阿利森：先验实在论、经验实在论和先验观念论

Transcendental Realism, Empirical Realism and Transcendental Idealism HENRY E.ALLISON University of California at Davis The debate regarding the interpretation of Kant’s idealism is usually seen as turning on the best way to understand his transcendental distinction between appearances and things in themselves: that it marks either a contrast between two types of thing (the ‘two-object’ or ‘two-world’ view) or one between two sides or aspects of ordi- nary empirical objects (the ‘two-aspect’ view).1 But, even though I have long been associated with the latter camp, I have also thought for many years that this is not the most helpful way to frame the issue. The problem lies in an ambiguity inherent in the two-aspect view. It can be understood either metaphysically, as a thesis about the kinds of properties attributable to empirical objects, that is, as a form of property dualism in which these objects are assigned both phenomenal and noumenal properties, or methodologically, as a contrast between two ways in which such objects can be considered in a philosophical reflection on the conditions of their cognition. Accordingly, I take the fundamental question to be whether tran- scendental idealism is to be understood in the latter way or as a form of metaphysical dualism (whether as a thing or a property dualism being a matter of relative indifference). And I have further thought that the best way of addressing that question is through a considera- tion of the view which Kant opposes to transcendental idealism, namely, transcendental realism. If this realism is identified with a particular metaphysical doctrine then transcendental idealism must be as well; but if, as I maintain, transcendental realism cannot be so understood, then neither can Kant’s idealism.2 I shall here argue somewhat obliquely for the latter alternative by exploring the nature of the difference between transcendental and empirical realism.3 The discussion consists of four parts and a brief appendix. The first part maintains that rather than being KANTIAN REVIEW, VOLUME 11, 2006 1 01 Allison KR.qxp 09/08/2006 16:30 Page 1 either a distinct ontological doctrine or a mere label for everything to which Kant was opposed, transcendental realism should be understood as the view that spatiotemporal predicates are appli- cable to things in general. Since this view is shared by all ontologies (at least all those with which Kant was concerned), transcendental realism is not committed to a particular ontology; but, since it is contrasted with an empirical realism, which limits the scope of these predicates to objects of possible experience, it is also not a vacuous label. And from this I conclude that transcendental idealism likewise does not constitute a distinct ontological posi- tion, but instead provides a radical alternative to ontology.4 The second and third parts deal with two possible objections to this reading. The first is that it conflicts with Kant’s official account, which charges transcendental realism with conflating appearances with things in themselves rather than with inflating claims about objects of possible experience into claims about things in general. By examining the relationship between the concepts involved, I show that there is no incompatibility between the two characteriza- tions. The second line of objection is that such a non-metaphysical interpretation of transcendental idealism trivializes it by reducing it to a recommendation of epistemological modesty. I respond by acknowledging that transcendental idealism, so construed, does consist essentially in such a recommendation but denying that this trivializes it. The point is further explored in the fourth part, which analyses Kant’s indirect argument for transcendental idealism drawn from his resolution of the mathematical antinomies and explores the therapeutic function of this idealism in his treatment of transcendental freedom. In the appendix, I surmise that one reason why many contemporary Kantians are dismissive of tran- scendental idealism is that they mistakenly assume that a central concern of the Critique is to establish a robust realism in the face of a radical sceptical challenge. I suggest instead that Kant argues from rather than to such realism and that his central concern is to limit its scope to objects of possible experience. 1. Kant defines transcendental realism in two places in the Critique and in each of them he contrasts it with transcendental idealism. In HENRY E. ALLISON 2 KANTIAN REVIEW, VOLUME 11, 2006 01 Allison KR.qxp 09/08/2006 16:30 Page 2 the first of these, he characterizes it as the view ‘which regards space and time as something given in themselves (independent of our sensibility)’ and suggests that such a realist ‘interprets outer appearances . . . as things in themselves, which would exist inde- pendently of us and our sensibility and thus would also be outside us according to pure concepts of the understanding’. By contrast, transcendental idealism is defined as ‘the doctrine that all appear- ances are to be regarded as mere representations and not things in themselves, and accordingly that time and space are only sensible forms of our intuition, but not determinations given for themselves or conditions of objects as things in themselves’ (A369).5 In the second, transcendental realism is accused of making ‘modifications of our sensibility into things subsisting in themselves, and hence makes mere representations into things in themselves’, whereas transcendental idealism affirms that ‘all objects of an experience possible for us, are nothing but appearances, i.e., mere representa- tions, which, as they are represented, as extended beings or series of alterations, have outside our thoughts no existence grounded in itself’ (A490–1/B518–19). Although it is much more explicit in the first, both accounts indicate that the dispute between the two forms of transcendentalism concerns primarily the nature of space and time and, as a consequence of this, the nature of things encountered in them – what Kant terms ‘appearances’. In essence, Kant’s charge is that by viewing space and time as ‘given in them- selves’, that is to say, independently of the conditions of sensibility, transcendental realism conflates spatiotemporal entities, which for the transcendental idealist are ‘mere appearances’, with things in themselves. Since any number of views might be characterized as regarding space and time as ‘given in themselves’ and since Kant himself accuses philosophies of many different stripes of conflating appearances with things in themselves, these characterizations of transcendental realism seem much too vague to define a metaphys- ical position with which transcendental idealism might meaning- fully be contrasted.6 Alternatively, if, as is sometimes done, tran- scendental realism is identified with the scientific realism of the Cartesians and Newtonians (roughly what Berkeley understood by ‘materialism’), then the situation seems even worse. First, it implies that Kant’s idealism is akin to Berkeley’s, something which Kant vehemently and famously denied. Second, it invites the familiar TRANSCENDENTAL REALISM, EMPIRICAL REALISM KANTIAN REVIEW, VOLUME 11, 2006 3 01 Allison KR.qxp 09/08/2006 16:30 Page 3 charge of neglected alternatives.7 For, clearly, Kant did not intend to suggest that such realism and his idealism exhaust the philo- sophical universe. As we shall see, this issue becomes particularly pressing in connection with Kant’s indirect argument for transcen- dental idealism through the negation of transcendental realism. Unless it is assumed that they are contradictory opposites, this argument cannot get off the ground. We thus appear to find ourselves confronted with a dilemma. If transcendental realism is seen as an amorphous, ill-defined meta- physical theory, it becomes difficult to see how it could contribute to an understanding of transcendental idealism; whereas if we identify it with a particular metaphysical theory it leads to the outright dismissal of an argument which Kant thought to be central to his project and, depending on the nature of this theory, perhaps to a highly unattractive view of transcendental idealism as well. My proposed way out of this dilemma is through a considera- tion of the difference between a transcendental and an empirical realism. In so doing, I hope to show that the former is not to be understood as a distinct metaphysical or, more properly, ontolog- ical theory, while at the same time giving it enough shape to preserve the significance of its contrast with transcendental idealism. The operative question is what makes transcendental realism transcendental, which, in turn, calls for an investigation of Kant’s multiple uses of this highly elusive, yet ubiquitous, term. But since anything approaching an adequate treatment of the topic is well beyond the scope of the present article, it must suffice to note that Kant seems to work with at least two competing conceptions of the transcendental. The first and featured conception is the one that he introduces in his stipulative definitions of ‘transcendental cogni- tion’ (A12/B25). According to these definitions, such cognition may (very roughly) be characterized as a second order activity concerned with an investigation of the a priori elements of our cognitive apparatus and the conditions and limits of our a priori cognition of objects.8 The second conception is the traditional one, according to which transcendental cognition is concerned with the nature of things in general, that is, with the subject matter of ontology or metaphysica generalis, as understood by the Wolffian school.9 And, just as Kant’s contemporaries were quite familiar with the latter but had great difficulty understanding the former, so HENRY E. ALLISON 4 KANTIAN REVIEW, VOLUME 11, 2006 01 Allison KR.qxp 09/08/2006 16:30 Page 4 for many present day readers the situation is virtually the reverse. Nevertheless, as we shall soon see, the recognition of the continued presence in the Critique of vestiges of this earlier conception is crucial for the understanding of transcendental realism and, there- fore, transcendental idealism as well.10 A good illustration of the juxtaposition of these two concep- tions of the transcendental is to be found in Kant’s remark concerning the meaning of ‘transcendental cognition’, a remark which Kant advises the reader to keep well in mind, since its import extends to everything that follows (A56/B80). The bulk of this remark is devoted to underscoring the distinction between tran- scendental and a priori cognition. Although the former is obviously a priori, it differs from ordinary, first order a priori cognition, such as is provided in mathematics, by the fact that it is concerned with the possibility of the latter. Our present interest, however, lies in the concluding and less frequently noted portion of this remark, where Kant adds parenthetically: Likewise the use [Gebrauch] of space about all objects in general would also be transcendental; but if it is restricted solely to objects of the senses, then it is called empirical. The difference between the transcendental and the empirical therefore belongs only to the critique of cognitions and does not concern their relation to their object. [A56–7/B81] Whereas the contrast between transcendental and a priori cogni- tion is a matter of level, that between a transcendental and an empirical use is a matter of scope, which is why it pertains to a critique of cognition. Accordingly, it is here that Kant’s employment of ‘transcendental’ overlaps with the traditional understanding of the term. Moreover, this is not an isolated occurrence or regression on Kant’s part to a pre-Critical standpoint, since, as we shall soon see, he systematically uses the term in precisely this way in contrasting a transcendental with an empirical use of the cate- gories and principles of pure understanding. In both cases, it amounts to a contrast between a generic use with respect to all objects and one restricted to a particular domain of objects, namely, phenomena or objects of possible experience. This has two implications, which are essential to the proper understanding of both transcendental realism and transcendental idealism. The first is that the difference between transcendental TRANSCENDENTAL REALISM, EMPIRICAL REALISM KANTIAN REVIEW, VOLUME 11, 2006 5 01 Allison KR.qxp 09/08/2006 16:30 Page 5 and empirical realism consists in the scope assigned to spatiotem- poral predicates rather than the degree or kind of reality attributed to them. In other words, it is not that empirical realism assigns a lesser degree of reality to such predicates, but merely that it restricts their applicability to the domain of possible experience. The second is that, though transcendental realism is committed to the proposition that spatiotemporal predicates are ontological in the traditional sense of applying with strict universality, it is not, as such, committed to any particular ontology of space and time. The latter point is reflected in the Transcendental Aesthetic, where Kant begins by raising the purportedly ontological question: What are space and time? Four possibilities are introduced. They might be: a) actual entities (substances); b) determinations of things (accidents); c) relations of things that ‘would pertain to them even if they were not intuited’; or d) ‘relations that only attach to the form of intuition alone, and thus to the subjective constitution of our mind, without which these predicates could not be attached to anything at all’ (A23/B37–38).11 The first three represent the traditional ontological options and, therefore, apply to things in general. And of these the second and third correspond respectively to the Newtonian and Leibnizian views, which were the main competitors at Kant’s time.12 The fourth is Kant’s Critical view and reflects his rejection of the whole ontological framework in which the question has traditionally been posed, one in which it is assumed that whatever status is given to spatiotemporal predicates they apply with strict universality. Thus, from Kant’s more compre- hensive point of view, the dispute between the Newtonians and the Leibnizians is reduced to a family quarrel. Seen in this light, Kant’s basic charge against transcendental realism (in all its forms) is that it erroneously assumes that spatiotemporal predicates, which he limits to the domain of possible experience, have an unrestricted scope. Moreover, since Kant’s doctrine of the ideality of space and time provides the warrant for his scope restriction, this further suggests that tran- scendental idealism, even as it appears in the Aesthetic, is best seen as a deflationary proposal rather than as an ontological thesis in direct competition with the various forms of transcendental realism.13 Otherwise expressed, inasmuch as the first three of the four possible accounts of space and time that Kant introduces exhaust the recognized ontological alternatives, the fourth, which HENRY E. ALLISON 6 KANTIAN REVIEW, VOLUME 11, 2006 01 Allison KR.qxp 09/08/2006 16:30 Page 6 represents Kant’s Critical position, might be regarded as proposing a radical alternative to ontology rather than, as it usually taken to be, a novel move within ontology. Specifically, the alternative is to consider space and time as ‘two sources of cognition’ (A38/B55), that is, as conditions of our cognition of things, rather than as themselves either things (substances), properties, or relations of things as such.14 Nevertheless, this does not suffice to establish Kant’s restriction thesis, since, for all that has been said so far, it might still be the case that space and time are conditions of the cognition of things in general. Thus, it is essential for Kant to demonstrate their connection with human sensibility. What must be shown is that the universality of space and time within human experience, a point on which the empirical and the transcendental realist agree, can be understood only on the assumption that their representations are a priori contributions of human sensibility, from which the scope limitation to objects of possible human experience (or at least to the experience of cognizers with our forms of sensibility) follows immediately.15 Whether Kant actually succeeds in showing this is another and more contentious question, which I cannot consider here.16 2. Even setting aside the question of the cogency of Kant’s arguments, however, it might be objected that this analysis misrepresents Kant’s actual accounts of transcendental realism, which make no explicit reference to things in general, but refer instead to things in themselves. In fact, rather than accusing transcendental realism of a simple scope confusion, we have seen that he charges it with the seemingly more heinous crime of making ‘modifications of our sensibility’ or ‘mere representations’ into things in themselves. How, then, it may be asked, are these accounts to be reconciled? Although dealing with this question will unavoidably require a digression from the main line of argument, my hopes are that this can be done fairly expeditiously and that it will eventually help to put the argument in a somewhat clearer light. To anticipate, I shall claim that even though the concepts of a thing (or something) in general and of a thing in itself occupy distinct spheres of philosophical TRANSCENDENTAL REALISM, EMPIRICAL REALISM KANTIAN REVIEW, VOLUME 11, 2006 7 01 Allison KR.qxp 09/08/2006 16:30 Page 7 reflection, they are related in such a way that if spatiotemporal (or, indeed, any) predicates were applicable to the former, they would ipso facto be applicable to the latter as well. As textual support for this reading, I shall rely primarily on Kant’s account of the putative transcendental use of the categories and their associated principles, by which he understands one with regard to the cognition of things in general.17 As was the case with spatiotemporal predicates, this is contrasted with an empirical use, which is restricted to the cognition of objects of possible experience or appearances. Although Kant defines the categories (nominally) as concepts of an object in general, he denies that they could have the former employment, since they would then lack the schemata necessary to link them up to the world.18 But he also suggests that if (per impossibile) they had such an employment, ‘as conditions of the possibility of things in general, they [could] be extended to objects in themselves (without any restriction to our sensibility)’ (A129/B178). Similarly, Kant states that, ‘The transcendental use of a concept in any sort of principle consists in its being related to things in general and in themselves’ (A238/B298), which suggests that he considered the relation of a concept to things in themselves to be a direct consequence of its relation to things in general and, therefore, as not requiring any further explanation.19 However, since this connection may not be as obvious to contemporary readers as it apparently was to Kant, it will be useful to take a closer look at the concepts involved. To begin with, we have seen that Kant takes the concept of a thing in general over whole cloth from the ontological tradition and that it encompasses whatever pertains to the thought of a thing as such. Accordingly, the project of ontology, traditionally understood, is to provide cognition of things by means of an analysis of this concept. And since this involves an illicit transcendental employment of the understanding, Kant proclaims famously that, ‘[T]he proud name of ontology, which presumes to offer synthetic a priori cognitions of things in general in a systematic doctrine . . . must give way to the modest one of a mere analytic