Printer Friendly

How does Kant prove that we perceive, and not merely imagine, physical objects? *.


IN THE REFUTATION OF IDEALISM and his notes to it in the B Preface, Kant frames his antiskeptical issue in terms of proving "the reality of outer sense," which requires proving that we perceive, not merely imagine, physical objects in space and time. (1) Kant's contemporary critic Maimon reasserted the Humean objection, that the appearance of physical objects in space and time is a deceptive illusion produced by our imagination. (2) The same kind of objection is made today, for example, by Stroud, both to Kant and to his recent expositors. Maimon's objection rests on serious misunderstandings of Kant's analyses and proofs, (3) and I agree with Stroud and Rorty that recent "analytic transcendental arguments" fail to rebut (in effect) Maimon's objection. (4)

Asking how Kant proves that we perceive rather than merely imagine physical objects in space and time, presumes that Kant does prove this. This I affirm. Affirming this, however, does hot presume that Kant proved it in precisely the way he proposed. I contend that Kant's proof succeeds in ways, and to an extent, that even Kant did not appreciate. In part, this is because his proof need not appeal to transcendental idealism. Indeed, parts of Kant's proof refute his key arguments for transcendental idealism. This paper epitomizes the key steps in Kant's unofficial, though sound, transcendental proof for the conclusion of his Refutation of Idealism: namely, "The mere, though empirically determined consciousness of my own existence proves the existence of objects in space outside me." (5)


The first important point, neglected by recent reconstructions, concerns method. Recent "analytic transcendental arguments" are, of course, analytic; they attempt to justify substantive, antiskeptical conclusions by analyzing the possibility of self-conscious experience. Yet Kant stresses that no analytic argument can justify any synthetic proposition a priori. (6) If recent notions of philosophical analysis are more expansive than Kant's, they are not expansive enough to support Kant's Refutation of Idealism.

Furthermore, "analytic transcendental arguments" take as their analysandum the possibility of consciousness per se. All such arguments are refuted by Rorty's objection, that
 Arguments of the Strawsonian type rest on considerations of
 which words can be understood independently of which other words.
 The relevance of these considerations vanishes if we admit the
 possibility of a being who could experience something as an X but
 could not use the word 'X' nor any equivalent expression. (7)

However, the possibility of Rorty's imagined being would not perturb Kant at all. Kant expressly aims at identifying the transcendental conditions necessary for the possibility of human self-consciousness, and more particularly, the transcendental conditions required for us to be aware of our existence as determined in time, that is, to be aware of some events happening before, during and after others. (8) To do this, Kant engages us with a series of wildly counterfactual thought-experiments designed to bring us to recognize some of our key cognitive capacities, and their attendant incapacities. Appreciating and assessing these thought experiments involves "transcendental reflection." (9)


Kant's analysis of space and time stresses the spatio-temporality of our forms of intuition and our use of concepts of space and time in order, inter alia, to make an important semantic point about determinate reference, that is, reference to any one particular, single item. Kant stresses our incapacity to represent to ourselves the absence of space and rime. Nor can we perceive space or time as such, though of course we can conceive of their being void, or even of their being absent. (10) Kant's point concerns a key feature of the representational capacities of human beings--of out representational capacities. Whether other beings (for example, of the kind Rorty imagines) have different representational capacities is irrelevant to understanding human knowledge. The positive implications of Kant's observations about our spatio-temporal representational capacities concerns an important semantic and cognitive insight that undergirds Kant's insistence on the distinction between, and the interdependence of, sensibility and understanding in human knowledge of the world.

One key semantic point, recognized by Kant, is that definite descriptions do not suffice for knowledge of particulars. Putative definite descriptions aren't self-identifying: they don't intrinsically reveal whether they are empty, uniquely satisfied, or ambiguous. Any reasonably specific description (sans token indexicals, covert or overt) may be satisfied by nothing or by several things. Specificity of description cannot guarantee particularity of reference. Whether a description is empty, definite, or ambiguous depends equally on the contents of the world. For human beings, the only way to pick out spatio-temporal particulars is by sensing them (directly or indirectly). For human beings, singular cognitive reference requires singular sensory presentation. Semantic reference to particulars requires token indexicals in some form, which can play their role in human cognition only in perceptual circumstances (which can include observational instruments). Perceptual circumstances, for human beings, are spatio-temporal circumstances. Identifying spatio-temporal particulars by sensing them involves, in part, identifying at least approximately the spatio-temporal regions they occupy. (11)

Out ineliminable recourse to spatio-temporal specification is reflected by recent analyses of the "character" of demonstrative terms, where such terms can be used or understood only by understanding the speaker-centered spatio-temporal reference frame they implicitly presuppose. (12) Conversely, for us, singular cognitive reference also re quires predication, namely, ascribing some at least approximately identified characteristics to any particular we sense, within some at least approximately determined spatio-temporal region. Moreover, predication and spatio-temporal determination are interdependent. (13) Kant's account of the conjoint cognitive functioning of human sensibility and understanding reaches this same conclusion (below, section 9).


Kant's proof succeeds to a greater extent than even he appreciated because it provides two sound, genuinely transcendental arguments for (not "from") mental content externalism. The first of these arguments turns on the following considerations: Kant's "formal" idealism requires that the matter of experience be given to us ab extra. This is a transcendental material condition of self-conscious experience. (14) Another transcendental material condition of self-conscious experience is the transcendental affinity of the sensory manifold. (15) Kant notes that any world in which human beings are capable of self-conscious experience is one that must provide us with a certain minimal, recognizable degree of regularity and variety among the contents of our sensations. In any world lacking this minimum degree of regularity and variety, we could make no judgments. Therefore, we could not identify objects or events; therefore, we could not distinguish ourselves from them; and therefore, we could not be self-conscious.

In this connection Kant argues that the fact that a complete sensibility and understanding are capable of associating perceptions does not of itself determine whether it is possible for appearances or perceptions to be associated. (16) If they are not associable, there may be fleeting episodes of empirical consciousness (that is, random sensations), but there could be no integrated, and hence no self-conscious, experience. In part this would be because those irregular sensations would afford no basis for developing empirical concepts nor for using categorial concepts to judge objects. (There could be no schematism and hence no use of Kant's categories in a world of utterly chaotic sensations.) In this regard, the necessity of the associability of the sensory manifold is a conditional necessity, holding between that manifold and any self-conscious human subject. Necessarily, if a human subject is self-consciously aware of an object (or event) via a manifold of sensory intuition, then the content of that manifold is associable. The associability of this content is its "affinity." Because it is necessary for the possibility of self-conscious experience, such affinity is transcendental.

Kant makes the transcendental status of this issue plainest in the following passage, though here he speaks of a "logical law of genera" instead of the "transcendental affinity of the sensory manifold":
 If among the appearances offering themselves to us there were
 such a great variety--I will not say of form (for they might be
 similar to one another in that) but of content, that is, regarding
 the manifoldness of existing beings--that even the most acute human
 understanding, through comparison of one with another, could not
 detect the least similarity (a case which can at least be thought),
 then the logical law of genera would hot obtain at all, no concept
 of a genus, nor any other universal concept, indeed no
 understanding at all would obtain, since the understanding has to
 do with such concepts. The logical principle of genera therefore
 presupposes a transcendental [principle of genera] if it is to be
 applied to nature (by which I here understand only objects that
 are given to us). According to that [latter] principle, sameness of
 kind is necessarily presupposed in the manifold of a possible
 experience (even though we cannot determine its degree a priori),
 because without it no empirical concepts and hence no experience
 would be possible. (17)

Despite Kant's shift in terminology, it is plain that the condition that satisfies the "logical law of genera" at this fundamental level is the very same as that which satisfies the "transcendental affinity of the sensory manifold." In the extreme case suggested here by Kant, where there is no humanly detectable regularities or variety within the contents of our sensory experience--call this "transcendental chaos"--there could be no human thought, and so no human self-conscious, at all. (18) Kant establishes this necessary transcendental condition for self-conscious human experience by identifying a key cognitive incapacity of ours: our inability to be self-conscious, even to think, even to generate or employ concepts, in a world of transcendental chaos. We can recognize Kant's insight only by carefully considering the radically counter-factual case he confronts us with: by recognizing how utterly incapacitating transcendental chaos would be for our own thought, experience, and self-consciousness.

This transcendental proof establishes a conditionally necessary constraint on the sensory contents provided to us by whatever we experience. Below a certain (a priori indeterminable) degree of regularity and variety among the contents of our sensations, our understanding cannot make judgments; consequently under that condition we cannot be self-conscious. (Above this minimal level of regularity and variety, there is then a reflective issue about the extent to which our experience of the world can be systematized.)

This condition is peculiar because it is both transcendental and formal, and yet neither conceptual nor intuitive, but rather material. The transcendental affinity of the sensory manifold is transcendental because it is a necessary a priori condition of the possibility of self-conscious experience. It is formal because it concerns the orderliness of the matter or content of sensation. However, ultimately it is satisfied neither by Kant's a priori intuitive conditions of experience, space and rime as forms of human intuition, nor by the a priori conceptual conditions of experience, our categories of judgment. As Kant twice acknowledges, its satisfaction is due to the "content" or the "object" of experience. (19) Because the matter of sensation is given us ab extra, we do hot and cannot generate it. Consequently, we also cannot generate or otherwise insure any degree of regularity or variety among the contents of our sensations. The contents of our sensations, along with their recognizable similarities and differences, must be given to us by something other than ourselves. Consequently, this is a genuinely transcendental argument for mental content externalism: we cannot be self-consciously aware of any purported "mental" contents without being aware of at least some "mental" contents that concern and derive from something other than and outside of us. (20)


Kant's semantic point about determinate cognitive reference, and his first proof of mental content externalism, are reinforced and angmented by his proof that we can only make legitimate causal judgments about spatio-temporal substances. This argument provides a second, stronger transcendental proof of mental content externalism. It proceeds in two steps: The Paralogisms of Rational Psychology prove that we cannot make any legitimate causal judgments about merely temporal objects or events, while the Analogies of Experience prove that we can make legitimate causal judgments only about spatio-temporal substances.

Kant contends that causality is strictly related to substance. (21) Kant argues in the Paralogisms (in both editions) against out knowledge of a substantial self, and he argues that in psychology we have no evidence of any extended substance. (22) If we have no evidence of a substantial self, then none of us can use any of the Principles of the Analogies to make judgments about ourselves. Thus we cannot justify any determinate causal judgments in psychology because we cannot identify any causally active substance(s) within the sole form of inner sense, namely time.

The main target of the Paralogisms, to be sure, is traditional rationalist psychology, (23) but even when stating this, Kant indicates an empirical aspect of his criticism: the concept of a simple nature cannot be a predicate in an objectively valid experiential judgment. (24) Kant quickly elaborates the empirical aspect of his criticism by criticizing any empirical use of the category of substance regarding either oneself or one's psychological states: the only empirically usable concept of substance is the permanence of an object given in experience, but no such permanence can be demonstrated in the case of the "I." (25) Kant argues that there can be no synthetic a priori principles about the soul at all, of any kind. Any rational doctrine of the soul, whether a priori or empirical, purports to make synthetic judgments a priori. Such judgments require intuitions as a judgmental connecting link, but there are no suitable intuitions to be round in inner experience, (26) because we intuit nothing permanent or abiding in inner sense. (27) Consequently, rational psychology is hot a doctrine but a discipline limiting our cognitive aspirations. (28)


An important though neglected feature of Kant's analysis of legitimate causal judgments is that we can only make such judgments about spatio-temporal substances. The importance of being able to identify "permanent" or abiding substances, that is, objects or events that persist through changes, and why we can only identify such substances within space and rime conjointly, are made evident by a widely neglected feature of Kant's Analogies of Experience. Kant's three Analogies form a tightly integrated set of mutually supporting principles, each of which can be used only together with the other two. (29) The First Analogy treats the persistence of substance through changes of state (transformations); the Second Analogy treats only rule-governed causal processes within any one substance; and itis the Third Analogy alone which treats causal interaction between any two (or more) substances. Kant is express about this. (30) Hence only with the Third Analogy does Kant respond directly to Hume's skepticism about our knowledge of causal powers, because only there does he defend a transeunt accourir of causality, the view that something in a causally active substance goes out beyond that substance to influence or causally affect something else, that is, to effect a change in a distinct substance--in brief, the thesis that all physical events have external causes. (31) Despite the complexities of these issues, the main point Kant makes about the necessarily joint use of the three principles of the Analogies can be summarized briefly.

Determining that we witness either coexistence or succession requires discriminating the one from the other, and both determinations require that we identify objects that persist through both the real and the apparent changes involved in the sequence of appearances we witness. We cannot directly perceive or ascertain either time or space (above, section 2), and the mere order in which we apprehend appearances does hot determine an objective order of objects or events. Consequently, given our cognitive capacities, we can determine which states of affairs precede which others, and which states coexist, only under the condition that we identify enduring substances that interact and thus produce changes of state in one another. The existence of identifying enduring substances is necessary for us to determine the variety of spatial locations that objects or events occupy, to determine changes of place (both local and translational motion), and to determine the nonspatial changes (transformations) objects undergo. To make any one such identification requires discriminating the present case from its possible alternatives, which requires conjoint use of all three principles defended in the Analogies. Failing to employ these principles successfully would leave us, as Kant says in the A Deduction, with "nothing but a blind play of representations, that is, less than a dream." (32)

That Kant is correct about these important theses can be seen by recalling Hume's perplexities in "Of Scepticism with Regard to the Senses," (33) and certain facts Kant notes about the requirements for our distinguishing the subjective order of apprehension from the objective order of events. Kant notes that apprehending the manifold features of a house is successive, although the features of the house exist concurrently. (34) Hume concurs, for when a porter delivered him a letter, he recognized that the porter climbed stairs that must still exist beyond the bounds of Hume's study, and that the door to his study must still exist behind his back, if he heard the porter's knock and the door's squeaky hinge as the porter entered. (35) The implications of Hume's observations are manifold. (36)

Note first that these observations acknowledge that we ascribe both perduring existence and causal properties to ordinary physical objects. Second, these ascriptions require concepts that cannot be defined in accord with Hume's own concept-empiricism, namely, the concepts "cause" and "physical object." (37) These concepts are thus a priori. In view of the widespread recent rejection of conceptempiricism, it is worth noting that Kant's analysis shows that these are very special a priori concepts, because their legitimate use is required for anyone to be self-conscious at all, and so to learn, define, or acquire any concept that requires experience for its meaning or acquisition. In brief, Kant's Categories count as what may be called pure a priori concepts.

Third, Hume notes that ascribing continued existence and causal properties to physical objects outstrips out sensory observations, as Hume understands them. (38) Nevertheless, ascribing these characteristics to physical objects is necessary in order to preserve the coherence of our beliefs about the world. (39) Hume finds such "coherence" too weak to justify trusting his senses. (40) Hume overlooked what Kant saw, namely, that the coherence of our beliefs about our surroundings is only the tip of the issue. At stake is their very existence, their very possibility. (41) Without the capacity to make causal judgments we could never "derive" (as Kant says) the subjective order of apprehension from the objective order of the world, (42) nor could we distinguish between our subjective order of apprehension and any objective order of things and the events in which they participate, (43) including those events called "perceiving" them. We could hot identify sensed objects at all, not even putatively; we could not identify the door on the basis of its squeak, nor could we identify ourselves as being aware of the door on the basis of its squeak. In practice, Hume clearly distinguished the subjective order in which his experiences occurred from the objective causal order of objects and events that gave rise to his experiences, though his epistemology cannot account for this ability. Kant's transcendental proofs concern hot merely the possession of certain concepts but their use in legitimate (that is, true and justified) cognitive judgments of these sorts. (44) (In this regard, motions of our own bodies alter our perspectives in ways required to distinguish the objective order of events from the subjective order of apprehension, as noted in Kant's example of viewing a house. (45)) Because we can use the categories of cause and substance only with regard to spatial objects and events, and because we can identify a temporal order of events only by correctly using the concepts of cause and substance, by which alone we can distinguish the subjective order of apprehension from the objective order of events, the objective order of events we identify must be a causal order of perceptible spatio-temporal substances.


I have alluded to the transcendental character of Kant's proofs of mental content externalism, and his proof that we can only make legitimate causal judgments about spatial substances (above, sections 4-6), without having yet sufficiently explained their transcendental character. Their transcendental character concerns their being formal cognitive conditions for the very possibility of self-conscious human experience, which can be known a priori, and from which follow other a priori knowledge. (46) Both of these features stem from the fundamental roles of such judgments in our self-ascription of our own experiences. Famously, Kant argues that each of us must be able to identify our representations as our own, "for otherwise I would have as multicolored, diverse a self as I have representations of which I am conscious." (47) Kant's term verschieden ("diverse") connotes either qualitative or quantitative distinctness. While not as strong, say, as verteilten ("distributed") Kant uses it here to contrast with the analytic unity of apperception, to emphasise the lack of such unity in the indicated circumstance, in which we would have, at most, only flickering moments of sensory consciousness, though (Kant argues) no self-consciousness (see above, section 4). Beforehand he says this directly, using a stronger term: "For the empirical consciousness which accompanies diverse (verschiedene) representations is in itself dispersed (zerstreut) and without connection to the identity of the subject." (48)

At an utter minimum, Kant's point is that, because sensory representations are fleeting, their mere occurrence does not suffice for us to identify or to recall them as our own. Each sensory representation is at best only a momentary bit of consciousness, and can neither provide nor serve as a consciousness (much less a self-consciousness) of any plurality of sensory representations. Being able to recognize any plurality of sensory representations as one's own requires intellectual recognition of that plurality of representations, and of oneself as aware of them. The recognition of such representations as one's own requires intellectual judgment. Being able to recognize a plurality of representations as one's own is necessary for gaining any stable knowledge--or even stable beliefs--about what we experience. The analytic unity of apperception, expressed by the "I think," requires for its possibility the synthetic unity of apperception through which a plurality of sensations are integrated together and recognized as one's own. (49) The transcendental affinity of the sensory manifold, that is, a humanly detectable degree of regularity and variety among the contents of what we sense, is a minimum condition for the possibility of any synthetic unity of apperception. Transcendental chaos (above, section 4) blocks the analytic unity of apperception because it blocks the synthetic unity of apperception. Transcendental affinity of the sensory manifold is thus a minimal condition required for our understanding to function, to develop or use concepts at all.

Moreover, the relevant kind of recollection of our own sensory states requires more than that some current state be caused by some prior, putatively recollected state. It requires that our present recollection be, and manifestly be, of a prior state of one's own. Hume's causal account of memory fails to meet this requirement. (50) This kind of recollection is required both for the recognition of any stable object or of any process (whether motion or transformation) over any period of time, however short, as well as for the recognition of any personal history of experiences, however brief or long, however haphazard or integrated it may be. Kant's point is that the mere occurrence of a recollection-impression within a bundle or the mere inherence of a representational state, the object of which happens to be past, within a Cartesian mental substance, do not suffice--for beings like us--to identify those states as out own and so to be able to base cognitive judgments on them.

The thought experiment signalled here by Kant's "otherwise" (51) is to reflect on the implications of our only having fleeting episodes of empirical awareness, that is, sensations, or analogously Humean impressions, which would indeed enable us only to have "as multicolored, diverse a self as I have representations of which I am conscious." Reflecting on this wildly counterfactual state of affairs underscores and should support our endorsment of Kant's conclusions that the analytic unity of apperception is necessary for any empirically determinate self-conscious experience we enjoy, and that this analytic unity of apperception is rooted in the synthetic unity of apperception. Through this synthetic unity alone we can grasp various sensory representations as belonging together in the perception of any one object or event, and through it alone we can grasp various sensory perceptions of objects or events as belonging to our own first-person experience and its history. (52)

The fundamental role of this synthetic unity of apperception for the possibility of the occurrence of any analytic unity of apperception, for any instance of the "I think," is supported by Kant's proof of the transcendental affinity of the manifold of sensory intuition (above, section 4) and his proof that legitimate causal judgments can only be made about spatio-temporal substances (above, sections 5 and 6). If either of these conditions fails to be satisfied, no human "I think" could occur, because conditions required for any synthetic unity of apperception would not be satisfied, in which case no apperception, no analytic unity of self-consciousness, could occur either. Consequently, both of these conditions are genuinely transcendental. (53) The fact that making causal judgments requires being able to identify particular causally active substances in space thus provides a second, stronger proof of mental content externalism. Understanding why this is the case again raises issues central to Kant's semantics.


Kant's complex semantics are based on his Table of Judgments. Fortunately, Kant's completeness proof for the Table of Judgments has been brilliantly reconstructed by Wolff, (54) which enables us to reconsider Kant's semantics and Transcendental Deduction much more carefully than heretofore.

Kant holds that our pure a priori concepts, the categories, have a logical significance independent of their schematization. This logical significance, catalogued in the Table of Judgments, (55) is enriched with a transcendental significance by relating pure concepts to the sensible manifold provided by our forms of intuition. (56) This is the task of Kant's Metaphysical Deduction of the Categories. However, this transcendental significance of the categories does not suffice for determinate cognitive reference to particulars. Determinate cognitive reference to particulars also requires one of two further steps: either the categories must be "schematized" in order to refer them to possible sensory appearances, thereby obtaining singular determinate reference only in connection with singular sensory presentation of spatio-temporal particulars; or the unschematized Categories can be referred to particular moral agents by using various principles of Kant's practical philosophy. (This second kind of singular reference is not relevant to the present topic and shall not be discussed further here.)

Kant closely associates significance (Bedeutung), sense (Sinn), and even content (Inhalt) with a concept's "connection" (Beziehung) or reference to objects, where this referentiality is secured via our forms of sensory intuition. (57) Kant's account of "objective validity" requires that, for any concept to be fully meaningful, it must be referable to possible or actual objects of human experience, where such "referability" is secured spatio-temporally, via our spatio-temporal forms of intuition and singular sensory presentation. This component of Kant's theory of semantic meaning concerns referentiality, not "empirical content" as understood by various empiricist theories of meaning, to which Kant's views have been erroneously assimilated. (58)

Kant's semantics explicitly proscribes both empirical and synthetic a priori knowledge of particular objects beyond the bounds of sensory experience. (59) When Kant states that a "merely transcendental" use of categories is "in fact absolutely no use," his full statement indicates that this uselessness pertains to determinate judgments about particular objects; that is, the transcendental use affords neither empirical nor synthetic a priori knowledge of particular objects. This is clear from the specific context, though Kant himself added in his Nachtrage the further clarification that this use is no real use "to know something." (60) Similarly, he clarified the meaning of his statement that no object is determined in the absence of the condition of sensible intuition, by adding "thus nothing is known." (61) The transcendental use of pure categories affords no knowledge, neither empirical nor synthetic a priori, of particular objects. This is the "transcendental use" of pure concepts Kant repeatedly criticizes and repudiates in his Critique of Pure Reason; this is the nerve of his critique of rationalist metaphysical pretensions to knowledge. Conversely, Kant's semantics affords genuine cognitive significance only when concepts are "connected" or referred to particular objects via singular sensory presentation, and thus provides for singular cognitive reference.


The importance of and the relations among these key points--namely, the spatio-temporal character of our representational capacities (above, section 3); Kant's first transcendental proof of mental content externalism (section 4); the restriction of legitimate causal judgments to objects and events in space (sections 5 and 6); the way in which the "I think" presupposes a synthetic unity of apperception (section 7); and the role of singular sensory presentation in genuine cognitive significance (section 8)--all converge in Kant's claim that "Thought is the act of relating given intuition to an object." (62) How these key points bear on Kant's claim can be understood by considering Kant's account of sensations and perceptual synthesis.

Kant espoused a sophisticated version of sensationism. (63) On Kant's considered view, outer sensations are hot themselves objects of self-conscious awareness (except under highly unusual circumstances), although they are basic events or processes of sensing. In Kant's usage, sensation (Empfindung, and its cognates) indicates a corresponding object or a reality (Real, Wirklichkeit). (64) Kant's view can be put adverbially: we sense (for example) greenly; we do hot sense green, although we sense green features of objects or, less commonly, of colored light; we sense "the real" that corresponds to sensation. Sensations, or acts of sensing, are momentary; only series of sensations are temporally extended. (65) We can have self-conscious experience of any object or event only insofar as we integrate a plurality of sensations when perceiving that object (or event), and only insofar as we judgmentally identify and integrate several of its sensed features. (66) Only this integration and judgmental articulation enables us either to experience or to know any particular object or event, by enabling us to exploit information about it provided through sensation. The synthesis that brings about the referential and representational role of sensations is a function of the kinds of judgments we as human beings can make. (67) Only the categories, which derive their functions of unity from our twelve basic forms of judgment, can guide our judgmental integration of sensations in our experience or knowledge of any objects or events. (68)

Kant's doctrines of perceptual and judgmental "synthesis" clearly identify what is now called "the binding problem" in neurophysiology of perception. The "binding problem" is actually a set of problems regarding the proper coordination of sensory, perceptual, or cognitive information within our neuropsychological processes of cognition. (69) To the transcendental power of imagination, Kant ascribed the proper coordination of sensations into percepts of particular objects or events. To cognitive judgments of the understanding, Kant ascribed the proper coordination of our recognition of individual features, aspects, or characteristics into the recognition of any one particular object or event. (70) If contemporary neurophysiology ascribes more integrative functions to out sensory apparatus than Kant allowed, this does not detract from Kant's keen recognition of a genuine problem, widely neglected by advocates of the "new way of ideas" and of sense data. Moreover, it detracts nothing from Kant's identification of a crucial problem involved in our explicit cognitive recognition that any one object or event displays a particular set of characteristics.

Because we cannot perceive either space or time as such (above, section 3), we cannot group apparent sensory qualities into properties of objects simply by their apparent spatio-temporal coordinates. We can only identify the spatio-temporal region occupied by any particular object by recognizing the spatio-temporal array of objects and events before us. Doing this requires identifying those objects as causally interacting substances that determine one another's locations, motions, and transformations (per above, sections 5 and 6). In this way, Kant's analysis reaches Evans's conclusion, that predication and spatio-temporal localization are mutually interdependent. (71) To this Kant adds: both of these coordinated forms of identification are parasitic on the causal order of physical events in space and rime, on the basis of which alone we can distinguish our subjective order of experience from the objective order of events (above, sections 5 and 6). Only by distinguishing these can we identify objects or events at all, and only by identifying them can we identify ourselves both as distinct from them and as aware of them. Out empirically determined self-consciousness (above, section 1) is precisely our awareness of ourselves as being aware of some events occurring before, during, and after others. For reasons summarized herein, Kant is right that this form of self-consciousness is only possible for us human beings on the basis of our consciousness of objects outside us in space. Consequently, anyone who is self-conscious enough to follow this line of reasoning, or even simply to raise skeptical questions, can know a priori that we have at least some knowledge of physical objects in our environs by understanding Kant's proof.

Kant's proof of the "reality" of perception is a transcendental proof of unqualified realism about molar objects in our physical environs. It is not a proof of some transcendentally qualified, merely "empirical" realism. In part this is because Kant's main arguments for transcendental idealism assume rather than prove that the transcendental conditions for self-conscious human experience can be satisfied or fulfilled only if transcendental idealism is true. This assumption is refuted by Kant's own transcendental arguments for mental content externalism (sections 4-6, 9), because this argument shows both that transcendental conditions for the possibility of self-conscious human experience can be satisfied by mind-independent factors, and that the transcendental affinity of the sensory manifold can be satisfied only by a mind-independent factor, namely, the degree of orderliness among the manifold of given material of sensation. This proof thus provides a sound version of the "neglected alternative" objection to Kant's arguments, and it thus provides a model for the construction of such objections regarding the satisfaction of the other transcendental conditions Kant identifies.


Skeptics and advocati diaboli may retort that this is a nice story but hardly a proof. The issue thus raised requires appreciating what Kant's proof achieves and what can properly be expected of philosophical proof. One of the deepest errors of "analytic transcendental arguments" has been to assimilate Kant's analyses to the Cartesian predicament Kant decidedly rejected. Kant is the first great non-Cartesian epistemologist. He decisively rejected Cartesianism in several ways. First, Kant rejects the Cartesian assumption, shared by Hume, that runs through the entire sense-data tradition, namely, that states of sensory consciousness (sensations) are automatically also states of our self-consciousness. This assumption, when conjoined with infallibilist assumptions about epistemic justification (see below), inevitably leads to the ego-centric predicament of insoluble Cartesian skepticism.

Kant also rejected Cartesianism, secondly, by espousing certain forms of externalism, hot only regarding mental content (above, section 4) and causal judgment (sections 5, 6, 9), but also regarding epistemic justification. Kant's transcendental conditions for the possibility of self-conscious human experience need only be satisfied for any human to be self-conscious; no one needs to know that they are satisfied in order to be self-conscious, nor does anyone need to know that they are satisfied in order to understand or to use Kant's proof. On the contrary, transcendental proofs work due to the converse relation between their ratio essendi and ratio cognoscendi: the satisfaction of the transcendental conditions of the possibility of self-conscious human experience is the ratio essendi of self-conscious human experience. Once Kant's proof establishes this, then anyone's actual self-conscious experience is--and if one understands Kant's proof, it is also known to be--the ratio cognoscendi of there being perceptible, causally active physical objects in one's surroundings.

A third aspect of Kant's non-Cartesianism is his clear recognition that any tenable epistemology requires some substantive premises that cannot be proven by purely deductive means and do hot pass the test of Descartes's evil deceiver. This is why Kant's method of transcendental reflection involves our reflecting on some carefully chosen, wildly counterfactual circumstances in order to identify some of our key cognitive capacities and their attendant incapacities (above, sections 2 and 4).

Skeptics and advocati diaboli dismiss premises that do not meet infallibilist standards. Kant, however, recognized infallibilist models of epistemic justification as the skeptical trap and philosophical pipedream they are. He understood very well the failure of Descartes's effort to refute skepticism moto geometrico. To this I add: Descartes's argument is infected, not by one, but by five distinct vicious circularities. (72) Kant was right to develop a radically non-Cartesian approach to skepticism and to the philosophical analysis of our empirical knowledge. Not only does Kant advocate a fallibilist account of empirical knowledge, (73) he advocates a fallibilist account of transcendental knowledge as well: establishing the basic inventory of our human cognitive capacities and incapacities is a collective undertaking, requiring constructive mutual assessment. (74) Any form of justification based on constructive mutual assessment is inherently fallibilist, because we human beings are inherently fallible.

Kant's non-Cartesian insights did not prevent him from also trying to prove his antiskeptical conclusions "apodictically." (75) Kant's model for this was the traditional model of a rational science that deduces every conclusion from rational first principles (scientia), as exemplified by Christian Wolff. (76) To fulfill this deductivist model, Kant proposed to establish his transcendental account of human knowledge in the Critique of Pure Reason, which Kant understood to require transcendental idealism. Kant proposed that transcendental philosophy would establish both the legitimacy of and the parameters for properly scientific (wissenschaftliche) metaphysics, which he duly published as The Metaphysical Foundations of Natural Science ("Foundations" for short) and The Metaphysics of Morals. In turn, the Foundations were to establish the a priori principles required to ground and justify empirical physics. This is a grand philosophical vision. Having examined it very closely, I submit that no one could better carry out this vision than Kant did. (77) However, this aspect of Kant's epistemology ultimately serves to undermine its own deductivist model of "scientific" justification, and thus to reinforce the fallibilist model of justification embedded in Kant's method of transcendental reflection.

Very briefly, Kant's transcendental idealism and his foundational sequence of transcendental, metaphysical, and empirical principles fails to prove the key causal principle central to the Critique of Pure Reason, namely, that every event has a cause. The problem is that Kant's analysis in the Critique of Pure Reason expressly addresses only the general causal principle, that every event has a cause. However, the causal principle actually required by the Analogies of Experience is the specific causal thesis, that every physical event has an external physical cause (transeunt causality). Kant distinguished these two principles only in the Foundations, where he also recognized that this specific causal thesis cannot be proven on transcendental grounds alone, for it also requires metaphysics. With this, Kant's foundational order of philosophical priority is jeopardized. However, careful examination of Kant's proof of the specific causal principle in the Foundations reveals that his key premise rests not on metaphysical analysis, but on our empirical ignorance of any instances of hylozoism. With this, Kant's foundational order of philosophical priority is destroyed, as are the deductivist, "scientific" aspirations embodied in Kant's grand vision of "scientific" philosophy. Neither Kant's transcendental idealism nor his deductivist model of rational, scientific knowledge can prove apodictically the causal principle we need and use, namely, that every physical event has an external physical cause. Transcendental idealism provides no answer to Hume's causal skepticism.


Does the failure of Kant's deductivist model provide aid or comfort to skeptics? No. An extension of Kant's new method of transcendental reflection, along the lines recommended herein, provides sufficient proof of the specific causal principle. In part, this is due to Kant's semantics (above, section 8): we can use the general causal principle in connection with (in Beziehung auf) particular objects only in those cases where we can refer the specific causal principle to spatio-temporal objects. Once the distinction between these two causal principles is recognized, Kant's Transcendental Deduction and Analogies of Experience can be revised accordingly, in part by highlighting the fallibilist aspects of Kant's methods, to provide a genuinely transcendental proof of the conclusion of Kant's Refutation of Idealism. This proof is strongly reinforced by Kant's two transcendental proofs of mental content externalism (above, sections 4-6, 8, 9). Kant's fallibilism, along with the failure of both Descartes and Kant's own deductivist efforts, help show that the infallibilist assumptions involved in global perceptual skepticism are far from innocent or inevitable assumptions. Indeed, they are themselves a key roadblock to understanding our empirical knowledge.

Global perceptual skepticism challenges the "whole of our perceptual experience." In the Transcendental Dialectic, Kant points out that this putative "whole of perceptual experience" is itself no object of perceptual experience. (78) No wonder it cannot be justified by recourse to perception! Furthermore, as a "whole," this alleged "whole of perceptual experience" is at best an Idea, in Kant's technical sense. More precisely, it is inherently a transcendent Idea, to which we can give no objective validity. Furthermore, the skeptical "hypotheses" used to generate this alleged "whole of perceptual experience" are all designed in principle to be cognitively transcendent; in principle, they cannot be verified or refuted by any empirical evidence or inquiry. Consequently, they are "hypotheses" in name only, and radically distinct in kind from genuine, empirically usable hypotheses.

Kant's criticisms of these skeptical strategies are underscored by his semantics of cognitive reference (section 8), which entail that none of these skeptical hypotheses, nor the alleged "whole of perceptual experience," admit of any determinate reference to any particulars we can identify. Finally, Kant's fallibilism, together with his transcendental proof that we can be self-conscious of our existence as determined in time only if in fact we are aware of, and have some knowledge of, spatio-temporal, causally active substances in our surroundings, block the skeptical generalization from occasional perceptual error to the possibility of universal perceptual delusion. It does so by demonstrating that any world in which we are altogether perceptually deluded is a world in which no human being can be self-conscious. In any such world, no human being can raise skeptical doubts. So if we are alert enough to raise skeptical doubts, a close study of Kant's transcendental proof of realism suffices to allay those doubts permanently. Global perceptual skeptics simply assume that we can be self-conscious without being conscious of anything outside our minds. Kant's transcendental proof of realism shows just how portentous this assumption is. If Kant is right, global perceptual skepticism rests on profound, even willful self-ignorance: the question "What can I know?" (79) is indeed closely connected to the question, "What is it to be human?" (80)

University of East Anglia

Correspondence to:,

* This essay is dedicated to Burkhard Tuschling, wegfuhrender Philosoph.

(1) Immanuel Kant, Kritik der reinen Vernunft (Critique of Pure Reason), Bxl-xli n., B276-7 n. Kant's first Kritik appears in Kants Gesammelte Schriften (hereafter, "Ak"), 29 vols., Koniglich Preussische Akademie der Wissenschaften (Berlin: G. Reimer, 1902--). The numbers in citations to this work indicate volume:page.lines; for example, Ak 3:137.5-13. Others of Kant's works are cited by the initials of their German titles. All translations from Kant's works are mine; the pagination from Ak is provided in all recent translations of his works.

(2) Solomon Maimon, Gesammelte Werke, ed. Valerio Verra (Hildesheim: Olms, 1965), 5:377-8, 386.

(3) For example, his writings show no trace of Kant's key doctrine of the transcendental unity of apperception. I found none, and neither did Achim Engstler, Untersuchungen zum Idealismus Salomon Maimons (StuttgartBad Cannstadt: Frommann-Holzboog, 1990), 94-5, 122-3.

(4) Recent reconstructions fail to engage the cote of Kant's proof, because they focus on out possessing the concept "physical object," or on out using it, though without requiring ourjustified or even correct use of it. Strawson (The Bounds of Sense [London: Methuen, 1966]), Rorty ("Strawson's Objectivity Argument," Review of Metaphysics 24 [1970]: 222, 224; "Verificationism and Transcendental Arguments," Nous 5 [1971]: 3-14), and Stroud ("Transcendental Arguments and 'Epistemological Naturalism'," Philosophical Studies 31 [1977]: 106, 110) focus too much on concept possession, and specify their "application" too vaguely, to capture the character and point of Kant's transcendental proofs. Similarly, Bennett's "Objectivity Argument" focuses on the "application" of concepts in a way that reflects rather than rejects Hume's analysis in "Of Scepticism with Regard to the Senses" (A Treatise of Human Nature, ed. David Fate Norton and Mary J. Norton [Oxford: Oxford University Press, 2000], bk. 1, pt. 4, sec. 2) because in Bennett's argument their "application" does not require their correct (truthful) application. See Bennett, Kant's Analytic (Cambridge: Cambridge University Press, 1966), 202-14; and "Analytic Transcendental Arguments," in Transcendental Arguments and Science, ed. Peter Bieri, Rolf-Peter Horstmann, and Lorenz Krtiger (Dordrecht: Reidel, 1979), 52-5. On the inadequacy of "analytic transcendental arguments," see Thomas Grundmann, Analytische Transzendentalphilosophie. Eine Kritik (Paderborn: Schoningh, 1994); David Bell, "Transcendental Arguments and Non-Naturalistic Anti-Realism," in Transcendental Arguments: Problems and Prospects, ed. Robert Stern (Oxford: Oxford University Press, 1999), 189-210; and Kenneth R. Westphal, Kant's Transcendental Proof of Realism (Cambridge: Cambridge University Press, 2004), chap. 1.

(5) B275; Ak 3:191.18-20. The full analysis is provided in Westphal, Kant's Transcendental Proof

(6) B263-5, B810; Ak 3:184.26-185.19, 509.24-510.25. See Manfred Baum, Deduktion und Beweis in Kants Transzendentalphilosophie (Konigstein: Hain bei Athenaum, 1986), 1, 175-81.

(7) Rorty, "Strawson's Objectivity Argument," 224, compare 231.

(8) More specifically, Kant's analysis seeks the transcendental conditions that make self-consciousness possible for finite beings possessing spatio-temporal forms of intuition and a discursive understanding, though human beings are the only instance of such beings we know of.

(9) Westphal, Kant's Transcendental Proof, chap. 1.

(10) A19-20, 22-3, 31, 172-3, 188, 214, 487/B34, 37-8, 46, 207-8, 214, 231, 261,515.

(11) The cognitive insufficiency of descriptions theories of reference was Kant's point of departure for the whole Critique. See Arthur Melnick, Space, Time and Thought in Kant (Dordrecht: Kluwer, 1989), 1-5, 25-6.

(12) David Kaplan, "On Demonstratives," in Themes from Kaplan, ed. Joseph Almog et al. (New York: Oxford University Press, 1989), 481-563; John Perry, "The Problem of the Essential Indexical," Nous 13 (1979): 3-21; and Gareth Evans, The Varieties of Reference (Oxford: Oxford University Press, 1982), chap. 6.

(13) Gareth Evans, "Identity and Predication," Journal of Philosophy 72, no. 13 (1975): 343-63.

(14) Henry Allison, Kant's Transcendental Idealism: An Interpretation and Defense (New Haven: Yale University Press, 1983), 250.

(15) A112-4.

(16) A121-3.

(17) A653-4/B681-2; Ak 3:433.14-29 (emphases added).

(18) Kant's argument about this "Logical Law of Genera" closely parallels his argument about the transcendental affinity of the sensory manifold: both concern the recognizable orderliness of what we sense, and the constitutive necessity of that orderliness for the very functioning of out understanding. This functioning is required for any synthetic unity of apperception, and thus is required for any analytic unity of apperception, that is, for the occurrence of any human "I think." There is, however, a difference between Kant's two cases: The Logical Law of Genera concerns objects, while Kant usually states transcendental affinity of the manifold of sensory intuition in terms of the contents of sensations (but see A90-1/B122-3). Plainly, if the Logical Law of Genera is satisfied, so is the transcendental affinity of the manifold of sensory intuition. However, perhaps there could be transcendental affinity among the manifold of sensory intuition only to the extent that there were humanly detectable regularities and variety among sensory contents, without our being able to identify objects in nature. To this extent, the Logical Law of Genera is a stronger principle. The extent to which the satisfaction of these two principles could in fact diverge is difficult to determine. Kant claims that failure to satisfy either principle has the same consequence, namely, human understanding simply could not function. In that case, there could be no synthetic unity of apperception, and so no analytic unity of apperception, and so no self-consciousness of the form expressed by "I think" (B131-9). The difficult point is to determine whether human understanding could function while using only the categories of quality and quantity; judgments using these two categories could potentially be made in circumstances that satisfied the transcendental affinity of the manifold of sensory intuition, though not the Logical Law of Genera. Resolving this issue would require minute investigation of Kant's Transcendental Deduction. Fortunately, two central points suffice here. First, both principles, the transcendental affinity of the sensory manifold and the Logical Law of Genera, provide transcendental proofs of mental content externalism, though of slightly different kinds. In either case, this is a major anti-Cartesian result. Second, Kant's antiskeptical transcendental proof of realism sans phrase needn't appeal to the bare possibility of the analytic unity of apperception. It can appeal to the perhaps stronger, certainly more explicit premise of Kant's Refutation of Idealism, that we are aware of our own existence as empirically determined in time (B275).

The substantive difference between Kant's two principles can be clarified by considering why he uses two designations for what is at bottom the same principle. The transcendental affinity of the sensory manifold concerns the bare minimum level of regularity and variety among the contents of our sensations that is required to enable us to identify kinds or genera at all. Once satisfied, there is then a reflective issue, addressed by Kant's Transcendental Law of Genera, concerning the extent to which the kinds or genera we identify can be systematized. Thus, satisfying the transcendental affinity of the sensory manifold is a precondition for our generating any empirical intuitions at all, while the reflective issue addressed by Kant's Transcendental Law of Genera presumes that we have sufficiently coherent empirical intuitions to identify spatio-temporal objects or events, where we try to systematize the characteristics of them we have identified. This contrast, however, did not preclude Kant from highlighting, in the passage just quoted, the transcendental, constitutive issue of the affinity of the sensory manifold while explaining the status of his transcendental law of genera.

(19) A112-3, A653-4/B681-2.

(20) Thus, transcendental proofs can justify conclusions much stronger than Rorty recognizes; he claims that the most they can show are interrelations among thoughts. See "Strawson's Objectivity Argument," 236; "Verificationism and Transcendental Arguments," Nous 5 (1971): 3-14. Part of why Kant fails to recognize his own achievement in this regard is that the transcendental affinity of the sensory manifold is a formal, transcendental, though material condition for the very possibility of self-conscious experience. The architectonic of Kant's transcendental idealism does not provide for such conditions. For a concise discussion of such issues, see Kenneth R. Westphal, "Must the Transcendental Conditions for the Possibility of Experience be Ideal?" in Eredita Kantiane (1804-2004): questioni emergenti e problemi irrisolti, ed. Cinzia Ferrini (Naples: Bibliopolis, 2004), 107-26. Significantly, the later Wittgenstein makes the same case for mental content externalism, though without invoking Kant's specific cognitive psychology. See Kenneth R. Westphal, "Kant, Wittgenstein, and Transcendental Chaos," Philosophical Investigations 28, no. 4 (2005): 303-23.

(21) B183, A182-84/B225-7, A204/B249; Ak 3:137.30-138.4, 163.1-32, 176.19-20.

(22) A381, B291, B293-4.

(23) A342/B400, B405-6; Ak 3:263.16-20, 266.16-25.

(24) A361.

(25) A349-50.

(26) A398-9; Ak 4:248.28-249.11; B421-2; Ak 3:275.13-20.

(27) A366; Ak 4:230.18-28; compare A349-50, A361, A381, A398-9, A402-3; Ak 4:221.1-15, 227.21-8, 251.12-20; B420; Ak 3:274.15-24.

(28) B421; Ak 3:274.36-275.4; compare B420; Ak 3:274.24-6; KdU [section] 89, 5:460.20-32.

(29) Paul Guyer, Kant and the Claims of Knowledge (Cambridge: Cambridge University Press, 1987), 168, 212-14, 224-5, 228, 239, 246, 274-5; Westphal, Kant's Transcendental Proof, [subsection] 36-9.

(30) B111; KdU, Ak 5:181.

(31) I retain the archaic spelling of "transeunt" because the Oxford English Dictionary indicates it is used precisely and exclusively in the sense here indicated.

(32) A112, Ak 4:84.30-1.

(33) Hume, Treatise, bk. 1, pt. 4, sec. 2.

(34) A190/B236.

(35) Hume, Treatise, bk. 1, pt. 4, sec. 2, par. 20.

(36) For discussion of this section of Hume's Treatise, see Robert Paul Wolff, "Hume's Theory of Mental Activity," in Hume: A Collection of Critical Essays, ed. Vere Chappell (New York: Anchor, 1966), 99-128; Barry Stroud, Hume (London: Routledge, 1977), 96-117; Norman Kemp Smith, The Philosophy of David Hume (London: Macmillan, 1941), 443-94; and Kenneth R. Westphal, Hegel, Hume und die Identitat wahrnehmbarer Dinge (Frankfurt am Main: Klostermann, 1998), [section] 4.

(37) Regarding "cause," see B240-1 and Lewis White Beck, Essays on Kant and Hume (New Haven: Yale University Press, 1978), 121-9; regarding "physical object," see Hume, Treatise, bk. 1, pt. 4, sec. 2, pars. 23-8. Stroud claims that Hume's appeal to propensities can be eliminated by replacing such talk with conditional regularities about the occurrence of certain "perceptions" in the mind, given certain series of other perceptions (Hume, 131). However, at best this provides only occasioning causes of the use of the concept "body," but accounts neither for the definition nor the origin of that concept. Moltke Grain overlooks Hume's recognition of the shortcomings of general principles of psychological association in accounting either for our concepts of or our beliefs about causal relations among physical objects ("The Skeptical Attack on Substance: Kantian Answers," Midwest Studies in Philosophy 8 [1983]: 366). Rorty likewise overlooks the problems Hume found in his study ("Strawson's Objectivity Argument," 209). Hume awoke Kant from his dogmatic slumbers only because he rethought Hume's first Enquiry deeply enough to recognize its implications for causality and especially for physical objects, which Hume developed only in the Treatise (bk. 1, pt. 4, sec. 2). It behooves Kant's critics to study Hume with equal care.

(38) Hume, Treatise, bk. 1, pt. 4, sec. 2, pars. 20, 22, compare par. 56.

(39) Hume, Treatise, bk. 1, pt. 4 sec. 2, pars. 18-21.

(40) Hume, Treatise, bk. 1, pt. 4 sec. 2, par. 56.

(41) This central feature of Kant's transcendental proofs is omitted by Stephen Korner ("Zur Kantischen Begrtindung der Mathematik und der Naturwissenschaften," Kant-Studien 56, no. 3/4 [1966]: 463-73; and "The Impossibility of Transcendental Deductions," in Kant Studies Today, ed. Lewis White Beck [LaSalle, Ill.: Open Court, 1969], 230-44), by Jay Rosenberg ("Transcendental Arguments Revisited," Journal of Philosophy 75, no. 18 [1975]: 611-24; and "Transcendental Arguments and Pragmatic Epistemology," in Transcendental Arguments and Science, 245-62), and by Robert Stern ("On Kant's Response to Hume: The Second Analogy as Transcendental Argument," in Transcendental Arguments: Problems and Prospects, ed. Robert Stern [Oxford: Oxford University Press 1999], 47-66). It is noted, however, by the following: Quassim Cassam, "Transcendental Arguments, Transcendental Synthesis and Transcendental Idealism," Philosophical Quarterly 37, no. 149 (1987): 355; and Barry Stroud, "Kant and Skepticism," The Skeptical Tradition, ed. Myles Burnyeat (Berkeley: University of California Press, 1983), 429; and Stroud, "Kantian Argument, Conceptual Capacities, and Invulnerability," in Kant and Contemporary Epistemology, ed. Paolo Parrini (Dordrecht: Kluwer, 1994), 248.

(42) A193/B238.

(43) A193-5/B238-9.

(44) See above, n. 4.

(45) A192/B237-8.

(46) B25, 40.

(47) B134; Ak 3:110.7-9; compare A111-12.

(48) B133; Ak 3:109.16-18.

(49) B131-9.

(50) Stroud, Hume, 124-6, 135-40; Keith Yandell, Hume's "Inexplicable Mystery"." His Views on Religion (Philadelphia: Temple University Press, 1990), 108-10.

(51) B134.

(52) B131-6.

(53) It may appear that [section] 20 of Kant's Transcendental Deduction tries to establish conditions for the possibility of human self-consciousness that are independent of and prior to the conditions for the possibility of self-conscious human experience. [section] 20 focuses only on conceptual transcendental conditions, and it does not consider the material transcendental conditions that are latent in Kant's account, especially of transcendental affinity. [section] 20 considers Anschauungen, not Empfindungen. Any one Anschauung already integrates Csynthesizes") some plurality of sensory Empfindungen. Hence, if there is any given empirical Anschauung (as [section]20 requires), there must be transcendental affinity of the sensory manifold contained within that sensory intuition. [section] 20 argues, in brief, that empirical intuitions must stand under the categories, because we have no other functions of unity that could possibly guide the synthesis required by or for any one empirical intuition, because synthesizing sensations into an intuition likewise requires that those sensations exhibit transcendental affinity. In these ways, the Transcendental Deduction requires the broader issues highlighted here.

[section] 20 may appear to focus on out concept of "cause" rather than "substance." However, it treats the Categories en bloc, and so includes "substance" as much as "cause," and it refers back to [section] 19 (as it should), where an example of a substance--a body--is a key illustration of Kant's point.

I cannot enter further into the details of Kant's Transcendental Deduction here. Sec Baum, Deduktion; Melnick Space, Time and Thought; Pierre Keller, Kant and the Demands of Self-Consciousness (Cambridge: Cambridge University Press, 1998); and Robert Greenberg, Kant's Theory of A Priori Knowledge (State College: Pennsylvania State University Press, 2001). Important cautions are issued by Guyer, Claims, and "The Transcendental Deduction of the Categories," in The Cambridge Companion to Kant, ed. Paul Guyer (Cambridge: Cambridge University Press, 1992), 123-60; and by Robert Howell, Kant's Transcendental Deduction (Dordrecht: Kluwer, 1992). I believe the Deduction must be thoroughly rethought in light of Michael Wolffs brilliant work: Die Vollstandigkeit der kantischen Urteilstafel (Frankfurt am Main: Klostemmnn, 1995); "Erwidemng auf die Einwande von Ansgar Beckermann und Ulrich Nortmann," Zeitschrift fur philosophische Forschung 52, no. 3 (1998): 435-59; and "Nachtrag zu meiner Kontrovers mit Ulrich Nortmann," Zeischrift fur philosophische Forschung 54, no. 1 (2000): 86-94.

(54) Sec references to Wolff in the preceding note.

(55) A79, 147/B104-5, 186; Ak 3:92.16-19, 139.11-37.

(56) A76-7/B102, A147/B186, A248/B305, A254/B309; compare B148-9, A181/B224; Ak 3:91.2-13, 139.25-9, 208.16-29, 210.35-211.14; compare Ak 3:118.7-16, 161.27-31.

(57) B300.

(58) For example, by Strawson, Bounds of Sense, 16, and by Eric Sandberg, "Thinking Things in Themselves," in Proceedings of the Sixth International Kant Congress, ed. Gerhard Funke and Thomas Seebohm (Lanham: University Press of America, 1989), 2.2:23-31. On the semantic sense of Kant's term Beziehung, see Greenberg, Kant's Theory, 57-67, 69-71, 119 n. 17, 187-8; and Robert Hanna, Kant and the Foundations of Analytic Philosophy (Oxford: The Clarendon Press, 2002), 83-95, 136-7.

(59) A247-8/B304-5; Ak 3:207.23-208.15.

(60) Ak 23:48.16-17; compare Beno Erdmann, Nachtrdge zu Kants Kritik der reinen Vernunfl (Kiel: Lipsius & Fischer: 1881), no. cxxvii.

(61) Ak 33:48.14.

(62) A247/B304; Ak 3:207.23-4.

(63) Rolf George, "Kant's Sensationism," Synthese 47, no. 2 (1981): 229-55; compare William Harper, "Kant on Space, Empirical Realism, and the Foundations of Geometry," Topoi 3, no. 2 (1984): 143-61.

(64) B34, 74, 182, 207, 209, 609, 751; A20, 166, 373-4.

(65) B209.

(66) George reminds us that in contemporaneous philosophical usage, Kant's related term "Erkenntnis" (in the distributed singular) designates cognitive reference to a particular object or event. See "Sensationism"; compare Kant's taxonomy of representations (A319-20/B376-7).

(67) Regarding "synthesis," see Baum, Deduktion; Paul Guyer, "Psychology and the Transcendental Deduction," in Kant's Transcendental Deductions, ed. Eckart Forster (Stanford: Stanford University Press, 1989), 47-68; Patricia Kitcher, Kant's Transcendental Psychology (Oxford: Oxford University Press, 1990); Howell, Kant's Transcendental Deduction; Andrew Brook, Kant and the Mind (Cambridge: Cambridge University Press, 1994); and Jay Rosenberg, Accessing Kant: A Relaxed Introduction to Kant's Critique of Pure Reason (New York: Oxford University Press, 2005).

(68) See, for example, Allison, Kant's Transcendental Idealism, 115-22, 173-94; Herbert J. Paton, Kant's Metaphysic of Experience, 2 vols. (London: George Allen & Unwin; New York: Humanities Press, 1936), 1:245-8, 260-2, 304-5, 2:21-4, 31-2, 42-65, 68-9; Hanna, Kant and the Foundations, 76-83; Wolff, Urteilstafel, 58-73; Greenberg, Kant's Theory, 137-57; and J. Michael Young, "Functions of Thought and the Synthesis of Intuitions," in The Cambridge Companion to Kant, 112-13.

(69) Adina Roskies, "The Binding Problem," Neuron 24 (1999): 7-125. This set of problems has only recently received attention from contemporary epistemologists; see The Unity of Consciousness: Binding, Integration, and Association, ed. Axel Cleermans (Oxford: Oxford University Press, 2003).

(70) A79/B105-06, B152, B162 n.

(71) Evans, "Identity and Predication."

(72) Kenneth R. Westphal, "Sextus Empiricus Contra Rene Descartes," Philosophy Research Archives 13 (1987-88): 91-128.

(73) A766-7/B794-5.

(74) Onora O'Neill, "Vindicating Reason," in The Cambridge Companion to Kant, 280-308.

(75) Bxxii, 765.

(76) Bxxxvi.

(77) Westphal, Kant's Transcendental Proof, chaps. 4-6.

(78) A483-4/B511-2.

(79) A805/B833.

(80) A805/B833, Logik, Ak 9:25. I am grateful to Robert Greenberg, Robert Howell, and Jeffrey Edwards for very helpful comments on earlier drafts of this paper.
COPYRIGHT 2006 Philosophy Education Society, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2006 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Westphal, Kenneth R.
Publication:The Review of Metaphysics
Geographic Code:4EUGE
Date:Jun 1, 2006
Previous Article:Self-consciousness and intersubjectivity.
Next Article:Immediacy and mediation in Schleiermacher's Reden uber die Religion.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters