It seems natural to think that the distinct immediacy and authority with which we know ourselves has to be due to a special theoretical capacity granting us a unique view of ourselves, unavailable to the gaze of others. The fact that we know ourselves in a more immediate way and with an authority that outranks the authority of others must have something to do with the privileged spectator position we have with regard to our own inner life, the stage of the ‘internal theatre’ to which we can direct our inner eye. As suggestive as this inner perception model may seem, the fundamental problems that it raises are well known. In recent years, Richard Moran has proposed a radical alternative to this approach, suggesting that the true root of these problems does not just reside in the misleading implications of the metaphors of an inner eye or an internal theatre, but in the even more general issue that this model misconstrues our self-knowledge by understanding it as a form of theoretical cognition. If we try to explain self-knowledge in terms of a ‘purely theoretical or spectator'sstance toward the self’, whatever its concrete form or shape, we distort the fundamental character of self-knowledge (Moran Reference Moran2001, 3): ‘Modelling self-consciousness on the theoretical awareness of objects obscures the specifically first-person character of the phenomenon’ (Moran Reference Moran2001, 32). To understand the immediacy and authority of first-person knowledge, we rather need to recognize the extent to which self-knowledge is due to a practical stance. That I have distinct access to my own beliefs and intentions is not due to the fact that I am in a privileged position to observe them, but due to the fact that I am the very person determining and maintaining them. My self-knowledge is thus not based on special evidence available only to me or an exclusive point of view. Rather, it is based on the special way in which the determination of my own attitudes is up to me and the point of agency that I have with regard to these attitudes. I know my ‘beliefs, desires, intentions’ (Moran Reference Moran2001, xxxiii) immediately and with authority, because I am the agent determining them. The peculiar features that distinguish knowledge of myself from knowledge of others are thus to be explained in terms of the specific agential character that self-knowledge involves. The agential stance that I have in view of my own attitudes becomes evident in the way in which I can identify them not merely by observing them and attributing them to myself as anyone else might, but by determining and avowing the.
There are a number of different inspirations Moran cites for this fundamental idea – the idea that self-knowledge is rooted in a specific form of practical self-relation – but one especially salient source is Kant.Footnote 1 Matthew Boyle (Reference Boyle2009) has suggested that returning to Kant may not only allow us to better understand the underlying intuition guiding Moran'saccount, but in addition will lead to a clarification that helps to defend Moran against an important objection, namely the worry that Moran'saccount cannot explain the privilege the first person has with regard to knowing one'sown basic sensations. Boyle reminds us that Kant distinguishes between two fundamentally different ‘kinds of self-knowledge: knowledge of ourselves through “inner sense” and knowledge of ourselves through “pure apperception.”’ Whereas knowledge through inner sense ‘gives us knowledge of our sensations, knowledge of ourselves as passive beings’, the self-knowledge granted by pure apperception ‘gives us knowledge of what we think and judge, knowledge of our own spontaneity’ (Boyle Reference Boyle2009, 133). The latter self-knowledge – that granted by pure apperception – could thus help to elucidate the agential mode in which we know our own beliefs and intentions on Moran's account. By distinguishing this type of self-knowledge from another type, Kant makes explicit that not all self-knowledge is of the active sort Moran is interested in. Kant thereby contests what Boyle has termed the Uniformity Assumption: the assumption that we need to ‘seek some common explanation of all of the cases in which we speak immediately and authoritatively about our own mental states’ (Boyle Reference Boyle2009, 141). In addition, Kant suggests that there is an interesting dependency relation between these two forms of self-knowledge such that we can see why the agential self-consciousness granted to us by pure apperception indeed deserves our primary attention – without it, we could not have any self-knowledge at all.
In what follows, I will not discuss Moran'sor Boyle'saccounts but follow the suggestion of taking a closer look at Kant'salleged distinction of different forms of self-knowledge. I think that it is a deeply intriguing idea that self-knowledge is based on a practical or agential self-relation, and I agree that Kant is a crucial source to consult when we want to get clear about this idea. As his conception of pure apperception indeed indicates, Kant suggests that our self-knowledge depends on an agential stance: on consciousness of ourselves as self-active, as spontaneous. However, I want to contest the idea that Kant'sdistinction of pure apperception and inner sense can be understood as pointing us to two self-standing types of self-knowledge. Boyle himself already draws on the fact that self-knowledge of our sensations presupposes self-consciousness of our thinking and judging I. Here I am more interested, however, in the reverse dependency. For Kant it is not the case that pure apperception as such gives one cognition of one'sself, ‘Erkenntnis seiner selbst’ (Kant Reference Kant and Timmermann1998, B158).Footnote 2 Although Kant'sconception of pure apperception indeed gives us an account of the awareness we have of our agential self, it renders problematic the idea that this awareness by itself amounts to knowledge or cognition of ourselves. In merely being conscious of my pure apperceptive activity, I do not yet cognize my ‘I’. In order for us to know not only that we think and judge, but what it is that we think and judge we are dependent on the mediation of inner sense. To yield cognition or knowledge of ourselves, our apperception is thus dependent on a given sensible manifold and on the articulation of our thoughts in inner sense.Footnote 3
Kant'ssignificance for the current debate is thus not that through his conception of pure apperception he has specified a form of immediate self-knowledge. Rather, his significance resides in the fact that he has characterized our self-knowledge as inherently double: it is based on two forms of self-relation that are in seeming tension with one another, but only jointly give us knowledge of ourselves. In what follows, I will (I) first retrace the way in which Kant distinguishes and relates pure apperception and inner sense. Next (II), I will highlight the way in which Kant'sdistinction implies that we are conscious of ourselves in two ways that, in the context of Kant'stheoretical philosophy, remain in a problematic tension. I will close (III) with a discussion of the way Kant tries to elaborate this tension through an account of practical self-consciousness in his conception of practical reason.
I.
Kant'smost extended discussion of consciousness of ourselves through inner sense and of consciousness of our ‘I’ through pure apperception can be found in the Critique of Pure Reason. This discussion is, however, not presented in terms of an independent theory of self-knowledge. Rather, Kant introduces these notions in his attempt to specify the transcendental conditions of possibility of cognition of objects in general. He first introduces his notion of an inner sense in the context of the Transcendental Aesthetic and specifies pure apperception in the context of the Transcendental Analytic. The Transcendental Aesthetic highlights that due to the nature of our intuition, consciousness of ourselves as an object of inner sense requires inner perception of a manifold, antecedently given sensibly in the subject and not self-actively produced. To intuit ourselves we have to affect ourselves; we cannot intuit ourselves as we would immediately and spontaneously think ourselves, but only according to the way in which we affect ourselves through inner sense.Footnote 4 The Transcendental Analytic, in turn, highlights the way in which the objectivity of our cognition ultimately depends on the original synthetic unity of pure apperception.Footnote 5 Only through the synthetic activity of conjoining representations and being conscious of my conjoining them (Kant Reference Kant and Kemp Smith1929, B133) – that is, only through the 'unity of consciousness in the synthesis' – can I ‘constitute the relation of representations to an object, and therefore their objective validity’ (Kant Reference Kant and Kemp Smith1929, B137). ‘Thoroughgoing identity of the apperception of a manifold... is possible only through the consciousness of this synthesis.’ (Kant Reference Kant and Kemp Smith1929, B133, emphasis added) Pure apperception is thus the vehicle or medium through which the objects of our cognition can take shape and appear to us. Kant thus does not focus on the way in which this pure apperception may give us access to ourselves, but rather the way in which it allows us to constitute our experience such that it gives us access to objects in general. As Kant writes in the Paralogisms: ‘We can thus say of the thinking “I”... that it does not know itself through the categories but knows the categories and through them all objects, in the absolute unity of apperception, and so through itself.’ (Kant Reference Kant and Kemp Smith1929, A402) Pure apperception is, in other words, what makes the mind transparent to the objects given in its cognition.Footnote 6
I recall this framing only to remind us that self-knowledge is not in and of itself Kant'stopic. He is not concerned to explain the immediacy or authority with which we speak about our mental attitudes or sensations, and he is not concerned with a comprehensive account of the different ways in which we relate to ourselves (as a thinking I, as a person, as a sensible being, as a body, etc.), although his account touches on all of these ways in which we might be an issue for ourselves. His immediate topic is the conditions of possibility of objective knowledge in general. And it turns out that the forms of intuition and the understanding that are required for knowledge of objects, in general, involve certain forms of self-consciousness. Given our forms of intuition, we can become a determinate object for ourselves through inner sense; and given the synthetic activity of understanding, our cognitive activity involves a pure form of self-consciousness expressed in the ‘empty representation’ or ‘mere consciousness’ I (Kant Reference Kant and Kemp Smith1929, B404), the ‘concept’ or ‘judgment’ ‘I think’ (Kant Reference Kant and Kemp Smith1929, B399). This is not to say that we cannot try to develop a comprehensive theory of self-consciousness from Kant'sreflections, Footnote 7 but it reminds us that this requires careful attention to the way in which self-consciousness becomes a topic for Kant. Given the distinction Kant makes between self-consciousness through inner sense and through pure apperception, leading to a consciousness of the ‘“I” as object of perception’ and the ‘“I” as subject of thinking’ (Kant Reference Kant and Louden2006, 7:134FN) respectively, it is suggestive to think that Kant'saccount implies the idea that there are two self-standing types of self-knowledge with distinct contents. Boyle suggests such an understanding when he reformulates Kant'sdistinction as one of two forms of self-knowledge, one concerning our sensations, the other regarding our thoughts and judgments.Footnote 8 As a closer look at Kant'sdiscussion makes clear, however, the types of self-consciousness are neither self-standing nor can they properly be distinguished simply in terms of these contents. The two types of self-consciousness rather involve two fundamentally different modes of self-relation that have to co-operate to produce substantial self-knowledge, be it of our sensations or our thoughts.
Let me briefly point out the way in which these two types of self-consciousness depend on one another in bringing forth self-knowledge. Kant distinguishes ‘inner sense’ or ‘empirical apperception’ on the one hand, from ‘pure’ or ‘transcendental apperception’ on the other (Kant Reference Kant and Kemp Smith1929, A107). Whereas empirical apperception is consciousness of myself according to the always changing ‘determinations of our state in inner perception’, transcendental apperception is a ‘pure, original, immutable consciousness’ of my ‘I’ (Kant Reference Kant and Kemp Smith1929, A107). Now, it is probably not very controversial that empirical apperception depends upon pure apperception, since it is Kant'sexplicit aim in the first Critique to show that pure apperception is a transcendental requirement for any form of objective knowledge. For our experiences in inner sense to amount to cognition of ourselves – cognition of an aspect of our empirical I, given as an object of our inner sense – our experiences require a certain type of unity that is based on the unity of pure apperception. In the absence of the original synthetic unity of apperception, no objective knowledge, be it about other objects in the world or myself, is possible. Kant makes this particularly clear in a letter to Marcus Herz from 1789 in which he contemplates a possible cognizer lacking pure apperception: If I abstract from the synthetic activity of my understanding and imagine myself to be a mere animal, I may still assume that the sense data given in my mental life could ‘carry on their play in an orderly fashion’, but in the absence of the synthetic activity of the understanding this could only happen ‘without my cognizing the slightest thing thereby, not even what my own condition is.’Footnote 9 Without the synthetic activity of the understanding, the data of sense could never represent objects: ‘They would not even reach that unity of consciousness that is necessary for cognition of myself [Erkentnis meiner selbst] (as object of inner sense)’ (Kant Reference Kant and Zweig1999, 11:52).
Let us, therefore, assume that there cannot be any knowledge of myself as an object of my inner sense without pure apperceptive consciousness of myself as the subject of my thoughts. This is a fact that the contemporary debate seems fully aware of, wherever it takes up Kant'sdistinction.Footnote 10 Less noted, however, is the extent to which that pure apperception is in turn also dependent on the given manifold of intuition. To elaborate this, it is helpful to take note of the distinction that Kant draws between the analytic unity of pure apperception, expressed in the ‘I think’ which must be able to accompany all my representations, and the synthetic unity of apperception which is presupposed by this analytic unity (Kant Reference Kant and Kemp Smith1929, B133). In general, an analytic unity can be understood as the unity of a one contained in many: a shared feature or mark common to a multiplicity of instances. A synthetic unity, by contrast, is the unity of many contained in one, a unity of conjunction.Footnote 11 The analytic unity of the ‘I think’ – the one trait accompanying all of my representations insofar as they are mine – presupposes the synthetic unity of apperception: my operation of self-consciously conjoining these representations in one consciousness. Kant writes:
The thought that the representations given in intuition one and all belong to me, is therefore precisely the same as the thought that I unite them in one self-consciousness, or can at least so unite them; and although this thought is not itself the consciousness of the synthesis of the representations, it presupposes the possibility of that synthesis. In other words, only in so far as I can grasp the manifold of the representations in one consciousness, do I call them one and all mine.’ (Kant Reference Kant and Kemp Smith1929, B134)
The simple unity of the I think thus does not precede the synthetic activity of the understanding, it rather results from the synthetic activity of my thinking something, actualizing the original synthetic unity of apperception. However, if the ‘I think’ indeed depends upon the possibility of such self-conscious synthesis, it seems dependent upon a given manifold that lends itself to this synthetic activity. Pure apperception cannot itself produce the matter for its synthetic activity, but has to rely on receptivity to attain the manifold it can unite:
For through the ‘I’, as simple representation, nothing manifold is given; only in intuition, which is distinct from the ‘I’, can a manifold be given; and only through combination in one consciousness can it be thought. An understanding in which through self-consciousness all the manifold would eo ipso be given, would be intuitive; our understanding can only think, and for intuition must look to the senses. (Kant Reference Kant and Kemp Smith1929, B135)Footnote 12
The point here is not just that pure apperception is dependent on the cognitive matter to be provided from elsewhere. Kant also wants to point us to the fact that pure apperception, taken in abstraction, grants us only a very peculiar access to ‘ourselves’ that Kant himself would not venture to call cognition of myself (Erkenntnis meiner selbst) and that does not yield substantive knowledge of myself.Footnote 13 For Kant, cognition by definition requires the unity of sensibility and understanding, intuition and concept (Kant Reference Kant and Kemp Smith1929, A51/B75). Kant does think that the understanding can be conscious of the mere unity of its action in abstraction from sensibility, that is to say: it can be conscious of the unity of its operation directly just by virtue of the thought ‘I think.’ With this consciousness abstracting from sensibility altogether, the understanding, however, does not cognize (erkennt) anything determinate of the putatively represented object (the ‘I’). In abstraction from all sensibility, all that the understanding becomes conscious of with regard to its synthesis is the mere form and the sheer unity of its action (Einheit der Handlung): The synthesis of the understanding is ‘nothing but the unity of the act, of which, as an act, it is conscious to itself, even without [the aid of] sensibility’ (Kant Reference Kant and Kemp Smith1929, B153). But such consciousness of the mere unity of the act – consciousness of the ‘I’ as the vehicle of all concepts – cannot be called cognition of myself. As Kant writes explicitly in § 25 of the deduction: ‘Consciousness of one'sself is... very far from being cognition of one'sself [Erkenntnis seiner selbst]’ (Kant Reference Kant and Kemp Smith1929, B158, transl. modified). And in the Paralogism section he underlines: ‘I do not cognize [erkenne] myself through being conscious of myself as thinking’ (Kant Reference Kant and Kemp Smith1929, B406, transl. modified), for I do not cognize or know any object by merely thinking, but ‘only in so far as I determine a given intuition with respect to the unity of consciousness in which all thought consists’ (Kant Reference Kant and Kemp Smith1929, B406).Footnote 14 The awareness of myself I gain by virtue of the transcendental synthesis of the manifold of the representations is not consciousness of myself as an object – neither an object of inner sense, nor a thing in itself – it is rather merely consciousness of the actuality of a distinct activity: ‘In the... synthetic original unity of apperception, I am conscious of myself, not as I appear to myself, nor as I am in myself, but only that I am.’ (Kant Reference Kant and Kemp Smith1929, B157, emphasis added) The consciousness of the ‘I think’ as such thus makes me aware that I think and that I am but does not amount to cognition of what I think or what I am.
This awareness is certainly not without import. It does not determine my existence such that I can cognize given features of myself, but it makes me aware of a distinct mode of my existence that distinguishes me: it reveals me as activity, as spontaneity, as a self-given unity. Moreover, it gives me access to the pure form of this fundamental activity that I am. But this spontaneous mode of existence and this pure form is not disclosed as a content of this consciousness which I cognize, but rather as the mode of this consciousness: as the agential form of my knowing things. It is through this active or performative character of this consciousness that this ‘I think’ in some very specific sense entails that I exist.Footnote 15 By means of the ‘I think’ I do not become aware of an object that I can know to exist, I rather become aware of the ‘I’ of the ‘I think’ as something that ‘exists in the deed’ (‘etwas, was in der Tat existiert’, Kant Reference Kant and Timmermann1998, B423FN). Pure apperception thus reveals to me the extent to which I, qua subject, exists as pure activity.Footnote 16 Kant explicitly rules out that the ‘I’ of the ‘I think’ is here given as an object of my cognition at all: it is neither given ‘as appearance, nor as thing in itself (noumenon)’ (Kant Reference Kant and Kemp Smith1929, B423FN); neither ‘as we appear nor as we are’ (Kant Reference Kant, Guyer, Bowman, Guyer and Rauscher2005, 364).
The self-consciousness of pure apperception thus gives me access to my existence in a certain distinct mode or voice, but it does not grant me knowledge of any determinate features that characterize me apart from my being a thinker.Footnote 17 As such, transcendental apperception only gives us consciousness of the form of the activity of the ‘I think’, but not knowledge of the characteristics of the ‘I’ as an object: Neither knowledge of the soul as the general type of object that may underlie this ‘I’, as the Paralogisms show, nor cognition of the particular self that I am. Although transcendental apperception puts me in a relation to myself that is indispensable for me to be a cognizer at all, including me being a cognizer of myself, on its own, it does not suffice to yield cognition of myself. In §25 of the deduction Kant gives the following explanation of why this is:
Now in order to cognize ourselves [zum Erkenntnis unserer selbst], there is required in addition to the act of thought, which brings the manifold of every possible intuition to the unity of apperception, a determinate mode of intuition, whereby this manifold is given; it therefore follows that although my existence is not indeed appearance (still less mere illusion), the determination of my existence can take place only in conformity with the form of inner sense, according to the special mode in which the manifold, which I combine, is given in inner intuition. Accordingly I have no cognition of myself as I am but merely as I appear to myself [ich habe also demnach keine Erkenntnis von mir wie ich bin, sondern bloß wie ich mir selbst erscheine]… Just as for cognition of an object distinct from me I require, besides the thought of an object in general (in the category), an intuition by which I determine that general concept, so for cognition of myself I require, besides the consciousness of myself, or the fact that I think, an intuition of the manifold in me, by which I determine this thought. I exist as an intelligence which is conscious solely of its power of combination; but in respect of the manifold which it has to combine I am subjected to a limiting condition (entitled inner sense), namely, that this combination can be made intuitable only according to relations of time, which lie entirely outside the concepts of understanding, strictly regarded. Such an intelligence, therefore, can cognize itself only as it appears to itself in respect of an intuition which is not intellectual and cannot be given by the understanding itself, not as it would cognize itself if its intuition were intellectual. (Kant Reference Kant and Kemp Smith1929, B157f.)
As Kant makes clear in this passage, pure apperception alone is insufficient for self-cognition. Besides the thought of myself, cognition of myself requires an intuition of the manifold in me. Only thereby can my thought of myself attain determinacy. Note that Kant is here not distinguishing pure apperception of our thoughts, which may not be cognition strictly speaking, from a different form of self-consciousness of other items – say sensations – which does constitute self-cognition. Rather, he considers the possibility that the synthetic activity of apperception itself may become known to us through inner sense: ‘this combination can be made intuitable according to relations of time, which lie entirely outside the concepts of understanding strictly regarded.’ The subject can thus cognize itself as an intelligence, but only ‘as it appears to itself in respect of an intuition.’ (Kant Reference Kant and Kemp Smith1929, B157f) This seems to suggest that Kant does not think that we have one way of accessing our thoughts and judgments and quite another one of knowing our sensations, one relying on pure apperception and one relying on inner sense. Whatever we become aware of in inner sense presupposes pure apperception to become an object of knowledge; and whatever we want to know through our power of thinking or judging can only be known ‘as it appears in respect of an intuition’ and by making itself intuitable according to relations of time.Footnote 18
II.
As we have seen, the two forms of self-consciousness – pure apperception and inner sense – need to co-operate in order to grant us cognition of ourselves. Whereas it is widely recognized that inner sense presupposes the synthetic unity of apperception, it is less commonly taken into account that pure apperception only leads to a determinate cognition of myself – instead of mere consciousness of my pure I – by reference to an ‘intuition of the manifold in me, by which I determine this thought.’ It is only when we realize this that we see the true problematic force of Kant'sfundamental distinction of the two modes of self-relation. Kant is not distinguishing two types and regions of self-knowledge that could happily co-exist alongside one another, one concerning our sensations knowable through inner sense and one concerning our thoughts and judgments knowable through the pure force of thinking. Rather, he describes two ways in which we relate to ourselves that need to co-operate for us to know anything about ourselves at all, be it about our sensations or about our thoughts and judgments.
Only now am I arriving at the aspect that interests me most in Kant'saccount: Kant does not only require that these two forms of self-relation should co-operate, he also characterizes them in a way that makes it hard to see how they can, in fact, give rise to a unified perspective on the same subject. On Kant'saccount, humans can only know themselves on account of a double consciousness of themselves: Consciousness of ourselves as the purely active, spontaneous thinkers of our thoughts and consciousness of ourselves as the objects of our inner sense. Only to the extent that we become an object to ourselves is there something to know at all. But to the extent that we become an object to ourselves we can only know ourselves as an appearance, and even more to the point: we become accessible to ourselves merely as the result of an activity, not as the activity itself. Even though it is our very own activity, we cannot cognize it directly. It thus seems that, on Kant'saccount, it is impossible for us to cognize ourselves as subjects. To cognize ourselves, we have to turn ourselves into objects. Yet, through the self-consciousness of pure apperception, we are aware of ourselves qua subjects. To unify both self-relations into a coherent form of self-knowledge we would need to find a way of cognizing ourselves in such a manner that, qua subjects, we know ourselves as objects. It is not immediately clear, however, how this should be possible. There is a deep sense in which it seems that ‘awareness of something as an object and awareness of it qua subject are mutually exclusive modes of awareness’, to borrow a formulation by Quassim Cassam (Reference Cassam1997, 5).
Recall once again the fundamental way in which Kant distinguishes, even opposes the form of pure apperception and empirical apperception such that it can give rise to the impression that these two forms of self-knowledge have different objects: As Kant proposes, pure apperception is a consciousness of the ‘I’ as the subject of thought, whereas empirical apperception makes me aware of the ‘I’ as the object of awareness (Kant Reference Kant and Louden2006, 7:134FN).Footnote 19 In pure apperception, I become conscious of what I do, whereas inner sense gives me access to what I undergo.Footnote 20 Pure apperception gives me awareness of the determining self, where I may cognize the determinable self through inner sense (Kant Reference Kant and Kemp Smith1929, B407; B157FN). And in pure apperception I become aware of something that exists in the deed without becoming conscious of it as either appearance or thing in itself (Kant Reference Kant and Kemp Smith1929, B423FN); in inner sense, I become aware of myself as a fact, knowing it inevitably only as an appearance. It thus seems that these two modes of self-awareness give me access to two fundamentally different types of being, such that the unity of the ‘I’ that we are conscious of in these two ways becomes mysterious. As Kant grants in the Anthropology, it at least seems that the ‘I’ here is ‘double’, ‘which would be contradictory’ (Kant Reference Kant and Louden2006, 7:134FN). What is this ‘I’ that is supposed to be both the subject of my thinking and the object of my perceptions, both only existent in the act and a resultant fact, both a spontaneous act and the being affected by such a spontaneous act?
Given this duplicity, it is not entirely clear in what sense we can identify the active subject of my thinking with the constituted object of my perceptions such that we can say that what I come to know in inner sense is my very own active self. Even though it is the indispensable condition of my cognitive activity, it seems that the ‘I’ of the ‘I think’ can only be known in a distorted form: I cannot cognize it on its own terms as the pure act that it is, but only come to know it by turning it into an object of knowledge, a determinable self, an appearance, and a fact. It thus seems that I cannot know myself directly to be the very sort of thing that makes me a knower. Or, put differently, to know myself, I have to become something else: I have to turn myself into a phenomenon, fall into nature, manifest myself as an object of my perception. To fully understand myself as the being that I am would therefore mean to realize that my self-awareness has an irredeemably double, even ironic structure; it is split between, and yet fated to hold together, my transcendental consciousness of myself as the thinker of my thoughts and my empirical knowledge of myself as the object of inner sense.Footnote 21
In Kant'streatment of these issues, it becomes clear that Kant has an acute sense of the possible tension between these two ways of relating to myself. It is, however, not entirely clear what he wants to make of this. He seems to oscillate between three strategies: (i) first, downplaying the tension such that it is not that mysterious after all how we come to identify the ‘I’ of pure apperception and the intuited ‘I’ of inner sense. (ii) A second, competing strategy we find in Kant is to embrace the very split as our true condition which defines us as human beings and demonstrates the truth of transcendental idealism. (iii) A third strategy is to embrace this split, but still recognize that something more needs to be said in order to make sense of this condition.
The first strategy turns on Kant neglecting the contrast of subject and object, activity and passivity, determining self and determinable self, deed and fact, and employing the distinction of thing in itself and appearance to make sense of the split. Since it is a general and familiar feature of Kant'saccount that we can only know appearances, it should not puzzle us that this is true for ourselves as well. To become an object of knowledge requires us to become an appearance and hence we cannot know ourselves as we are but can only know ourselves as appearance. In this way, Kant hopes to resolve the apparent ‘paradox’ of inner sense (Kant Reference Kant and Kemp Smith1929, B152f.). However, this response fails to address the fact that the problem, in this case, is much more specific: In order to appear, the ‘I’ has to take on the shape of an object being thought. But we are at the same time aware of the ‘I’ as existing as a pure activity. How is it, then, that ‘I, as intelligence and thinking subject’ can in fact truly ‘cognize myself as an object that is thought’ (Kant Reference Kant and Kemp Smith1929, B155, emphasis added, transl. modified)? Appearance in this specific instance seems to verge on dissimulation or distortion since it portrays what is the thinking subject as merely an object being thought. The ‘I’ as intelligence does not actually seem to appear, but rather manifests itself in an indirect form that presents us with a being of a fundamentally different kind. The object of inner sense is not so much the appearance of the intelligence but rather its symptom or trace.Footnote 22 If we are, however, required to identify this intelligence and its symptom, this puts us under a restriction far more difficult to meet than the idea that we should identify a thing and its appearance as being of the same subject. I am not sure if Kant really captures the contrast that he has described, when he suggests that the contrast between the two I'sis one merely in form, not in content.Footnote 23 In addition, it is puzzling how Kant could think that the distinction between appearances and things in themselves could be helpful here given that he had explicitly argued that the ‘I’ of the ‘I think’ is not a thing at all – neither appearance nor thing in itself – but something that exists only in the deed. In what sense can we take this I that is a pure deed and the active synthetic origin of all our cognition to be identical in content with the thing that we grasp in empirical apperception as an object of cognition?
These difficulties suggest that it may not work to downplay the tension. One should rather embrace it as part of the distinctive finite character of human cognition. Kant'sway of arguing for this is his insistence that the human cognizer is fated to become conscious of itself in this double way since intellectual intuition for us is not even a conceivable possibility.Footnote 24 As we do not have intellectual intuition the ‘I think’ cannot lend manifold content to itself but has to rely on sensible intuition. This does not only mean that the ‘I’ cannot create the world by simply thinking it. It also implies that the ‘I’ can only come to know itself in a distinctly indirect or mediated fashion. To become something known the ‘I’ needs to affect its inner sense and appear as an object of inner sense. It thus seems that by virtue of lacking intellectual intuition human cognizers are incapable of an immediate form of self-cognition.
This limitation quite naturally produces the urge to supersede the Kantian framework by questioning whether intellectual intuition is indeed impossible for us. This is a route that the German Idealists explored in various ways. Fichte, for example, thinks that to give a full account of the ‘I’ requires us to see that a certain form of intellectual intuition underlies its originary form. He repeatedly raises doubts as to whether Kant was indeed justified in denying us any form of intellectual intuition. Fichte'spoint, however, is not to suggest that we can, in fact, create the world just by thinking it and that our intellect is thus indistinguishable from a divine intellect. His suggestion is actually much more internal to Kant'sown project: he thinks that what Kant calls pure apperception is a form of original self-consciousness that is itself evidence of the actuality of an intellectual intuition in human cognition.Footnote 25
Be that as it may, Kant himself certainly continues to deny that there is anything close to intellectual intuition in our theoretical cognition. For Kant himself, this lack points us to the limitations that any form of theoretical cognition imposes upon our capacity for self-knowledge. According to the theoretical model, knowledge requires a subject thinking an object given to it. In order to know myself, I need to become an object to myself, thereby effectively separating myself from myself, the object known from the subject knowing this object. I can only know the active subject obliquely through the object known. If we try to apply this theoretical model to the original self-consciousness of pure apperception itself, it seems obvious that it is inadequate to capture its structure. All of this might suggest that in order to make sense of our self-relation, we need to turn to practical cognition instead. In practical cognition – in determining myself through knowledge of what I should do and who I ought to be – I may become an object to myself in a different sense and I may gain a different understanding of the original self-consciousness of the ‘I think’ (or rather: the ‘I will’). Kant suggests that our practical form of cognition indeed gives us access to an original self-consciousness ‘through which our reality would be determinable, independently of the conditions of empirical intuition’ (Kant Reference Kant and Kemp Smith1929, B430). Kant insists that in the practical context he is still not attributing intellectual intuition to us, Footnote 26 but he is specifying a type of cognition that comes close to a functional equivalent. Consider this passage from the First Critique:
Should it be granted that we may in due course discover, not in experience but in certain laws of the pure employment of reason – laws which are not merely logical rules, but which while holding a priori also concern our existence – ground for regarding ourselves as legislating completely a priori in regard to our own existence, and as determining this existence, there would thereby be revealed a spontaneity through which our reality would be determinable, independently of the conditions of empirical intuition. And we should also become aware that in the consciousness of our existence there is contained a something a priori, which can serve to determine our existence – the complete determination of which is possible only in sensible terms – as being related, in respect of a certain inner faculty, to a non-sensible intelligible world. (Kant Reference Kant and Kemp Smith1929, B430f.)Footnote 27
This brings us to a third possible strategy: embracing the split without denying that a different way of relating to it or of putting it to work is needed in order for it to make sense. With this we return to the very basic idea with which I started: that self-knowledge is fundamentally practical. Even though there already is an agential aspect in pure apperception, in the context of Kant'stheoretical philosophy this agential aspect seems curtailed, since theoretical cognition always includes a peculiar disjunction between the subject and the object of cognition. This agential aspect can thus only become manifest in the indirect form of the consciousness of an object symptomatic of, but at the same time reifying and obfuscating, the pure self-activity of the ‘I’. The only paradigm of an object of knowledge that is available in theoretical cognition seems inadequate to the thing to be known. The question is whether practical cognition offers a different understanding of the object of cognition and the disjunction of subject and object of knowledge, which affords us a different sense of the I'sactivity and a fuller conception of how this activity can be known.
III.
At crucial points in the Groundwork and in his Critique of Practical Reason, Kant returns to the thought that the ‘I’ is accessible to itself in two fundamentally different ways. The ‘I’ knows itself through its appearances in inner sense and thus knows itself as an object in nature; in virtue of pure apperception, however, the human being becomes aware of ‘what there may be of pure activity in him (what reaches consciousness immediately and not through affection of the senses)’ (Kant Reference Kant and Gregor1996a, 4:451; cf.Reference Kant and Gregor1996, 5:114). This gives the human being reason to consider itself as an intelligence. In his practical philosophy Kant shifts the perspective on the relation of these two stances toward the self in three fundamental ways: first (i) he does not limit his consideration to the way in which the appearance of ourselves necessarily falls short of giving us adequate cognition of the ‘I’ of pure apperception, but explores the way in which the ‘I’ of the ‘I think’ grants us a standpoint from which we may transcend and practically liberate ourselves from our own appearance. Secondly, (ii) Kant re-describes the ‘I’ that we do not cognize but already become conscious of through pure apperception as our ‘proper self’ (Kant Reference Kant and Gregor1996a, 4:452, emphasis added). Only as an intelligence – only as the ‘invisible self’, as the second Critique has it (Kant Reference Kant and Gregor1996, 5:162) – am I my proper self; as a human being, I am a merely derivative ‘appearance of myself’. And thirdly (iii), the unity of these two selves is neither downplayed as unproblematic nor simply stipulated as necessary, it is presented as a task. We thereby not only grasp the proper self as the source of the determination of the ‘appearance of ourselves’, but also disclose the task of bringing forth a type of appearance that can be understood as expressive of the activity that underlies it. In Kant'spractical philosophy, we are thus not simply left with the disjunction of my intelligible and my empirical self, the ‘I’ of pure activity and its reification as a sensible object, but confronted with the task of their mediation. It is from the point of view of his practical philosophy that we see Kant developing a response to the question how spontaneity itself may appear and how we may know of ourselves, qua subjects, as objects.Footnote 28
To enact these three shifts, we need to gain a more determinate consciousness, if not knowledge, of our proper self. The awareness of the ‘I’ of pure apperception might open up the possibility of a different standpoint, but it gives us no positive determination of this stance. It is the crucial advancement of practical over theoretical philosophy that it provides us with ‘a fact that points to a pure world of the understanding and, indeed, even determines it positively and lets us cognize something of it, namely a law.’ (Kant Reference Kant and Gregor1996, 5:43) This gives us a more robust way of understanding the way in which our empirical self is an appearance of our proper self not just in the sense that it is a derivative effect of it, but also in the sense that it can be grasped as its expression. And finally, Kant'spractical philosophy tries to bring into view the operations that bring about an empirical order in which the proper self might manifest itself in such a way that it actualizes itself in it.Footnote 29
The more determinate consciousness of the proper self becomes available to us in the form of the fact of reason: the consciousness of the moral law by means of which we finally become aware of a positive determination of the intelligible realm. For us to know this law, our faculty of reason again has to affect us; this affection does not result in mere facts of inner sense, but in what Kant calls the sole factum of reason, which we subjectively grasp by means of a distinctive kind of feeling Kant calls Achtung. This feeling, according to Kant, allows us to become aware of ourselves not only as the being affected but also as therein affecting ourselves.Footnote 30 The factum of reason that gains subjective reality by means of this feeling is neither a datum nor a construct of reason, but the very deed of reason itself. That we do not perceive the deed of reason as a dead fact or object is plain from the form our awareness of it takes. It presents itself as a task to us, as what is to be done. In my moral consciousness, the deed of reason thus does not present itself as a fact but as a command calling for further deeds.
According to the second Critique, this law requires us ‘to furnish the sensible world, as a sensible nature (in what concerns rational beings) with the form of a world of the understanding’ (Kant Reference Kant and Gregor1996, 5:43). Applying this general understanding of the task of the moral law to the problem of the self, it seems to require that one inform one'sempirical self in such a manner that it becomes the expression and the reflection of the very form of a proper self. I take this not only to mean that the empirical self should abide by rules given from somewhere else but requiring that the empirical self takes on the very shape of the proper self: the very shape of an active self. This would seem to require that it would need to reflect its inherent dependence on, and determination by, the pure practical self in the very mode in which it is an appearance and an object. This presents us with the task of rearticulating the empirical self as the manifestation of the very difference between proper and empirical self, as the re-entry of this difference into the world of appearance. Note the fact that Kant ties our capacity of reason to our ability to distinguish ourselves from ourselves: As he writes in the Groundwork, ‘the human being... finds in himself a capacity by which he distinguishes himself from all other things, even from himself insofar as he is affected by objects, and that is reason’ (Kant Reference Kant and Gregor1996a, 4:452, first emphasis added), and the Prize Essay seconds that by pointing to the unexplainable but undoubted fact that ‘I, who think, can be an object (of intuition) to myself, and thus distinguish myself from myself’ (Kant Reference Kant, Allison, Heath, Hatflield, Friedman, Allison and Heath2002a, 20:270). Expressing a proper self thus seems to require an empirical self that internalizes this self-distinction.Footnote 31
How exactly to think of this and how to relate this form of ethical elaboration of one'sself-difference back to one'stheoretical self-relation is far from fully developed in Kant'spractical philosophy, and only partially advanced in his third Critique. My complaint here is not that Kant has failed to demonstrate how we can ultimately overcome the duplicity, but merely that he has not told us enough about putting it to work. If Kant'saccount is right, simply superseding the double sense of myself might well amount to losing myself. It is thus key to acknowledge that Kant is committed to the idea that the unity of our self-knowledge can only be the unity of a difference, and that even the practical determination of our existence does not proceed by means of intellectual intuition, but has to go through the trouble of taking up and transforming the sensible world such that it becomes expressive of its rational ground. In other words, to require that Kant should tell us more about the relation of the proper and the empirical self cannot mean that the distinction between pure and empirical self-consciousness, and between the theoretical and the practical articulation of this difference, should be extinguished. It must rather be sublated: We need to understand pure apperception in a way that is reflective of its dependence on empirical apperception; and awareness of my own appearance is to be conceived in such a way that it does not become the reification or obliteration of the active self but a way for its activity itself to gain objective consistency.Footnote 32
We started with the diagnosis that modeling self-knowledge on ‘theoretical awareness of objects’ obscures the distinctive character of self-knowledge and suggested that we have to understand self-knowledge in practical terms instead. To the extent that transcendental apperception on Kant'saccount is awareness of myself as pure activity, it is natural to assume that we may be able to derive such a practical understanding of self-knowledge from Kant'snotion of pure apperception. I have tried to cast some doubt on this hope by investigating Kant'saccount of pure apperception. I have first tried to show that pure apperception alone yields no cognition of myself and that self-cognition on Kant'saccount seems to require the co-operation of pure apperception and inner sense. Secondly, I have suggested that in the context of Kant'stheoretical philosophy this co-operation seems to leave us with a split sense of our own selves. A sense of the practical that is derived from pure apperception alone is thus revealed to be insufficient. It is a form of the practical still tied to a theoretical framework, governed by an irreducible gulf between subject and object. In the third section, I have suggested that we can think of Kant'spractical philosophy as a response to this problem, developing a form of cognition by means of which the subject actualizes itself objectively and in knowing its practical objects can come to know itself qua subject. The self-knowledge delineated here is practical, but not in the sense of an immediate awareness of my pure activity; it is practical in the deeper sense of coming to know myself through my deeds.
Acknowledgments
I had the great opportunity to discuss versions of this paper at Ryerson University, the University of Leipzig, and the University of Essex. Many thanks to the participants for their perceptive comments. They have largely helped me clarify my tentative thoughts on the issues discussed above. Special thanks are due to Cristóbal Garibay-Petersen, Thomas Land, Dirk Setton, and Alexey Weißmüller for instructive comments on an earlier draft.
