Kant calls “empirical” and “pure” apperception. Kant also calls the latter “a priori” and “transcendental” apperception. Perhaps one can say that there are two (very) broad schools of thought on how to understand pure apperception. On one, it is a sui generis kind of (reflective) consciousness, to be distinguished from (but nevertheless probably closely related to) other forms of consciousness; on another, pure apperception is to be understood as something more homely and less spooky-sounding. Some have claimed that “pure apperception” is, for Kant, just synonymous with “consciousness.” Others have wanted to read instances of the expression as elliptical for Kant’s principle of the transcendental unity of apperception or as elliptical for one’s belief in this principle. Perhaps there are other ways to make this notion homely, too.

By contrast, I favor the first approach. Many philosophers prior to Kant and many since have believed that consciousness is inherently reflective. One way one might try to pitch this belief is as follows. If one is conscious of, e.g., the tree, then one must also be, but in a different way from the way one is conscious of the tree, conscious of oneself (perhaps conscious of oneself perceiving the tree). I think that Kant was trying to capture something like this in his notion of pure apperception.

I billed this post as about Patricia Kitcher’s work on Kant, and so it is. She has written extensively on Kant’s philosophy of mind, but it seems to me that until relatively recently, she was pessimistic about the intelligibility of Kant’s notion of pure apperception. Here she is in Kant’s Transcendental Psychology (1990):

“A priori consciousness” suggests a special type of consciousness. It is hard to see what this could be, however. The only awareness that we have of our states [according to Kant?] is through inner sense (104-05).

In a more recent essay, however—“Kant on Self-Consciousness”¹—she seems far more sympathetic:

I believe that Kantian Self-Consciousness [i.e., transcendental self-consciousness] involves a necessary consciousness of mental activities that are required for cognition, but a consciousness that does not enable the subject clearly to see these activities for what they are (346).

I quite like this essay, and I largely agree with this more recent analysis of what transcendental self-consciousness is. But two things about Kitcher’s article struck me as curious. Perhaps they are related. First, while Kitcher does note that this essay constitutes a modification of views she has presented in the past, she does not include among these presentations, e.g., her book. But surely she would retract the first remark that I cited above. If not, I’d be interested to know how the latter remark is continuous with the first.

This is a minor point. More noteworthy is how she goes on to describe transcendental self-consciousness.² I have in mind in particular such claims as that it is “an obscure consciousness of [one’s] own creative acts.”³ What does it mean to say that this self-consciousness is “obscure”? Interestingly, she appeals to Leibniz to clarify—e.g., his petites perceptions. These representations are supposed to be obscurely conscious states (of, e.g., the noise of a single wave among thousands). And in Kant’s Anthropology, Kant also endorses what are, in effect, petites perceptions.

I find this appeal to Leibniz both interesting and curious. I would have thought that Leibniz with petites perceptions and Kant with his remarks on this matter in the Anthropology were making room not for obscurely conscious representations but for non-conscious representations. Part of Leibniz’s contribution to philosophy of mind, long before Freud was born, was his opposition to Descartes and Locke’s position that mentality is essentially conscious.⁴ His petites perceptions are his chief examples. If they are conscious, even obscurely so, then this is quite compatible with the position that consciousness is the mark of the mental.

My questions, then, are as follows. If the foregoing is right, does it amount merely to a terminological quibble? If so—if “obscure consciousness” is just non-conscious representation—then it seems that Kitcher is actually committed to the position that transcendental apperception is non-conscious representation of one’s mental activity (something similar is the position of certain higher-order theorists of consciousness). But then this phenomenon would no longer be well described as a kind of self-consciousness (although it would be a kind of self-representation). On the other hand, perhaps my remarks constitute more than a terminological quibble. But then, I wonder whether the passages to which Kitcher turns to support her analysis of “obscure consciousness” are truly appropriate to pick out the kind of phenomenon (obscure consciousness—in particular, obscure self-consciousness) that she wants to explain.

¹ The Philosophical Review 108, No. 3 (July 1999): 345-386. Links to this essay are to its location on JSTOR. You’ll need access to JSTOR to access this article.

² At the end of her paper, Kitcher argues that this kind of consciousness is genuinely a form of consciousness of the self.

³ 375. This claim in particular should be contrasted to Kitcher’s earlier contentions that Kant retracts as quickly as he endorses “synthesis watching.”

⁴ E.g., see Alison Simmons, “Changing the Cartesian Mind: Leibniz on Sensation, Representation and Consciousness,” The Philosophical Review 110, No. 1 (Jan. 2001): 31-75. The link is to this essay’s location on JSTOR.