KIYOSHI CHIBA | Kants Ontologie der raumzeitlichen Wirklichkeit | Walter de Gruyter 2012


By Kiyoshi Chiba

The central question of my book Kants Ontologie der raumzeitlichen Wirklichkeit is: ‘Is Kant’s Transcendental Idealism a kind of realism or idealism?’ Realism is, roughly, the doctrine that objects exist independently of us (e.g. of our representation, consciousness, knowledge/knowability, cognitive capacity etc.). Idealism, by contrast, claims that objects are in a certain sense dependent on us; a prominent example is phenomenalism, a version of which claims that spatiotemporal objects are logical constructions out of sensations.

Due to the widespread bad reputation of idealism in general, many scholars today aspire to liberate the Kantian philosophy from its idealistic elements as much as possible. This intention can be discerned not only in outright anti-idealistic interpreters such as Paul Abela (2002), Arthur Collins (1999) and Rae Langton (1998), but also in advocates of the so-called two-aspect interpretation such as Lucy Allais (2004, 2007), Karl Ameriks (1992, 2006), Tobias Rosefeldt (2007) and Allen Wood (2005). It is often thought that such an anti-idealistic reading is the only possible way to understand Kant’s philosophy in a philosophically meaningful manner. I do not agree with this assessment and stand rather on the side of idealistic interpretations, such as those of Richard Aquila (1983), Jonathan Bennett (1966, 1974) and James Van Cleve (1999).

The opposition ‘realism/idealism’ as provisionally formulated above must be made more precise. For this purpose, I adopt the conceptual apparatus which Michael Dummett has provided for the analysis of realism and its opposite ‘anti-realism’. My book has two aims: First, to show that Kant’s Transcendental Idealism advances an anti-realist ontology of spatiotemporal reality, and second, to elucidate the aspects of Kant’s anti-realist ontology with the help of Dummett’s conceptual apparatus.

My book has three parts. Part I (Chapter 1–3) has a clarificatory task. Often an exegetical debate is confused because the various positions are not presented in precise enough terms. Therefore, before addressing Kant’s text, I clarify my own position and situate it in the context of current interpretations of Transcendental Idealism.

Realism and Anti-realism

In Chapter 1, I explain the terms ‘realism’ and ‘anti-realism’. Realism in the Dummettian sense is primarily a verification-independent truth-conception, i.e., the truth of statements is independent of the possibility of our verification of them. This leads to a cognition-independent ontology of objects. By contrast, anti-realism is a negation of realism and therefore leads to some version of verification-dependent truth-conception and cognition-dependent ontology. Due to this definition of ‘anti-realism’, there is no middle path between realism and anti-realism. Thus, even the ‘internal realism’ of Hilary Putnam and e.g. the ‘pragmatism’ of Charles S. Peirce are classified as anti-realism.

In Chapter 2, I aim to clarify what realist and anti-realist interpretations of Kant’s Transcendental Idealism respectively amount to. The latter embrace every kind of interpretation which ascribes to Kant some kind of anti-realism of spatiotemporal objects; the so called ‘phenomenalistic’ and also the ‘coherent-theoretical’ interpretations are examples of this position (among other variations).

A terminological proviso would be in order here. The claim that Transcendental Idealism is a kind of anti-realism (in the Dummettian sense) of spatiotemporal objects does not conflict with Kant’s well-known statement that Transcendental Idealism is an empirical realism. Contrary to the view held by some interpreters, Kant himself defines empirical realism solely as an epistemological thesis, more concretely, as the anti-sceptical thesis that we can get secure knowledge about the existence of spatial objects, without specifying in what sense the existence of spatial objects is to be understood (see A370, A371 and A375). By contrast, realism and anti-realism in Dummett’s sense are both ontological positions, as was explained above. So, there is no inconsistency in saying that Kant affirms empirical realism and denies realism in Dummett’s sense.

Realist interpretations, which ascribe to Kant a kind of realism of spatiotemporal objects, have also many variants. The most typical and important version of this is the realist two-aspect interpretation which ascribes to Kant the view that:

Spatiotemporal objects, or more precisely, things we cognise as spatiotemporal, exist independently of our cognition; only the ways they spatiotemporally appear to us are dependent on us, that is, on the forms of our sensibility and understanding.

This position is realist in Dummett’s sense, because it conceives of the existence of spatiotemporal objects as cognition-independent. I reject this interpretive option, not because of some peculiar problems—e.g., it appears to assume a kind of isomorphism between appearances and things in themselves—but because of its commitment to the realism of spatiotemporal objects.

The conceptual scheme realist/anti-realist turns out to be useful also to clarify current interpretations of Transcendental Idealism. In Chapter 3, I argue for three points: (1) The two-world interpretation comports well with the anti-realist interpretation. (2) The metaphysical two-aspect interpretation leads to a realist interpretation, more concretely, the realist two-aspect interpretation introduced above. (3) The methodological two-aspect interpretation turns out either to be inconsistent in itself (e.g. the well-known variants by Henry Allison [1983, 2004] and Gerold Prauss [1974]) or, if consistently developed, to lead to a radical anti-realism, which denies the cognition-independent reality of any kind, and is just for this reason unattractive both exegetically and systematically.

In my book I defend an anti-realist two-world interpretation. It seems at first glance that the problem of the two-world/two-aspect interpretation is unsolvable, especially because good textual supports are found for both sides. Taking this circumstance into consideration, I do not directly tackle this problem, but rather solve first the question whether Kant’s Transcendental Idealism is a realism or an anti-realism of spatiotemporal objects. Once this point is settled, we are in a better position to decide on the two-world/two-aspect issue.

Justifying an Anti-Realist Interpretation

In Part II of my book, I provide, on the basis of the aforementioned conceptual clarifications, a defence of my anti-realist interpretation through a concrete analysis of relevant texts in Kant. First, I discuss especially the positive textual supports for the anti-realist interpretation. Then, in Chapter 4, I investigate the Antinomy-chapter of the first Critique, in particular the indirect proof of Transcendental Idealism, which says:

Transcendental Realism inevitably leads to the antinomy; this can be avoided only by Transcendental Idealism. Hence, Transcendental Idealism is true, on Kant’s implicit assumption that Transcendental Realism and Transcendental Idealism are contradictory positions.

Subsequently, in Chapter 5, I investigate the Fourth Paralogism in the A-version of the first Critique, in particular, the argument which I call the ‘indirect justification’ of Transcendental Idealism, which is supposed to prove that:

Transcendental Realism inevitably leads to Sceptical Idealism (i.e., scepticism which doubts the existence of spatial objects); this can be avoided only by Transcendental Idealism. Hence, Transcendental Idealism is to be preferred.

These investigations show that it is Transcendental Realism’s commitment to realism (in the Dummettian sense explained above) that is responsible both for the antinomy and for Sceptical Idealism. Therefore, Kant’s Transcendental Idealism, which aims to avoid these, must be a kind of anti-realism. What I aim to show in these two chapters (4 & 5) is not just that the aforementioned passages can be interpreted anti-realistically, nor even that such an interpretation is plausible. I intend rather to establish definitely, against any kind of realist interpretation, that Kant’s arguments in those passages are consistently understandable only in the framework of an anti-realist interpretation. True, to show this is a tall order and for this purpose a detailed investigation is required, which I provide in the book.

In Chapter 6, I investigate the aspects which seem to support the realist interpretation. First, I look at the Refutation of Idealism in the B-version of the first Critique, which is often cited as decisive evidence for Kant’s commitment to the realism of spatiotemporal objects. I show that the Refutation is neutral with respect to realism/anti-realism. Namely, the argument of the Refutation refutes Sceptical Idealism without considering whether the existence of spatial objects, which the Refutation aims to establish, should be understood realistically or anti-realistically.

The most serious obstacle for the anti-realistic interpretation is Kant’s two-aspect language (the type of expression which Stephen Barker called “language of appearing”, contrary to the “language of appearance”).[1] I show first that this language in fact leads to a realist two-aspect theory of spatiotemporal objects. However, Kant’s commitment to anti-realism cannot be denied, as was shown in the previous chapters of my book (see also Kant’s use of the two-world language, which suggests an anti-realism of spatiotemporal objects). These results together lead to the conclusion that Kant himself swayed between a realist two-aspect and an anti-realist two-world theory (not just language) without noticing their incompatibility.

This conclusion forces us to shift the target of interpretation from pure exegesis to rational reconstruction, which brackets apparent inconsistencies in Kant’s terminology and tries to make Kant’s arguments as a whole more coherent. With this approach, I investigate the passages where the two-aspect language is especially dominant, i.e. in the Transcendental Aesthetic and in Kant’s doctrine of the double character of the self in his theory of freedom. I show that although these passages make a realistic interpretation plausible, the arguments which are presented in these passages are compatible with the anti-realist interpretation. (The same does not hold for the realist interpretation regarding the Antinomy-chapter and the Fourth Paralogism.)

The investigation of Part II as a whole shows that Kant’s Transcendental Idealism can be adequately interpreted only in an anti-realist manner, insofar as it should be understood as a coherent doctrine at all. Furthermore, regarding the two-aspect interpretation, we can now maintain that its most persuasive option, the metaphysical two-aspect interpretation, is to be rejected because of its commitment to a realist interpretation. The methodological two-aspect interpretation, which was shown to be unattractive because, if consistently developed, it negates the possibility of cognition-independent reality, will be rejected finally when it is shown—which I do in Chapter 8 of my book—that Kant’s theoretical philosophy requires the existence of cognition-independent entities. In the following investigation I elaborate a version of the anti-realist two-world interpretation.

It has been shown so far in the book that Kant’s Transcendental Idealism is committed to an anti-realism of spatiotemporal reality. This result is important for current Kant interpretations, but in itself not very informative for the explanation of Transcendental Idealism, since the term ‘anti-realism’ here refers only to a negation of realism. A large range of possible options is still available. It is obvious that Kant does not accept an extremely subjectivist position such as the position that spatiotemporal objects exist only as long as they are actually perceived by us. How then should the central thesis of anti-realism, namely that the existence of objects is dependent on our cognition, be understood in the Kantian framework?

In Part III of my book, I elucidate further the concrete contents of Kant’s anti-realist ontology of the spatiotemporal reality with the aid of the Dummettian conceptual apparatus, above all that of mathematical intuitionism. Especially the following two topics are discussed: (1) Which concrete version of anti-realism regarding spatiotemporal reality is adequate for Transcendental Idealism? (2) How do spatiotemporal objects as ‘appearances’ relate to the so-called ‘things in themselves’?

Anti-Realism and Spatiotemporal Reality

Regarding the first topic (Ch. 7), I begin with examining the following three versions of anti-realism which are often ascribed to Kant but turn out to be inadequate for the Kantian anti-realism:

Temporal Version: A statement about spatiotemporal objects is true at moment t iff it has been verified at t. In ontological terms: The world itself, and not just our knowledge of it, is gradually enlarged in accordance with the progress of our inquiry.

Peircean Version: A statement is (atemporally) true iff it would be verified if one were in the ideal state of information, where all that is humanly knowable were to be completely known.

Standard-Phenomenalistic Version: A statement is (atemporally) true iff it would be verified if one were put in proper circumstances. (A ‘proper circumstance’ for the statement “While Caesar crossed the Rubicon, he sneezed three times” is e.g. the place and time in which one could directly observe Caesar crossing the Rubicon.)

The problem of these options is, above all, that they are incompatible with Kant’s solution of the antinomy in the ninth section of the Antinomy-chapter. Taking this point into account, I work out an empirical counterpart of the atemporal truth-conception of mathematical intuitionism. This version of anti-realism proves to suit Kant’s solution of the antinomy well.

Intuitionistic anti-realism, however, cannot readily be applied to empirical discourse, since it assumes the so-called ‘monotonicity’ (which says that once verified, a statement remains verified eternally), as is typically seen in intuitionistic logic. Regarding this there are two problems: (a) In the empirical domain it is rather normal that verifications which have once been achieved will get lost irrecoverably in the passage of time. (b) Empirical verifications are defeasible without exception, i.e., it is always possible that they will be dismissed as insufficient or false in the progress of inquiry. By considering these problems, I elaborate the above-developed intuitionistic anti-realism step by step, in order to reach an atemporal version of anti-realism based on Crispin Wright’s notion of superassertibility.

Appearances, Things in themselves, and Affection

In Chapter 8, I deal with the second topic, i.e., the relation of the spatiotemporal objects to the so-called ‘things in themselves’, especially with respect to the problem of affection. I defend the theory of transcendental affection, i.e. affection by things in themselves. First, I show that the transcendental affection in Kant’s original form cannot be sustained and needs modification. Mostly—except for a few passages such as A288/B344, A358, A359 and in the case of self-affection—Kant talks about the affection as a three-place relation, which has as its relata: (1) an affecting thing, (2) an affected subject and (3) a produced sensation, and in particular assumes that affection is a kind of causal relation in which the affecting thing and the affected subject are numerically distinct. However, this claim is, strictly speaking, not required by Kant’s doctrine of receptivity, and also conflicts with Kant’s well-known thesis of the inapplicability of categories—in this case, especially those of plurality and causality—to things in themselves.

Taking these problems into consideration, I propose a modified theory of transcendental affection. I agree with Kant that his theoretical philosophy, above all, its central doctrine of the receptivity of our experience, requires not only meaningful reference to things in themselves, but also the existence of something that is independent of our cognition, especially of our spontaneity, and which prompts sensations in our mind. My modified reading, however, does not demand that this ‘something’ be a numerically distinct entity from the affected subject. As a result, the modified theory is free from the traditional criticism of applying the categories of causality and plurality to things in themselves. That is, my reading does not entail, for example, that a thing in itself which is numerically distinct from us, exercises a causal influence on us.

There are a number of possible problems with my modified theory of transcendental affection. The first problem concerns that this theory does make a positive claim about the existence of things in themselves. Does this not clash with the unknowability of things in themselves and with the inapplicability of categories to them? I tackle this problem by investigating Kant’s arguments for the unknowability and the inapplicability theses. Another problem concerns an apparently realist implication of transcendental affection. It seems that transcendental affection is incompatible with the core idea of the anti-realist claim that objects are cognition-dependent, since transcendental affection appears to have the following consequences:

(1) Spatiotemporal objects and things in themselves are isomorphic, and the differences on the side of the former are grounded on those on the side of the latter.

(2) Although formal properties of spatiotemporal objects (such as their spatiotemporality) depend on the formal conditions of our cognitions, the concrete state of the spatiotemporal world (i.e., what kind of spatiotemporal objects there are and how they concretely exist) is determined, before every actual realisation of our cognition, by the way how things in themselves which affect us are.

It turns out that the first thesis is indeed incompatible with anti-realism concerning spatiotemporal objects. However, my modified theory of transcendental affection can be defended by showing that the first thesis does not really follow from the assumption of transcendental affection. The second thesis, on the contrary, has to be acknowledged even from the anti-realist’s point of view, because it is based on the idea that truth is not affected by spontaneous decisions on the side of cognising subjects. This idea is so deeply rooted in our conception of truth that no theory, whether anti-realist or not, can abandon this idea without losing its plausibility. Rather, it must be reconsidered whether the second thesis does in effect necessarily conflict with the core idea of anti-realism. I show that this is not the case and, in particular, that the second thesis is compatible with the intuitionist versions of anti-realism.

[1] S. Barker, ‘Appearing and Appearances’, in L. W. Beck (ed.), Kant Studies Today (La Salle, Ill.: Open Court), pp. 274–89.

© 2013, Kiyoshi Chiba

Kiyoshi Chiba is Associate Professor of Philosophy at the University of Yamagata, Japan. Chiba obtained his Ph.D in Philosophy in 2010 from the University of Bonn, Germany. His dissertation “Kants Ontologie der raumzeitlichen Wirklichkeit. Versuch einer anti-realistischen Interpretation der Kritik der reinen Vernunft”, the book under discussion here, was published by Walter de Gruyter in 2012 in the series Kant-Studien Ergänzungshefte and can be purchased here.