Kant's transcendental functionalism.
Separating Kant's epistemology and metaphysics from his psychology has a long tradition, which can be traced back to neo Kantianism, (6) but this antipsychological interpretation has become particularly influential since Peter Strawson called for an abandonment of "the imaginary subject of transcendental psychology" in favor of "the strictly analytical argument" for the necessary structure of objective experience. (7) By contrast, functionalist interpretations attempt to revive the "subjective" side of Kant's Critique, at the cost of surrendering certain central tenets of transcendental idealism. Yet, I think both Strawson and the functionalist interpreters fail to do justice to Kant's strategy to reveal the necessary structure of cognition and reality by analyzing the faculty of cognition (Erkenntnisvermogen). The way that Kant characterizes cognitive faculties such as sensibility, understanding, and apperception does give prima facie evidence for a functionalist interpretation. The problem is how a functionalist interpretation can be brought into line with Kant's overall position of transcendental idealism. In this paper I will develop an interpretation called transcendental functionalism, which explains Kant's way to account for the objective structure of reality based on a theory of necessary cognitive functions.
While ordinary functionalism of the mind describes mental states in terms of their functional roles, transcendental functionalism determines what functions the mind has to realize if it is to be capable of objective cognition. It is not directly about the human cognitive system, neither at the phenomenal nor the noumenal level, but about an abstract functional structure that Kant refers to as the transcendental subject. Accordingly, Kant's cognitive psychology does not aim directly at factual knowledge of the human mind's workings, but rather, by analyzing the concept and structure of objective cognition, to reveal the cognitive functions that are necessary for all potential cognizers including, but not limited to, human beings. It is a theory of rational cognition as such, which at the same time accounts for the basic structure of reality that can be cognitively accessible to any finite rational agents.
This paper begins with an analysis of Kant's faculty of cognition, followed by an examination of the apparent conflicts of the ordinary functionalist interpretations with Kant's transcendental idealism (I). Kant's faculty of cognition represents a theoretical construct of the transcendental subject, but if human beings can be qualified as rational cognizers, they must have somehow realized the functional constraints that define the transcendental subject (II). It is important for a coherent interpretation of Kant to distinguish between two ways of viewing the faculty of cognition: it can either be considered transcendentally as a set of abstract functional constraints that are valid to all potential cognizers, or empirically as a particular realization by a complicated system of mental operations that take place in time (III). Only the former belongs properly to Kant's project of grounding the necessary structure of cognition and reality, although Kant did discuss empirical cognitive functions in the Critique, especially in the subjective deduction of categories in the A-edition. The transcendental-functionalist framework developed in this paper offers a better explanation of Kant's distinction between the subjective and the objective deductions of categories, as well as the reasons for Kant to adopt another strategy in the B-deduction (IV).
Kant's Faculty of Cognition. Kant's faculty of cognition consists of several distinctive powers: the basic faculties of sensibility, understanding, and reason, with imagination and the power of judgment as mediating powers among them. " Kant defines each faculty by its specific functional role. Meerbote calls Kant's theory "a cognitive psychology" or "a faculty psychology which speaks of capacities and abilities of various sorts which are needed for empirical cognition." (9) Consider the two constitutive faculties for empirical cognition: sensibility refers to "the receptivity of our mind to receive representations insofar as it is affected in some way" and understanding to "the faculty for bringing forth representations itself, or the spontaneity of cognition." (10) Sensibility and understanding are recognized as two indispensable and mutually irreducible stems of human cognition, because each plays a distinctive and indispensable function in making cognition. Similarly, other cognitive faculties such as imagination, the power of judgment, and reason are conceived in terms of their respective functional role in cognition, and they are unified as a cognitive system by functions such as synthesis and apperception. These functional descriptions not only frame Kant's view of the cognitive mind, but also determine the basic tenets of his epistemology and metaphysics. However, problems arise immediately when one attempts to locate the faculty of cognition in Kant's dichotomy of noumena and phenomena.
Kant's epistemology grounds the objectivity of cognition at the cost of the incognizability of things in themselves. Our cognitive forms determine the structure of objects only insofar as they are appearances whose reality depends, in a certain sense, on our faculty of cognition. The problem, however, is whether the cognitive faculty itself belongs to the phenomenal or the noumenal mind. Either way seems to lead to a dead end. Let us begin with the noumenal interpretation. Kant's Critique often gives the impression of describing a system of hidden psychological structures and processes that construct appearance out of the manifold given from things in themselves. Sensibility and understanding could not be faculties of the phenomenal mind because the latter is itself an appearance existing in the spatiotemporal world. The faculty of cognition would thus refer to the functions and operations of the noumenal self that synthesize the formless manifold into the spatiotemporally ordered appearance.
The difficulty of this noumenal interpretation is apparent. Kant rules out the possibility of cognizing things in themselves, and the noumenal incognizability principle applies to the cognizer itself just as much as to external objects. Even the cognizer can be cognized only as an empirical object. Kant rejects the Cartesian presupposition that a thinking subject has an immediate and indubitable cognitive access to itself. Even though Kant's transcendental deduction grants a pure representation "I think" to accompany all representations of a thinking subject, he considers it fallacious to infer from the transcendental apperception any rational knowledge about the thinker. In order to make legitimate applications of categories such as substance and causality, the sensible conditions of their applications must be fulfilled because the categories would otherwise have no objective meaning. If the noumenal interpretation maintains that things in themselves affect the noumenal mind to produce appearances, this description already involves an illegitimate employment of the category of causality. Causal properties cannot be meaningfully employed to describe noumena and their relations, and thus things in themselves cannot be said to exert any causal influence on the sensibility. Even if Kant allows the application of an unschematized category of causality beyond the empirical realm, it would mean something different from the proper causal relation that applies only to temporal events. It makes no sense to talk about synthetic operations or cognitive processes of the noumenal mind, since operations and processes are temporal-causal concepts that are applicable only to phenomena.
If cognitive powers such as sensibility, understanding, and synthesis were to describe functions and operations of the noumenal mind, the Analytic would be describing nontemporal and noncausal features of something that is shown to be theoretically incognizable by the Analytic. (11) Thus, Kitcher's verdict seems inevitable:
If the phenomenal-noumenal distinction is exclusive and exhaustive, then transcendental psychology must be about the phenomenal self, and so empirical, for the straightforward reason that no positive doctrines can be noumenal. Although it provides only a highly abstract, functional description of a thinking self, the description is still positive. (12)
Proponents of the functionalist interpretation view the cognitive functions and operations discussed in the Critique as features of the phenomenal mind, which itself exists in the spatiotemporal world. They tend to value Kant as an important forerunner of cognitive science and attempt to rediscover Kant's insights into the structure and workings of the human cognitive system. However, this phenomenal interpretation surely has not been acceptable to Kant. It blurs fundamental differences between his philosophical inquiry and empirical scientific studies of the mind. If Kant's aim was to search for a priori conditions for the possibility of objective cognition, an empirical study of the human mind could not serve this aim, since it would at best discover factual knowledge of the structure and workings of the human cognitive system. (13)
As the phenomenal mind is part of the empirical world, the development of its characteristics is governed by the same set of natural laws that determine the evolution of all organisms. Human beings are equipped with a highly sophisticated cognitive apparatus that receives and processes information in an intelligent way, but this intelligence system is the result of a long process of biological evolution, in which certain cognitive functions have been naturally selected for their adaptive advantages. Human beings have evolved the five senses of sight, hearing, smell, taste, and touch, but this is merely a contingent fact. We could well have developed other senses such as the ability to perceive ultrasonic signals as bats do. Knowledge about the structure and characteristics of the senses is unavoidably a posteriori and contingent.
If cognitive functions such as sensibility and understanding merely describe factual characteristics of human beings, then they would be contingent features of a biological species. Our possibilities of perceiving the world are certainly constrained, to a certain extent, by the existing senses, but they are not, as sensibility and understanding, supposed to be the a priori conditions that account for the possibility of objective cognition. Kant was well aware of the difference. (14) He pointed out that we cannot account for the necessity of causal laws by the fact that our understanding always synthesizes representations according to causal schemas, even if these schemas turn out, as a matter of fact, to be universal. Factual knowledge about the information-processing mechanism of the human mind cannot explain the objectivity of human cognition and even less the structure of empirical reality. Kant himself explains the crucial point as follows:
If ... the categories were ... subjective predispositions for thinking, implanted in us along with our existence by our author in such a way that their use would agree exactly with the laws of nature along which experience runs (a kind of preformation-system of pure reason), then ... in such a case the categories would lack the necessity that is essential to their concept. For, e.g., the concept of cause, which asserts the necessity of a consequent under a presupposed condition, would be false if it rested only on a subjective necessity, arbitrarily implanted in us, of combining certain empirical representations according to such a rule of relation. (15)
In short, the cognitive structures that form an account for the possibility of objective cognition cannot be contingent features of the phenomenal mind.
A more fundamental problem is that the phenomenal interpretation fails to acknowledge that the conditions of the possibility of appearance cannot themselves be an appearance. (16) Kant's conclusions cannot be achieved by describing mental processes or operations of the phenomenal mind, since all such processes or operations take place in time. The formal conditions that account for empirical reality cannot themselves be temporal processes nor anything determinable in time. Time is the form that has to be accounted for and cannot be taken for granted. It does not make sense to explain why things must be arranged in a temporal order by describing processes in which the temporal order has already been assumed. Simply put, a temporal entity or existence cannot constitute a condition of temporality; or in Kant's words, "the subject, in which the representation of time originally has its ground, cannot thereby determine its own existence in time." (17) Taking Kant's faculty of cognition as a system of cognitive processes or operations that are executed by the phenomenal mind, or the human brain, would mistake Kant's project for "a certain physiology of the human understanding (by the famous Locke)." (18)
As a functionalist interpreter who identifies Kant's psychology with the phenomenal mind, Kitcher is aware of the conflict with Kant's metaphysics, but she chose to bite the bullet, giving up Kant's metaphysical position and the ideality of time in particular:
The various activities that are described in the Deduction's account of how the mind influences (or might influence) what we know can only be understood temporally. They are processes and so take time. According to the Aesthetic, however, the mind's activities produce time. So they cannot take place in time.... Under these circumstances I see no choice but to reject the metaphysical claim. (19)
Kant aimed to account for the objectivity of cognition by grounding the structure of empirical reality in the cognitive subject. If the form of time could not be attributed a priori to the subject, then the temporalized schemas and the synthetic principles of experience would be devoid of the objectivity that Kant attempted to establish. The ideality of time is so crucial to Kant that the whole Copemican revolution would stand or fall with it. Rejecting the ideality of time would be, as Kitcher admits, "a drastic move within Kant's system." (20) Functionalist interpreters accept the phenomenal interpretation in order to revive important insights from Kant's psychology, but they do so at the cost of abandoning central tenets of Kant's epistemology and metaphysics. Yet, this cost, as I will show, is not necessary.
The Transcendental Subject and Its Empirical Realizations. Kant has a systematic picture of the human mind, which is divided into the faculties of cognition, desire, and the feelings of pleasure and displeasure according to their different functions. These three basic faculties correspond roughly to the themes of the First, Second, and Third Critiques respectively, with a comprehensive discussion of the human mind being presented in the Anthropology from a Pragmatic Point of View. Even if the forms of cognition and reality depend on the structure of the human mind, their dependence is only restricted to part of the mental powers. There is no necessity that the human mind must consist of exactly these three basic faculties, and the possibility of cognition does not presuppose the tripartite faculty either. (21) Even within the cognitive powers there are many human features, such as the five senses, that are irrelevant to the conditions of the possibility of objective cognition. If sensibility and understanding are two indispensable roots of cognition, they must be necessary in a relevant sense. The five senses belong to factual but contingent features of the human cognitive system, while the faculties of sensibility, understanding, and reason cannot be mere factual characteristics if they are the very conditions that not only make objective cognition possible but even confer objectivity to empirical reality. The former types of cognitive functions describe features of the human being as a biological species and thus belong to the empirical psychology that is discussed by Kant in the Anthropology, whereas the latter constitutes a peculiar Kantian discipline in the first Critique that can be called transcendental cognitive psychology.
While empirical psychology describes factual characteristics of the human mind as a phenomenal being in the spatiotemporal reality, transcendental cognitive psychology does not refer to any concrete entity, neither the phenomenal nor, if any, the noumenal mind, but to an abstract entity which represents a logical or conceptual unity of the functional requirements for anything that can be qualified as a cognizer. Kant refers to it as the "I, or He, or It (the thing), which thinks" or "a transcendental subject of thoughts = x." (22) The transcendental subject, as I have argued in another paper, (23) is not anything to which the category of existence can be applied, as it is only a theoretical construct as the hypothetical bearer of the faculty of cognition. Transcendental cognitive psychology is thus a theory about an "imaginary subject" that is equipped with all the necessary functions for objective cognition. Yet, if there is anything in the world that is capable of cognition, it must be a particular realization of the transcendental subject. Analyses of cognitive faculties such as sensibility and understanding are not straightforward descriptions of the human mind, but rather descriptions of the functional constraints that apply to all potential cognizers, including, but not limited to, humans. Robert Hanna is right in saying that "Kant's transcendental psychology is a non-naturalistic a priori theory of any actual or possible mind possessing a unified system of innate cognitive capacities just like ours--whether that creature happens to be biologically human or not." (24)
In criticizing the psychological interpretation of Kant's transcendental deduction, Paul Guyer makes a similar point: "Kant's premise is not a psychological claim but a basic constraint on any system for synthesizing data that are only given over time." (25) The constraints are not general facts about human psychology but conceptual truths or functional requirements for any potential cognizer, including, as Guyer suggests, even computers or forms of artificial intelligence. (26) Although Kant often addresses philosophical problems from a human perspective, claiming that all basic philosophical questions relate to the question "what is man?", (27) his epistemology and metaphysics cannot be understood as anthropological. Kant's faculty of cognition represents an abstract model of cognitive functions that are necessary for any cognizer. It can be compared to the description of a Turing machine, which is an abstract device to represent mechanical procedures or algorithms. The Turing machine does not refer to anything existent in the real world, although it is presented as a physical device that manipulates symbols on a strip of tape according to a table of rules. It is only a theoretical construct, but every concrete computer can be viewed as a particular realization of it. Similarly, Kant's transcendental cognitive psychology specifies the necessary functional constraints for all potential cognizers, of which the human cognitive system is just a particular realization. If there are Martian scientists, they must somehow, in their own way, have materialized the same set of functions.
Strawson was not mistaken in calling Kant's transcendental psychology an "imaginary subject," but he failed to recognize the crucial importance of this theoretical construct to Kant's epistemology and metaphysics. If the functional requirement of the transcendental subject is valid to all potential cognizers, then the required functions are not contingent human features but necessary conditions of the possibility of cognition. They are necessary in the sense that if any creature is qualified as a cognizer, it must have adequately realized them. It is possible that the world has never produced any cognizer at all. This can be simply because no creature has ever evolved a cognitive apparatus that fulfills the necessary functional requirements, although the world could have been cognizable if any creature had developed an adequate cognitive apparatus. However, the failure to produce cognitive agents could lie more fundamentally in the constitution of the world itself. If the world were organized in such a way that the spatiotemporal and categorial forms of cognition could never find appropriate applications, then Kant would refer to this as the case in which objects did not conform to our cognition. (28)
There is no metaphysical necessity that the world must be organized in a way that matches with the forms described by the transcendental cognitive psychology. The world is not necessarily cognizable, but Kant's Copemican revolution amounts to showing that necessarily, if the world is cognizable, then it must conform to the transcendental cognitive functions. It is, as Guyer points out, a conditional, relative, or hypothetical necessity instead of an unconditional and absolute one. (29) Kant does not start from nowhere; instead he takes certain basic features of cognition for granted and searches for the necessary conditions of their possibility. The project of transcendental cognitive psychology is to determine the necessary functions that make cognition possible. If human beings turn out to be qualified cognizers, this implies not only that the human cognitive system has sufficiently realized the required functions, but also that the empirical world in which human beings live is in fact organized in such a way that the spatiotemporal and categorial forms of our cognition can legitimately be applied to objects. Kant's psychology is thus essentially connected with his epistemology and metaphysics.
However, it is to be noted that the cognitive powers discussed in the Critique can mean two different things. In the first place, they refer to the necessary cognitive functions that are ascribed to the transcendental subject and account for the forms of cognition and reality, but they can also describe the particular realizations of functional constraints in human beings. In Kant's terms, they can be considered either transcendentally or empirically. This twofold perspective is explained by Kant most clearly in the following passage in reference to the three basic cognitive powers:
The possibility of an experience in general and cognition of its objects rest on three subjective sources of cognition: sense, imagination, and apperception-, each of these can be considered empirically, namely in application to given appearances, but they are also elements or foundations a priori that make this empirical use itself possible. Sense represents the appearances empirically in perception, the imagination in association (and reproduction), and apperception in the empirical consciousness of the identity of these reproductive representations with the appearances through which they were given, hence in recognition. (30)
Here Kant uses the concepts of sense and apperception instead of sensibility and understanding, but let us set aside the terminological issue. What is important is Kant's distinction between two ways to consider the three basic cognitive powers. Sense/sensibility, imagination, and apperception/understanding can be considered, on the one hand, transcendentally as a priori conditions of cognition and, on the other hand, empirically in application to given appearances.
The empirical and transcendental aspects of a cognitive power share the same functional characteristics. Kant calls a cognitive power empirical when it is considered "in application to given appearances," that is, when it is no longer an abstract function but a real mental process of the phenomenal mind that processes information in a specific way. The transcendental functions serve as a foundation for their empirical counterparts, in such a way that
pure intuition ... grounds the totality of perception a priori; the pure synthesis of the imagination grounds association a priori; and pure apperception, i.e., the thoroughgoing identity of oneself in all possible representations, grounds empirical consciousness a priori. (31)
What Kant calls the a priori grounding relation is, in my transcendental-functionalist interpretation, the concrete realization of a transcendental cognitive power by its empirical counterpart. While a transcendental cognitive power stipulates abstract conceptual requirements, its empirical counterpart realizes the functional constraints in appearances within the spatiotemporal and causal framework of empirical reality.
Transcendental versus Empirical Cognitive Psychology. Failing to distinguish transcendental cognitive functions from their empirical realizations, functionalist interpreters such as Kitcher mistake what is described in the Critique for the temporal-causal processes performed by the phenomenal mind--a mistake that leads them to reject Kant's theory of the ideality of time. However, the fact that Kant's transcendental cognitive functions cannot be analyzed in temporal-causal terms has also been brought up as a challenge to functionalist interpretations of Kant. As far as transcendental cognitive functions are concerned, Matt McCormick is right in saying that "[f]or Kant, our account of the processes of mind is not and cannot be a causal account of the relationships between mental states since causal ordering of events is a result of cognitive processing." (32) McCormick argues that Kant cannot be a functionalist because functionalism is built upon what Kant aims to explain, as it "presupposes a causal mechanism of some sort and seeks out the exact arrangement of causally related components." (33) It is true that many forms of functionalism in the contemporary philosophy of mind define mental states in virtue of their causal relationships, but this is not an essential feature of functionalism. Functionalism is not bound to define mental states in causal terms; the input-output relationships can be characterized abstractly, for example, in terms of machine tables. 31 The fact that Kant's faculty of cognition cannot be defined in causal terms does not speak against a functionalist interpretation of Kant; it rather indicates that Kant's functionalism is of a peculiar sort.
Certainly, for Kant, the cognitive functions that account for the objectivity of temporal-causal relations cannot themselves be defined in proper causal terms; instead, they must be defined by abstract logical concepts. Kant does allow a nonempirical, preschematized use of the concept of causality, which is the pure category of causality signifying a logical relation between ground and grounded. (35) Admittedly, this use of the concept of causality sometimes creates the wrong impression that Kant conceived the necessary cognitive functions as temporal processes or causal operations, but a careful distinction between transcendental and empirical cognitive psychology can help clear up the confusion.
The difference between transcendental and empirical cognitive powers can be illustrated by Kant's idea of the schematism of categories. Pure categories are fundamental concepts of understanding which determine the basic forms of thought and do not involve the temporal content of sensibility, but in order to become applicable to empirical objects, categories must be transformed into temporal schemas. The purely logical content of a category is enriched with temporal determination to make it homogeneous to sensible intuitions. In the category of causality, for example, the logical relation between conditions and conditioned is schematized to become applicable to temporal relations between successive events as cause and effect. A logical order becomes a rule-governed temporal order. Similarly, cognitive faculties taken transcendentally are not temporal-causal processes but purely conceptual determinations that describe the functional constraints for all finite cognizers.
These functions can be realized by different cognizers in multiple ways, but if time is a universal form of intuition, not restricted to the human sensibility, then the different ways of realization must share the form of temporality, fulfilling the functional requirements by specific temporal processes and causal operations. Thus, the relation between transcendental cognitive faculties and their direct empirical realizations resembles the relation between pure categories and their respective temporal schemas. Transcendental cognitive faculties such as intuition, imagination, and apperception are formal-functional structures, while their empirical realizations in perception, association, and recognition are cognitive mechanisms of the mind that process information in a temporal-causal fashion. Detailed mechanisms of empirical realizations could vary significantly across different species of cognizers, but the basic structures in any empirical cognizer are determined by their transcendental grounds. (36)
Take sensibility and understanding as examples. They are two independent and indispensable faculties of cognition because cognition depends on both the functions of referring to objects and applying concepts to them. Kant defines the function of intuition as the capacity to "relate to objects." (37) Taken transcendentally, it does not refer to any particular causal process of the mind through which objects are perceived, but merely a logical relation that stands between a mind and its objects. The Aesthetics argues that the possibility of relating to objects has to rely on the spatiotemporal forms of intuition. This amounts to saying that relating to any particular object means, in its most fundamental way, the possibility of identifying or locating it within a spatiotemporal framework. This logical relation must be realized by causal mechanisms through which the phenomenal mind can become related to the object by being affected by it. The empirical realization of intuition involves psychological processes that Kant characterizes as sensation: "The effect of an object on the capacity for representation, insofar as we are affected by it, is sensation. That intuition which is related to the object through sensation is called empirical." (38) Empirical or sensible intuition is a causal process of affection, that is, "real affection" (39) that takes place between empirical objects and a phenomenal mind, whereas intuition as such is an abstract relation that is not described in causal terms. (40)
The above analysis also applies to the faculty of understanding, which consists basically of the function of using concepts in judgments. The Analytic may be seen principally as explaining the conditions for concept application, thereby arguing for the universal validity of categories. The most fundamental condition in this regard is what Kant calls the transcendental, pure, and original apperception, that is, the self-consciousness that produces "the representation I think, which must be able to accompany all others and which in all consciousness is one and the same." (41) The transcendental apperception represents the ultimate logical unity of a transcendental subject, to which all representations of a cognizer must be ascribable. It is thus "the thoroughgoing identity of oneself in all possible representations." (42) However, this transcendental unity or identity is not anything that a cognizer can be introspectively conscious of because it is neither given in empirical intuition nor determinable in time. In the phenomenal mind, there must be an empirical apperception which realizes the logical requirement of unity through particular psychological processes. As the schematized counterpart of a logical unity, empirical apperception is nothing but an identity over time. The transcendental apperception of an abstract cognizer must be realized by concrete cognitive mechanisms through which an empirical cognizer in spatiotemporal reality can determine its own identity despite its changing internal states over time. In this sense, the logical unity is "schematized" into an empirical identity over time. (43)
The empirical identity that belongs to a thinker exists in spatiotemporal reality, while the transcendental or logical unity is only an abstract condition for every potential thinker. The transcendental apperception "I think" is not a factual description of my thinking activity but merely a "logical function," (44) which serves as a unifying condition for all representations that can be attributed to a thinking subject. In other words, the pure "I think" describes neither the fact that I am thinking, nor even anything about me. I can certainly use the statement "I think" to describe the fact that I am now thinking. However, as Kant explains, "the proposition 'I think,' insofar as it says only that I exist thinking, is not a merely logical function, but rather determines the subject (which is then at the same time an object) in regard to existence, and this cannot take place without inner sense." (45) There are two different senses in which the phrase "I think" can be used. Kant considers ordinary usages of the proposition "I think," including the Cartesian cogito, as an empirical statement that describes the mental state of an existing "I," but in Kant's transcendental cognitive psychology, the "I think" is taken only problematically. (46) Taking the "I think" problematically means engaging in the analysis of the abstract structure of a hypothetical entity. Although this imaginary transcendental subject does not exist anywhere, it constitutes the condition that accounts for the structure of empirical existence. (47)
Subjective versus Objective Deduction. A significant difficulty in understanding Kant's Critique is that the transcendental and empirical aspects of cognitive functions are not always clearly separated, and the failure to draw a consistent distinction has led to confusions, even by Kant himself. Kant was not always consistent in distinguishing transcendental from empirical cognitive functions. The problem is most obvious in Kant's attempt to prove the objective validity of categories. Kant almost completely rewrote the chapter on the transcendental deduction of categories in the B-edition. He was certainly unsatisfied with the A-deduction in certain aspects, for otherwise he would not have troubled himself with a new version. Nevertheless, even in the A-edition, Kant also expressed reservations about the deduction. In the A-Preface, Kant draws a distinction between objective and subjective deductions, a distinction which was dropped in the B-edition. This distinction has caused considerable controversy among Kant scholars, and the transcendental-functionalist framework outlined above can offer a better explanation of Kant's distinction as well as the reason for a fundamental revision in the B-deduction. Let us start with Kant's remark on the deduction in the A-preface:
This inquiry [the deduction], which goes rather deep, has two sides. One side refers to the objects of the pure understanding, and is supposed to demonstrate and make comprehensible the objective validity of its concepts a priori; thus it belongs essentially to my ends. The other side deals with the pure understanding itself, concerning its possibility and the powers of cognition on which it itself rests; thus it considers it in a subjective relation, and although this exposition is of great importance in respect of my chief end, it does not belong essentially to it.... In view of this I must remind the reader in advance that even in case my subjective deduction does not produce the complete conviction that I expect, the objective deduction that is my primary concern would come into its full strength. (48)
Kant has not stated explicitly which parts of the deduction form the subjective and objective sides respectively, but Kant scholars generally agree in identifying the subjective deduction with the second section of the A-deduction (49) and the objective deduction with the third section. (50) The subjective deduction is mainly about the doctrine of threefold synthesis. (51) The fact that Kant discards a detailed analysis of the threefold synthesis in the B-deduction is consistent with the reservations he expressed in the A-preface. (52) However, it has been widely debated as to why Kant considers the subjective deduction inessential to his chief end and how the subjective deduction differs from and is related to the objective one.
Kant values the objective over the subjective deduction because the former deals with the objects of cognition, aiming to establish the objective validity of categories, whereas the latter investigates the subjective sources of cognition. Brook summarizes the difference as follows:
Since the objective deduction is about the conditions of representations having objects, a better name for it might have been, "deduction of the object". Similarly, a better name for the subjective deduction might have been "the deduction of the subject" or "the deduction of the subject's nature." (53)
Accordingly, the objective and the subjective deductions would represent two quite different types of inquiry: the former that deals with objects is an analytic, epistemological project, while the latter that deals with the subject's nature is psychological or generative in character. The subjective deduction can offer us insights into the construction of the human mind, but it may not help explain the necessary conditions of objective cognition. This interpretation is shared by many Kant scholars including Strawson and Kitcher, (54) although, as mentioned above, Strawson sees Kant's major contribution in the analytic project, whereas the functionalist interpreters tend to emphasize Kant's insights into human psychology.
This standard interpretation is problematic for a number of reasons. Above all, it does not do justice to Kant's general strategy to account for the objectivity of cognition. The basic idea of Kant's Copemican revolution consists in the novel attempt to ground the necessary forms of objective cognition on conditions that are ascribed to the cognitive subject. The major task of the deduction of categories is to show "how subjective conditions of thinking should have objective validity, i.e., yield conditions of the possibility of all cognition of objects." (55) The deduction could never be successful without dealing with "the pure understanding itself, concerning its possibility and the powers of cognition," (56) which is supposed to belong to the subjective deduction. Although a detailed discussion of the threefold synthesis is discarded in the B-deduction, the doctrine itself is still maintained in the B-edition. (57) In fact, Kant's analysis of the subject's nature is not confined to the subjective deduction; even the objective sides of the A-deduction and the B-deduction also revolve around the subjective faculties of sensibility, imagination, and understanding as well as the activities of synthesis, judgment, and apperception. The deduction is, in a sense, unavoidably psychological. (58) The difference between objective and subjective deductions cannot, as Brook formulated, be reduced to a difference between deductions of the object or the subject. It must lie, rather, in their different ways to deal with the subjective conditions of objective cognition.
According to Norman Kemp Smith, the objective and the subjective deductions focus on different aspects of experience in the following ways:
In the subjective deduction experience is chiefly viewed as a temporal process in which the given falls apart into successive events, which, in and by themselves, are incapable of constituting a unified consciousness. The fundamental characteristic of human experience, from this point of view, is that it is serial in character. Though it is an apprehension of time, it is itself also a process in time. In the objective deduction, on the other hand, the time element is much less prominent. Awareness of objects is the subject-matter to which analysis is chiefly devoted. (59)
For Kemp Smith, the distinctive feature of the subjective deduction is temporality in a double respect: it accounts for our consciousness of time by analyzing mental processes that are themselves temporal in character. The objective deduction, by contrast, addresses our consciousness of objects. Yet Kemp Smith suggests that not only is the subjective deduction, despite Kant's reservation, essential to the transcendental project, but the consciousness of time is even more fundamental and indubitable than the consciousness of objects. (60) Kemp Smith thus considers the A-deduction superior to the B-deduction and rejects the neo-Kantian (and thus the Strawsonian) attempt to separate Kant's epistemology from all psychological considerations. (61)
Kemp Smith correctly identifies the key characteristic of the subjective deduction, but he fails to explain the complex situation, that is, why Kant on the one hand is convinced of his theory of the threefold synthesis, but on the other hand concedes that it does not belong essentially to his chief end. The framework of transcendental functionalism offers a more reasonable explanation: Kant had to discard the subjective deduction in the B-edition, not because it deals with the cognitive system of the subject, but rather because it treats the cognitive system as a series of temporal operations. Although both the objective and the B-deductions also deal with the subject's faculty of cognition, they treat it as a formal-functional structure instead of a series of temporal processes. As explained above, describing temporal processes cannot account for the temporal character of empirical consciousness. Surely, every individual cognizer finally realizes the necessary cognitive functions by mental operations that take place in time, but a story about the empirically realized functions does not belong properly to the transcendental account of cognition. Now, Kant's ambivalence toward the subjective deduction is due to the subtle characteristics of its content: Although mental or psychological operations do not constitute conditions that account for the objectivity of categories, a sufficiently abstract description of temporal processes is not just empirically or contingently true of human beings, but a priori true of any potential cognizer, provided that time is a universal form of intuition.
In the previous section, I compared the relation of transcendental cognitive functions to their empirical counterparts with the relation of pure categories to their temporal schemas. Yet, it has to be noted that empirical realization can be analyzed at different levels of abstraction, involving different degrees of empirical contingency. The faculty of sensibility, for example, is realized in human beings through a complex sensory system. Kant divides sensibility into inner and outer senses, and the latter further into the vital and the five organic senses of touch, hearing, sight, taste, and smell. The faculty of outer sense, taken transcendentally, is the capacity to relate to external objects, and its empirical realization, at its most abstract level, is a certain temporal-causal process through which external objects become related to an empirical mind. The fact that the faculty of outer sense in humans is realized by the five senses is a biological and contingent fact. Martian scientists may have other senses than ours, but still they must have the capability of the outer sense to access information from the external world. While concrete empirical realizations of a transcendental function may vary among different species of cognizers, a sufficiently abstract description of these empirical realizations in temporal-causal terms is true to all.
In the framework of transcendental functionalism, a more refined distinction should be made among transcendental cognitive functions, their direct empirical counterparts, and their concrete realizations in individual cognizers. The threefold distinction can be illustrated by the relations among pure categories, their temporal schemas, and specific empirical concepts. For example, while the pure category of causality is a conceptual relation between conditions and conditioned, its temporal schema is a rule-governed succession of events. Both the pure category and the temporal schema of causality are, according to Kant, a priori and universally valid, but applications of the causal schema to specific natural phenomena are not. Causal laws of specific natural phenomena are discovered a posteriori and subject to empirical verification. In the same way, specific psychological mechanisms that realize a transcendental cognitive function can be studied only empirically, even though a highly abstract description of the basic temporal-causal processes that realize the transcendental function can be achieved by philosophical reflection.
Applying the above distinction to the problem of subjective deduction, it is not difficult to see why Kant considers the subjective deduction inessential to his project despite being convinced of its validity. The theory of threefold synthesis in the subjective deduction is an abstract description of the temporal realization of the necessary functions of sensibility, imagination, and understanding. The deduction is subjective in the sense that it describes psychological mechanisms in temporal terms, although the description is given at such an abstract level that its truth is not restricted to human psychology. Kant indeed has good reasons to include the empirical, psychological story in the Critique in order to illustrate how the abstract functional structure has to play out at the phenomenal level, but at the same time he runs the risk of misleading readers to a psychological or phenomenal interpretation of the faculty of cognition.
Despite its a priori and universal validity, describing psychological mechanisms is not a way to explain the objective form of reality. Therefore, the subjective deduction, even if it truly describes the basic psychological process of all potential cognizers, can be rightly accused of psychologism, if it is understood as a deduction of the universal validity of categories. This possible misunderstanding finally led Kant to discard a detailed discussion of the threefold synthesis in order to avoid the charge of psychologism. The B-deduction is depsychologized, replacing much of the discussion of the synthesis in terms of temporal-sequential processes by an analysis of the structure of combination. The depsychologization consists in the detemporalization in the formulation of the necessary cognitive functions and structures. By restricting the analysis of the B-deduction to nontemporal, conceptual terms, Kant focuses on what is really essential to his chief end, that is, the functions that can account for the necessary structure of the cognizable reality. (62)
The Chinese University of Hong Kong
(1) Wilfrid Sellars, Essays in Philosophy and Its History (Dordrecht: Reidel, 1974), 62-90.
(2) Andrew Brook, Kant and the Mind (Cambridge: Cambridge University Press, 1994); Andrew Brook, "Kant and Cognitive Science," in The Prehistory of Cognitive Science, ed. Andrew Brook (Basingstoke, England: Palgrave Macmillan, 2007), 117-36; Patricia Kitcher, "Kant's Real Self," in Self and Nature in Kant's Philosophy, ed. Allen W. Wood (Ithaca, N.Y.: Cornell University Press, 1984), 113-47; Patricia Kitcher, Kant's Transcendental Psychology (Oxford: Oxford University Press, 1990); Ralf Meerbote, "Kant's Functionalism," in Historical Foundations of Cognitive Science, ed. J.-C. Smith (Dordrecht: Reidel, 1989), 161-87; Thomas Powell, Kant's Theory of Self-Consciousness (Oxford: Clarendon Press, 1990); Jay Rosenberg, The Thinking Self (Philadelphia: Temple University Press, 1986).
(3) Brook, Kant and the Mind, 1.
(4) Ibid., 11.
(5) Kitcher, Kant's Transcendental Psychology, 140-41.
(6) Lanier R. Anderson, "Neo-Kantianism and the Roots of Anti-Psychologism," British Journal for the History of Philosophy 13, no. 2 (2005): 287-323.
(7) Peter F. Strawson, The Bounds of Sense: An Essay of Kant's Critique of Pure Reason (London: Methuen, 1966), 32. In a later article Strawson relativized his criticism of Kant's psychology. See Strawson, "Sensibility, Understanding, and the Doctrine of Synthesis: Comments on Henrich and Guyer," in Kant's Transcendental Deductions: The Three Critiques and the Opus Postumum, ed. Eckart Forster (Stanford: Stanford University Press, 1989), 77.
(8) Kant follows Alexander Baumgarten to relegate sensibility and imagination to the lower faculties of cognition (AA 7:140-43, 196; AA 9:36), whereas under-standing, judgment, and reason are classified as higher faculties 0CPR, A130/B169; AA 20:201). References to Kant's Critique of Pure Reason (CPR) are to the pagination of the first (A) and/or the second edition (B). References to other works of Kant are given with volume and page numbers of the Akademie-Ausgabe (AA). Immanuel Kant, Gesammelte Schriften, ed. Koniglich Preufiische Akademie der Wissenschaften (Berlin: Walter de Gruyter, 1900-). All English translations of Kant's works are from Immanuel Kant, The Cambridge Edition of the Works of Immanuel Kant, ed. Paul Guyer and Allen W. Wood (Cambridge: Cambridge University Press, 1992-).
(9) Meerbote, "Kant's Functionalism," 161.
(10) CPR, A51/B75; A19/B33.
(11) Some Kant interpreters defend the noumenal interpretation by restricting Kant's noumenal incognizability thesis. Karl Ameriks, for example, claims that "Kant does not mean to block all kinds of knowledge of things in themselves, but only certain types." Karl Ameriks, "Kant and Short Arguments to Humility," in Interpreting Kant's Critiques, ed. Karl Ameriks (Oxford: Oxford University Press, 2003), 143. More specifically, Ameriks suggests that "the restraints of the Paralogisms ... may be meant primarily to restrict only fairly determinate (i.e., positive and specific) claims about noumenal individuals. Thus it still may be possible to allow some knowledge about mind, simply in the sense of truths that apply to one's intrinsic individual being, as long as this involves merely such indeterminate claims as that being is non-spatio-temporal. Similarly, it may be possible to allow some quite determinate knowledge about mind in another sense, as long as this involves merely the transcendental structures of our experience." Karl Ameriks, Kant's Theory of Mind: An Analysis of the Paralogisms of Pure Reason (Oxford: Clarendon Press, 2000), 8. However, if the transcendental structures of experience represent knowledge of the noumenal mind, then we would have a good deal of substantial knowledge about things in themselves. This seems hardly reconcilable with Kant's repeated emphasis of the incognizability of things in themselves.
(12) Kitcher, Kant's Transcendental Psychology, 22.
(13) Strawson criticizes the phenomenal interpretation as follows: "The workings of the human perceptual mechanism, the ways in which our experience is causally dependent on those workings, are matters for empirical, or scientific, not philosophical, investigation. Kant was well aware of this; he knew very well that such an empirical inquiry was of a quite different kind from the investigation he proposed into the fundamental structure of ideas in terms of which alone we can make intelligible to ourselves the idea of experience of the world." Strawson, The Bounds of Sense, 15.
(14) Kant states clearly that "[t]aste and colors are by no means necessary conditions under which alone the objects can be objects of the senses for us. They are only combined with the appearance as contingently added effects of the particular organization. Hence they are not a priori representations, but are grounded on sensation, and pleasant taste is even grounded on feeling (of pleasure and displeasure) as an effect of the sensation." CPR, A28-29.
(15) CPR, B167-68.
(16) Henry E. Allison, "On Naturalizing Kant's Transcendental Psychology," in Idealism and Freedom: Essays on Kant's Theoretical and Practical Philosophy (Cambridge: Cambridge University Press, 1996), 65.
(17) CPR, B422.
(18) CPR, Aix. Strawson calls it the ad hominem objection: "The theory of synthesis, like any essay in transcendental psychology, is exposed to the ad hominem objection that we can claim no empirical knowledge of its truth; for this would be to claim empirical knowledge of the occurrence of that which is held to be the antecedent condition of empirical knowledge." Strawson, The Bounds of Sense, 32.
(19) Kitcher, Kant's Transcendental Psychology, 140-41.
(20) Ibid., 141.
(21) For example, damage of the prefrontal cortex of the human brain can lead to the loss of emotional responses without substantial impairment of cognitive powers. See Antonio Damasio, Descartes' Error: Emotion, Reason, and the Human Brain (New York: G. P. Putnam, 1994), 34-79.
(22) CPR, A346/B404.
(23) Chong-Fuk Lau, "Kant's Epistemological Reorientation of Ontology," Kant Yearbook 2 (2010): 134-38.
(24) Robert Hanna, Kant and the Foundations of Analytic Philosophy (Oxford: Clarendon Press, 2001), 35-36.
(25) Paul Guyer, "Psychology and the Transcendental Deduction," in Kant's Transcendental Deductions: The Three Critiques and the Opus Postumum, ed. Eckart Forster (Stanford: Stanford University Press, 1989), 65.
(27) AA 9:25.
(28) CPR, Bxvi.
(29) Paul Guyer, Kant and the Claim of Knowledge (Cambridge: Cambridge University Press, 1987), 54-57; 121-24.
(30) CPR, A115; compare A94.
(31) CPR, At 15-16.
(32) Matt McConnick, "Questions about Functionalism in Kant's Philosophy of Mind: Lessons for Cognitive Science," Journal of Experimental and Theoretical Artificial Intelligence 15, no. 2 (2003): 256. 34
(33) Ibid., 264.
(34) Machine-state functionalism appeals extensively to the model of Turing machine as a hypothetical device, with which I also compared Kant's faculty of cognition. See Hilary Putnam, "The Nature of Mental States," in Mind, Language, and Reality (Cambridge: Cambridge University Press, 1975), 429-40.
(35) AA 5:195.
(36) In explaining his machine-state functionalism, Hilary Putnam draws a similar distinction between "two possible descriptions of the behavior of a Turing machine--the engineer's structural blueprint and the logician's 'machine table.'" Putnam, "Minds and Machines," in Mind, Language, and Reality, 372. The latter refers to the Turing machine as an abstract device and the former to its physical realization. The twofold distinction also applies to human psychology: "The 'behavioristic' approach ... aims at eventually providing a complete physicalistic description of human behavior, in terms which link up with chemistry and physics. This corresponds to the engineer's or physicist's description of a physically realized Turing machine. But it would also be possible to seek a more abstract description of human mental processes, in terms of 'mental states' (physical realization, if any, unspecified) and 'impressions' (these play the role of symbols on the machine's tapes)--a description which would specify the laws controlling the order in which the states succeeded one another, and the relation to verbalization (or, at any rate, verbalized thought)." Ibid., 372-73.
(37) CPR, A19/B33.
(38) CPR, A19-20/B34.
(39) AA 15:165.
(40) If affection is considered a proper causal process, then it can be about only the relation from empirical objects to an empirical perceptive mind, but not between things in themselves and sensibility.
(41) CPR, B132.
(42) CPR, A116.
(43) Interesting is Kant's idea that the identity of an empirical cognizer or a phenomenal self over time cannot be sustained by a single, thoroughgoing identical representation because what "is customarily called inner sense or empirical apperception" is, according to Kant, "forever variable; it can provide no standing or abiding self in this stream of inner appearances." CPR, A106-07.
(44) CPR, B143; B428.
(45) CPR, B429.
(46) CPR, A 347-48/B 405-06.
(47) Kant's concept of transcendental subject is similar to Wittgenstein's notion of metaphysical subject in the Tractatus: "there really is a sense in which philosophy can talk about the self in a non-psychological way.... The philo-sophical self is not the human being, not the human body, or the human soul, with which psychology deals, but rather the metaphysical subject, the limit of the world--not a part of it." Ludwig Wittgenstein, Tractatus Logico-Philosophicus, trans. David F. Pears and Brian F. McGuinness (London: Routledge, 2001), [section]5.641, p. 70.
(48) CPR, Axvi-xvii.
(49) CPR, A95-114.
(50) CPR, A115-30.
(51) CPR, A97.
(52) Nathan Bauer recently proposed an alternative reading, identifying the subjective deduction with the third section of the A-deduction and the objective with the second section. Nathan Bauer, "Kant's Subjective Deduction," British Journal for the History of Philosophy 18, no. 3 (2010): 433-60. However, Kant's abandonment of the theory of threefold synthesis in the B-deduction speaks strongly against his interpretation.
(53) Brook, Kant and the Mind, 106.
(54) Strawson, The Bounds of Sense, 31-32; Kitcher, Kant's Transcendental Psychology, 65.
(55) CPR, A89-90/B122.
(56) CPR, Axvi.
(57) CPR, B104.
(58) Kitcher, Kant's Transcendental Psychology, 65.
(59) Norman Kemp Smith, A Commentary to Kant's "Critique of Pure Reason" (Basingstoke, England: Palgrave Macmillan, 2003), 239-40.
(60) Ibid., 241.
(61) Ibid., 242; xlix.
(62) The work on this paper was supported by a grant from the Research Grants Council of the Hong Kong Special Administrative Region (Project No.: CUHK 446912/12H).
Correspondence to: Chuong-Fuk Lau, Department of Philosophy, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong.
|Printer friendly Cite/link Email Feedback|
|Publication:||The Review of Metaphysics|
|Date:||Dec 1, 2014|
|Previous Article:||Experiencing tradition versus belonging to it: Gadamer's dilemma.|
|Next Article:||Quietism or description? McDowell in dispute with Dreyfus.|