Kant's Transcendental Deduction

In this impressive effort, Henry E. Allison not only provides a very extensive treatment of the Transcendental Deduction (TD) in both editions of the Critique of Pure Reason, but also examines what he regards as all relevant background writings by Kant, including material from the 1760s, the Dissertation, the notes from the “silent decade” and the Prolegomena. His careful treatment of these supporting texts should be useful not only to those concerned with the TD but also for scholars with wider interests in Kant's outlook.For the most part, Allison conducts his inquiries in Kant's own terminology, offering clarifying explanations and critical revisions. Despite the apparent psychological nature of most of the main notions in the TD, such as apperception, synthesis, and the faculties, he continues to distance himself from psychologically oriented approaches. And although he seems to think that what he is presenting should be relevant to contemporary philosophy, no attempt is made at relating these notions to either current philosophy of mind or cognitive science.The reader should also be aware that this work presupposes some familiarity with Allison's previous writings on the first Critique.Allison continues to maintain that Kant's view that knowledge involves not only sensory experiences but also the application of concepts, which he calls “the discursivity thesis,” creates an epistemological difficulty that the TD is meant to resolve. He takes the text at A89–90/B122 to be a definitive characterization of this problem, where Kant states that the categories “do not represent to us the conditions under which objects are given to us in intuitions.” We thus have the issue of “how subjective conditions of thinking could have objective validity.” Allison refers to this throughout in an almost sardonic manner as “the specter that Kant is seeking to exorcise.”In examining section 2 of the A version, he detects a central role for the notion of representation, specifically, the concept of an object in general, which he, correctly I believe, understands as consisting of the categories. He reads Kant as holding that the categories are second-order concepts that are “a set of a priori rules through which representations are related to an empirical object” (232). Allison thinks that the categories are required because the representation of an object as a particular kind presupposes that it “also be recognized as an empirical object tout court” (231).Allison takes section 3 of the A TD to involve the presentation of Kant's “arguments in a systematic form” (243). He focuses on the second part, the argument from below, but is unable to find a sustained line of reasoning. He later characterizes this material as involving the intent “to introduce the reader, step by step, to a radically new philosophical project” (433).The B TD is interpreted as containing a different strategy, which consists of a single proof in two steps. The first part seeks to establish that any representation that is part of the objective unity of apperception is “related to [the concept of] an object and thus necessarily stands under the categories” (329). The second part is read as an attempt at establishing that the categories are also conditions of nondiscursive experiences, which would connect them with the a priori forms of space and time. Here, Allison continues to maintain that the term object has two meanings in this material, expressing “intentional object” in the first part and “ordinary (physical) object” in the second part. While he seems to think that the first step is successful, he finds that the second only achieves its goal for the first six (“mathematical”) categories; he suggests that argumentation from either the Prolegomena or from the Analytic of Principles might complete the task.Underlying Allison's reading is a novel idea that forms a running theme, namely, that what Kant is trying to establish in the TD is “a normative demand on every cognizer” that “imposes a conceptual constraint on how the human understanding must unify the data provided by the senses [via the categories] . . . if it is to relate them to an object” (113). These epistemic conditions are thought to be required to apply to every empirical judgment, thus achieving the supposed goal of prohibiting lack of agreement between sensibility and the understanding. This view is also an attempt to explain what sort of necessity Kant is seeking to establish for the categories, since, as Allison rightly asserts, it is not logical necessity.The primarily textual basis that Allison uses for this reading is the passage in the Prolegomena that equates objective validity and necessary universality while discussing the grounds of judgments of experience versus judgments of perception. It is also associated with Kant's view that concepts are rules, which, according to Allison, means that the rules “involve a kind of epistemic right to expect that whatever objects fall under the rule contain the properties that are expressed in the concept-rule and to require that other cognizers, who possess the concept, will agree with one's judgment” (237). And he sees this idea of a normative demand in a passage in the A TD that we will consider shortly.However, Allison does not convincingly show that the TD involves this notion. There is no explicit textual basis in the Critique for the idea of a normatively necessary epistemic demand, let alone one connected with the grounds of judgment. Tellingly, when examining the introductory sections of the TD, Allison acknowledges that Kant's example of necessary causation that is used to illustrate the need for a deduction does not reflect the normative necessity that is supposedly at issue (237). Nor does this notion actually play much of a role in Allison's treatment of Kant's reasoning. He attempts to remedy this absence by reminding the reader of normative demands whenever Kant mentions objective validity. But that just makes it seem superfluous. Additionally, his final summary (435–6) reveals that he does not manage to directly connect normative necessity with the key issue of representation.Similarly, Allison is unable to connect the above characterization of concepts, which is also textually unsupported, with the TD's main reasoning. To achieve that, as I have explained elsewhere, we should instead examine Kant's explanation that concepts are unifying functions. The application of a concept to an experience unifies that experience with any experiences that share the features specified by the concept. This is a necessary rather than a contingent unity—conceptualization is not association. Allison mentions this “analytic unity,” as Kant calls it (258), but not does not appreciate its role in the argumentation; I believe that it works as follows: In order to produce such unity, via recognition of multiple instances of a concept, the experienced instances must be part of the same, thus necessarily unified, consciousness (A103). Since we can arbitrarily apply concepts to experiences, consciousness as a whole must be necessarily unified. So there must be ur-unifiers—nonempirical, thus a priori, concepts/functions that unite consciousness, namely, the categories. And, since we are aware of everything in consciousness, there must be a priori awareness of this unity, hence, transcendental apperception.Turning now to the passage that Allison reads as an avocation of a normative demand, we find Kant, after having examined the unity of consciousness, now considering representation. He states that “the relation of all cognition to its object carries something of a necessity with it . . . since insofar as they are to relate to an object our cognitions must necessarily agree with each other in relation to it, i.e., they must have that unity that constitutes the concept of an object” (A104). I suggest that this necessity is not normative, but conceptual. If we take this to be about the concept of having an object, that is, representing something, and also assume, as Kant seems to, that all conscious states represent, then, since they are united under the concept “representation(al),” they must share features that make all of them representational. What, then, could all conscious states necessarily have in common? Several lines later, Kant tells us that “the unity that the object makes necessary can be nothing other than the formal unity of consciousness in the synthesis of the manifold of the representations” (A105). So he is claiming that the (“formal”) functions/concepts that unite consciousness and thus allow for the general application of concepts are also what make conscious states representational—that give them objects, i.e., references. This is to say that unification by the categories constitutes mental representation.If this can be sustained, then we know, a priori, that anything that we are able to represent, thus any physical object, must accord with the categories—a tidy objective deduction. It follows, contrary to empiricism, that our perceptual systems must include processing (“synthesis”) that builds this unity into our perceptual representations (“intuitions”). A secondary project thus emerges, namely, to provide broad psychological explanations, guided by philosophical argumentation, to outline how nonempirical (“pure”) perceptual operations produce the requisite unity. This removes the worry about possible incompatibility between the faculties, reinforcing the objective deduction. Many of the passages that Allison focuses on in the A edition, specifically part 3, as well as the second portion of the B version, seem to be concerned with this psychological (“subjective”) deduction. So, like many other commentators, he is attempting to wring epistemological results out of material that is primarily psychologically oriented.Along with an unsupported normative reading and focus on the wrong passages, there is also a foundational problem that hampers Allison's investigation. While he does not make it explicit here, he continues to impose a perceptual direct realist view on Kant. Although Allison, correctly I think, determines that the notion of representation is at the heart of the TD, he is prevented from grasping the reasoning I have sketched, since, for the direct realist, perception is not representational. We thus find him wondering why showing that “by means of [the categories] an object can be thought” (A97) should be a difficult task (205). By contrast, in reading Kant as a perceptual representationalist we can appreciate why the question of how we are generally able to represent the world arises. And, as I have indicated, this allows for a much simpler and clearer understanding of Kant's main line of argumentation, which does not seek to demonstrate agreement between the understanding and sensibility, but rather provides a sweeping a priori justification of the validity of the categories for anything we represent.So, while I find many lesser points of agreement along with some new insights in Allison's careful treatment of Kant's texts, his overall reading of the TD appears to be mistaken

Analytic philosophy · Calculus (dental) · Contemporary philosophy · Epistemology · General interest · Transcendental number · Mathematics · Medicine · Philosophical Ethics and Theory · Philosophy