SILVAN IMHOF | Der Grund der Subjektivität. Motive und Potenzial von Fichtes Ansatz | Schwabe, 2014 


By Dietmar H. Heidemann and Oliver Motz

One of the greatest virtues one can praise a book on transcendental philosophy for is the way in which the author manages to interweave and balance systematic and historical considerations. In the case of Silvan Imhof’s Der Grund der Subjektivität, this concerns not only the style but just as much the essence of what is advanced in its 267 pages: Imhof wants to read Fichte’s account of subjectivity—the book’s central topic—not as a completely independent systematic enterprise that would take the absolute ‘I’ as an unquestioned, paradigmatic starting point (pp. 21–2), but rather as a very specific response to a very specific problem: the sceptical inquiries of Maimon and Aenesidemus Schulze directed at Kant and Reinhold.

Part I of the book, entitled ‘Die skeptizistische Herausforderung der Transzendentalphilosophie’, therefore gives a historically grounded but at the same time very systematically argued account of two sceptical worries that Imhof then later portrays as being the motivation for Fichte’s new account of subjectivity: The problems of facticity and application. The former is Imhof’s label for Maimon’s contention that transcendental philosophy—while it may very well show that we must necessarily believe our representations to be referring to actual being—lacks the means to establish their factual referring (p. 27). The problem of facticity, in other words, questions the objective validity of representations, including the categories. The problem of application, on the other hand, is to explain where this deficit of transcendental philosophy stems from, namely the heterogeneity of the two stems of cognition established by Kant’s First Critique. If sensible intuition and understanding, i.e. the matter and form of cognition, are of completely different origin, there is no way to establish a priori the applicability of concepts to sensible data (p. 38). Any a posteriori appeal to actual pieces of cognition would be to beg the question that transcendental philosophy, according to Imhof’s Maimon, is supposed to answer.

Portraying furthermore the general approach of Reinhold’s more systematic take on both of these problems as well as the reasons for its failure, Part I of Imhof’s book draws a clear and instructive picture of the systematic and historical context that nurtured the development of Fichte’s early Science of Knowledge. Even by itself it would make for an excellent read for anyone seeking to understand the state of early transcendental philosophy prior to Fichte. One should note at this point that Imhof has obviously put considerable effort into his text in order to make it readable and accessible even to those who might not be overly familiar with the historical debates he draws on not only in this first part.

1. Conditioned and Unconditioned Positing as Presupposing

Part II is intended to show the potential of Fichte’s approach with regard to the challenges laid out in Part I. Imhof’s overall line of reasoning proceeds in three steps. First, he seeks to establish a catalogue of criteria that an account of subjectivity must fulfil in order to be in any position to overcome the problems of facticity and application. Second, he shows the deficits of what he calls “Tatsachenmodell der Subjektivität”, i.e. the model or conception of subjectivity as a fact: As long as subjectivity is understood as an external relation among things it will always be, Imhof claims, susceptible to sceptical attacks. In a third step Imhof tries to show how these deficits call for a different ontological status of the subject, namely not as a thing, but as an acting.

During the first of the aforementioned steps, Imhof gives an intriguing reading of §1 of the Foundations of the Entire Science of Knowledge: He chooses to reframe Fichte’s take on the sceptical challenge in a terminology of presuppositions in a rather straightforward Strawsonian sense. This reframing plays a crucial role in his overarching line of argument for the superiority of the concept of Tathandlung over the model or conception of subjectivity as a fact.

It is not entirely clear whether Imhof’s approach as a whole hinges on the validity of this exact reframing or whether another theory and terminology could serve the same purpose. Still it seems worthwhile to examine to what extent this reconstruction actually manages to give a clearer picture of Fichte’s exposition of his first principle in §1 of the Foundation. This relatively short piece of text is notorious for its interpretative difficulties attributed by numerous scholars to flaws in Fichte’s text itself. Hartmut Brands (2003), for instance, holds that it is impossible to give an interpretation of the beginning of the Foundation that would at the same time be adequate and consistent.[1] Criticism of a reconstruction such as the one undertaken by Imhof may thus turn out to actually be a criticism of Fichte’s text itself rather than of its reconstruction.

Accounts of transcendental arguments as exploiting presuppositions made by statements the truth of which no one would deny have a rich tradition.[2] While Imhof does not explicitly subscribe to a specific theory of presuppositions, it can be taken from certain remarks (such as footnote 46, p. 97) that he draws on the concept as introduced by Strawson (1950) and more concisely formulated in Strawson (1952) as:

[A] statement S presupposes a statement S′ in the sense that the truth of S′ is a precondition of the truth-or-falsity of S. (1952:175)

Whenever Imhof speaks of presuppositions in general (e.g. p. 102), that is, without special regard to Fichte or scepticism, these presuppositions are existential: A sentence about an object presupposes the existence of that object. The mere presupposition made by a statement Imhof also calls mere intentional reference, while factual reference means the case where the presupposition is met, that is, the intentional entity actually does exist. Only in this latter case can the sentence have a truth value. The statement:

<KING> The present King of France is wise

presupposes the existence of a present King of France, and its having a truth-value depends on whether this presupposition is met, that is, whether there actually exists a present King of France.

2. Sceptical Challenge in the Terminology of Presuppositions

At no point in his book does Imhof explicitly reframe the sceptical challenge in this terminology of presuppositions, even though he has recently done so in Imhof (2016). But even without this further reference there can be little doubt as to how Imhof sees the problem of facticity, i.e. the sceptic’s doubt concerning the objective validity of cognition, in terms of presuppositions. Since he considers the objective validity of judgements as a matter of their referential success, this referential success is here equated with the factual truth of the presupposition made by a given judgement.[3] The sceptic’s worry then is as follows:

1. The existential presupposition of a judgement, i.e. its merely intentional reference, cannot guarantee its factual reference, for the latter depends on something independent of the former.

2. But a sentence (or judgement) can only be true-or-false, i.e. objectively valid, when it factually refers to something (p. 102).

3. Therefore, there can be no a priori, and a fortiori no transcendental, argument for the objective validity of cognition.

Imhof’s sceptic concedes that transcendental philosophy manages to establish the necessity of intentional reference to what our judgements are about in order for them to be true or false, i.e. their making certain presuppositions. But the problem of facticity persists, for the sceptic denies that transcendental philosophy offers any way to ensure a priori that this reference is also factual, i.e. that these objects do actually exist as required by the presupposition.

3. Positing=Presupposing?—A Terminological Worry

A closer look at Imhof’s reconstruction of Fichte’s exposition of the first principle in the Foundation seems to reveal a problem that, while perhaps terminological, may very well point to a deeper problem in Imhof’s take on §1 of the Foundation. Imhof’s central, and quite promising, idea here is to understand Fichte’s talk of ‘positing’ as “making an existential presupposition” (p. 102), which is to offer a whole set of interpretative devices for Fichte’s numerous uses of this term. Imhof backs up this conceptual equation by a reference to Kant’s Doctrine of Elements in the Critique of Pure Reason (B625), where Kant does in fact speak of ‘positing’ in a way that may be understood as presupposing in a Strawsonian sense. Already at this stage, however, it seems questionable to take all instances of positing to be cases of presupposition. For ‘positing’—in Kant as much as in Fichte—would not only apply to presupposed existence, but also to explicitly existential judgements such as “There is a present King of France”. Here the existence of a present King of France, while posited, is certainly not presupposed but asserted.

A more worrisome terminological inconsistency that seems to challenge at least a universal equation of positing and presupposing, arises where Imhof applies this reading to Section 2 of Fichte’s exposition of the first principle (SW I,39). Fichte here famously attributes a faculty of “unconditioned positing” to everyone who accepts the certainty of the general form of judgement, A=A.[4] Imhof accordingly introduces two terms to further qualify different kinds of positing understood as presuppositions, namely ‘conditioned’ and ‘unconditioned’ (Chapter 1, Section 2). Imhof surely wants these terms, ‘conditioned’ and ‘unconditioned’, to be contradictory qualifications, such that any given presupposition is either conditioned or unconditioned. It appears, however, that these terms figure in at least two different oppositions.

Let us now see how Imhof introduces this pair of terms: Fichte, he claims, takes on the sceptical challenge as presented above, by attacking premise (1) of the above-presented argument, that is, by producing “a special case of positing understood as existential presupposition” (p. 102), where the factual reference of a judgement is automatically given together with its intentional reference. The first principle is supposed to be that case. Since he chooses to call this special case “unconditioned positing”, it would seem obvious to reserve the term ‘conditioned’ for the ordinary cases, in which the conditions for the presupposition’s facticity are independent of the mere making of that presupposition (as in <KING>). Let this be opposition #1 between the terms ‘conditioned’ and ‘unconditioned’, where these terms qualify the epistemic relation between intentional and factual reference.

In contrast to this first opposition, however, Imhof’s subsequent analysis of the presuppositional structure of A=A, the general form of judgement, suggests a second, quite different opposition between these concepts. The general form of judgement, Imhof’s Fichte claims, does not presuppose the existence of A, but rather that

<PSP> “If A exists, then A exists.” (SW I, 93; Fichte 2003:94)

Imhof stresses that this explication given by Fichte is not a logical transformation of the general form of judgement (A=A), but rather a semantic explication of its conditioned presupposition (p.  103, 103n.57). Here it seems that Imhof does not use the term ‘conditioned’ in the sense of opposition #1. His stressing of the distinction Fichte makes in this context between “There is an A” and <PSP> suggests that ‘conditioned’ here means the conditional structure of <PSP>: If…then. In this picture, then, the opposition between conditioned and unconditioned presuppositions would not be a matter of the external, epistemic relation between the statement (here A=A) and its presupposition (as in opposition #1), but rather an internal relation within the presupposition itself, namely its conditional form. Imhof’s use of the term ‘conditioned’ thus appears to be equivocal: in opposition #1 it is contrasted with unconditioned presuppositions, the facticity of which is guaranteed by their mere being made; but on this second opposition it means the conditional relation between two existential statements within a presupposition (“If A exists, then A exists”) as opposed to the simple structure of “There is an A”.

Now one might take this observation to be a merely superficial problem that does not seriously affect Imhof’s analysis as a whole as long as it could be fixed by a more differentiated terminology. One could try to distinguish his two uses of the words ‘conditioned’ and ‘unconditioned’ by qualifying them as epistemic (opposition #1) and internal (opposition #2); from here on we shall also use numerical indices to indicate how the terms are to be understood. Then one would take it that <PSP> is internally conditioned, but the question remains whether it is so epistemically with respect to A=A. This latter question now is of importance when Imhof feeds the conditionedness of <PSP> into the following argument:

Steht die Wahrheit des Identitätssatzes bereits dann fest, wenn die Existenz von A nur präsupponiert oder bedingt gesetzt wird, hängt die Wahrheit des Identitätssatzes nicht von der Existenz von A bzw. von der Faktizität des Bezugs von ‘A’ ab. Fichte schließt daraus, dass noch eine weitere, nun aber unbedingte Setzung, also eine erfüllte Präsupposition mit dem Satz verbunden sein muss. (p. 103)

Assume “conditionally” (bedingt) is to be understood here in the sense of opposition #1. In that case Imhof here finds himself saying that A=A presupposes the existence of A just as <KING> presupposes the existence of a present King of France. This is of course in blatant contradiction with what Imhof correctly takes Fichte to be saying, namely that A=A does not presuppose the existence of A (be it conditionally1 or unconditionally1). What, according to Imhof’s Fichte, A=A does presuppose is <PSP>, which is conditional2. It appears that Imhof here falls prey to his equivocal use of the word ‘conditioned’.

Now, to be fair, Imhof himself actually expresses this very contradiction in the quote above. The general form of judgement, A=A, (conditionally) presupposes the existence of A, that is, its capability of having a truth-value depends on the existence of A; but its truth does not depend on the existence of A. Surely, Imhof cannot have overlooked that taken at face value this seems contradictory, since whatever is necessary for the truth-or-falsity of a judgement is a fortiori necessary for its truth. So it would seem that at this point Imhof really uses ‘conditioned’ in the internal sense (#2). But then, again, what about the judgement’s (<PSP>) relation to A=A?

What would it mean anyway for the judgement A=A to presuppose the judgement “If A exists, then A exists” (<PSP>)? The question is far from trivial since the paradigmatic cases of existential presuppositions, such as in <KING>, are simple cases of existential statements with no further logical structure as the implication in <PSP>. Applying an understanding of presuppositions as sketched above, Imhof’s Fichte would claim here that in order for A=A, a tautology, to be either true or false it must be the case that <PSP>, another tautology, is true. This is however far less obvious than it is in the paradigmatic cases and would be in further need of clarification as to why exactly there should be a presuppositional relation between these judgements rather than some kind of implication. In footnote 57 on p. 103 Imhof gives a hint as to what he may have in mind here. He mentions a rule of transformation that would allow for A=A to be alternatively expressed as the following implication:

<ALT> If something is A, then it is A.

This judgement, unlike <PSP>, is not supposed to reveal a presupposition of A=A, but is supposed to be the result of a merely formal transformation. But even in this case it is hard to see how this implicative predication would presuppose the truth of the equally implicative existential judgement “If A exists, then A exists” (<PSP>). Imhof does little to render this alleged presuppositional relation intelligible. This seems to be reason enough to question whether there is such a relation between <ALT> and <PSP> altogether.

It must be stressed here that the problems discussed above concern only the first part of Imhof’s analysis of §1. His reconstructive approach fares much better where he analyses the implicative structure of <PSP>, the connection Fichte labels as ‘X’, as in turn presupposing the existence of a certain instance of A which is going to be the absolute ‘I’. If one were to assume a presuppositional relation between A=A and <PSP>, this particular instance would have to be a presupposition of the second order, that is, a presupposition made by a presupposition (<PSP>) of A=A. Nothing in Imhof’s text suggests that this is what he has in mind. It would then make more sense to understand the existence of X as a direct, first-order presupposition made by the general form of judgement. After all, a connection like X could be taken directly from the implicative form of <ALT>.

While certain details in Imhof’s account remain unclear, the general line of reasoning embraced with the terminology of presuppositions is sound and illuminating on many occasions. His analysis of the particular instance of the general form of judgement “I am I” as unconditionally1 presupposing the existence of the ‘I’ by virtue of its mere form is one of them. Equally convincing is the catalogue of criteria that is the result of his analysis (pp. 106–10): The subject of the intended special instance of A=A must be (a) presupposed by the mere connection within the judgement, (b) this presupposition must be factual and (c) this facticity must be guaranteed by the mere presupposition. Only an account of subjectivity as fact-act (Tathandlung), he then goes on to argue, meets all of these criteria.

4. Transcendental Philosophy, Discursivity and Intellectual Intuition

Subjectivity as fact-act is a systematic element of Part III of the book, which is devoted to intellectual intuition (intellektuelle Anschauung). Here the author addresses the problem of application, that is, the problem of the two stems of cognition in Kant (see above, at the outset of this commentary), and once again argues for the claim that in order to account for transcendental philosophy the focus must be on scepticism. Fichte is primarily, although not exclusively, occupied with sceptical attacks on knowledge or cognition. He takes up, Imhof argues, a generic problem already familiar to Kant. Now Imhof’s overall strategy is to attribute to Kant what he calls the “Tatsachenmodell der Subjektivität”, that is, the model or conception of subjectivity as a fact. For the very reason of this ill-conceived model or conception Kant is not able to rebut sceptical attacks against the Transcendental Deduction. By contrast, Fichte is able to defend the Deduction’s main claims with the help of the concept of Tathandlung (fact-act), i.e. intellectual intuition, a cognitive capacity Kant rejects on principal grounds (pp. 145–7). Since Fichte sees himself as a follower of Kant, the crucial question then is, as Imhof correctly points out, how he can show that intellectual intuition is compatible with Kant’s transcendental idealism (p. 149).

It is uncontroversial that Kant does not allow for intellectual intuition because human intuition is sensible and cannot merge with conceptual capacities. The crucial point is ‘discursivity’ (pp. 164–6). Cognition, for Kant, is discursive, and therefore intuition and understanding cannot cooperate in a non-discursive way. In order to demonstrate that this is one of the major deficiencies of Kant’s philosophy Imhof turns the tables and argues that, on his reading of Fichte, intellectual intuition is a precondition of discursive cognition such that in order to have discursive cognition cognisers must have intellectual intuition. In the broad sense of the word, Imhof’s argumentative strategy is transcendental since his aim is to show that any discursive cogniser who denies intellectual intuition must presuppose it given the discursivity of his or her knowledge claims. This is a striking claim. In what follows the focus therefore is on Imhof’s presentation of this strategy, that is, on his understanding of (Kantian) discursivity and the connection he establishes between (Fichtean) intellectual intuition, on the one hand, and (Kantian) concept and intuition, on the other.

According to Imhof, Kant holds the “principle of discursivity” (Diskursivitätsprinzip) which is the thesis that cognition is discursive because sensibility and understanding are two heterogeneous faculties that are both indispensable for cognition, that is,

dass für alle sachhaltige Erkenntnis sowohl der Verstand als auch die Sinnlichkeit erforderlich sind. Alle Erkenntnis ist diskursiv, weil Sinnlichkeit und Verstand zwei heterogene Vermögen sind, die jedes für sich keine Erkenntnis generieren, aber je einen konstitutiven und unverzichtbaren Beitrag zur Erkenntnis leisten. (p. 163)

Imhof backs his claim up with the well-known passage from A51/B75–6 on the blindness of intuition and emptiness of concepts. This passage, however, does not mention discursivity. What Imhof seems to have in mind is what Henry Allison calls the “discursivity thesis”. According to Allison, human cognition is discursive because “it requires both concepts and sensible intuition” (Allison 1996:6). For Allison the “discursivity thesis” is a claim about the composition of cognition whereas, for Kant, discursivity is actually a feature of the kind of concepts the human understanding is able to form and apply, namely discursive concepts. Therefore the human understanding itself is a discursive cognitive faculty, not an intuitive one (cf. KU, AA 5:406).

Thus, unlike Imhof’s contention, from the fact that human cognition is a combination of sensibility and understanding, or intuition and concept, respectively, it does not follow that it is discursive. It is discursive because the concepts the human mind can form and apply are discursive. A discursive concept is a universal (abstract) representation (repraesentatio universalis), that is, an analytic identity that must be conceived as a composite of marks. The concept ‘book’, for instance, comprises those marks that are characteristic of books such that a person possesses the concept ‘book’ if she apprehends that composite of marks in the correct way. The concept ‘book’ then allows for the subsumption of a given intuition under it, which can be expressed in the singular judgement “This is a book”. Since (human) concepts are discursive they do not directly refer to objects. For as general representations they cannot individuate singular objects by themselves but only with the help of sensible intuition. Discursive concepts rather distribute the common marks they represent over the particular. In that respect they are distributive representations that contain an infinite number of representations of different possible representations as their common mark under themselves such that conceptual relations are logical relations according to species and kind (A25/B39–40).

Discursivity therefore means that given particulars or individuals must be analysed in order to find out about what is common to them. The common mark that is abstracted by way of comparing particulars or individuals is a composite of partial representations, that is, a concept as the analytic identity of what is otherwise different (A32/B48). Hence, discursivity does not pertain to (sensible) intuition, for intuition is a singular representation, directly referential and does not imply the logical relation of species and kind. That is to say, the features characteristic of discursivity cannot be found in intuition, be it sensible or intellectual intuition, for with respect to their formal structure concept as repraesentatio universalis and intuition as repraesentatio singularis are mutually exclusive kinds of representation.

Imhof would probably not disagree on these points since he implicitly acknowledges the differences between the discursivity of concepts and the non-discursivity of intuition (cf. pp. 164–6). Nonetheless his claim is that (Fichte’s) intellectual intuition is a necessary presupposition of discursivity, or more precisely of the application and objective reference of concepts and implicitly of the justification of the objective validity of concepts and judgements (p. 172).

Now the difficulty, it seems to me, is the following: If Fichte’s intellectual intuition is genuine intuition, then everything that is characteristic of a discursive concept cannot be found in intellectual intuition. According to Imhof, intellectual intuition is both, a presupposition of discursivity as well as a constitutive element of discursivity or discursive cognition. As he describes it, Kant’s critique of the possibility of intellectual intuition concerns the kind of cognition that refers to objects independently of sensibility such that a concept refers directly to objects (pp. 173–4). As we have just seen, this is not the crucial issue for Kant. The point rather is that if one accepts that human concepts are discursive or at least that discursivity is indispensable for human cognition, as Fichte does on Imhof’s reading, then it is hard to see how intellectual intuition can be connected with or even be present in cognition as the compound of concepts and sensible intuition (p. 174). For we cannot even in principle describe how non-discursive (intellectual) intuition is constitutive for discursivity because the concepts we thereby use are discursive but would need to be non-discursive in order to be able to somehow grasp the case of an intuition that allows for direct reference of concepts. Discursive concepts, however, cannot do this job.

Imhof might object that this critique does not hold because Kant himself conceives of cognition as the combination of a non-discursive (intuition) and a discursive (concept) element. It cannot be denied that, for Kant, cognition is the product of the cooperation of intuition and concept. However, this kind of cooperation is to be conceived in terms of the subsumption of an intuition under concepts and cannot be understood as a kind of merging such that intuitive representations transfer into conceptual, discursive representations or the like. No doubt the subsumption of intuitions under concepts through the transcendental imagination is a significant challenge that Kant tries to meet with the help of the transcendental schematism. But the transcendental schematism does not lead to the transition of intuition into concept. Since, according to Imhof, Fichte’s intellectual intuition is a presupposition and even a constitutive element of all discursive cognition (“konstitutiver Bestandteil aller diskursiven Erkenntnis”, p. 173), something discursive (cognition) obviously proceeds from something non-discursive, or to put it more mildly, something non-discursive has the capacity to ground something discursive.

It is hard to see how this is possible even in the single case where it holds in Fichte, that is, the fact-act of the self-positing ‘I’ (cf. pp. 138, 205). Imhof claims that intellectual intuition only occurs in combination with sensible intuition and concepts and can therefore be isolated like Kant isolates sensibility and understanding, or intuition and concept respectively, from cognition (p. 174). This claim is rather dubious. For even if it could be isolated, how can a discursive cogniser describe intellectual intuition and its cognitive activity by means of discursive concepts? One option Imhof would have at this point is to argue that Fichte’s intellectual intuition is completely different from what Kant calls ‘intellectual intuition’. But this is not what Imhof considers, and for good reasons, because in that case it would make little sense to fruitfully relate Fichte’s conception of intellectual intuition and Kant’s critique of intellectual intuition. There are certainly relevant differences, for example, that Fichte seems to believe in the possibility of actual awareness of intellectual intuition whereas Kant conceives of it as a ‘pure’ cognitive capacity, and that Kant’s critique addresses the rationalist conception of intellectual intuition which Fichte rejects. But what both agree on is that intellectual intuition is non-discursive such that the representations intellectual intuition has are somehow directly referential.

For the sake of Imhof’s argument, let us assume that the isolation of intellectual intuition succeeds. What do we have in this case? According to Imhof, what we have achieved is intellectual intuition as pure acting that is directed towards something existing in itself although not as an object or thing (in itself) independent of consciousness in the Kantian sense. It is for this reason that Fichte’s intellectual intuition does not violate the critical restriction Kant imposes upon cognition. For, first, intellectual intuition does not deliver cognition of objects that exist independently of consciousness, and, secondly, it only relates to that acting which makes consciousness and hence experience possible (pp. 177–8). In light of what has been said above about the issue of discursivity, this reasoning is unconvincing. For intellectual intuition clearly violates Kant’s critical restriction of cognition because Fichte claims that it is a non-discursive cognitive capacity that is supposed even to constitute experience, which for Kant clearly amounts to discursive cognition. Imhof does not spell this out, that is, he does not show in detail how by a mere acting of the self-positing ‘I’ or the fact-act (intellectual intuition) discursive cognition (experience) is constituted.

On the other hand, Imhof is quite explicit about the theory of subjectivity Fichte holds. Accordingly, Fichte’s theory of subjectivity is “Handlungsontologie” (p. 179), an ontology of acting, the view that the self-positing ‘I’ is not an acting thing but a pure acting. In order to convince the reader that only the ontology of acting can provide reasons for why intellectual intuition constitutes discursivity, he contrasts it with Kant’s theory of self-consciousness. As already mentioned, Imhof attributes to Kant what he calls the model or conception of subjectivity as a fact. In particular, the model implies that the self-conscious ‘I’ is a thing and that self-intuition or recognition of the ‘I’ itself is not possible. This cognitive restriction with respect to the ‘I’ results from Kant’s view that the ‘I’ in itself, the transcendental ‘I’, affects the phenomenal ‘I’ (pp. 180–99, 219, 223). Imhof discusses, in some detail, this view as well as Kant’s theory of apperception in the Transcendental Deduction (pp. 180–203). This discussion is illuminating as such, although one might find certain claims less than clear, e.g. to attribute to the ‘I think’ the role of an intentional or propositional operator (“Rolle eines intentionalen oder propositionalen Operators”, p. 198), which remains largely unexplained.

More importantly, against the backdrop of his metaphysical interpretation of Kant’s theory of self-cognition, the author points out that this theory is threatened by sceptical attacks because Kant maintains the existence of an ‘I’ in itself that affects the phenomenal ‘I’. It is the uncognisable ‘I’ in itself (“Ich an sich”) that the sceptic can easily put in doubt and that Fichte finds a highly implausible ontological supposition (p. 216). Fichte’s main objection to Kant’s theory of subjectivity is, according to Imhof’s reconstruction of the argument, that Kant holds an ontological view of subjectivity that grounds subjectivity on the metaphysical model of affection (p. 217). The Kantian model of affection, however, belongs to the traditional ontology of things (“Ding-Ontologie”, p. 220) that Fichte rejects. Because of the shortcomings of this model Kant’s theory fails and Fichte succeeds in replacing it with his theory of subjectivity as fact-act and intellectual intuition. According to Imhof’s interpretation, Fichte’s theory has the advantage that the subject knows of itself as acting subject, that is, as producing subject and produced subject at once (pp. 221–2). The ‘I’s consciousness of itself, i.e., intellectual intuition, already is cognition (p. 221, cf. pp. 228–9). Imhof’s overall argument then asserts that Fichte’s intellectual intuition as presupposition and constitutive element of discursivity holds because unlike in Kantian self-consciousness it is not ungrounded in a vulnerable ontology of things but in an ontology of the acting ‘I’. Kant’s false ontological commitments simply result in the unacceptable model of the affecting, unknowable ‘I’ in itself. Once these ontological commitments are given up we shall succeed, like Fichte, in putting in place what Kant rejects: intellectual intuition as the epistemic ground of cognition.

This line of argument seems problematic for several reasons. First, as already mentioned, intellectual intuition is non-discursive and therefore cannot constitute discursivity. Second, Imhof’s overall argument depends on the view that Kant champions the model of metaphysical affection. Since this model cannot avoid sceptical conclusions with respect to the existing ‘I’ in itself it fails, as does the entire Transcendental Deduction. And because the Transcendental Deduction fails Kant’s cognitive restriction fails too. His critique of intellectual intuition is therefore unjustified and can be overcome by Fichte’s theory of fact-act. The strategy then seems to be that Kant fails because the model of metaphysical affection is fundamentally misconceived. But is it really the case that Kant advocates metaphysical affection? As §24 of the Transcendental Deduction clearly shows it is the understanding that affects the inner sense of the subject. The understanding, however, cannot be conceived of as the existing ‘I’ in itself or transcendental ‘I’. It is the cognitive faculty of concept formation that has no existence in or outside space and time but is a merely logical principle of thought. It is for this reason that Kant does not make any ontological claims with respect to the understanding, transcendental apperception or pure self-consciousness. Imhof is, of course, right that Kant introduces the concept of the intellectual consciousness that I am or exist (pp. 205ff.). This concept, however, does not refer to an existing ‘I’ in itself but seems to be a remnant of Kant’s Cartesianism that is as such incompatible with his Critical philosophy. Hence, affection or self-affection cannot count as an indicator of what Imhof calls subjectivity as a fact that is to be replaced by Fichte’s conception of fact-act.

There might be other difficulties involved in Kant’s theory of self-consciousness. The author’s strategy to reintroduce intellectual intuition as a presupposition of discursivity because the Kantian metaphysical conception of affection and theory of self-consciousness fails is ultimately unsuccessful.

Although the book makes several claims that seem to be difficult to defend, Imhof makes a valuable contribution to a crucial problem in Fichte’s idealism and in classical German philosophy. The book is well-written, clearly structured, and always focused on arguments. It not only provides an excellent overview of Fichte’s account of the self-positing ‘I’ and related aspects but also offers an original interpretation of that account. There is much to be learned from this book.

Invited: 21 May 2015. Received: 27 February 2017.


[1] See also Rohs (1991:51).

[2] For a critical overview and approach see Grundmann (1994), ch. 3.

[3] For the sake of simplicity Imhof chooses to exclude the distinction between the levels of language (sentences) and thought (judgements) from his considerations (p. 97, n. 46). Even though there might be reasons to think that some of the problems indicated here stem from precisely this omission, it cannot be explored here in greater detail. Following Imhof’s example, the terms ‘sentence’ and ‘judgement’ will be used interchangeably here as well.

[4] Imhof repeatedly refers to this judgement as the ‘law of identity’, while he definitely, and correctly, understands it as the form of predication S is P (see p. 96, n.41).


Allison, H. (1996), Idealism and Freedom. Essays on Kant’s Theoretical and Practical 
Philosophy (Cambridge: Cambridge University Press).

Brands, H. (1982), Cogito ergo sum: Interpretationen von Kant bis Nietzsche (Freiburg: Alber).

Fichte, J. G. (2003), Science of Knowledge,  trans. and ed. P. Heath and J. Lachs (Cambridge: Cambridge University Press).

Grundmann, T. (1993), Analytische Transzendentalphilosophie. Eine Kritik (Paderborn: Schöningh).

Imhof, S. (2016), ‘Einsturz und Neubau. Fichtes erste Grundsatzkonzeption als Antwort auf den Skeptizismus’, Fichte-Studien 43:52–70.

Rohs, P. (1991), Johann Gottlieb Fichte (München: Beck).

Strawson, P. F. (1950), ‘On Referring’, Mind 59 (235): 320–44.

—— (1952), Introduction to Logical Theory (London: Methuen).

© Dietmar H. Heidemann & Oliver Motz, 2017.

Dietmar Heidemann is Professor of Philosophy and Chair in the Institute of Philosophy, University of Luxembourg. He specialises in Kant and German Idealism, epistemology, philosophy of mind and subjectivity, and metaphysics. He has published widely on Kant and German Idealism. Heidemann is the publisher and editor of the annual journal Kant Yearbook (2009–) and is both a Member of the Board of the Internationale-Kant-Gesellschaft and a member of the Kant Kommission der Berlin-Brandenburgischen Akademie der Wissenschaften. In the latter capacity he currently works on the new Akademie-Ausgabe of the Kritik der reinen Vernunft.

Oliver Motz is pursuing a Ph.D. in philosophy at the University of Luxembourg. He specialises in transcendental philosophy, especially Fichte, and the epistemology of mediation (conflict resolution). 

Website (Heidemann)