COREY DYCK | Kant and Rational Psychology | Oxford University Press 2014


By Toni Kannisto

Corey Dyck’s Kant and Rational Psychology is something of a game changer. It is a prime example of a historically informed reading of Kant that completely revises how we ought to understand the aims, progression, and details of his arguments. The Paralogisms chapter of the Critique of Pure Reason—which refutes rational psychology or metaphysics of the soul—is infamous for its complexity and convoluted nature, riddled as it is with apparently inessential and confusing side remarks and with excursions that seem to contradict the main arguments. By carefully analyzing the conception of soul and rational psychology from Christian Wolff and the Wolffian tradition through the pre-Critical to the Critical Kant, Dyck succeeds in outlining a new thread of argumentation in Kant that utilizes and explains those hitherto neglected, shunned, or even ridiculed passages. This feat alone makes Dyck’s book an essential reading for anyone interested in, and working with, Kant’s philosophy.

Yet the purpose of this essay is to let the author meet a critic, and in that spirit I will challenge Dyck’s interpretation of transcendental illusion and argue that it faces severe systematic and textual headwinds as well as overextends the reliance of Kant’s critique on the Wolffian tradition. Rather than merely outlining these problems, I will take on the more difficult task of sketching—within the limited space of this essay—how a different interpretation of transcendental illusion avoids most of them. This interpretation, the defence and development of which would have to be taken up on another occasion, would to my mind accommodate most of Dyck’s original insight and to that extent vindicate Dyck’s core theses. An exception would be his denial that the soul is cognized in inner experience and his reading of the positive utility of the idea of soul as a regulative principle, both of which are unconvincing. Due to Dyck’s detailed treatment of Kant’s already complex Paralogisms, it is exceptionally challenging to formulate a systematic critique of his view, and I can only hope to have correctly presented it so as to, at the very least, prompt Dyck to further clarify, specify, and defend his novel interpretation.


The driving force of Dyck’s interpretation is his claim that so-called rational psychology, a term coined by Wolff, does not seek to establish a metaphysical doctrine of the soul through pure reason alone but “relies essentially upon empirical psychology” (p. 9) that makes experience its touch-stone (e.g., pp. 22, 28, 36, 45). Dyck calls this rational psychology “broad” in contrast to the “narrow” and more traditional rational psychology that “would proceed completely independently of experience” (p. 9). Against the standard view that takes Kant to criticize only the latter, Dyck argues convincingly that broad rational psychology is at least as much Kant’s target and that the full breadth and force of his critique in the Paralogisms can only be understood against this backdrop (pp. 3, 8, 18).

However, in seeking to correct the imbalance in existing interpretations that focus on narrow rational psychology, I am afraid that Dyck has replaced it with an imbalance of his own that pushes narrow rational psychology too far to the side-lines. This is not just a matter of textual focus—which may out of necessity be restricted to making the case for the importance of broad rational psychology—but a matter of consistency as well: to my mind, Dyck’s reading of the illusion and error in the broadly rationalistic psychologies of the Wolffian tradition does not always sit well either with the alleged illusion and error in narrow rational psychology or with Kant’s own writings on the matter.

Dyck’s interpretation is grounded on the surprising claim that the transcendental illusion that underlies the metaphysical error of rational psychology would be that “the I of the I think” seems “to be given empirically” in “inner intuition” (p. 15)—or that the “soul [were an] empirically given object of inner sense” (p. 208) (cf. pp. 17, 99ff., 103, 166, 184, 215, 218, 226). Although there are ways to interpret and amend Dyck’s claim so that at least some of its problems can be circumvented, as it stands his view clashes with numerous passages where (the Critical) Kant explicitly claims that in fact the ‘I’ or the soul is given as well as cognized empirically through inner sense, inner perception, and inner experience. Consider e.g. the following passage from the B-Deduction:

But how the I that thinks is to differ from the I that intuits itself (for I can represent other kinds of intuition at least as possible) and yet be identical with the latter as the same subject, how therefore I can say that I as intelligence and thinking subject cognize myself as an object that is thought, insofar as I am also given to myself in intuition, only, like other phenomena, not as I am for the understanding [i.e., as a noumenon] but rather as I appear to myself, this is no more and no less difficult than how I can be an object for myself in general and indeed one of intuition and inner perceptions. But that it really must be so can be clearly shown […]. (B155–6, trans. amended)

The ‘I’ is cognized, although only as an appearance and not as a thing in itself (see also B157–9). Similarly in the passages[1] that Kant added to the B-edition Transcendental Aesthetic, he says that “[c]onsciousness of oneself (apperception) is the simple representation of the I”, and he confirms that since apperception (mere thinking) does not “[i]n human beings” amount to “intellectual” “inner intuition”, this “consciousness requires inner perception” (B68). On this account, the mind “intuits itself […] as it appears to itself, not as it is” (B69). Kant’s subsequent clarification leaves no doubt: that this “self-intuition of the mind” represents the mind “as it appears [erscheint]” is “not to say that these objects (of outer and inner sense) would be a mere illusion [Schein]” (B69). Directly contradicting Dyck’s claim that it is an “illusion that the I or soul is empirically given” (p. 217), Kant then concludes:

Thus I do not say […] that my soul only seems [scheint] to be given […]. It would be my own fault if I made that which I should count as appearance [Erscheinung] into mere illusion [Schein]. (B69–70)

Sometimes Dyck claims that the illusion is the apparent givenness of the soul specifically as the intellectual, unconditioned, simple, unified, persisting and substantial subject (e.g., pp. 17, 103, 219). Since it is definitely true that none of these predicates can pertain to the empirically given soul (e.g., A399), perhaps Dyck should be read not as claiming that the soul is not given at all but that it is not given in the right way.[2] But Dyck does quite categorically maintain even in the case of empirical psychology that “the empirical concept [of soul…] could only be a concept of the subject drawn from the illusory appearance of the soul as given in inner experience” (p. 224).[3] Since the metaphysical predicates listed above play no role in empirical psychology, Dyck must mean that the soul is an illusion tout court.


Kant begins the Transcendental Dialectic precisely with a discussion of the difference between appearance and illusion and emphasizes that in the Dialectic the topic is illusions, not appearances—which are legitimate objects of cognition (A293/B349–50). Furthermore, Kant explains that the Dialectic treats transcendent ideas of pure reason, which are grounded on a principle that deliberately “takes away [the] limits [of possible experience]” and “indeed bids us to overstep them” (A296/B353). To the ideas “no congruent object can be given in the senses” (A327/B383); their objects are objects “of pure understanding” (A327/B384).[4] Since transcendental illusion is effected by these principles, it seems implausible at the outset that it could consist in objects of the ideas even just seeming to be sensibly and empirically given. Indeed, whenever Kant says that the ideas are only ideas and do not denote an actual object, this supposed object is a thing in itself.[5] This suggests that the illusion is that the object of the idea would exist as a thing in itself, not as a seemingly empirical object or appearance. One also wonders how Dyck would account for the practical use of these ideas, which certainly points to noumena.

I do not ultimately understand what Dyck really means when he claims that the illusion is that the unconditioned, substantial soul or I seems to be empirically given in inner intuition, perception, or experience. Senses cannot distinguish between the conditioned and unconditioned—this is a matter of reason (A307–8/B363–4)—so how could something appear in experience as if it were unconditioned? (cf. A326/B383, A307–8/B363–4). Or how could it seem that something is given as a substance in the senses, when substance as a category is not given in experience at all but put into it through the synthetic activity of our understanding? Kant agrees with Hume that we do not and could not perceive causality, so nothing can really appear in perception as if it were a cause—rather, Kant claims to have shown, causality is thought in the object. How can reason in general force inner sense to seemingly perceive something it cannot even in principle perceive?

So when Dyck makes the repeated claim that the illusion is that e.g. “the soul should appear to be the persisting object of inner experience” (p. 221), the question arises: if there is nothing persisting in inner experience (A381, B412–13), how could the soul (even illusorily) seem to be perceived as something persisting in it? The role of persistence is, as Dyck recognizes (p. 98), that it allows us to apply the concept of substance to outer appearances: by perceiving that things retain some properties across time, we are allowed to judge that there is an underlying substance to which the properties belong as accidents (A182–8/B224–32). Exactly because, according to Kant, we do not perceive anything persistent in inner sense (in pure time), persistence cannot ground the judgement that a substance underlies inner representations (A365). Kant’s point does not seem to be that we perceive something persistent where there really is none, but that we do not find any perception of something persistent in inner sense.

Dyck does refer to passages that he takes to confirm his view, the “most unequivocal” (p. 99) of them being in the Prolegomena:

Now it does seem [scheint] as if we have this substantiale in the consciousness of ourselves (i.e., in the thinking subject), and indeed have it in immediate intuition; for all the predicates of the inner sense are referred to the I as subject, and this I cannot again be thought of as the predicate of some other subject. It therefore seems [scheint] that in this case completeness in referring the given concepts to a subject as predicates is not a mere idea, but that the object, namely the absolute subject itself, is given in experience. But this expectation is disappointed. (AA 4:334; trans. amended, first emphasis mine)

As the emphasized sentence shows, however, the reason for it appearing as if there were something substantial in apperception is that the subject to which all the predicates of inner sense refer cannot itself be thought as a predicate. Thinking is what reason does. The object as “an object of inner sense” (ibid.) is indeed given (this is not the illusion), but its substantiality is inferred. Thus this passage need not be understood as claiming that the substantiality of the soul appears as if it were given in inner experience, but only that the soul is given in inner experience and then thought by reason as substantial etc., i.e., as the absolute subject.

In fact, I think Dyck’s best reference (pp. 12, 85, 166, 229) is to the passage in the A-Paralogisms where Kant states that “nothing is more natural and seductive than the illusion of taking the unity in the synthesis of thoughts for a perceived unity in the subject of these thoughts” (A402). It does here indeed seem that the illusion results in a “perceived unity”, i.e., in the soul seemingly appearing as a unity in perception. As far as I can tell, however, this is the only passage that does so unequivocally, and since in other passages the category (here: unity) is applied by demand of reason and not by warrant of perception, this passage (even if we grant Kant consistency) could still be read along these lines. I do perceive myself (in inner sense), and in apperception I think myself as a unity, so I take the perception of myself to be a perception of a unified self (since apperception teaches me that the ‘I’ of the ‘I think’ is a unity). This is not to say that I perceive the unity but that I perceive something that I take to be a unity due to a demand of reason.[6]

Be that as it may, Dyck’s claim about illusion is in a way the opposite of what I take Kant to say. Kant’s point, to my mind, is not that we seem to perceive something abiding in inner sense whereas nothing really does, but that we really do not perceive anything that would abide in inner sense yet we nonetheless have to think that something does. This latter is the illusion, that there would be a substantial unified simple self grounding all appearances of inner sense despite the fact that we perceive none—because reason demands it. Persistence is just the ground of cognition of substances in objects of possible experience, and thus the soul could be a substance in itself even if we cannot cognize its persistence.


While transcendental illusion is inevitable and necessary, the metaphysical error grounded on the illusion is avoidable. Transcendental illusion moreover has a positive utility in facilitating the application of the regulative principles of reason to experience; metaphysical error occurs when this subjective necessity of the illusion is mistaken for objective cognition of an object. Thus although the error may manifest itself in a variety of ways, the illusion—when correctly interpreted in its positive significance—is always the same. This is another challenge for Dyck’s conception of transcendental illusion, as it does not fit the narrowly rationalistic psychologist in any clear way: How can the attempt to proceed independently of all experience be grounded on an illusion that the soul is given in experience? It would rather seem that, on the contrary, he is free from this (allegedly inevitable) illusion and even avoids the metaphysical error. Since the illusion is natural and unavoidable, it should be the same for both narrow and broad rationalists (indeed for all rational thinkers).

Dyck does recognize the problem that it is, “to say the least, not immediately clear” (p. 229) how this illusion is compatible with the previous chapters. While he tackles the problem in the Conclusion (cf. also p. 91), the section is frankly too short and too detached from the rest of the treatise to be fully convincing. According to Dyck, the “dialectical illusion” of the narrowly rationalistic psychologist is “the temptation to mistake the indeterminate character of the consciousness of existence included in the I think for a cognition of the existence of the soul as a purely intellectual object” (p. 229). As a “variation[] of the same general form of transcendental illusion […] the narrowly rationalistic psychologist overlooks the empirical character of the I think” (p. 230). But it is indeed difficult to see how this could be even a version of the illusory appearance that the soul is given in inner experience—even though I agree with Dyck that the narrowly rationalistic psychologist does commit this error.

Indeed, I think the value of Dyck’s contribution is that he formulates correctly and precisely the error that ensues when the rationalistic psychologist misunderstands the illusion. It is just that by reading Kant and his critique through the Wolffian tradition of rational psychology Dyck seems to end up interpreting the general illusion in terms of the particular error of this type of rational psychology. Thus, this novel and insightful analysis of the error and its resolution—long neglected by Kant scholars—is unfortunately transformed into an interpretation of the illusion that ought to, but does not seem to, apply to the narrowly rationalistic psychologist as well. Fortunately for Dyck, remedying this would mostly require only changing the letter, not the spirit of his work.

I believe it is possible to identify a more general sense of illusion that would ground the error of both the broadly rationalistic psychologist (what Dyck calls illusion) and the narrowly rationalistic psychologist. Transcendental illusion can be taken to involve just some realization of the idea of soul: the illusion is that the soul as the object of the idea exists (A339/B397).[7] The idea of soul is a concept of reason, an idea of the unconditioned or absolute subject ‘I’ that is inferred from the conditioned ‘I think’ of apperception (A334/B391, A342–3/B400–1). The illusion of rational psychology is, then, that there exists such an unconditioned ‘I’. Because the rationalist—like every thinker—is subject to this illusion, out of subjective necessity he seeks to determine this purported object of the idea through the metaphysical predicates of the categories: as a unified, simple substance (A344–7/B402–6). The difference between the broadly and narrowly rationalistic psychologist would only consist in how great a role they are willing to give experience in determining the object of the idea they believe to exist—whether they determine this soul broadly both as an object of experience and reason or narrowly as merely an object of reason.

Irrespective of whether the rational psychologist relies on experience or not, error ensues, for it is a mere illusion that the object of the idea of soul exists, and it is therefore futile to attempt to determine it metaphysically. The object of inner experience cannot be the object of the idea because it cannot (under the conditions of sensibility) be the bearer of the intellectual predicates attributed to this idea—it cannot, among others, be a substance, for there is nothing abiding in inner experience, and persistence is the necessary condition of the empirical application of the category of substance (A399, A405). It is just the soul as it appears, and there are no grounds for holding that its properties would be the same as the (supposed) properties of the idea of soul. While the intellectual object, in turn, could, to be sure, instantiate the predicates of the idea of soul if there were such a thing, in the Paralogisms Kant precisely argues that the very inferences designed to demonstrate this are formal fallacies (A345–6/B403–4). These two ways are, then, two different manifestations of the error resulting from the same underlying illusory assumption: that the soul as an object of the idea exists.

I think Dyck’s insight into the broadly rationalistic psychologist’s error can be retained if we understand the latter as taking the appearance of the self (an empirical, given object of inner experience) to instantiate the metaphysical properties (unity, simplicity, substantiality, immortality etc.) that are only really thought in the rational idea of the self (the object of which would be a noumenon and not a phenomenon). It is not that the substantial soul seems to be given in inner perception but that the appearance of the soul that is really given in inner perception is taken for a substance, which amounts to a seeming of the rational kind: the rationalist does not perceive wrongly but thinks wrongly.

Dyck also correctly identifies the error of the narrowly rationalistic psychologist (although he calls it illusion). The latter takes the apodictic inferences of pure reason that utilize the idea of soul to demonstrate that the soul really is a simple, unified substance. But, as Dyck points out, in this the rationalist “mistakes the indeterminate consciousness of our existence that is only possible through abstraction from the empirical manifold for a cognition of the existence of the thinking subject itself” (p. 230, cf. B426–7). This indeterminate consciousness is the ‘I think’ of apperception (B157–9, A342–3/B400–1, B422–3n.) that is the “sole text of rational psychology” (A343/B401). It expresses an “empirical but in regard to all kinds of intuition indeterminate proposition” (B421) that indicates a “simple and in content for itself wholly empty representation I” (A345–6/B404). Yet owing to the illusion that the object of the idea exists, the narrowly rationalistic psychologist mistakes this empty thought for a representation of the putative object of this idea.[8]

These two errors are grounded on the same illusion. The ‘I think’ expresses an act of thinking (apperception) and indicates that something thinks. This “I, or He, or It (the thing), which thinks” is the wholly indeterminate “transcendental subject of thoughts = x” (A346/B404): in apperception “I am conscious of myself not as I appear to myself, nor as I am in myself, but only that I am” (B157, cf. B157–8n.). There are two ways to determine what it is that thinks, which correspond to two concepts of the soul.[9] One is the inferred concept or idea of reason through which the ‘I’ is identified as the unconditioned or absolute subject that grounds all conditioned acts of thinking. The other is the empirical concept of soul through which the ‘I’ is experienced in inner perception. The metaphysical predicates of rational psychology can justifiably be attributed to the idea of soul, but this indicates only that reason has to think the ‘I’ through these predicates, not that the I exists with them. The empirical soul does exist, since I intuit it in inner experience, but experience cannot justify attributing the metaphysical predicates to it.

Now, a transcendental realist combines the two and says: the ‘I’ that I intuit is the same ‘I’ that I think and will therefore have the properties that reason thinks in it. Although Kant agrees that the ‘I’ is the “same subject” (B155), they still “differ” (ibid.) because there are restricting conditions on how the ‘I’ can be intuited, and so as an object the empirical soul is only the ‘I’ as it appears under these conditions, whereas as an object the transcendent idea of soul oversteps these conditions and points to the ‘I’ as it is in itself. Therefore, although the ‘I’ that both concepts seek to determine is the same subject (the indeterminate thinker), it is determined as an object in different ways, and so we are not warranted in predicating of the ‘I’ as it appears that which reason would think in the ‘I’ as it is in itself.[10]

The reading sketched here has the benefit of coupling the errors of rational psychologists with Kant’s accusation that they subscribe to transcendental realism—a term that surprisingly only occurs twice in passing in Dyck’s book (pp. 192, 194n.49). A transcendental realist takes the objects of experience (appearances) for things in themselves. That the broadly rationalistic psychologist does so is clear enough: if the self as it appears in experience were a thing in itself, then it would have to be the same self as the one thought through reason. If the narrowly rationalistic psychologist similarly takes the self that he experiences (a phenomenon) for the same self as that which he thinks (a noumenon), then although he would avoid grounding the metaphysical properties of the soul on experience, he would ground the very existence of the soul itself on experience of the phenomenal self—after all, he takes the phenomenal soul to be the same soul as the noumenal one.

A transcendental realist would not question whether the object he experiences through inner sense is the same object that he thinks through reason. What he does question is whether (a) the metaphysical properties of the self can be discovered in experience a posteriori and whether (b) reason succeeds in attributing these properties to this self a priori. The broadly and narrowly rationalistic psychologists differ not in their belief that there is a soul but in what they take to be sufficient grounds for metaphysical knowledge about this soul, the former admitting (some) experience, the latter only rational demonstration.


Although my reading of the illusion requires some dismantling of Dyck’s edifice, it does not go against the gist of his view, and in no way does it affect his excellent exposition of the relevance of broadly rationalistic psychology to Kant or the value of his thorough and original treatment of the four paralogisms. Yet Dyck’s interpretation of the positive utility of the idea of soul as a regulative principle is hereby set in dubious light. Since, as I showed, Kant believes that the soul really is given as an appearance, Dyck’s claim that only the “illusory appearance [of the soul] as the empirically given object of inner sense” (p. 208) can ground the positive use of the idea of soul stands in need of revision. Indeed, there are further and greater worries with this view as well.

First, as mentioned above (II), the principles of pure reason are transcendent, they “incite us to tear down all those boundary posts [of possible experience] and to lay claim to a wholly new territory that recognizes no demarcations anywhere” (A296/B352).[11] The objects of the ideas (if any) should therefore be noumena, not objects of experience or even something that just seems to be such: “A concept made up of notions, which goes beyond the possibility of experience, is an idea or a concept of reason” (A320/B377). According to Kant, the (direct) “objective use of the pure concepts of reason is always transcendent” (A327/B383). The idea as a focus imaginarius that according to Kant facilitates the regulative use of the principles of pure reason “lies entirely outside the bounds of possible experience” (A644/B672)—Kant even says here that the illusion is the “deception, as if these lines of direction [that converge in the focus imaginarius] were shot out from an object lying outside the field of possible empirical cognition” (ibid.). How could the illusion then be, as Dyck would have it, that the soul appears to be not outside of possible experience but indeed in inner experience, empirical rather than transcendent or noumenal?

Second, as Dyck argues, the specific use of the idea of soul is to facilitate the use of the “idea of a fundamental power […] to reduce [the] apparent variety [of the powers of the human mind] as far as possible” (A649/B677). In this way it becomes “more probable” that these different powers are “expressions of one and the same power” (ibid.). Dyck claims that only the illusion that the self appears in inner experience can direct us to seek this common power underlying apparently distinct powers and so to seek the systematic unity of inner appearances. Importantly, however, this “unity of reason is merely hypothetical” (A649/B677). Since the soul does appear, in Dyck’s reading the unity of reason would be a fact, not hypothetical. Even if he means that the illusion is the soul’s appearance specifically as a substance etc., the unity of reason could not be a mere hypothesis either, since it would instead simply be false, as it is impossible for the soul to appear in inner perception as if it were substantial.

I cannot here tackle all of the extensive textual evidence that Dyck provides. In any case, the evidence remains circumstantial, for in the passages from the Critique Kant simply does not make any explicit references to the soul having to be given specifically in inner experience for the idea to function in its positive regulative role. (It is admittedly true that due to the convoluted nature of the Paralogisms, hardly any interpretation of it can boast explicit textual evidence.) Kant merely claims that the illusory object of the idea of soul—as the unconditioned simple, unified substance—is required for the systematic unity that reason seeks to attain in “explaining the appearances of the soul” (A682/B710). Because Dyck reads these passages through his general view of illusion, although this provides a novel and intriguing (albeit difficult to grasp) interpretation, it is far from “supplying indirect evidence” (p. 17) to the general view—it would rather seem to become the first victim of the problems of the latter.


Although Dyck is absolutely correct in emphasizing the centrality of Kant’s discussions about the empirical concept of soul—a topic long neglected in Kant scholarship—I cannot help but feel that this focus comes at the expense of the bigger picture. I do not find it convincing that Kant would have refuted broad rational psychology by showing that the soul is an inferred rather than an empirical concept (pp. 71, 75, 79–81). As I have shown, Kant seems to have both an empirical and a rational, inferred concept of soul. The soul is not an object of either concept exclusively: as an appearance it is an object of inner experience; in itself it is an intellectual object of the idea of soul.[12] The error of the broadly rational psychologist would rather seem to be that he fails to recognize that these two concepts originate in different faculties. The ‘I’ as I appear to myself originates in inner experience and sensibility—it is given—whereas the ‘I’ as I am in myself originates in reason—it is inferred. Both are attempts to determine the wholly indeterminate and formal ‘I’ of the ‘I think’ of apperception.

Now Kant’s main philosophical question is what we can cognize about either of these concepts a priori. Whether a concept appears in experience is not decisive: as Kant makes it clear in the very first sentences of the Introduction to the Critique, “[t]here is no doubt whatever that all our cognition begins with experience”, but while therefore “all our cognition commences with experience, yet it does not on that account all arise from experience” (B1). This is the question of a priori justification: e.g., causality may de facto first appear in experience, but that there is causality at all is not justified by experience. For Kant a judgement can be analytic (and hence a priori) despite consisting solely of empirical concepts (AA 4:267)—such as “bachelors are unmarried”. Similarly we can have synthetic a priori judgements with empirical concepts, e.g., “the solar eclipse has a cause”, which employs the category of cause.

Since the fact that “bachelors are unmarried” consists of empirical concepts does not prevent it from being an a priori judgement, the question whether the soul is an empirical concept does not decide whether there can be synthetic a priori judgements about it. How we de facto come to have the concept of soul (its empirical deduction) is beside the real issue of how and under what conditions we can justify the use of this concept—or its different versions—to objects (its transcendental deduction). For this reason it does not matter as much as Dyck seems to think whether the concept of soul is empirical or inferred. What matters is whether we can make synthetic a priori judgements about either of their objects.

By fixating on whether the soul is an empirical concept, Dyck neglects Kant’s more pertinent question of whether we can have a priori knowledge of the soul. As I have indicated above, Kant does grant that experience offers us a posteriori cognition of the soul as it appears and that reason determines a priori the properties we must think in the concept of the soul as it is in itself. This is why he claims that the tenets of rational psychology are apodictic yet analytic or identical judgements (B407–9), i.e., analytic a priori, whereas the judgements of empirical psychology as synthetic a posteriori truths are insufficient for metaphysics (A381–2). The problem is that by not distinguishing between appearances and things in themselves, the transcendentally realistic rational psychologist (of whatever stripe) takes the two (putative) objects—the ‘I’ of inner experience and the ‘I’ of the idea of soul—to be the same object, so that whatever reason determines about the latter must pertain also to the former. But since these properties can pertain only to the soul in itself and the latter is just its appearance, the two cannot be combined in a metaphysical doctrine of the soul.

Contra narrow rational psychology, reason alone cannot ground the objective truth of its judgements about the idea of soul without given intuitions (only their logical and analytic truth); contra broad rational psychology, experience cannot ground the objective truth of reason’s judgements about the idea of soul either, for although it does provide us with an intuition of the soul, it does so only about the soul as it appears, not as it is in itself. The rationalistic psychologists did not discover this error, for they took appearances to be things in themselves, thus conflating the intuition of the soul with the idea of the soul. And the reason why they never critically assessed the grounds of this identification is that they were under the transcendental illusion that the soul as an object of the idea exists to begin with. Had they noticed the illusion, they might have asked whether the assumption that this object exists could be based on experience or on reason, and through that come to see the groundlessness of the identification of the ‘ ‘of inner experience with the ‘I’ of reason and ultimately the groundlessness of rational psychology itself—as Kant eventually did.


In this critical essay I have presented some systematic problems, unclarities, and shortcomings of Dyck’s Kant and Rational Psychology. These do not undermine the essential core of Dyck’s contribution, namely the historical narrative of how and in what form rational psychology came to be the target of Kant’s criticism. Dyck’s treatment of the Wolffian tradition and the four paralogisms is excellent. He shows that Kant’s critique is far more sophisticated than previously assumed and yet succeeds in holding the threads together well enough to weave a clear picture. Dyck’s work suggests that much of the difficulty in decoding the Paralogisms stems from attempts to isolate a single line of criticism where there really are two. Although I think Dyck himself puts too much emphasis on one of these lines, the mere revelation that there are two distinct lines gives us hope that we may arrive at a cogent interpretation. It is with this hope in mind that I have targeted Dyck’s general interpretation rather than his individual arguments and sketched a reading—that can hardly be properly defended here—that to my mind takes us closer to exposing a unified illusion behind both broad and narrow rational psychology. It remains to be seen whether I have, to Dyck’s mind, made headway in amending his view or if I have, through misunderstanding, only managed to take steps back to the traditional readings that he rightly criticizes.


[1] I wish to thank Jonas Indregard for pointing out these passages to me.

[2] That is, while Dyck may be incorrect about the ‘self’ or even the ‘I’ suitably interpreted, it would still be correct to say that the ‘I’ of specifically the ‘I think’ (or the soul) cannot be cognized. But the quoted passages (and others) do in fact have Kant say that the I that is cognized via inner sense is the same I that is (merely) thought in the ‘I think’—and Kant also calls it a soul. It is, however, true that we do not cognize the I in pure thought or via apperception alone and that we also do not cognize the soul as a thing in itself. We cognize the thinking I as it appears in inner perception, not as it is in pure thought, and we cognize the soul as it appears in experience, not as it exists in itself. The doctrine of the soul as it appears Kant calls the empirical doctrine of the soul, and in both the doctrine of the soul and in the doctrine of the bodies (physiology of inner sense and outer sense, respectively) “much can be cognized empirically” (A381). Dyck does not seem to make these distinctions or have room for them.

[3] Dyck’s recurring expression “illusory appearance” is very unfortunate, for it meshes together illusion and appearance so that it is difficult to ascertain whether he takes the soul to be an illusion or an appearance. Context and other, less ambiguous formulations suggest, however, that he means that the illusion is that the soul appears—i.e., that the soul is not really an appearance but only seems to be one.

[4] See also e.g., B68, A310–11/B367, A313–14/B370–1, A320/B377, A326/B383, and B429–30 Dyck discusses this briefly on p. 82.

[5] See e.g., A327/B384, A350, A358–60, A405, B409–10, B429–30, A673–4/B701–2.

[6] This is a subtle but important point: since I know e.g. that every alteration is caused, when I perceive alteration, I perceive something caused, but I do not hereby perceive causation itself or know that the alteration is caused because I would perceive its cause.

[7] For similar characterizations of the illusion, see e.g., A339/B397, A384–5, A396, B426.

[8] Although Dyck expresses reservations as to whether it is likely that the rationalist conflates the mere empty representation of a transcendental subject of thoughts “with a putative object of intellectual cognition” (p. 228), the rationalist of this kind does in fact seem to take necessary propositions to indicate their objectivity, as Henry Allison has argued (H. Allison, Kant’s Transcendental Idealism [Yale UP, 2004], p. 340). Thus if the rationalist takes his inferences to establish necessary truths (which Kant grants him), he would also (illegitimately) take them to be about objects (something Kant does not grant) (B407–9).

[9] Although this is certainly an intricate point that splits hairs so that even Kant is unclear about his distinctions (cf. A341/B399), Kant does indicate that of the ‘I’ of apperception “one cannot even say that it is a concept, but a mere consciousness that accompanies every concept” (A346/B404). When one seeks to determine this indeterminate consciousness, one does so through concepts (as predicates in judgements), and here we can distinguish between the indeterminate non-concept ‘I’ and the attempted determination of it through the empirical and rational concepts of soul. “I am a soul” expresses a judgement where the ‘I’ is the as such indeterminate subject and the “soul” is the determinate concept that is predicated of the former.

[10] I cannot within the confines of this essay defend this interpretation in any detail. This two-fold attempt is to my mind most explicit in Kant’s distinction between the analytic and synthetic procedure of rational psychology, the latter corresponding to narrow rational psychology and the former to broad rational psychology (B416–22). In B428–30, Kant seems to distinguish two possible routes to determining the ‘I’ of ‘I think’ (as subject and object), first in pure thinking in which case “I represent myself neither as I am nor as I appear to myself” (B429) and then through inner sense, in which case “to cognize [the ‘I’] as noumenon […] is impossible, since inner empirical intuition is sensible” (B430).

[11] Dyck discusses this on p. 82, but he makes very little of the distinction between transcendental and transcendent principles. He does interpret “transcendent“ in a different way than I do, but his reference is to Kant’s unpublished note from 1774, and in any case he only uses the term in relation to the Second Paralogism and even then does not contrast it with “transcendental” (p. 123–5).

[12] The empirical soul is the soul only as it appears just like every empirical object is for Kant an appearance. That the soul as an appearance cannot be cognized in the same way as outer appearances is not due to it being less an appearance and more an illusion but due to Kant’s intricate philosophy of time and space—something Dyck discusses in depth in his book. An inner object is purely temporal and affords far less by way of cognition than a spatiotemporal object.

© Toni Kannisto, 2015.

Toni Kannisto obtained his Ph.D. in Philosophy from the University of Oslo, Norway, in 2012 and is currently postdoctoral fellow there. His areas of specialization are Kant and metaphysics and its history. Among his publications are ‘Positio contra complementum possibilitatis: Kant and Baumgarten on Existence’, forthcoming in Kant-Studien, and ‘Three Problems in Westphal’s Kant’s Proof of Transcendental Realism‘ (also in Kant-Studien). He is currently writing, together with Camilla Serck-Hansen, a book on Kant’s Transcendental Dialectic.