Notes on <Kant and the Critique of Pure Reason> - The Meaning of Transcendental Idealism
- Kant’s empirical realism: the nature of appearance
- The existence of things in themselves
- Things in themselves and appearances
- The transcendental ideality of the self
- Entering into, and remaining within, the Kantian system
Perhaps the most intriguing single comment made about transcendental idealism is the famous remark of Kant’s contemporary Jacobi that, year after year, he had been forced in confusion to recommence the Critique because he had found himself unable to enter into the system of Kantian philosophy without the presupposition of the thing in itself, and yet, with that presupposition, unable to remain within it.
Kant asks us to think of objects of our perception as mere subjective determinations of our being, and yet at the same time as the product of our being affected. The existence of things in themselves is necessary for transcendental philosophy to uphold the objective validity of our thought about the world. But transcendental idealism also informs us that things in themselves, and the mode in which they affect us, are utterly unknown to us. with which admission, Jacobi contends, Kant’s system becomes unintelligible, for it simultaneously affirms and denies that we have access to objects that make impressions on our senses.s
Kant’s empirical realism: the nature of appearance
Kant’s account of empirical objects is interpreted as assigning to them no deeply different ontological status from the purely phenomenal objects of Berkeley’s idealism. Kant’s emphasis on the strictly transcendental sense in which empirical objects are to be identified with appearances (A29-30/B45) may suggest a difference of kind between his idealism and Berkeley’s.
Kant’s response
Kant replies that his differences from Berkeley are, Kant states, that he affirms, whereas Berkeley denies, the existence of things in themselves, and that Berkeley reduces experience to ‘mere illusion’, as transcendental idealism does not.
The second difference claimed by Kant may seem unfair to Berkeley, who explicitly denies that his account renders the perceptual world illusory. Kant’s point is that for Berkeley the content of experience is not constrained by anything at all, and so is tantamount to illusion.
Kant claims that appearances are grounded on the subject with respect to their form, and on things in themselves with respect to their existence.
Kant’s non-phenomenalism and Berkeley’s transcendental realism
What does establish conclusively Kant’s difference from Berkeley is the consideration that Kant’s ontology of empirical objects is not, and could not be phenomenalistic.
It is true that, for Kant, all objects lie within the framework of experience, But this does not make them ontologically of a kind with Berkeleyan ideas. Being subject to the condition of experienceability is not the same as being composed of experiences in any sense. Cognition of inner experience, sensory states of the subject, would precede that of outer objects, from which subjective states would be independent.
To the extent that Berkeley acknowledges cognition to be something in its own right over and above mere sensing, there is no dependence of objects upon cognition as such for Berkeley: ideas exist fully formed in being sensed. Kant, by contrast, asserts the fundamental dependence of objects on the conditions of our cognizing (judging) them. The understanding’s employment of rules in a priori synthesis cannot, therefore, be compared with the phenomenalist’s logical construction of objects.
The conception of appearances as phenomenalistic constructions belongs not to transcendental idealism, but to transcendental realism. Phenomenalistic entities are objects whose being is exhausted in their immediate sensory appearing. Purely phenomenal entities like mental images are of course admitted by Kant as contents of the empirical world, but they do not play any role in his account of what empirical objecthood as such amounts to.
The concept of representation itself, in transcendental discourse, does not refer to mental items in the sense of the empirical contents of minds considered as empirical particulars, but to subjective elements of cognition considered transcendentally, i.e. as making cognition of objects possible. Kant’s description of appearances as nothing but representations and thus determinations of the self, is not an ontological classification of any sort, and cannot be interpreted as assigning them a mental status. In this way, Kant’s conception of empirical reality avoids the straight choice of Locke or Berkeley.
The two senses of appearance
- Appearance O: objects of representation
- Appearance R: representation
Appearance O and appearanceR may be understood as referring to the empirical object considered at different stages in Kant’s analysis of the process of cognition: appearance R is the sensible appearing that provides an intuitive datum for cognition, and appearance O is the object cognized as so appearing. Understanding converts (‘objectifies’) appearances R into appearances O.
The existence of things in themselves
The second component of the ‘standard picture’ of transcendental idealism, on account of which it encounters criticism of the sort of which Jacobi provides an early and forcible example, is Kant’s affirmation that things in themselves exist.
The contradiction of the claim is that we have knowledge of things of which we have no knowledge. However, we know things in themselves only in so far as we know that something not constituted by the forms of our sensibility must occupy the conceptual space outside experience.
Finding grounds for Kant’s assertion
It is clear that there can be nothing internal to sensation considered empirically that implicates a relation to things in themselves, since Kant conceives sensation as preceding all representation, and in any case as known to us empirically only in so far as it corresponds to the matter of appearance. Also the thing in itself plays no role in the constitution of objectivity: its place is taken by the transcendental object, a concept which we employ without thereby positing the existence of anything transcendent.
Kant may seem to be arguing that the existence of things in themselves follows from the concept of appearance: that things in themselves exist because appearances are necessarily (by virtue of their concept) of things which cannot be identified with appearances, and so must be identified with things in themselves. But this cannot be right, since although Kant’s concept of appearance certainly implies the concept of the thing in itself, and there are objects satisfying the concept of appearance, none of this implies that there are any objects satisfying the concept of the thing in itself.
Alternatively, Kant may seem to be basing the existence of things in themselves on a causal inference - from the existence of appearances as effects, to that of things in themselves as their causes. This cannot be right either, however, since the assumption that appearances are effects of anything at all presupposes exactly what needs to be established. Kant also has argued that deployment of the causal principle outside the sphere of experience is illegitimate.
Alternatively, we can suppose that the existence of things in themselves follows somehow from the consideration that there must be an ultimate end to the explanation of things, and that the realm of appearance is not ultimately self-explanatory. But, on the face of it, to say any of these things would be, in Kant’s terms, to base the existence of things in themselves on reason’s demand for the unconditioned, or to seek to derive ontological conclusions from the principle of sufficient reason in Leibnizian fashion, and the Amphiboly and Dialectic disallow anything of the sort.
Transcendental idealism without the existence of things in themselves
Given the above difficulties we might wonder what is lost if we remove things in themselves. Advocates of this move claim that a conceptual, not an ontological, contrast of things in themselves and appearances is needed to express Kant’s insights concerning the nature and scope of our knowledge. The concept of the thing in itself is interpreted as never referring, and as serving exclusively negative purposes. It informs us about how and what we do know, by telling us how and what we cannot know.
If this view is coherent, then, even with the existence of things in themselves subtracted from the ontological picture, transcendental idealism survives as the doctrine that all that can be known to exist are appearances. Jacobi is wrong that the existence of the thing in itself is required to uphold the objectivity of our perceptions.
One way of upholding Kant’s claim for the existence of things in themselves is provided by the ‘two conception’ view of things in themselves and appearances, to be explained later.
Things in themselves and appearances
Appearances of things in themselves
In considering our situation transcendentally, we discover that appearances are merely ideal, as well as being grounded in things in themselves, there is a sense in which appearances carry a reference beyond themselves. There is thus a sense in which appearances are, in addition to being grounded on things in themselves, appearances of things in themselves.
Two objects versus two conceptions
Does Kant envisage things in themselves and appearances as composing two worlds, in the sense of two ontologically distinct sets of objects? Or is the language of things in themselves and appearances an expression of two points of view on a single set of objects?
The concept of thing in itself is intended to refer to the same objects that we know as appearances, but under a non-empirical description. This two concept view grants the concept of the things in itself a positive, referring use.
Two object view is saying that appearances and the thing in itself are two ontologically distinct objects. Two object theorists may claim that Kant talks as if things in themselves and appearances were two aspects of one thing, only in order to bring out their incompatibility and thereby underline the necessity of distinguishing them ontologically.
Those who take the two object view typically (though not exclusively) go on to charge transcendental idealism with incoherence, on account of the difficulties. Consequently, those who defend Kant’s doctrine of things in themselves tend to be proponents of the two conception view.
There is nothing at issue between the two views from the perspective of a finite cognitive subject engaged in transcendental reflection: aside from special contexts like human freedom, it is a matter of indifference whether one says that there is one world conceived in two ways, or two worlds.
The transcendental ideality of the self
Kant, as we have seen, holds that transcendental ideality extends to the self as much as it does to things distinct from it.
There are two problems, first, it would seem that by Kant’s own account we have much knowledge of a determinate but non-empirical kind about the self, suggesting that here at last we must regard ourselves as penetrating to the nature of something as it is in itself. There are two sense of self, self qua transcendental subjectivity or condition of objects, and the self qua thing in itself or noumenon. Both are non-empirical, but it is to the former that Kant means to ascribe a priori synthesis. Ultimately, all it amounts to is the fact that the subject has priority in the subject-object relation, that the supreme condition of this relation is itself subjective.
The second problem arises because of the undeniable deep differences between inner and outer sense. Inner sense has no manifold of its own, and it’s “material” consists only of outer objects. This makes it hard to see how an ideality-implying story may be told about inner sense.
Kant also mentioned that ideality of self is confirmed in the second Critique: even the thinking subject is in inner intuition a mere appearance to itself, gets its full confirmation in the Critique of Practical Reason, and that so thoroughly that one would have to arrive at it even if the former had never proved this proposition at all.
The self in the system of experience
Progress can be made by considering the place of the self in Kant’s theory of experience as a whole. Although the transcendental ideality of the self does not follow from self-affection merely qua action of the understanding, it does follow from the objectification of the self in the empirical order which that action entails. Confining self-knowledge to the self qua appearance is an essential concomitant of Kant’s Copernican programme since otherwise it treats self as a privileged real object and the theory becomes a novel transcendental realism.
Entering into, and remaining within, the Kantian system
We have seen that Jacobi’s contention that the Kantian system incorporates a paradox is not justified. There is no contradiction in saying that the objects of our perception are inside us in the transcendental sense, and that they presuppose something outside us of which we can have negative and existential knowledge.
Kant’s picture appears paradoxical to Jacobi because he does not hold apart the two levels, empirical and transcendental, and consequently mistakes the relation between things in themselves and appearances for a relation within the empirical realm.
The peculiarity of our cognitive situation, according to Critical philosophy, is that we can grasp the perspective which conditions our knowledge only by referring to a point of view outside it, of which we can form a conception, but which we cannot occupy.
What at root separates transcendental idealism from its critics is thus a meta-philosophical difference. Transcendental realism holds that human knowledge can account for itself, and in so doing knows reality: it assumes that the fundamental conditions of human knowledge are identical with constituents of reality, and that there is nothing necessary for cognition that cannot itself be cognized.
Transcendental idealism, by contrast, holds that cognition is subject to conditions that cannot themselves be cognized in the same sense in which objects are cognized, and so that human knowledge can account for itself only by referring outside its sphere.
There is a difficulty in entering into the Kantian system, in so far as it contradicts the expectation that our cognitive situation will be elucidated in a non-perspectival manner. This difficulty is, however, acknowledged and explained within the system itself, in terms of the task of thinking the perspective that constitutes our identity as subjects. And because the difficulty is dissolved on the inside of the Kantian system, Jacobi should have found no problem in remaining within it.