Hegel and Formal Idealism

Abstract

I offer a new reconstruction of Hegel's criticism of Kant's idealism. Kant held that we impose categorial form on experience, while sensation provides its matter. Hegel argues that the matter we receive cannot guide our imposition of form on it. Contra recent interpretations, Hegel's argument does not depend on a conceptualist account of perception or a view of the categories as empirically conditioned. His objection is that, given Kant's dualistic metaphysics, the categories cannot have material conditions for correct application. This leads to subjectivism in the content of experience: the subject is given an implausibly strong role in determining what is the case. Hegel's own absolute idealism solves this problem.

I. Introduction¹

Kant described his philosophy as formal idealism (B159n).² He held that both the subject and external reality contribute to experience. We give experience its form: its ordering in space and time and under the categories. By producing in us a manifold of sensation, external reality provides experience with its matter. Formal idealism thus uses the distinction between form and matter to give both idealism and realism their due.

In his 1802 essay Faith and Knowledge (Glauben und Wissen), Hegel wrote of Kant's view: ‘A formal idealism which in this way sets an absolute Ego-point and its intellect on one side, and an absolute manifold, or sensation, on the other side, is a dualism’ (78/333). Hegel held that the relation between the categories and the matter of intuition was unintelligible in formal idealism. Too much of experience is imposed by the subject, leading to ‘subjectivism’ or ‘psychological idealism’.

While it is often noted that Hegel rejects Kant's central use of the form-matter distinction—namely, the distinction between concepts and intuitions—Hegel's argument for this position remains obscure. This is largely because the mature Hegel was content to state his objections in summary fashion, without detailed engagement with Kant's texts.³ Hegel's detailed argument occurs in early texts like FK and the Differenzschrift (DFS), which have received much less philosophical attention. The present paper aims to fill this gap.

My reading of the argument against Kant in FK differs from the two major extant readings by Robert Pippin and Sally Sedgwick. Unlike these authors, I do not take Hegel's argument to turn on a conceptualist account of perception or a view of the categories as empirically conditioned. Rather, Hegel aims to show that Kant's metaphysics leads to a problem in his theory of cognition. In particular, Kant's dualistic treatment of the subject and external reality has the consequence that the categories lack application-conditions which are responsive to the features of the manifold to which they are applied—in short, material conditions for correct application. On my reading of the argument, absolute idealism is a natural response.

After briefly sketching some Kantian doctrines (II), I state the central problem (III), contrasting my view with those of Pippin and Sedgwick. I then (IV, V) reconstruct Hegel's argument against Kant in FK. Finally, (VI) I assess two lines of response and (VII) explain how absolute idealism solves the problem.

There are two points I should make at the outset. First, this is early Hegel. Faith and Knowledge predates the Phenomenology of Spirit and the logico-metaphysical system set out in the Science of Logic and the Encyclopaedia. While I will point out some continuities in what follows, the positive views Hegel held in 1802 should not simply be imputed to his later self. Second, and despite the first point, there is no reason why Hegel should have abandoned the negative side of his early critique as he developed his system.⁴ The argument I draw from FK aims to show that Kant's account fails even by its own lights. It remains of interest in assessing the prospects not only of Kant's particular variety of formal idealism but of other attempts to follow the same path between idealism and realism.⁵

II. Kant: apperception and the manifold

To make sense of Hegel's argument we need some basic Kantian doctrines in view.⁶ For Kant, cognition requires bringing together two sorts of representations: concepts and intuitions (A50/B74). These arise from distinct faculties. Concepts originate in the understanding's spontaneous activity, while intuitions are (in beings like us) receptive, the product of sensibility. Concepts and intuitions come together in a process of progressive applications of form to matter. Sensations are ordered in space and time into intuitions; intuitions are bound together by the categories into cognitions; and cognitions are unified, incompletely, by reason's search for fundamental laws of nature.

Sensation, ‘the matter of sensible cognition’ (A50/B74), is the ‘effect of an object on the capacity for representation’ (A20/B34). External reality impinges on our sensibility and produces sensations; these sensations are then arranged in space and time, the forms of sensibility, giving rise to empirical intuition.⁷ Intuitions are singular representations (Allais 2015: 147). All sensations are ordered in time, the form of inner sense, while some sensations are also ordered in space, the form of outer sense. We also have pure or ‘formal’ intuitions, which are free from sensation and are of space and time alone.

Kant often refers to intuition as a manifold, which suggests a collection of unconnected impressions.⁸ In particular, we cannot receive in intuition a representation of a complex as complex:

[T]he combination (conjunctio) of a manifold in general can never come to us through the senses, and therefore cannot already be contained in the pure form of sensible intuition. […] we can represent nothing as combined in the object without having previously combined it ourselves […] among all representations combination is the only one that is not given through objects. (B130)

This spells out the sense in which sensibility gives us a manifold: it is a set of categorially unconnected representations which has to be ‘gone through, taken up, and combined’ (A77/B102) by the activity of the understanding. Call the view that we do not receive categorial form in intuition ‘perceptual atomism’.⁹

Concepts are general representations; they have a unifying function. This is bound up with their nature as exercises of spontaneity: the unity of a concept is the unity of an act. Kant tells us that ‘concepts [rest] on functions. By a function, however, I understand the unity of the action of ordering different representations under a common one. Concepts are therefore grounded on the spontaneity of thinking’ (A68/B93). Like intuitions, concepts are empirical or pure. Empirical concepts such as ‘red’ are abstracted from experience. Pure concepts or ‘categories’ are not: they are imposed by the subject. These categories—which include unity, reality, substance, causation and possibility—provide the basic framework for experience of an objective world (Strawson 1990: Part II, ch. II).

The basic principle of our spontaneity, in its theoretical aspect, is the original synthetic unity of apperception (B134n). It tells us that for any representations P and Q, if they are my representations then I must be able to say ‘I think P’ and ‘I think Q’; and for this to be possible, I must be able to combine them in one unified experience, to ‘call them all together my representations, which constitute one’ (B135; cf. Messina 2014: 12). The basic ways that we combine representations are the categories (B144). Unity under the categories is, in turn, what gives our representations relation to an object (B137).

The act of ‘putting different representations together’ under the categories is called synthesis (A77/B103). In Kitcher's words, the function of synthesis is to ‘construct unified representations’ out of the unconnected manifold which we receive through the senses (1994: 81). Kant tells us that the a priori synthesis which unifies the manifold is carried out by the imagination, referring to it as the ‘productive synthesis of the imagination’ (A118) or the ‘transcendental synthesis of the imagination’ (B151). This synthesis in turn makes possible the (correct or incorrect) application of the categories in empirical judgment. Finally, reason seeks further unity by binding cognitions together under higher and higher laws of nature (A644/B672).

Kant's idealism is built around this account of cognition. I said above that the account can be seen as a process of progressive applications of form to matter. Form and matter, for Kant, have different metaphysical origins: the basic matter of experience, sensation, results from affection by an external reality, while all the varieties of form—space and time, the categories, and the Ideas of reason (A645/B673)—are the contribution of the transcendental subject. The reason why we cannot apprehend things as they are in themselves, but only as they appear to us, is that our cognition depends on sensible intuition. Our experience is spatiotemporal because our forms of sensibility are space and time; nothing in external reality requires precisely these forms, and Kant is clear that, for all we know, other creatures could have other forms (B150). In this sense the things we cognize are merely ‘ideal’.

III. The guidance problem

In this section I will sketch the problem animating Hegel's arguments in FK and contrast my interpretation with those of Pippin and Sedgwick.¹⁰

Hegel argues that there is no way of making sense of how the categories and the manifold of intuition come to bear on each other.¹¹ Kant's theory requires the manifold to play a role in determining the content of experience, while also being formed by spontaneity. But the categories and the manifold are conceived of as alien to one another, such that the manifold cannot play this role. That this is Hegel's worry can be seen in passages like the following:

Identity of this formal kind finds itself immediately confronted by or next to an infinite non-identity, with which it must coalesce in some incomprehensible way. On one side there is the Ego […] with its synthetic unity which, taken thus in isolation, is formal unity of the manifold. But next to it there is an infinity of sensations […] (FK: 76/332)

That is, the pure concepts of the understanding have to unify the products of sensibility, but precisely because the two factors are so different, there is no intelligible explanation of how they jointly give rise to the content of experience. And if the manifold of intuition cannot play a role in determining the content of experience, then that content is determined only by the subject—a position Hegel calls ‘psychological idealism’ or subjectivism. Let me explain.

In a priori synthesis we combine (or ‘synthesize’) the manifold of intuition in accord with the categories. For short, say that we combine the representations by giving them ‘categorial form’. Here is a simplified example to illustrate the principle.¹² Say I am given a series of representations—P, Q, R, S—in intuition. I combine them as follows: P and Q are representations of a single object, as are R and S. The object P + Q is causing some effect in the object R + S. All the representations thus form part of a single unified experience. Synthesizing the representations in this way means that certain things are true: namely, that there are these two objects in the world, with this causal relationship.

Now, it is implicit in Kant's account that there must be multiple logically possible ways of combining the deliverances of sensibility. (It might be logically possible, say, to combine the given representations as being of three objects, P, Q + R and S.) For to hold that there is only one logically possible way of synthesizing a manifold would be, in effect, to hold that subjects receive the manifold as already categorially unified, contrary to perceptual atomism. This fact leads to the requirement which drives Hegel's argument, which Pippin has called the ‘guidedness of empirical knowledge’ (1982: 46), and which I will refer to as guidance.¹³

Guidance comprises two conditions. First: among the possible ways of combining the manifold in a priori synthesis, some must be correct, while the others are incorrect, in virtue of features of the manifold which hold independently of how it is in fact combined. If this were not so, what is true and false, and which objects existed within experience, would depend solely on my spontaneity. The range of possible combinations—i.e. possible facts—would still depend on the manifold, but the choice among these possible facts would be made by spontaneity. This is subjectivism.

Second: whatever features of the manifold make some way or ways of combining it correct also explain why I combine the manifold correctly in a priori synthesis. Rejecting this condition leads to incoherence: while there would be standards for the correctness of combinations of the manifold, those standards would not play any role in explaining why experience turns out to be combined in a correct rather than incorrect way.¹⁴

Let us return to the simplified example to see what this entails. There must be something in the manifold of P, Q, R, S that justifies and explains the subject's act of combining them as of P + Q and R + S, and no similar justification for combining them in at least some other ways. Recall, however, that combination cannot be given in intuition: the manifold does not contain the relation P + Q. Nor can it contain some other relation between P and Q, say P * Q, to guide our synthesizing activity. For if the *-relation means that the subject must unify the relata into P + Q, then it amounts to receiving a representation of combination in intuition.

If the manifold is to justify and explain our combining it in some ways and not others, then, there must be something about P and Q which makes them well-suited for being combined, in a way that Q and R are not. In particular, it must be the case that the category under which P and Q are combined is associated with an application-condition which is satisfied by P and Q, but not satisfied by Q and R. This is where Kant's dualism of subject and external reality becomes problematic. Given this dualism, Hegel will argue, it is not clear how elements of the manifold can have features capable of guiding the categories; equivalently, it is not clear how the categories can have application-conditions responsive to any features of the manifold, that is, material conditions of correct application.¹⁵

Let me briefly comment on how my account differs from recent interpretations by Pippin and Sedgwick.¹⁶ Pippin identifies the concept-intuition theory as the target of Hegel's critique, but emphasizes that ‘what Hegel is out to criticize is not the distinction itself, but the way Kant understands the nature of concept-intuition unity in knowledge claims’ (2005: 26). While Hegel agrees that knowledge involves both concepts and intuitions, ‘he is objecting to a “mechanical” opposition in favor of an “organic” role for the imagination in understanding the relation between intuition and concept’ (2005: 28). But what does it mean for the relation between category and intuition to be ‘organic’ rather than ‘mechanical’? For Pippin, to treat the relation as mechanical is to treat the relata as separable; to treat the relation as organic is to treat them as inseparable. On Pippin's reading, then, Hegel's critique of Kant prefigures recent arguments by conceptualist theorists of perception against non-conceptualist theorists (Pippin 1989: 28–31). The solution is to adopt the conceptualist view that the manifold is ‘already conceptually articulated’ (Pippin 2005: 34; see also Schulting 2016; Wretzel 2018).

On my view, Hegel's objection does not depend on whether Kant was a conceptualist or a non-conceptualist. The deep dualism in Kant is not the epistemic distinction between categories and intuitions, but the metaphysical distinction between the subject and external reality.¹⁷ In so far as both conceptualists and non-conceptualists are committed to this metaphysical picture, they both fall within the scope of Hegel's argument. In particular, as I will argue below, a conceptualist account of the manifold of intuition is subject to Hegel's critique in so far as the categorial form of the manifold fails to be determined by features of external reality. The solution to the problem, therefore, is not to modify Kant's epistemology; it is to modify Kant's metaphysics. While Pippin is rightly concerned to avoid interpreting Hegel's theory as a sort of pre-Critical atavism, Hegel's critique has a more significant metaphysical dimension than is apparent in his reading.¹⁸

Sedgwick (2012) also reads Hegel as objecting to the concept-intuition theory. She observes that Hegel is not calling into question Kant's ‘insistence that our discursive form of understanding must rely on sensible intuitions as well as concepts’ (2012: 57); he is not proposing to collapse intuitions into concepts or concepts into intuitions. Rather, Hegel questions what Sedgwick calls the ‘heterogeneity’ or ‘externality’ of concept to intuition. This externality has two aspects (2012: 87): (i) that ‘concepts are not already given with the a posteriori matter of empirical intuition but must be contributed by us’ and (ii) that ‘the conceptual form we contribute cannot be known to reflect the nature of the given content itself’ (2012: 118). Together, Sedgwick argues, these claims, which Kant inherits from empiricism (1996: 571; 2012: 89), lead him to subjectivism. In my view, while Sedgwick's interpretation comes close to articulating Hegel's objection, the exact nature of the problem remains unclear in her account.

The claim (i) that for Kant, concepts are contributed by the subject, rather than given with the matter of intuition, is correct if we take ‘concepts’ to mean pure concepts. But this claim is a pillar of Kant's theory of cognition and not something that ought to worry a Kantian. The core of the problem must therefore be (ii): that the conceptual form the subject contributes cannot be known to reflect the nature of the given content.¹⁹ But it is not wholly clear what it means for form to reflect the nature of content, why it fails to do so in Kant's theory, or how this leads to subjectivism.

One thing this might mean is that the activity of giving form to the manifold gives us knowledge about the manifold or about the manifold's cause. For example, if I sort a pile of chess pieces into black and white, this form-giving activity reflects the nature of the content; if I sort them into two groups by whim, this reflects nothing about the content. If that is right, then, for form to reflect the nature of content, the content must be such that it determines, for a given form, whether the form is being applied correctly, and explains why the form is being applied correctly. In my terms, the matter has to be able to guide the application of form.

If this is the right reading of Sedgwick, then we agree on the core problem about form and matter. However, we differ on the root of the problem and on Hegel's solution to it. On my reading, the root of the problem is that the subject is too isolated from the manifold; the solution, as I will show, is a prior unity of subject and manifold. On Sedgwick's view, the problematic gap between form and content in Kant is a result of ‘the assumption that reason […] is a transcendent power’ (2012: 95), in other words ‘that some of our ideas or concepts are formed in complete abstraction from any input from the realm of the empirical’ (2012: 97). The solution is to recognize that all concepts are historically and empirically conditioned, such that critique, in the Kantian sense, is impossible: ‘it is not possible for us to abstract to a meta-level form of inquiry that in no way reflects our debt to the ordinary as well as scientific practices of our day’ (2012: 159).²⁰

I have two concerns about this reading. First, in so far as Hegel treats the categories as empirically conditioned, he is left without an answer to the question (how is synthetic a priori knowledge possible?) which motivated Kant's anti-empiricism about the categories in the first place. Second, the idea that Hegel rejects such meta-level inquiry is in some tension with the project of the Logic, which seems to acknowledge no debt to history, empirical knowledge or everyday practices of inquiry. These concerns are of course not enough to justify rejecting Sedgwick's account, but are perhaps enough to justify considering another.

IV. Absolute antithesis: Hegel's criticism of Kant

In this section I begin my interpretation of Hegel's criticism of Kant in FK.

Hegel criticizes Kant for setting up ‘an absolute antithesis’ (FK: 78/334): both the manifold of intuition and the categories are in different ways abstract and indeterminate. The manifold is difference without unity; the categories, or the ‘abstract Ego’, are unity without difference, ‘empty identity’ (71/328). Underlying Hegel's concern is the principle that for a representation to have determinate content it must have both form and matter. As Inwood writes, ‘Pure, contentless form and pure, formless matter are, on Hegel's view, the same: a wholly indeterminate abstraction’ (1992: 110). For Kant, however, the basic factors of experience are exactly these: the categories, which are purely formal, and the manifold, which is solely material. The indeterminacy of the two sides is what leads to the problem about their relation. I will begin with the manifold and then consider the categories.

IV.i. The manifold

Hegel sees the manifold of intuition as an ‘infinite non-identity’ (FK: 76/332), a collection of categorially unconnected atoms. It appears, he writes, that ‘the manifold of sensibility, empirical consciousness as intuition and sensation, is in itself something unintegrated, that the world is in itself falling to pieces’ (FK: 74/330). But it is not clear that Hegel wholly disagrees with this picture. In EL, he describes the sensory in atomistic terms: the representations we receive in sensation and intuition ‘constitute a manifold in terms of their content, but equally by virtue of their form, i.e. by virtue of the status of being outside one another as is characteristic of sensoriness’ (EL: §42). So while Hegel sees this conception of intuition as one that Kant takes over from empiricism, he does not reject all features of it.²¹ He agrees that the categories are not contained in the sensory manifold; this reflects the superiority of reason over the senses: ‘one does not hear the universal and one does not see it; instead, it is only for the spirit’ (EL: §21A). The problem is not that the manifold ought to itself contain categorial form, but that it is not able to constrain our application of the categories.

The manifold as we receive it lacks categorial form, or what Hegel calls ‘objective determinateness’ (FK: 74–75/330–31).²² This means that the manifold is indeterminate: ‘once it is abandoned by the categories, this realm cannot be anything but a formless lump’ (FK: 76/332). The elements of the manifold are disconnected from one another. It contains difference or matter. But there are no relations between the elements, so there is no unity or form. In intuition alone there is ‘no relative identity of unity and difference’ (FK: 70/327). It follows that the manifold lacks determinate content.

But is this not exactly Kant's point, when he says that intuitions without concepts are blind (A51/B75), that on their own they lack relation to an object? If this is all Hegel means, then he has not disagreed with Kant. In fact, Hegel means something stronger: the manifold is indeterminate with regard to the categories in that it cannot guide our application of categories to it. There are many ways the manifold could be bound into objects; these various ways form a space of possibilities; and the manifold does not exclude any of these ways as being incorrect.

To properly evaluate this criticism we need to consider the categories. Before this, I want to raise an objection to Hegel's account. Hegel seems to assimilate intuition to sensation. He talks of ‘intuition and sensation’ or ‘feeling and intuition’ in the same terms, characterizing both as indeterminate. A Kantian might object that while sensation is genuinely formless, Kant does not claim that it has representational content—it merely provides the material for intuition. By contrast, in intuition the sensory matter is ordered in space and time. This ordering, some Kantians may add, is not due to the subject's synthetic activity, but is simply given in virtue of our forms of sensibility. The manifold of intuition is therefore not indeterminate.

There are two ways Hegel might respond. First, he might offer a ‘conceptualist’ reading of Kant, that is, deny that the manifold can have even spatiotemporal unity without prior synthesis by the subject.²³ It seems that Hegel does read Kant (in particular, the B160 footnote) this way: ‘the original synthetic unity of apperception is recognized also as the principle of the figurative synthesis, i.e. of the forms of intuition; space and time are themselves conceived as synthetic unities, and spontaneity […] is conceived as the principle of the very sensibility which was previously characterized only as receptivity’ (FK: 69–70/327). Such a reading supports the claim that before spontaneity does its work, the manifold is formless.

But, second, Hegel does not need to make this claim here.²⁴ A better response, for present purposes, is that the Kantian's objection misses the point. Hegel is concerned about the manifold's lack of categorial form. This is consistent with the manifold's having some non-categorial form. Such a form is only relevant in so far as it can guide the application of the categories: the Kantian would need to claim that, prior to any synthesis, we receive intuitions of unified things moving through space, etc. Such a reading is possible (Allais 2015: chs. 7, 11). But it is in serious tension with the perceptual atomist claim that we cannot receive representations of complexes as such in intuition. If the Kantian does not want to take such a strongly non-conceptualist position, then the spatiotemporal form of intuition is not guiding the categories. Then, for Hegel, the difference between intuitions and sensations does not matter: they are both formless with respect to the categories.²⁵

IV.ii. The categories

Hegel's portrayal of the categories mirrors his depiction of the manifold. If the manifold is difference without unity, then the categories are unity without difference: like the manifold, but for the opposite reason, they are abstract and indeterminate.

Before looking at Hegel's view, there are a few things to note. First, Hegel's target varies. In different places in FK he criticizes the Ego, the I, the unity of self-consciousness, the concept and the understanding, all in basically the same terms. This sliding is, however, justified, as the targets are closely related: the I or Ego is the unifying subject of experience; the understanding is the faculty of concepts; the pure concepts are the form of the understanding and the basic modes by which the I unifies representations. Here I will focus on the I and the categories.

Second, Hegel distinguishes ‘what Kant calls the faculty of the original synthetic unity of apperception from the Ego which does the representing and is the subject—the Ego which, as Kant says, merely accompanies all representations’ (FK: 73/329). This distinction between the abstract Ego and the original synthetic unity of apperception is part of Hegel's general strategy in FK of defending the good Kant against the bad one. It is often unclear whether Hegel is offering his own views or articulating the ‘speculative’ ideas at which he thinks Kant was grasping. For clarity, I will not assess whether the distinction between the Ego and the original synthetic unity of apperception can be made out.²⁶ But the distinction is helpful to keep in mind in reading Hegel's discussion.

Hegel refers to the I as ‘empty identity’ (FK: 71/328) or ‘pure unity’ (FK: 73/330). In other places it is ‘an absolute Ego-point and its intellect’ (FK: 78/333) or ‘the absolute abstract unity’ FK: (85/338). But what is objectionable about this? Sedgwick (2012: 139) suggests the following: ‘[a]s independent from common reality, our concepts are taken to owe nothing of their nature and origin to objects known, to the process of knowing, to the relation of the knower to what is known. […] [S]uch concepts or categories are brought to experience by thinking and knowing subjects.’ This may be taken to suggest that Hegel holds, unlike Kant, that our categories are derived from sense-experience, not solely from the understanding. I do not think this can be right as a reading of Hegel. It is reminiscent of the empiricism Hegel consistently rejects, and it calls into question our ability to have synthetic a priori knowledge. This view could also be undermined by the kinds of arguments Kant makes for the a priori status of space (A24/B39). Like space, the categories cannot be derived from experience of objects, because they are preconditions for such experience.

A better reading is that guidance requires an affinity between concepts and the manifold they organize. ‘In isolation the pure concept is the empty identity. It is only as being relatively identical with that which it stands against, that it is concept; and it is plenished only through the manifold of intuition’ (FK: 70–71/328). A pure concept in isolation from the manifold it organizes lacks determinate content. But this sounds again like a conclusion Kant would accept—after all, concepts without intuitions are empty. Kant agrees that in some sense the categories lack content, and that they therefore need to be spelled out, or schematized, in terms of sensible objects (B149; A140/B179). If Hegel thinks he is making an objection against Kant, then, he must mean something stronger. Perhaps this. It is unclear how the categories can have material conditions for their correct application. Hegel's objection here deals with the same problem as his objection to the manifold, but from the opposite direction. The problem with the manifold is that, if it is formless, it is not clear how it can guide category-application. The problem with the categories is that it is not clear how they can relate to the matter of intuition, such that it can constrain their use.

It is easy to see how empirical concepts can have sensory conditions for right or wrong application. If we learn the concept ‘red’ by having red things pointed out to us, then the look of red things can partly constitute the concept. Then whether ‘red’ is correctly applied to X can depend on X's similarity to paradigmatic red things. By contrast, it is harder to see how pure concepts, which are not derived from experience, can have such conditions. But if they do not have such conditions, it is not clear how the manifold can guide the way we apply them.

This interpretation explains why Hegel directs his ire so often at the ‘abstract Ego’ rather than the categories themselves. The I or Ego is the source of the categories, and its metaphysical independence from the external reality that grounds sensation is what gives rise to the problem about how intuition can guide the categories. This point also clarifies the place of Hegel's critique in early German Idealism. If Kant's attempt to justify pure knowledge was vitiated by his dualistic conception of subject and reality, then it makes sense that post-Kantian idealists like Fichte, Schelling and Hegel would replace this dualism with a single principle from which both subject and reality are meant to derive.²⁷ While there is disagreement about the nature of the principle and the details of the derivation from it, it is common ground that there must be some such principle if we are to avoid Kant's error. I will return to this below.

The place where Kant tries to solve the problem is the Schematism: ‘how is the subsumption of [intuitions] under [categories], thus the application of the category to appearances possible, since no one would say that the category, e.g., causality, could also be intuited through the senses and is contained in the appearance?’ (B177/A138). Hegel sees the Schematism's importance and agrees that the imagination is crucial here. Why, then, does he think that Kant failed to solve the problem?

V. Formal identity: the role of the Schematism

In this section I set out the role of the Schematism in addressing the guidance problem and explain why, for Hegel, this solution is unsatisfactory. Before considering the Schematism, however, let us take stock. So far, there is an ‘absolute Ego-point and its intellect on one side, and an absolute manifold, or sensation, on the other side’ (FK: 78/333). Both sides are indeterminate. The manifold does not come to us as categorially unified; its points do not determine the way they should be connected up in synthesis. The I is independent of the manifold, such that its pure concepts have no necessary reference to the manifold's points, as they would if they were empirically derived; it is therefore not clear how the manifold can determine the correct application of the concepts. Hegel now argues that, because of this problem, the two sides cannot come together in the right way. The relation between the I and the manifold is ‘formal identity’ rather than ‘absolute identity’ (see Harris 1983: 45). By this Hegel means that the categories can intelligibly relate to the form of intuition, but not to its matter, sensation.²⁸

Kant holds that transcendental logic is able to specify application-conditions for the categories a priori (A135/B174–75). For each category, there is a corresponding time-determination or ‘schema’ which provides a criterion for the application of the category to intuition. Schemas allow the temporal form of the manifold to guide the application of the categories. As Bell (1987) has argued, the Schematism does not posit additional rules to govern concept-application, but rather involves an aesthetic response to the manifold of the sort that Kant describes in the CPJ; in Kant's words, a schema is in between a concept and an image (A141/B180).

I can now state Hegel's objection to the Schematism. The schemas give conditions for category-application in terms of the form of inner sense, time; they do not give conditions for category-application in terms of the matter, sensation. Aesthetic response is to ‘either shape or play’ (CPJ: §14, Ak. 5:225). But the temporal relations between empirical intuitions do not exhaust the differences in matter between them. There are similarities and differences among sensations which are not reflected in time-determinations.²⁹ Indeed, since time is something we contribute, it is not clear that temporal relations reflect any differences in what causes our sensations.³⁰ It follows that if our spontaneity is guided only by time-determinations, the sensory matter is not playing a role in guiding the synthesis: ‘the heterogeneity of appearance has here been left out’ (FK: 76/331). Next to the ‘formal unity of the manifold’, i.e. the formal aspects of the manifold, which have an intelligible relation (‘formal identity’) with the categories, there is an ‘infinity of sensations’ which is not doing any work.

But sensation is what we ultimately receive from external reality—from affection by things in themselves. In this way sensation is transcendentally objective. ‘[T]hat which corresponds to the sensation in [objects as appearances] is the transcendental matter of all objects, as things in themselves’ (A143/B182). So if the sensory matter of intuition is not guiding our application of the categories, it looks like the way we categorize the manifold is not constrained by what we receive from outside, but is up to us. It seems that categorial form is subjective:

[T]hat the manifold of sensibility, empirical consciousness as intuition and sensation, is itself something unintegrated, that the world is in itself falling to pieces, and only gets objective coherence and support, substantiality, multiplicity, even actuality and possibility, through the good offices of human self-consciousness and intellect. (FK: 74/330)

In fact, experience does have determinate content. If Kant is correct, this content arises from the coming-together of the categories and the manifold. But the way the manifold is synthesized depends not on the matter, but only on what we contribute by ordering the sensations in time and then applying the categories in virtue of time-relations: ‘the thing in itself becomes object in so far as it obtains from the active subject some determination which for this reason alone is one and the same in both of them’ (FK: 75/331). The form we contribute does not reflect the matter we receive. This form, however, determines which objects we experience and what causal relations they stand in—that is, it determines, to a large extent, what is true and false. In this way idealisms which settle for formal identity between I and manifold are ‘forever sliding into […] psychological idealism’ (FK: 76/332).³¹

VI. Two Kantian responses

In this section I consider two Kantian responses to Hegel's criticism.

VI.i. The metaphysical response

Hegel's criticism depends on the claim that the manifold is too indeterminate to constrain our application of the categories to it, but this is to take the manifold in isolation from its source in things in themselves. Things in themselves are perfectly determinate. The Kantian should not claim, here, that things in themselves fall under the categories, but rather that things in themselves have determinations isomorphic to the categorial features of appearances.³² For any fact about appearance X's being a substance (etc.), there is a corresponding fact about some set of things in themselves—call them X's ‘supersensible ground’. The determinacy of the supersensible ground finds expression in the manifold and guides the application of the categories.

Hegel anticipates this thought: ‘the incomprehensible determinateness of the empirical consciousness comes altogether from the things in themselves’ (FK: 74/330). He responds that the realm of things in themselves, because it falls outside the categories, must be a ‘formless lump’:

Objectivity and stability derive solely from the categories; the realm of things in themselves is without categories […]. The only idea we can form of this realm is like that of the iron king in the fairy tale whom a human self-consciousness permeates with the veins of objectivity so that he can stand erect. But then formal transcendental idealism sucks these veins out of the king so that the upright shape collapses and becomes something in between form and lump, repulsive to look at. (FK: 77/332)

In other words, the realm of things in themselves is indeterminate. It is not clear, however, that Hegel is entitled to this claim. A Kantian can maintain that, as things in themselves are beyond our knowledge, we have no justification for deciding one way or the other whether they are determinate.³³

But this response may not solve the problem. As long as we do not know that things in themselves are determinate, then we do not know that synthesis is guided by anything in the manifold. So subjectivism may, for all we know, be true. In this sense ‘the objectivity of the categories in experience’ is still ‘something contingent’ (FK: 77/332). Hegel's objection thus anticipates his later critique of the unconceptualized contingency that remains in Kant's account (Ng 2009). Such contingency makes it impossible to fully explain how knowledge is possible, and thus leaves us exposed to scepticism (Bristow 2007: chs. 2 and 3).

Waiving this issue, there is a further problem with the Kantian response. It is unclear how the (quasi-)categorial form of things in themselves could guide our synthesis. Hegel's discussion of the Schematism shows that, while the form of the manifold can guide the application of the categories, its matter cannot play a similar role.³⁴ And to hold that the form of the manifold reflects information about things in themselves would threaten the ideality of time. Even if things in themselves are determinate, unless we can explain how this determinacy is expressed in the manifold in such a way as to guide synthesis, this determinacy is beside the point.

We might try to strengthen the Kantian response by characterizing the categories as ‘response-dependent concepts’. A response-dependent concept is one whose correct application is defined in terms of the circumstances under which a (well-functioning, appropriately situated, etc.) subject will apply it. We might say, for example, that ‘red’ applies to those things that a well-functioning subject, trained in the use of the concept, will classify as red. Perhaps Kant can hold that while there is no account of how we apply the categories to the manifold, we do, undeniably, apply them, and (supposing our faculties are working properly) whatever in the manifold inclines us to apply a given category will count as a correct condition for its application.

I do not think that response-dependent concepts can coherently be invoked here. As Wright has argued, an account of a concept as response-dependent requires that we have an independent and non-trivial grasp of the conditions under which the subject's response counts as appropriate. These will usually involve conditions of the subject and of the environment. But a Kantian is in no position to specify such conditions, since we have no access to our cognitive faculties except through reflection on their activity of forming experience. We can give no content to the idea of our transcendental faculties malfunctioning. And we can therefore give no content to the idea that the categories are correctly applied to whatever in the manifold induces their application by a well-functioning set of faculties.

VI.ii. The epistemological response

Hegel is demanding an account of how the manifold of intuition guides the application of the categories in synthesis. A Kantian may reject this explanatory demand on the basis that Hegel is asking about something beyond our cognitive reach.³⁵ The very point of Kant's Copernican Revolution (Bxvi) is that we cannot understand a priori cognition if we assume that cognition must conform to external reality, rather than that external reality conforms to our cognition. But to ask about guidance just is to ask about how our cognition conforms to external reality. A variant of this response is that the guidance problem is the product of an incoherent attempt to look at our relation to reality from ‘sideways-on’. In McDowell's words, ‘pertaining to things as they are given to our senses is as much objectivity as we can intelligibly want’ (2009a: 78). The fact that we can know things only in so far as they conform to the conditions for our knowing them is a truism; there is no room for a further account of how those conditions pertain to reality.

It is at least consistent for a Kantian to refuse to give such an account. The central result of the Critique of Pure Reason is that synthetic a priori cognition can do nothing more than ‘to anticipate the form of a possible experience in general’ (A246/B303; cf. B166); in other words, all we can have a priori is knowledge of the necessary features of appearances. The account that Hegel is after is clearly synthetic and a priori, but it is not about the necessary features of appearances. Hegel is asking about the workings of the synthesis that gives rise to appearances: a process of which we cannot have any intuition, and therefore cannot have cognition.

However, there are two problems with this response. The first is that rejecting Hegel's explanatory demand comes at a significant cost. Refusing to say how, or whether, synthesis is guided by the manifold means refusing to spell out the sense in which external reality provides the matter of experience, while the subject contributes its form. And this use of the form-matter contrast is essential to Kant's attempt to distance himself from Berkeley. In a letter to J.S. Beck, for example, Kant wrote that an attempt to assimilate the two ‘does not deserve the slightest attention. For I speak of ideality in reference to the form of representations; but they interpret this to mean ideality with respect to the matter […]’.³⁶ If the claim that the subject contributes only the form of experience cannot be intelligibly explicated, then neither can Kant's attempt to distance himself from Berkeley. In other words, Kant's dialectical context creates direct pressure to take the sideways-on view that McDowell would disavow. Kant has to insist on the formal character of his idealism, but it is not clear that this expression has any sense in McDowell's account.

The second problem with the epistemological response becomes clear when we ask about the status of Kant's own account, which includes claims about the nature and sources of sensation, intuition, pure and empirical concepts, and the process by which these come together. These are not matters of which we have an intuition (Young 1994: 332–33). It seems that there are two possibilities here. One is that Kant's own account has the status of cognition. In this case, the account Hegel is seeking would have the same status as Kant's own, and would not be beyond our reach. For Hegel is asking about the relation between intuition and category in synthesis—matters which are at the heart of Kant's account in the First Critique. The other possibility is that the account Hegel seeks could not have the status of cognition. But then it is not clear how Kant's own theory can have this status.³⁷ This is problematic: the fundamental defect of pre-Kantian metaphysics is meant to be that its claims do not have the status of cognition. These are well-known problems, and I will not attempt to decide between these options here (see Moore 2012: ch. 5). The point is that, however the problems are solved, the issues that Hegel is asking about will have the same epistemic status as core Kantian claims, such that a Kantian cannot consistently assert the doctrines set out in the CPR while refusing to address Hegel's criticism.

VII. Absolute identity: Hegel's alternative

I want to end by commenting on the general conclusions to be drawn from the above argument, in particular how Hegel's theory of an ‘original identity’ solves the problem he identifies in Kant's account.

VII.i. Deflationary and inflationary solutions

Kant attempted a purely formal idealism, on which we are the source of the form of experience, while things in themselves are the source of its matter, namely sensation. Such a position was meant to justify our claim to synthetic a priori knowledge while maintaining a role for external reality. FK aims to show that such a position cannot be sustained. We cannot make sense of the matter of experience if it is alien to our forms of cognition, and we cannot make sense of those forms if they lack an inherent connection to their matter.

These problems may bring to mind Davidson's (1984) argument against ‘the very idea of a conceptual scheme’. Davidson argues that we cannot make sense of a dualism of scheme and content—that is, of a conceptual scheme with which we subjects organize some unorganized content. He urges us to simply think in terms of ‘unmediated touch with the familiar objects whose antics make our sentences and opinions true or false’ (1984: 198; cf. Lear 1984: 242). One conclusion we could draw from Hegel's argument, then, is that we should give up on idealism: forget the idea that experience has a form which we bring to it, that this allows us a priori knowledge, that the form may or may not correspond to reality, etc. On this view we should excise the parts of Kant which make claims from a transcendental point of view—the ideality of space and time, the existence of things in themselves, the possibility of other forms of sensibility. For these are the sorts of claims that lead to subjectivism. This is the conclusion favoured by McDowell and by Sedgwick, who insists that reason cannot ‘transcend the realm of the empirical’ (2012: 94) as Kant thinks it can.³⁸ Both are responding to Hegelian concerns.

It is interesting, then, that this is not the conclusion Hegel draws in FK. Instead he says that—contra Kant—we can have ‘cognition of the Absolute’ (FK: 68/326) and that the key to this is to recognize the ‘absolute identity of the heterogeneous’ (FK: 72/328) from which ‘the Ego as thinking subject, and the manifold as body and world first detach themselves’ (FK: 71/328). He thinks that Kant should have accepted absolute idealism rather than the merely formal kind. That is, Hegel does not want to discard the idea that the objects of ordinary experience have a form which we can know about a priori; instead, he says that this form reflects the nature of the content. Not because we receive categorial form with the sensory matter, but because one thing, an ‘original identity of opposites’ (FK: 70/327), that is, of matter and form, gives rise to the sensory manifold and the I with its categories.

In short, Hegel is proposing an inflationary rather than deflationary answer to the guidance problem: rather than cutting away or reading down some of Kant's apparent theoretical commitments such that the problem never arises, he proposes to add an additional commitment which makes a solution to the problem possible. In the next section, I consider how to interpret his difficult claims. Before this, let me state the conditions that such an interpretation must meet. First, to solve the guidance problem, it must explain why the categories have material conditions for correct application. Second, to be dialectically apt as a response to Kant, the account must be knowable within Critical strictures.

VII.ii. The identity of subject and external reality

Hegel's positive view in FK involves an ‘original identity’ between the subject and external reality. To begin interpreting this claim, it is useful to consider how one might explain why two things, such as the subject and external reality, share a form. In a recent paper, Thomas Hofweber writes that there are ‘two straightforward ways in which this correspondence could be explained, based on two different directions of what is explanatorily more basic’: either external reality imposes a form on the subject, or the subject imposes a form on external reality (2019: 704). The former mode of explanation involves an empiricist account of concept-acquisition, while the latter leads to subjective idealism. But Hofweber leaves out a third option: a common cause of the subject and external reality, which, being the source of both, explains why they share a form. The explanatory strategy is familiar: the fact that B and C are both derived from A can explain why B and C share some feature.

This is the option Hegel takes up in FK. He proposes that there is a single principle which is metaphysically prior both to the subject and to the external reality which is cognized by the subject: the ‘original identity’ or ‘absolute synthesis’ from which ‘the Ego as thinking subject, and the manifold as body and world first detach themselves’ (FK: 71/328). He takes this principle to be what Kant calls the ‘original synthetic unity of apperception’, distinguishing this from the finite ‘Ego’ which ‘does the representing and is the subject’ (FK: 73/329).³⁹ Drawing on Schelling, Hegel also refers to this principle as the subject-object, giving rise to subjective and objective subject-objects: ‘For absolute identity to be the principle of an entire system it is necessary that both subject and object be posited as subject-object’ (DFS: 155/63; Ng 2020: 72ff).

The principle is ‘original’ in the sense that it is prior to the subject and external reality: it ‘must be conceived, not as produced out of opposites, but as a truly necessary, absolute, original identity of opposites’ (FK: 70/327); it is not ‘inserted between an existing absolute subject and an absolute existing world’, but is ‘that out of which subjective Ego and objective world first sunder themselves’ (FK: 73/329). By contrast, in Kant's account, the subject and external reality come together in the process of cognition, but there is no supposition that they have a common origin. The common origin proposed by Hegel makes possible a solution to the guidance problem, as I will explain.

First, the reason why our categories can be applied correctly or incorrectly is that the source of the manifold has a form. There is a right way to reconstruct a shattered pot because it was whole before it was broken. In the same way, the unity of the principle from which the manifold is derived explains why there is a correct way of unifying the manifold under the categories. Second, our categories have material, not merely formal, conditions of correct application because they are derived from the same source as the manifold. Such a connection between a concept and its instances can explain why the concept's application-conditions reflect features of the instances: this is, for example, the idea underlying abstractionist accounts of concept-formation. Here, if the subject (whose understanding is constituted by these categories) is generated out of an original identity of matter and form, then this explains why those subjects have material application-conditions. As a result, both conditions of guidance are satisfied: there are correct and incorrect ways of combining the manifold, and whatever features of the manifold make some ways of combining it correct explain why we combine it correctly in a priori synthesis.

Then, when we give form to the matter we receive, this corresponds to the original form that the matter had. When the subject is affected by external reality, giving rise to sensations ordered in space and time, this ordering reflects differences in the matter of intuition;⁴⁰ when this manifold is synthesized under the schematized categories, the resulting categorial form corresponds to the form which the matter originally had. The application of the categories to the manifold is understood by Hegel as a return to unity; in it, ‘knower and known are necessarily one’ (DFS: 164/70). As Hegel writes (EL: §42A1) “it is not” the subjective activity of self-consciousness that introduces absolute unity into the manifoldness. This identity is, rather, the absolute, the true itself. It is, so to speak, the benevolence of the absolute to release the individualities to their self-enjoyment, and this absolute drives them back into “the absolute unity”. Hegel suggests that the return to unity occurs inadequately in intuition (‘in it the identity […] is totally immersed in the manifold’ (FK: 70/327)) and more adequately in judgment (‘the original identity appears in consciousness as judgment’ (71/328)).⁴¹

Is this view consistent with Critical strictures about the limits of knowledge? Clearly Hegel's view requires us to say more than Kant does about the origins of the subject and of the external reality that gives rise to sensation; his attempts to attribute his view to Kant are hard to credit. But if Hegel is right, the existence of the original identity, and its priority to the subject and external reality, are required if knowledge is to be possible: a view, like Kant's, which leaves out the original identity is unable to explain the return to unity in cognition, finding only formal identity. The existence and priority of the original identity are, therefore, arrived at by a recognizably Kantian style of transcendental argument, and are consistent with Critical strictures to the extent that Kant's own commitments are.⁴² They are conditions for the possibility of knowledge; if we suppose that knowledge is possible, we are required to accept that these conditions hold.

Thus, Hegel's account in FK provides the outline of an alternative to Kant's view which solves the guidance problem and which is (in so far as it is required to solve that problem) not merely an exercise in pre-Critical metaphysics. That said, the account remains merely an outline. Hegel tells us little in FK about what exactly the original identity is, and in what sense it is prior to, or the source of, the subject and external reality. What is the process by which they come to exist?

It is not clear that Hegel can say more without the significant methodological developments of the Phenomenology and the Logic, in particular the role of contradiction in driving the system forward. But the explanatory structure first set forth in FK is one which will recur throughout Hegel's work: first, an original, synthetic unity; then a breaking-apart into subject and manifold; and third, a return to unity as the subject cognizes the manifold, guided by its original form. In this sense, the guidance problem and the explanatory structure it requires lead out to Hegel's mature system (Longuenesse 2007: 190ff).