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

By Kiyoshi Chiba

I want to thank Henny Blomme for the time and effort he spent on my book, and for his insightful comments (see here). His objections could be roughly divided into two groups: one concerns my distinction between realism and anti-realism, and the other concerns my interpretation of things in themselves. I will reply to them in turn.

1. The Distinction between Realism and Anti-Realism

One of Blomme’s main objections is that my distinction between realism and anti-realism is not useful, or even superfluous, for understanding Kant’s position. This objection is stated most concretely towards the end of Blomme’s critique (hereafter BC for short):

With respect to Kant’s theoretical philosophy, the disjunction that Chiba uses (“the existence of objects is independent from our cognition or the existence of objects is dependent on our cognition”) is not exclusive as long as it is not specified what is meant with object and what is meant with existence. But if Chiba would specify in which Kantian sense he wants the terms ‘existence’ and ‘object’ to be understood, I surmise that it wouldn’t appear to be very relevant anymore to want to decide whether Kant was a realist or an idealist, because then one would always end up with the interpretation of but ‘one aspect’ of his theory.

Blomme’s main claims are: (1) My distinction between realism and anti-realism is not exclusive as it stands, since it allows for some cases which are both realism in some sense and anti-realism in another sense. (2) If my distinction is further elaborated to make it exclusive, i.e., if it is amended by a specification of the sense in which (and for which objects) Kant’s Transcendental Idealism (TI) is an anti-realism and in which sense it is not, then the task to decide whether Kant’s TI is a realism or anti-realism loses its significance, since whichever conclusion we may get, the conclusion does not concern Kant’s philosophy as a whole but only one aspect thereof.

I will discuss claim (1) later. I would like to show first that even if (1) is true, (2) is not. Suppose that my distinction is really non-exclusive, and that it is amended in the suggested way, and furthermore that it follows from this new distinction that Kant’s TI is an anti-realism in sense A and a realism in sense B. I see no reason for saying that in this case the conclusion is uninformative and the distinction itself is superfluous. Even though it is true that in this case each of the senses A and B concern only one aspect of Kant’s philosophy, yet if the new distinction is exclusive and exhaustive, we indeed have grasped Kant’s position as a whole by taking both senses into account. At any rate, if we can specify correctly and precisely in which sense Kant’s TI is an anti-realism and in which sense it is not, this is doubtlessly a progress in understanding Kant’s theoretical philosophy as a whole. I do not understand how one can reasonably deny this.

Perhaps Blomme thinks that the aim of my investigation in my book (hereafter KORW for short) was simply to put some label on TI. This is suggested by the opening sentence of BC which states that the “goal” of my book is to answer the question whether TI is a realism or idealism. I would like to use this opportunity to clarify my intention: A labelling-game is not what I aimed for. The conceptual scheme of ‘realism/anti-realism’ should rather be a means with which to explain Kant’s position clearly and exactly.

Let me consider now Blomme’s rejection of the traditional schemes of the two-world and two-aspect interpretation (see Section 2 of BC), which is related to this topic. He points out first that there are important problems with these interpretive options, and especially that the distinction between two-world and two-aspect interpretation does not “entail a perfect disjunction”. From this he jumps to the conclusion that the distinction between two-world and two-aspect interpretation should be abandoned completely, without trying to consider whether this distinction perhaps can be elaborated further into a useful conceptual scheme for understanding Kant’s position, just as I tried to do in KORW.

It may appear that there is only a difference in mentality between us, namely that I prefer using old tools and amending them when their application faces difficulties, whereas Blomme prefers abandoning them. One might think that mentality is not a proper object of a rational discussion. However, I would like to point out here that such a difference in mentality has a considerable effect.

Blomme’s way of arguing definitely has a certain advantage: If a philosopher rejects a traditional distinction, she can easily show the prominence of her own position; it is not rare in philosophy that one boasts that one’s position ‘overcomes’ a traditional opposition. This way of arguing also has, however, a disadvantage that can’t be easily dismissed. It cuts off one’s position from fruitful outcomes of past discussions. I do not think that one can reasonably claim that there has been no substantial insight in the traditional debates in Kant-interpretation, for example, those about whether Kant’s distinction between appearances and things in themselves concerns distinct entities or aspects of one and the same thing. This question is important, and one must consider it before presenting one’s own position. Even if one rejects both alternatives in the end, the old scheme proves its significance just in that one elaborates and clarifies one’s own position against the old scheme. Just for this reason we should not abandon traditional distinctions or schemes so quickly, but first try and improve them as much as possible. As for the distinction between the two-world and two-aspect interpretations, I think it is a bit premature to say that the potential of this distinction has been exhausted completely. And, let me stress that this can be true even if there is no complete agreement between interpreters about what the two-world and the two-aspect interpretations really are.

In KORW, I did not simply repeat the ‘standard’ definition of these interpretive options, but rather proposed my own definition to make these options useful for my interpretation:

The opposition between the two-world and the two-aspect interpretation is about the question whether ‘appearance’ and ‘thing in itself’ are two aspects of one and the same particular thing (which we cognize empirically) or not. The two-world interpretation (which is sometimes called ‘two-object interpretation’) answers this question in the negative and claims that ‘appearance’ and ‘thing in itself’ are two distinct entities (or at least not two aspects of one and the same thing) (KORW, p. 67, italics added by me here).

Admittedly, this definition is not the same as the one that is widely accepted. The reason why I proposed this definition is that I believe that the essential opposition between advocates of what is generally called the ‘two-world’ and ‘two-aspect’ interpretation concerns the question whether ‘appearance’ and ‘thing in itself’ refer to aspects of one and the same particular thing or not. And I think this definition makes the two-world and the two-aspect interpretation indeed exclusionary or contradictory oppositions.

Now, if Blomme wants to refute my claim and to maintain that the distinction of these interpretive options does not make up an exclusive disjunction, he must consider my definition of them. But he fails to do so. See his following statement:

In my view, Kant’s theoretical philosophy does neither allow us to affirm the existence of a ‘second world’ (that is, a ‘world’ of things in themselves) […] (BC Section 2; my italics).

However, I explicitly stated in KORW, p. 68 that two-world interpretations in my definition need not assume the existence of things in themselves. I also explicitly requested the reader to understand the term ‘two-world interpretation’ in KORW just as I defined it (ibid.).

Let us then consider Blomme’s claim (2) that my distinction between realism and anti-realism is not exclusive, in that it allows some cases which are both realism in some sense and anti-realism in another sense. He argues for this claim in two places: (a) in the last paragraph of Section 1 (“This brings me to my next point  […]”) and (b) in Section 3. I will discuss them in turn.

(a) In the last paragraph of Section 1, Blomme mentions what I called the “determination thesis”:

How spatiotemporal objects are constituted (in other words, which propositions relative to spatiotemporal objects are true and which are false) is already determined before all actual realisation of cognitions, through factors which are independent of such realisation (KORW, p. 383, trans. mine).

I argued in KORW, p. 384–9 that this thesis seems to amount to realism about spatiotemporal objects, but in fact it does not, since it is compatible with some versions of anti-realism, when ‘realism’ and ‘anti-realism’ are taken in my sense. Now, Blomme does not examine this argument and simply asserts his opinion:

That this thesis is incorporated in the anti-realism that Chiba takes to be Kant’s position shows in my opinion how right Kant was to speak both of (some kind of) idealism and of (some kind of) realism with regard to the existence of spatiotemporal objects (italics added).

Well, since he does not follow my definition of realism and anti-realism, he can freely say that Kant’s position is some kind of idealism and some kind of realism in his sense. However, since he discusses neither my definition nor my argument, his objection has nothing to do with my claims. Moreover, he does not determine concretely what kind of realism and what kind of idealism he thinks that Kant’s position is, especially regarding the determination thesis. This is presumably due to his sceptical attitude towards the interpretive relevance of the distinction between realism and idealism in general. However, as long as he does not clarify his point, it is not possible to evaluate how much of a challenge his criticism poses to my conception.

(b) In Section 3 of BC, Blomme criticises my argument for transcendental affection (i.e., affection through things in themselves). Strangely, what he examines is not what I advanced as my argument, namely “argument T” (KORW, p. 351ff.), but what I presented as an insufficient one (the “Kantian” argument). This point, however, does not affect what I will discuss in the following.

Blomme’s argument here—as far as the alleged inexclusiveness of my distinction between realism and anti-realism is concerned[1]—can be summarised as follows: My thesis is that Kant’s TI is a kind of anti-realism about spatiotemporal objects, that is to say, TI regards spatiotemporal objects as dependent on our cognition. But I overlook, according to Blomme, that in the Kantian scheme there are two different kinds of spatiotemporal objects, namely phenomena and mere appearances: The former are what have “been determined with respect to all categories” (Section 3 of BC) so that they are dependent on our cognition; the latter are “undetermined object[s] of empirical intuition”, that is, “merely the empirical manifold in intuition, as being still undetermined with respect to the dynamical categories” (ibid.), so that they cannot be said to be dependent on our cognition in the full sense. Thus Blomme concludes:

As a consequence, I cannot agree with the formulation of Chiba’s leading thesis, namely that the fact whether spatiotemporal objects exist or do not exist independently of our cognition determines whether Kant is realist or anti-realist.

This argument deserves serious consideration; and I am very thankful to Blomme for this objection. I confess that I did not consider the problem of mere appearances when I wrote KORW, and additionally that I still have no elaborate interpretation of this topic. I suggest only the following points as a reply: (α) This problem may demand some revisions of my main thesis (that Kant’s TI is a version of anti-realism regarding spatiotemporal objects) but no essential modification. (β) I doubt that the “mere appearances” constitute a determinate class of spatiotemporal objects, if they are properly understood. That is to say, it may be that they are not objects that should be ontologically distinguished from phenomena, but rather denote an intermediate phase in the process of cognition, so that they are distinguished from phenomena only epistemologically.

(α) I think that Blomme overestimates the significance of his objection. If the characterisation of mere appearance as “undetermined object of an empirical intuition” (A20/B34) should be taken literally—although this is not uncontroversial, as will be argued below—mere appearances are objects of a very peculiar kind. They are certainly not objects of outer experience, but neither objects of inner experience (because for something to be an object of inner experience, it must be determined in a temporal relation to objects of outer experience; see Refutation of Idealism in the B-version of CPR). This means that mere appearances do not belong to the single universal objective time which we construct together with objects of outer as well as inner experience. This means further that mere appearances do not belong to the spatiotemporal world which we cognise. If we acknowledge these things as entities that are ontologically distinguished from determined spatiotemporal objects, they must be in another time. Is it, however, acceptable at all in the framework of Kant’s philosophy to assume something like ‘another time’?

Now, if Blomme is right in assuming that mere appearances are ontologically distinct entities, I have only to modify my main thesis in the following way: Kant’s TI is a kind of anti-realism of spatiotemporal objects qua objects of outer or inner experience. I believe that this modified thesis is still informative for understanding Kant’s ontology of spatiotemporal reality and, more importantly, still very controversial in current Kant interpretation. Mere appearances are odd entities in any case. I think it would be an exaggeration to say that an interpretation loses its point completely if it cannot explain the ontological status of such odd entities.[2]

(β) We have to ask here, however: Is it really correct to say that mere appearances constitute a class of objects which are spatiotemporal but independent of our cognition? I hereafter refer to this claim as ‘(P)’. I am not sure about the answer to this question, but certain about the following two things: First, it is highly doubtful that (P) is really the case, and secondly, Blomme advances no argument for this (at least in BC).

As for the first point, the above consideration has already given a good reason to doubt (P). If mere appearances do not belong to the single universal objective time, what time do they belong to? Is there another time which is dependent on the form of our intuitions and yet separated from objective time?

There is another question, too: Are mere appearances sensations or some undetermined objects which are spatial entities and thus distinguished from sensations? In the first case, I do not think that they would be ontologically distinguished from objects of inner experience. True, they are not yet determined, but they are nevertheless determinable in a certain way—call it “F-way”, and this means in the systems of anti-realism, which I presented as candidates for Kantian anti-realism in Chapter 7 of KORW (i.e., ZN, ZR+, ZN+, ZN+ mit SA and ZR+ mit SA), that they are ontologically determined in the F-way, before we actually cognise them as F. In this case, mere appearances are ‘undetermined’ only in the epistemological sense that they are not yet determined in our cognition, that is, not yet cognised concretely. In the second case, we can construct a similar argument: ‘undetermined’ spatial objects are not ontologically distinguished from objects of outer experience, and they are ‘undetermined’ only in the epistemological sense that they are not yet determined in our cognition. In each of these cases we need admit neither odd entities which are spatiotemporal but independent of our cognition, nor something like ‘another time’ which is separated from the single universal objective time.

The discussion so far is not conclusive, I confess, but it would be enough to show the following: It is not obvious that the case of mere appearances as ‘undetermined’ objects is a counter-example for my main thesis that Kant’s TI is a kind of anti-realism of spatiotemporal objects. If one wants to claim as Blomme does, one must provide additional arguments. And I can find no such argument at least in BC.

2. Things in Themselves

Section 2 and 3 of BC deal with the problem of things in themselves. I have already discussed the main point of Section 3, so it is only Section 2 that remains to be discussed. Unfortunately, Blomme’s objections in Section 2 fail to relate to my claims and arguments. In the following, I first discuss his main argument and explain how his objections miss the target, and subsequently comment on some particular points.

The main argument of Section 2 could be summarised as follows: [Step 1] In the Amphiboly-appendix, Kant maintains that form precedes matter, as far as appearances are concerned. [Step 2] It follows from this that it is not the case that there must be some ‘transcendental matter’, that is, things in themselves, underlying appearances. [Step 3] Like “almost all scholars who tried to give a valid interpretation of the status of things in themselves”, I overlooked this point and falsely concluded that there must be things in themselves.

Contrary to what Blomme assumes, I completely agree with Step 1. As for Step 2, I think that Blomme’s argument for this step is flawed (I discuss this point later), but I agree with his conclusion that the existence of things in themselves cannot be proved simply via the argument that appearances require things in themselves underlying them. This explains why Step 3, and hence Blomme’s entire argument, does not refute my specific claim and argument for the existence of things in themselves. Nor did I argue, contrary to what Blomme assumes, that “there must be some kind of ‘transcendental matter’, some kind of ‘determinable something’ that precedes the forms of intuition”. The following point has nothing to do with my argument for the existence of things in themselves either: “[N]o thing in itself is needed for the possible realisation of the forms of intuition as forms of empirical objects” (my italics). The topic of space and time as forms of our intuitions is irrelevant for my argument in question.

Blomme may respond: What “almost all scholars”, me included, have overlooked is not just the point that space and time as forms of our intuitions do not require the existence of something like ‘transcendental matter’—everyone would accept this, I think—but rather the fact that Kant indeed maintains in the Amphiboly-appendix, for whatever reason he may have, that we need not assume the existence of things in themselves. If this is what Blomme really intends to say, we find further textual support for this claim, for example, in A286ff./B343 from the Amphiboly-appendix, and also A252ff., B308ff. and A254–6/B310–12 from ‘Phenomena and Noumena’.

I did not ignore these passages; I even discussed them on p. 367ff. of KORW. After I examined what argument Kant advances for this claim and whether this argument is persuasive, I concluded that this claim is not well-grounded, and less persuasive and less significant than other factors which affirm, even require, the existence of things in themselves. These other factors do not only include individual passages, but also, and more importantly, one of the central elements of Kant’s theoretical philosophy, the receptivity of our empirical cognition. I would like to stress that when dealing with Kant, one should not treat a few individual passages as if they were alone decisive for a particular interpretation.

Now, let us move to two particular points in Section 2 of BC.

(1) As I mentioned above, I think that Blomme’s argument for Step 2 is flawed. His argument is surprisingly simple: “[The insight that form precedes matter regarding appearance] entails that the all too natural thought, that some ‘transcendental matter’ has to lie at the ground of appearances, is false” (my italics). I do not see any entailment relation here, nor does Blomme explain why and in what sense the insight in question entails the “all too natural thought”. In any case, more argument and explanation is necessary for justifying this step.

(2) Blomme argues that he advocates an interpretation of things in themselves which combines the interpretive options which I called (I) and (IIa) (see KORW p. 335ff.), despite my announcement that such a combination is impossible. Unfortunately, Blomme misunderstands my distinction from the outset. He describes option (I) as claiming that “[i]t is essentially an object of thought (in this case, the possibility that things in themselves exist has to be excluded)”. In fact, the possibility in question is not excluded; what is excluded is rather the possibility that such a thing in itself exists independently of our cognition. I made this point explicit in footnote 471 (p. 335), but this point had been stated already in the discussion of the methodological two-aspect interpretation on p. 77ff.

I am not sure about which option Blomme’s interpretation of things in themselves amounts to. It depends on how he answers the question: Does God—as a thing in itself—exist independently of our cognition (whether theoretical or practical), if it exists at all? If he answers this question affirmatively, then he is bound to accept option (IIa), if negatively, the option (I). Anyway, the following point is noteworthy: One can be agnostic about the existence of God, but not about the question just presented, because this question concerns what one means by saying that God exists (or does not exist).

In this reply, I have focused mostly on the insufficiencies of Blomme’s arguments. I think, however, that these insufficiencies are mainly due to the limit of the length of his critique and he has more detailed arguments in mind or in his forthcoming book (Blomme 2014). I thank Henny Blomme for his examination of my book and stimulating objections.



Blomme, H. (2014), Kant et la matière de l’espace. Le problème d’une fondation transcendantale de l’expérience extérieure, Hildesheim/New York: Georg Olms Verlag.


[1] The main aim of Section 3 of BC is to criticise what I called the “Kantian argument” for transcendental affection. The central point of Blomme’s objection against this argument can be summarised as follows: This argument assumes that every spatiotemporal object exists independent of our cognition, but in fact, there are some exceptions, namely mere appearances as undetermined objects, which are indeed spatiotemporal but not dependent on our cognition (because they have not yet been determined through cognition). The main point of this objection is that the existence of mere appearances makes my interpretation invalid, like the main point of Blomme’s objection against my distinction between realism and anti-realism which is discussed in the main text.

[2] Furthermore, it would be possible for me to respond to Blomme’s objection by slightly modifying the definition of realism and anti-realism in the following way: Realism claims that objects exist independently of our cognition and consciousness; anti-realism denies this. However, I doubt that such a modification is really necessary, as will be argued below.

Acknowledgements: I am very grateful to Martin Sticker for his thoughtful comments on my arguments as well as for his advices for improving the text.

© 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.