I. Introduction

SO FAR I have supported the claim that Kant regards empirical concepts as composed of an imaginative material that functions as an essential ingredient in the very experiences of which those concepts are predicated. The imaginative material is a manifold of anticipations and retentions, although Kant often adopts a more noematic perspective. When he does so, he tends to suggest that the most basic material for conceptual synthesis is rather a manifold of possible “appearances,” or perhaps even the parts of space and time themselves. In any case, we have not yet carefully examined the sense in which imaginative material is to be regarded as subject to the imposition of “forms,” by virtue of which it is finally embodied in full-blown concepts. To this extent, we still fall short of what Kant regards as full-blown “consciousness.”

The forms in question can only be what Kant calls “forms of judgment”—the ones that he identifies, in the Metaphysical Deduction, on the basis of an examination of the “logical function of the understanding in judgments” (A70/B95). As we saw in the preceding chapter, this implies that the forms of judgment cannot be mere forms for the connecting of concepts into judgments. In the privileged case (if there are indeed others) of judgments of intuition, they can involve nothing other than “that form which the understanding is able to impart to representations”—that is, to intuitions—themselves (A56/B80):

The same function which gives unity to the various representations in a judgment also gives unity to the mere synthesis of various representations in an intuition. . . . The same understanding, through the same operations by which in concepts, by means of analytical unity, it produced the logical form of a judgment, also introduces a transcendental content into its representations, by means of the synthetic unity of the manifold in intuition in general. (A79/B104-5)¹

The upshot, we have seen, is not simply the constitution of a judgment. It is the original formation of a concept for “use” in a judgment.

But how are we to understand this type of “constitution”? How can Kant suppose that a body of material, potentially ingredient in an intuition,² is able to become ingredient in a predicative “term” that one merely “applies” to that intuition among others? Kant’s view appears to be that consciousness itself is the form required for the purpose:

[C]onsciousness in itself is not a representation distinguishing a particular object, but a form of representation in general, so far as it is to be entitled cognition; for of it alone can I say that I am thereby thinking something. (A346/B404)

[T]he merely subjective form of all our concepts [is] consciousness. (A361)

More specifically, what is in question is some special kind of unity of consciousness:

The word ‘concept’ might of itself suggest this remark. For this unitary consciousness is what combines the manifold, successively intuited, and thereby also reproduced, into one representation. (A103)

This thoroughgoing synthetic unity of perceptions is indeed the form of experience; it is nothing else than the synthetic unity of appearances in accordance with concepts. (A110)

But this unity of possible consciousness [of objects as appearances] also constitutes the form of all cognition of objects. . . . [T]he mode in which the manifold of sensible representation (intuition) belongs to one consciousness precedes all cognition of the object as the intellectual form of such cognition, and itself constitutes a formal cognition of all objects a priori, so far as they are thought (categories). (A 129)

[Categories] are nothing but forms of thought, which contain the merely logical faculty of uniting a priori in one consciousness the manifold given in intuition. (B305)

This unity of consciousness (of the connection of our representations) is as much in us a priori as the foundation of all concepts, as the form of appearance is as the foundation of intuitions.³

A judgment is the representation of the unity of consciousness of various representations or the representation of their relation, so far as they constitute a concept.⁴

In other words, full-blown “consciousness” and the relevant “unity” of it are not simply the upshot of conceptualization. They are what constitute concepts in the first place.

It is also clear that Kant regards the relevant unity of consciousness as inseparable from “self”-consciousness. The link appears to lie in his claim that the requisite unity of consciousness does not involve a consciousness of objects, or appearances, simpliciter, but of something that he calls the synthesis of appearances:

The original and necessary consciousness of the identity of the self is thus at the same time a consciousness of an equally necessary unity of the synthesis of all appearances according to concepts. . . . (A 108)

Assuming that synthesis involves mental activity, the suggestion may seem helpful. It may seem implausible to reduce self-consciousness to the consciousness of a “unity” (or even possible unity) among objects or appearances. But it may not seem implausible to reduce it to the consciousness of some kind of mental activity (or possible activity). For example, self-consciousness might be supposed reducible to a consciousness of the possibility of appropriately justified “self-ascription.” However, I shall try to take Kant more literally in his own suggestion that original self-consciousness is consciousness of oneself directly in one’s consciousness of objects or appearances.

It is important to be clear that the suggestion would not contradict Kant’s insistence, contra Leibniz and Wolff, that the difference between intuition and conception is not reducible to a difference in the clarity of one’s consciousness of objects (cf. A43/B60-1). This is because, according to suggestion, the relevant consciousness of objects or appearances is itself to be explicated in terms of a mode of coming to consciousness of one’s own imaginative condition, insofar as the latter is embodied in intuitions. In other words, any relevant difference in the “clarity” of one’s consciousness of objects or appearances is itself to be explicated in terms of an irreducible relationship (of intentional correlation) between a certain mode of consciousness of objects and a mode of self-consciousness. To that extent, one might say both that the distinction between intuition and conception, hence original self-consciousness, is explicable directly in terms of a consciousness of objects or appearances and that it is explicable—only apparently to the contrary—in terms of a function whereby one is originally conscious of portions of one’s own imaginative condition. (The latter suggestion might then help make sense of Kant’s account of empirical concepts in terms of “abstraction.”)⁵

It is Kant himself who says that any truly cognitive apprehension does not merely entail, nor is it merely entailed by, some mode of self-consciousness. Rather, it is originally constituted through the effecting of the latter, and indeed through the effecting, thereby, of some special mode of “clarity” in consciousness:

All our cognition has a twofold reference: first a reference to the object, second a reference to the subject. In the first respect it is referred to representation, in the second to consciousness, the general condition of all cognition in general.—(Strictly, consciousness is a representation that another representation is in me.) . . . Variation in the form of cognition rests on [emphasis added] a condition that accompanies all cognition, on consciousness. If I am conscious of the representation: then it is clear; if I am not conscious of it, obscure. . . . Since consciousness is the essential condition of logical form in all cognition, logic can only be concerned with clear, not with obscure representations.⁶

This is not the only passage in which Kant equates consciousness (of objects) with the mere “clarity” of representations to their subject.⁷ The only way to avoid the apparent implausibility in this is, I suggest, to develop a sense in which the latter is not, as the claim might suggest, a consciousness of representations as distinct from a consciousness of ordinary objects or appearances. Rather, it must be a consciousness that is constituted in one’s consciousness of objects or appearances in the first place.

However we spell it out, it is the primary aim of the Deduction to establish at least a necessary connection between a more “subjective” (and so regarding one’s own states?) and a more “objective” (and so regarding objects or possible objects?) notion of what needs to get “synthesized,” in order to constitute a truly cognitive consciousness of objects or appearances. In these terms, the strongest form of such an endeavor might then appear to be the one that attempts to show that each of these is a necessary condition of the other. For example, we might try to show that the ability to be conscious of oneself as a subject is a necessary condition of the ability to recognize (or misrecognize) objects, and that the latter is also a necessary condition of the former. Correspondingly, a weaker approach in one direction, but stronger in the other, might attempt to show that at least the ability to achieve verifiable knowledge of oneself, as a subject, is a necessary condition of the ability to recognize or misrecognize objects, and that the latter is also a necessary condition of the former. Or reversing the permutations: perhaps the ability to be merely conscious of oneself as a subject is a necessary condition of the ability to achieve a verifiable knowledge of objects, and— more significantly—perhaps the latter is also necessary for the former. Whatever the variation, Kant’s contribution might be supposed to lie in the establishment of an external, albeit logical, connection between “consciousness” of self and of objects or their aspects.⁸ I want to suggest that he is aiming at a much closer connection.

II. Consciousness and Concept Formation

When we consider what we mean by consciousness, in particular by the consciousness of “objects,” then it is worth bearing in mind that we are considering something of which it has only in fact been recently permissible to speak, in either philosophical or ordinary English. The situation has been similar in French. In the traditions of Locke and Descartes, philosophical usage has only gradually extended the domain of “consciousness” beyond the apprehension of one’s own individual existence, together with the apprehension of whatever additional data might be comprised in the latter; the development of ordinary usage has been even more gradual. From antiquity through the 1700s, a more outward-looking notion was, to be sure, current as well. But it applied at most to one’s (co-) awareness of facts or of states of affairs, not to one’s awareness of objects or of the world.⁹ The earliest Oxford English Dictionary entry that concerns “external objects” is from Pope’s Rape of the Lock (1712-14), although there may be earlier candidates in Dryden.¹⁰ Neither Bailey¹¹ nor Johnson¹² recognizes the usage.

Though troubling to say it implies that the view was contested in his day, Reid puts it bluntly: “It is improper to say, I am conscious of the table which is before me. I perceive it, I see it, but I do not say I am conscious of it.”¹³ The Oxford English Dictionary classifies the usage as poetic. French, or at least academic French, simply ignores the option. Into the 1930s, it regards even the notion of conscience of oneself, and conscience of one’s inner states, as purely philosophical.¹⁴ Coste, in translating Locke, had to coin a term for the purpose, as did Leibniz in his commentary on Locke¹⁵ The earliest citation that I have found of an extension of the term to the apprehension of “external phenomena” dates from 1864 in French.¹⁶

These facts should lead one to wonder whether something might not be involved in the notion of a “consciousness” of objects that is not easily capturable in less “poetic” talk about cognition or perception of objects—merely supplemented by the apparently additional fact that one needs to have a “consciousness” (a higher-order cognition or perception?) of that very cognition or perception.¹⁷ In any case, Wolff had, by 1719, already established the notion of Bewusstsein as readily applicable to the consciousness “of ourselves and of other things [Dinge].”¹⁸ In the light of this, it may be all the more notable that Kant himself remains disinclined to speak of one’s Bewusstsein of objects (whether of Dinge in general, of Gegenstände, or of Objekte). He tends to do so only in contexts that emphasize that the objects in question are appearances or “representations.”¹⁹ As we have already seen, Kant even appears to offer a definition, according to which the proper objects of consciousness are one’s own internal states. All of this may seem to provide difficulty for my reading. But as I have also noted, it is hard to know what to make of Kant’s official “definition,” in the light of his own view as to the role of consciousness in the constitution of cognition. Indeed, in the very passage in which he proposes it, and in which he claims that one’s cognition of objects is constituted through some mode of consciousness of internal states, Kant illustrates what is supposed to be thereby constituted by referring to one’s Bewusstsein that something is a house.²⁰ (That Kant in general insists on favoring self-consciousness, when he speaks of Bewusstsein, may of course be explained in a number of ways. As suggested earlier, it might simply reflect his fear of being taken to espouse the view of Leibniz and Wolff, that the difference between a mere intuition and a true cognition of objects lies in an irreducible difference with respect to one’s clarity of consciousness of the latter.)

At least in particular cases, there seem to be readily available alternatives to talk about the consciousness of objects or appearances. Kant himself simply speaks of the cognition (Erkenntnis) of such things. Apart from that, where he is specifically concerned with Bewusstsein, he seems to be concerned with one’s consciousness, not of objects or appearances as such, but of the synthesis of objects or appearances (or else with one’s consciousness of the unity of such synthesis). But the difficulty is that there are a number of things that might be meant by the latter.

The following passage makes it clear that, whatever kind of consciousness, and of unity of consciousness, is essential to Kant’s account of cognition in the Critique, it can only be comprehended in terms of an account of imaginative anticipation and retention in experience, that is, in terms of an account of Reproduktion. The passage is from the third of the sections that the first-edition Deduction devotes to the problem of synthesis, “Synthesis of Recognition in a Concept”:

If we were not conscious that what we think is the same as we thought a moment before, all reproduction in the series of representations would be useless. For it would in its present state be a new representation which would not in any way belong to the act whereby it was to have been successively generated] zu dem Aktus, wodurch sie nach und nach hat erzeugt werden sollen]. The manifold of the representation(s) would never, therefore, form a whole, since it would lack that unity which only consciousness can impart to it. . . . The word ‘concept’ might of itself suggest this remark. For this unitary consciousness is what combines the manifold, successively intuited, and thereupon also [dann auch] reproduced, into one representation. This consciousness may often be only faint, so that we do not connect it with the act itself, that is, not in any direct manner with the generation of the representation, but only with the outcome. But notwithstanding these variations, such consciousness, however indistinct, must always be present; without it, concepts, and therewith knowledge of objects, are altogether impossible. (A103-4)

This passage may appear to deal with the problem of unity of consciousness solely in terms of a problem concerning the ascribability of representational states to subjects. In these terms, the relevant “manifold of representations” will presumably just be all such states, considered in relation to single subjects.²¹ Alternatively, it might be all of the possible appearances that correspond, as correlates, to those states. In either case, the “one” to which the “many” of the manifold is supposed to relate is simply a single subject of self-ascribable states. Unlike the first-edition Deduction, the second may appear to deal exclusively with this aspect of the problem of “unity of consciousness.” But there, as we have seen, this was not in fact Kant’s primary concern. As I hope to show, it is different in the present case as well. Kant’s argument is of course meant to bear on the problem of self-ascription. But the crux of the argument is that a concept of self-consciousness defined solely in terms of this notion can at best be derivative from a more basic concept of self-consciousness.²²

Kant does not, in the present passage, raise a general question concerning the possibility of the self-ascription of representational states. He raises a question concerning a more fundamental ability. The latter is the ability to anticipate, as an integral part of the conceptualization of appearances, states that may eventually be self-ascribable states. (Alternatively, he raises a question concerning the anticipability of the correlative appearances.) From the context, it is also reasonable to suppose that the role of “retention” is to be interpreted in this light. The discussion of retention must, of course, be relevant to the problem of one’s ability to self-ascribe past states.²³ But it cannot in the first instance merely bear on that question. As the quoted passage suggests, the primary question concerning “retention” is just as much a question concerning the notion of anticipation, so far as it is relevant to conception. Of course, as I have been reading Kant, the primary question concerning both must bear, not merely on a subject’s ability to anticipate and retain, but on the way in which conceptual states themselves anticipate and retain.

With this distinction in mind, Kant’s suggestion might be deeper than it first appears to be. For example, he might be suggesting the necessity of states that anticipate states that in their own turn embody, beyond their apprehension of a given appearance or appearances, a self-representation as having been anticipated in the first place: in other words, states that anticipate the possibility of their own retention. We shall see later that something like this—together with even more complex modes of “higher-order” anticipating—will in fact be required for a full elaboration of Kant’s view.

In any case, we shall be concerned only with a particular manifold of eventually self-ascribable representations. We shall be concerned only with those that are anticipated and retained by any given representation in its own right, insofar as the latter is regarded as subject to conceptualization. Given this, the most natural suggestion regarding the relevant “unity of consciousness” is that it is not in fact a unity that bears on representations simply by virtue of their being all (at least in principle) ascribable to a single subject; rather, it must be whatever kind of unity is required for the forming of concepts out of representations. (It helps to recall that it is a question concerning the nature of concepts that provides the focus for Kant’s formulations throughout this section. In line with this, Kant concludes his own elucidation of what the synthesis of “recognition” adds to that of “reproduction” with the conclusion that “the concept of [any] number is nothing but the consciousness of this unity of synthesis” (A103; emphasis added). At A107, similarly, he concludes that the transcendental unity of consciousness is the a priori ground of all concepts, not simply the ground of the knowledge to be gained with the help of concepts. I have commented on similar passages in the second-edition Deduction.)

The passage quoted above (from A103-4) clearly suggests an interplay between some kind of structure first introduced by means of a truly intellectual “consciousness” and some kind of structure already available through the purely imaginative capacity for Reproduktion. In that passage, Kant refers to the “act whereby” certain representations were “to have been successively generated.” Presumably, this involves just the sort of anticipative capacity with which Kant had been earlier concerned: that is why Kant speaks of representations that were (already) to have been generated, i.e., were already anticipated as possible, independently of a contribution from the higher powers now to be examined. But Kant now says that this anticipative capacity is insufficient for the conceptualization of appearances. What is needed, he seems to say, is what can only be provided, not simply by one’s ability to anticipate and retain, but by a certain mode of consciousness of whatever is anticipated or retained. To apprehend an appearance in a truly conceptual way, one must also have some kind of consciousness of those additional appearances, anticipated in the act of apprehension in question, precisely as appearances that are therein anticipated.

One may wish to attempt a simplification by reducing both of these elements in terms of the representational powers of already constituted conception. For example, one might take the “anticipations” ingredient in conceptualized apprehension simply to be judgments, implicit in such apprehension, bearing on the possibility of additional apprehension. And one might take whatever “consciousness” those acts essentially involve to be constituted by those same judgments, insofar as these are indeed implicit in the apprehension in question. But quite to the contrary of this, the proposed account calls for both a level of consciousness and a kind of anticipation and retention, ingredient in conceptual acts, that is (as it were) not yet conceptually judgmental at all. The full-blown act of conceptualization can only be what issues from the “application” of the former to the latter.

If we follow the proposed approach, then we will need, strictly speaking, to distinguish three possibilities, whereas we have so far recognized only two. We have, of course, the two possible cases: (a) where primitive imaginative anticipations and retentions are ingredient in intuitional states, but are “not yet” subjected to the forms of consciousness required for full-blown conceptualization, and (b) full-blown conceptualization in regard to imagined or perceived objects or appearances. (We may ignore the case of possible intuitions whose material is even more primitive than that of (a).) But to the extent that (b) is possible in the first place, a third and half-way case ought also to be possible: (c) where an intuitional state embodies the anticipation or retention of the mere possibility of states of type (b). That would be less primitive a phenomenon than (a), but would still fall short of an actual instance of (b). The question would then seem unanswerable: whether for a genuinely “conceptual” mode of consciousness we should strictly require (b), or at most something like (c). In any event, if states of type (b) are the paradigmatically “conceptual” ones, states of type (c) are presumably available only to creatures capable of the former states as well. I assume that Kant simply failed to distinguish between these two notions.

We might choose to regard instances of (c)-representation as involving a kind of “pre-predicative judgment. ” The title of judgment would of course associate them with states of type (b); the absence of appropriate intellectual form would associate them with (a). But then why not have felt free, pace Kant, to regard instances of (a)-representation as modes of “judgment” or even of “predication” in the first place—ones that are simply on a more primitive level than either (b) or (c)? What is important, in pursuing such issues of phenomenological taxonomy, is not to be misled by mere questions of terminology, and to remain clear about one thing: that whatever the relevant differences are, between the more and the less primitive levels, they could not merely involve differences, for example, in degree of complexity with regard to whatever type of anticipation and retention is already in question on the most primitive level. Some difference in “form” must also be involved.

With respect to the requisite notion of “form,” it may in fact be possible to adopt an approach somewhat different from what I have so far suggested, consistently with the general line of this study. To do this, we might try to avoid the introduction of an irreducible form of consciousness, whose function it is to provide for the eventual formation of concepts. Instead, we might simply make use of our two notions of relatively primitive anticipation and retention, on the one hand, and of some function whereby this more primitive material gets embedded in sets of higher-order anticipations and retentions. By the latter, again, we can not mean sets of more complex anticipations and retentions. What we must mean is sets of anticipations and retentions that do not merely anticipate and retain correspondingly possible appearances (or correspondingly possible intuitions), but rather correspondingly possible appearances (or intuitions) qua appropriately relatable to these (actually given) appearances (or intuitions)—or perhaps even qua appropriately relatable to these (actually given) anticipations and retentions themselves. Here, the “forms” required for the constitution of full-blown conceptions, out of manifolds of primitive anticipations and retentions in intuition, would themselves simply be sets of appropriately higher-order anticipations and retentions. This is not quite the same as the account that I am proposing. But it is possible that, here too, Kant himself was not sufficiently clear as to the alternatives.

Once we have granted anticipation and retention on the more primitive (purely “animal”) level, and introduced the possibility of their ingredience in intuition itself, then, in at least one respect, it would be a triviality to grant the possibility of “higher-order” anticipations and retentions in intuition. For example, it would be a triviality to grant the possibility, even on a purely “animal” level, of the anticipation or retention of the very satisfaction of some anticipation or retention. (In a sense, to anticipate something and to anticipate satisfaction of that very anticipation are just the same thing. That is presumably what constitutes the triviality in the case.) But that kind of “higher-order” anticipation and retention is not, of course, what I had in mind in the preceding paragraph. I pursue this point in more detail later. It should become clear, at least by then, why the proposed account would not in any case run afoul of Kant’s own insistence that, with respect to consciousness, “there is no gradation leading from animals to men.”²⁴

The alternative still requires attention to the specific problem of “consciousness,” but at a different point in our analysis. In particular, we would need to be careful not to think that the role of self-consciousness could be accounted for simply in terms of our recognition of the “higher-order” anticipations and retentions in question. In those terms, we could suppose ourselves to have done some justice to the problem of self-consciousness only to the extent that the ingredience in intuition of any kind of anticipation and retention was already conceded to involve some degree of “consciousness” to begin with (i.e., some degree of consciousness of possible objects or appearances, or at least of possible intuitions). After all, if the ingredience of those lower-level anticipations and retentions did not by itself amount to some level of consciousness in the first place, then how could the mere addition of “higher-order” anticipations and retentions just by itself imply that the subject was now self-“conscious”? By hypothesis, we would have to admit that the subject was now in some way self-anticipative. But why regard self-anticipation as a mode of self -consciousness—so long as anticipation was not by itself conceded to involve any mode of consciousness to begin with?

This second way of reading Kant—according to which the intellectual “form” required for the formation of concepts is merely a kind of higher-order anticipation and retention—would therefore require the supposition that the mere ingredience of such material in intuition already amounts to a genuine level of intuitional consciousness. In principle, I see no reason why Kant could not have conceded the point. In that case, however, it would be the mere form of intuition itself, at least as applied to a body of appropiate material, that constitutes “consciousness” as such. The (not necessarily chronologically subsequent) introduction of some higher form could then at most be constitutive of a higher level of consciousness, including, of course, self-consciousness, as Kant is concerned with that notion. As it happens, the Critique itself only occasionally suggests that the act of conceptualization, together with the self-consciousness from which it is inseparable, involves a different “level” of consciousness from what sensibility already involves. But it is a suggestion that Kant in fact repeats in many of his reflections outside the Critique.²⁵ Indeed, as we have already seen, he sometimes goes so far as to equate the peculiar quality of intellectual consciousness with nothing more than a difference in clarity, with respect to what is already available as material on lower levels.

With respect to his explicit comments on anticipation in judgment, Kant’s own examples are in an additional way unclear or misleading as well. This is because they too specifically focus on the case of the eventual satisfaction of a subject’s anticipations. Kant considers the case, for example, of a subject who is counting:

If we were not conscious that what we think is the same as we thought a moment before, all reproduction in the series of representations would be useless. . . . If, in counting, I forget that the units, which now hover before me, have been added to one another in succession, I should never know that a total is being produced through this successive addition. . . . (A 103)

Here, the emphasis is on the ability to recognize that a subsequently apprehended appearance, or a subsequently experienced representation, is related in a certain way to an earlier one. The problem with this is that our concern needs to be with the element of conceptualization that is directed toward given appearances, and with the structure of the anticipations necessarily involved in it, quite apart from whether those anticipations are ever in fact satisfied by subsequent apprehension. What is important is not the satisfaction, but the anticipation of it. So far, what I have argued is simply that the latter is provided by a capacity for Reproduktion that is incorporable into intuitions independently of a capacity for conceptual “recognition.” The latter is then to be constituted by the (not necessarily chronologically subsequent) imposition of appropriate form on the “reproductive” material in question.

We might want to adopt a purely dispositional approach, or even some kind of functionalist approach, to the higher-order “consciousness” of potential satisfaction, supposed to be necessarily ingredient in the conceptualization of appearances. Doing so, at least, need not be incompatible with Kant’s insistence on something more than purely animal anticipation. For this, we might simply regard the relevant dispositions (or the relevant functional states) as involving dispositions to judge in the full-blown sense (or as states that are apt to generate fullblown judgments, under the appropriate circumstances). It may also seem implausible to insist on anything more than this, by way of an actual “consciousness” of potential satisfaction, insofar as that consciousness is supposed to be ingredient in conception as such. In any case, Kant himself acknowledges that we ought not to connect the relevant consciousness “only with the outcome” of that act whereby some additional representation “was to have been successively generated”; we must rather connect it “with the act itself, that is . . . with the generation of the representation” in question (A103-4). In saying as much, he is presumably not making the trivial point that, in order to conceptualize an appearance as what one had all along anticipated, one must indeed have anticipated that appearance all along. His point must rather be that anticipations of the relevant sort—however otherwise construed—must involve the anticipation, not simply of appearances, but of the eventual apprehension of appearances. Indeed, as we shall see, they must involve the anticipation of the apprehension of appearances qua eventually apprehensible precisely as thus anticipated. Kant’s point must be, further, that while this in turn needs to build upon animal anticipation (and “retention”), ingredient in intuitions themselves— whatever its nature—it needs also to involve a higher-order transformation of such material. It may have been Kant’s uncertainty as to the nature of the foundational material that led him to blur the distinction between the problem of the structure of anticipation as such and that of the possibility of “recognizing” cases of its satisfaction or frustration.

Now we may return to our question concerning the distinction between two notions of unity of consciousness. It is obvious that such a distinction plays a role in the Deduction. What is less clear is what it amounts to. As we have already seen, Kant himself distinguishes a transcendental, and “original,” unity of consciousness—or unity of “apperception”—from the merely empirical unity involved in the succession of representational states in a subject (A 107; cf. B139-40). It is of course clear that the mere succession of states in a subject is not the same as the subject’s consciousness of those states as its own. Hence it could not possibly account, for example, for the subject’s consciousness of its own anticipations as eventually satisfied or satisfiable. So what more is needed?

We might be tempted to suppose that what is lacking needs to be directly supplied by one’s capacity for judgment, namely, by one’s capacity for judgment in regard to oneself as a subject of experiences, and by whatever additional judgments this in turn entails. What would be “transcendental” about the capacity in question, beyond the fact that it is a necessary condition of cognition, might then be taken to be that it involves the thought of a unifying center for experiences that is not itself an experienceable item in its own right. It is at best an object of judgment:

This original and transcendental condition is no other than transcendental apperception. Consciousness of self according to the determinations of our state in inner perception is merely empirical, and always changing. No fixed and abiding self can present itself in this flux of inner appearances. . . . What has necessarily to be represented as numerically identical cannot be thought as such through empirical data. (A 107; cf. B133-4)

There is no doubt that Kant regards original, transcendental self-consciousness as something purely intellectual, somehow superadded to a body of more primitive (merely “empirical”) material. The second edition seems to make the point more plainly than the first:

It must be possible for the ‘I think’ to accompany all my representations; for otherwise something would be represented in me which could not be thought at all, and that is equivalent to saying that the representation would be impossible, or at least would be nothing to me. That representation which can be given prior to all thought is entitled intuition. All the manifold of intuition has, therefore, a necessary relation to the ‘I think’ in the same subject in which this manifold is found. But this representation is an act of spontaneity, that is, it cannot be regarded as belonging to sensibility. (B131-2).

The crucial question is what all of this is supposed to involve.

In the next section, we shall more fully explore the question of “intellectual” as opposed to “sensible” self-consciousness. For now, we need at least to be clear concerning the central problem in any purely judgmental approach to the “I think.” The difficulty in the suggested approach is simply this: that to the extent that the “I think” involves purely judgmental “self-ascription”—whether or not the subject in question is itself an experienceable item—the capacity for adding the “I think” to a manifold of given material could not possibly play the role that Kant has in mind for it. That is, it could not play the role of a kind of self-consciousness that is supposed to be an essential part of the constitution of predicative consciousness in the first place. It could not do so precisely because the “self-consciousness” in question would already be a mode of predicative consciousness.

We need to keep the distinction between “modes” of self-consciousness in mind, when we consider the apparently bi-conditional structure of Kant’s argument in the Deduction. As we have seen, Kant maintains that the act of conceptualizing any given intuition (whether sensory or merely imaginative) involves a kind of self-consciousness that in some way rests on more primitive imaginative structures in intuition, but also goes beyond them by means of the “addition” of distinctively intellectual functions. In other words, any truly conceptualized instance of the consciousness of objects entails an instance of self-consciousness. This is the first part of the apparently bi-conditional claim. In it, the “consciousness” of objects in question is, of course, essentially predicative: it involves the “application” of concepts to intuited objects. Kant’s claim is then that this presupposes a kind of self-consciousness that rests on intellectual functions. But it is important to be clear that this is not at all the same as saying that it presupposes a kind of self-consciousness that is predicative in nature.

The ambiguity is disguised by Kant’s use of the term apperception and by his use of the formula “I think.” As we have seen, the former term might be employed specifically in order to designate some kind of self-consciousness. But Kant also uses it (though only when he regards apperception as “empirical”) as nothing other than one’s recognition (or misrecognition) of objects:

[A]pperception [represents appearances] in the empirical consciousness of the identity of the reproduced representations with the appearances whereby they were given, that is, in recognition. (A115)

In these terms, we could put the first part of Kant’s bi-conditional claim by saying that “apperception” (of objects) presupposes “apperception” (of oneself).²⁶ This may encourage the supposition that the latter, like the former, is predicative in regard to its own “object,” namely, the self. A similar point goes for Kant’s use of the formula “I think.” That the formula must be able to accompany representations, so far as they are truly one’s own, may mean that one must always be able to conceptualize one’s own intuitions in some way. But it may also mean something else, namely, that one must always be able to be “conscious” of oneself as conceptualizing intuitions (or as able to). Both of these need to rest on intellectual functions. But that does not imply that the latter is predicative (in regard to oneself) in anything like the way in which the former is (in regard to objects of possible intuition).

Now for the completion of Kant’s bi-conditional claim: including the part that goes from self-consciousness to consciousness of objects. According to the latter part, self-consciousness is not only necessary for, but also sufficient for, the consciousness of objects. But in fact, Kant appears to go even farther than this. For he appears to have some kind of constitutive identity in mind, not simply a relation of condition to conditioned:

Thus transcendental unity of apperception forms out of all possible appearances, which can stand alongside one another in one experience, a connection of all these representations according to laws. . . . The original and necessary consciousness of the identity of the self is thus at the same time a consciousness of an equally necessary unity of the synthesis of all appearances according to concepts, that is, according to rules, which not only make them necessarily reproducible but also in so doing determine an object for their intuition, that is, the concept of something wherein they are necessarily interconnected. (A108 [emphasis added]; cf. A122)

The transcendental unity of apperception is that unity through which all the manifold given in an intuition is united in a concept of the object. (B139; emphasis added)

This consciousness of my existence in time is bound up in the way of identity] identisch verbunden] with the consciousness of a relation to something outside me. (Bxln)

In other words, the consciousness of my existence is at the same time an immediate consciousness [zugleich ein unmittelbares Bewusstsein] of the existence of other things outside me. (B276)

[W]e can only determine ourselves in time, insofar as we stand in relation to things outside of ourselves and therein consider ourselves [und uns darin betrachten]. . . .²⁷

There are at least two issues that might be raised by such claims. Most often, again, they seem to be taken as bearing on the possibility of self-ascription. This, in turn, might be taken in either a stronger or a weaker sense: either as concerning the possibility of self-ascription as such, or, more specifically, as concerning the possibility of knowing the truth of self-ascriptions. In either case, Kant’s formulations in the quoted passages should appear to us extravagant. One might take the possibility of self-ascription, or at least of self-knowledge, in some way to presuppose a consciousness, or even a genuine knowledge, of objects (and vice versa). But it is difficult to see how either could be as intimately one, as Kant suggests they are, with the very acts whereby objects are known in the first place.

The situation appears considerably different, once we suppose that Kant is talking, not about self-ascription and the self-knowledge that can rest on it, but rather about the specific kind of self-consciousness that needs to have been in question to begin with—namely, the kind that needs to be regarded as formative of any kind of “ascription” in the first place. Kant’s point would in that case be that this kind of self-consciousness is not merely presupposed by, nor does it merely presuppose, a consciousness of objects. It is rather what constitutes such consciousness.

This is not to say that Kant has no concern with the conditions of empirical self-ascription as such. The second-edition “Refutation of Idealism” is, for example, specifically concerned with it. But we need to see that there are two different ways in which one might try to show that self-ascription entails the consciousness of objects. One of them, apparently inspired by Wittgenstein, would differ from the one that I am attempting to elucidate. But it is perhaps more or less what Kant has in mind at some points, for example, in the “Refutation of Idealism.” It attempts to discover necessary conditions of the possibility of self-ascription independently of the fact that self-ascriptive consciousness is a derivative mode of self-consciousness in the first place. In Kant’s view, as I have proposed, self-ascriptive consciousness is derivative from a pre-predicative mode of self-consciousness. The latter in turn does not merely entail, but is one with, consciousness of objects.

III. Determinate and Indeterminate Self-Consciousness

I postpone to the next chapter further reflection on judgmental “form” as such. But before we attempt to do some justice to this notion of self-consciousness, as in some way “one” with the consciousness of objects, we need to consider the relationship between another pair of distinctions that are central to Kant’s reflections. These are the distinctions between (relatively) indeterminate and determinate self-consciousness, on the one hand, and between purely intellectual and genuinely empirical self-consciousness, on the other. It will be important to see that, despite the empirical status of the latter, even empirical self-consciousness is, in its original structure, in a way “indeterminate” in nature. Apart from the problem of interpreting Kant, some reflection on these issues should help us to appreciate the bearing of Kant’s theory on some issues of recent philosophical concern.

It is evident that there are various ways in which one might be “determinately” conscious of oneself as an individual distinguishable from others. But there also seem to be modes of self-consciousness that do not involve this determinacy. Under certain circumstances, it seems, one may be self-aware but unaware as to who (or even what) one is in particular; one’s mode of self-consciousness is strangely indeterminate. This has suggested to some that one’s true (or “absolute”) self is really supra-individual. Certain aspects of Kant’s thought have in fact been supposed to imply this, at least by some of his idealistic successors. Kant may in any case have been the first to appreciate the corresponding difficulties for any account of the consciousness of oneself as an individual.

Supposing that there is a distinct individual that one is, it remains a question what it could possibly be to be conscious of oneself as being that individual. Could there be, for example, some special way (say, α) in which one always perceives or describes oneself, such that correctly taking oneself to be a particular individual, as perceived or as described in any other way, simply amounts to correctly judging that the two modes of perception or description are modes of perception or description of a single individual? The problem with this seems obvious. It leaves us in the dark as to what could be involved in perceiving or describing oneself in the allegedly special way in the first place. After all, if to take the individual in question to be oneself to begin with were merely to take what is perceived or described in a certain way to be the individual that is also perceived or described as α then we either begin an infinite regress or come close to making self-consciousness the consciousness of the truth of a mere tautology.²⁸ Thus the ability to identify oneself as a particular individual seems to demand the synthesis of two almost antithetical modes of self-consciousness; only one of them seems directly to involve the determinate consciousness of an individual that is either perceptually or descriptively distinguishable from others.

Kant’s reflections on the problem of self-consciousness are not usually read as indicating any particular positive view regarding the solution to this problem: as I shall call it, the problem of “self-determination.”²⁹ Clearly, it is not his primary aim to provide a solution. At most, as some have argued, Kant is trying to show that a certain condition—namely, consciousness of objects—is necessary for any kind of self-determination.

The question remains why Kant supposed the condition to be necessary. On some interpretations, its necessity is seen to lie in a purely negative consideration concerning self-determination. It is seen to lie in the fact that one is not really “given” any object in the first place, “identifiable” as oneself. What one is given, in the relevant sense, is simply a sequence of representations that are ascribable to oneself. This in turn seems to imply that awareness of oneself, as a subject of representations, could originally consist in nothing other than some mode of consciousness of those representations themselves, or least in some mode of consciousness of some connection or synthesis among them.³⁰

This may suggest that Kant’s conclusions concerning the necessity for a consciousness of objects indeed embody a positive doctrine concerning the nature of self-determination. However, the condition as so far stated, and the reasoning presumed to lead to it, are compatible with a number of different approaches to that question. For example, they are compatible with the view that one is originally aware of oneself, as the subject to which certain representations are ascribable, simply by means of a definite description: e.g., by means of the description the subject that stands in relation R to these representations or the subject that stands in relation R to these objects (for some unspecified relations). They are also compatible with a denial of that view. For example, they are compatible with the view that consciousness of oneself as the subject of representations simply is the consciousness of certain relations among those representations themselves.³¹

Whatever Kant’s positive account of self-determination, it is important to see that he regards the latter as not only possible, but also as compatible with, and even as plainly requiring, that one be able to “identify” oneself as a particular spatiotemporal individual:

[T]he thinking being (as man) is itself likewise an object of the outer senses. (B415)

We are to ourselves object of outer sense from the start [vorher], because otherwise we would not perceive our place in the world and be able to intuit ourselves in relationship with other things. . . . I am myself an object of my outer intuition in space and without this could not know my place in the world.³²

On the other hand, as already suggested, it is likewise Kant’s view that, in some other sense, it is impossible to identify or “determine” oneself in any way at all.

One is at most conscious that there is an individual who one is; at worst, one is conscious of nothing more than a mere “synthesis of representations”:

I call them one and all my representations. . . . This amounts to saying that I am conscious to myself a priori of a necessary synthesis of representations. . . . (B135)

[I]n 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. (B157)

I exist as an intelligence which is conscious solely of its power of combination. . . . (B158)

The consciousness of myself in the representation ‘I’ is not an intuition, but a merely intellectual representation of the self-activity of a thinking subject. (B278)

The I of apperception, and therefore the ‘I’ in every act of thought . . . [is] the poorest of all representations. (B408)

The “I,” the universal correlate of apperception and itself merely a thought, designates as a mere prefix a thing of indeterminate signification, namely, the subject of all predicates without any condition to distinguish this representation of the subject from that of a something in general, namely, substance; by the expression “substance,” one has no concept as to what this substance is.³³

In any case, there is at least some sense in which it is possible to “determine” oneself as a particular spatiotemporal individual. What’s more, this possibility is a necessary upshot of whatever it is that requires a consciousness of objects in the first place, as a condition of “self-consciousness” for Kant.³⁴ So the original question remains: What can a self-determination of this sort possibly amount to? What can it possibly be, for Kant, to take some spatiotemporal individual as “oneself”? As powerful as it might be, the mere conclusion that a consciousness (or even full-blown knowledge) of objects is necessary for any kind of self-determination is neutral with respect to the question. At least, it is neutral so long as it amounts only to the demand that, since one is initially “given” nothing beyond representations, self-consciousness needs to rest on the awareness of relations among the latter, in particular on such as satisfy the conditions of a consciousness of objects.

It is perhaps arguable that Kant did not intend to shed more light than this on the problem of self-determination. Perhaps, that is, he was only concerned with a certain necessary condition of the latter. But it is difficult to believe that the question did not have a more substantial place in Kant’s reflections. Obviously, Kant thought there was something problematic, even paradoxical, in the very idea that we are determinable as particular spatiotemporal objects.³⁵ (And this is so quite apart from commitment, grounded elsewhere in his thinking—for example, in his ethical theory—to the postulation of a personal identity on a purely noumenal level.) On the one hand, self-consciousness can only be, for Kant, the consciousness of a “synthesis of representations”—not the consciousness of any sort of object at all. On the other hand, the very reflections that lead Kant to this conclusion also lead him to the conclusion that one must be self-determinable as a spatiotemporal object. It would be incredible if, in the light of this, Kant did not regard his reflections as in some way clarifying what it is to be aware of an object as “oneself’ in the first place.

Now I have so far been emphasizing a contrast between self-consciousness as the consciousness of some kind of “synthesis,” on the one hand, and self-consciousness as the consciousness of oneself as a particular spatiotemporal individual, on the other hand. But the point is to see that we really need a threefold distinction. We can see this by noticing, first of all, that at least insofar as the former is the consciousness of synthesizing activities, it must be regarded by Kant as purely intellectual in nature. Sometimes, Kant says as much: “original” or “transcendental” self-consciousness is a purely intellectual sort of consciousness (cf. B157-8, B277n, B278). Obviously, whatever sort of consciousness this is, it must be distinguished from a determinate consciousness of oneself as a spatiotemporal individual. But the crux of the issue is precisely that Kant also sees the need for a consciousness of oneself that is neither purely intellectual in nature nor the determinate consciousness of a spatiotemporal individual:

The I think’ expresses an indeterminate empirical intuition, i.e. perception (and this shows that sensation, which as such belongs to sensibility, lies at the basis of this existential proposition). But the ‘I think’ precedes the experience which is required to determine the object of perception. . . . An indeterminate perception here signifies only something real that is given, given indeed to thought in general, and so not as appearance, nor as thing in itself (noumenon), but as something which actually exists, and which in the proposition, ‘I think,’ is denoted as such. (B422n)

The proposition I think,’ or I exist thinking,’ is an empirical proposition. But empirical intuition, and consequently the object thought as appearance, underlies [liegt . . . zum Grunde] such a proposition. . . . The proposition ‘I think,’ insofar as it amounts to the assertion, ‘I’ exist thinking,’ is no mere logical function, but determines the subject (which is then at the same time object) in respect of existence. . . . (B428-9)

On the one hand, as we can now see, Kant is prepared to acknowledge a form of self-consciousness that is not purely intellectual. Thus it does not, for example, merely contain the “completely undetermined concept of a thinking being in general,” one where “all that I really have in thought is simply the unity of consciousness, on which, as the mere form of knowledge, all determination [of the subject in question] is based” (B426-7). To the contrary, the consciousness in question is a consciousness of oneself in particular, hence a “determinate” consciousness. Furthermore, insofar as it is specifically grounded in sense perception, it must be a consciousness of oneself as a phenomenally identifiable, if not already identified, individual: the “real” that is given is not merely that of a noumenon. On the other hand, and despite all of this, the self-consciousness in question remains in a way “indeterminate.” Presumably, this can only mean that it is not the consciousness of any particular object, identifiable as “oneself.” Thus even ordinary empirical self-consciousness, of “oneself’ as someone in particular, can be either a determinate or an indeterminate self-consciousness. Apparently, it can be a genuine empirical self-consciousness, of oneself in particular, but still not the consciousness of a determinate object identifiable as oneself.³⁶

It is reasonable to conclude that this notion of an indeterminate but empirical self-consciousness—and neither that of a purely intellectual consciousness of some kind of “synthesis,” nor that of empirical self-ascription—must be the focus of our attempt to clarify Kant’s view. But at the same time, we must develop a sense in which self-consciousness, on its most basic level, is indeed nothing other than the consciousness of a “synthesis of representations”:

[T]his unity of consciousness would be impossible if the mind in knowledge of the manifold could not become conscious of the identity of function whereby it synthetically combines it in one cognition. . . . For the mind could never think this identity in the manifoldness of its representations, and indeed think this identity a priori, if it did not have before its eyes the identity of its act, whereby it subordinates all synthesis of apprehension (which is empirical) to a transcendental unity, thereby rendering possible their interconnections according to a priori rules. (A108; cf. B132-4)

We have, of course, already seen why self-consciousness needs to rest on, or even to be one with, the consciousness of a synthesis of representations. The reason is that nothing else is available, as a unitary object of immediate awareness, in all cases in which one is in fact self-conscious. In particular, “No fixed and abiding self can present itself in this flux of inner appearances” (A 107; cf. B133). If anything unitary, and yet essentially self-like, is to be found in the flow of representations, it would thus apparently need to be something definable, not in terms of detectable components of that flow, but only in terms of unitary activities, necessarily directed or directable at such components. But how do we get from here to a correlative apprehension of objects?

The inference would seem to rest on Kant’s theory concerning the nature of object-concepts. According to this theory, any truly conceptualized apprehension of objects is itself merely the apprehension of syntheses of representations:

But it is clear that, since we have to deal only with the manifold of our representations, and since that x (the object) which corresponds to them is nothing to us—being, as it is, something that has to be distinct from all our representations—the unity which the object makes necessary can be nothing else than the formal unity of consciousness in the synthesis of the manifold of representations. It is only when we have thus effected [bewirkt] synthetic unity in the manifold of intuition that we are in a position to say that we cognize the object. (A 105)

This concept [of objective reference] cannot contain any determinate intuition, and therefore refers only to that unity which must be met with in any manifold of cognition which stands in relation to an object. This relation is nothing but the necessary unity of consciousness, and therefore also of the synthesis of the manifold, through a common function of the mind, which combines it in one representation. (A 109; cf. B137)

But if this is Kant’s approach, then he appears to be involved in a fallacy. For even if (“original”) self-consciousness is no more than the consciousness of a synthesis of representations, and the consciousness of objects is no more than the consciousness of such synthesis as well, this does nothing to show that the unity in question in one of the two cases is in any way correlative with the kind of unity that is in question in the other. In the light of this difficulty, it might seem reasonable to return to a suggestion that I have already rejected. It might seem reasonable to return to the suggestion that, whenever Kant appears to equate self-consciousness with the consciousness of syntheses of representations, he only means to be saying, first, that self-consciousness rests on the ability to ascribe a manifold of states to a single enduring self and, second, that a necessary condition of such self-ascription is the ascription of predicates to objects, and vice versa. (And we might then regard such consciousness as purely “intellectual,” simply in the sense that it rests, not on the consciousness of an intuitable object identifiable as oneself, but simply on that of the rules or criteria for the type of ascription in question.)³⁷

IV. Self-Consciousness and the Imaginative Content of Perception

I want to elaborate an alternative reading. The alternative involves a very different approach to the “manifold” of representations and its “synthesis.” I have partly prepared the way for it, by distinguishing between manifolds of representational states, anticipable and retainable as co-ascribable to single subjects, and manifolds of representational states anticipable and retainable by other states as co-ascribable to single subjects. But we now need to introduce an additional distinction. It is the distinction between the anticipation and retention of states by other states, appropriately related to the given ones, and the anticipation and retention of states anticipated and retained as states that the given ones might themselves actually become (or that might themselves actually have become the given ones).

This puts the point, as it were, in terms of the possible “stretching” of states of consciousness into the future (or from out of the past). But it will be crucial to see that the point might be put in terms of the possible stretching of the (“immediate”) objects of consciousness themselves, that is, of “appearances.” For this, what we require is simply the supposition that the immediate objects of intuitional consciousness are just as literally stretches of time (whether actual or possible stretches) as they are actual or possible regions of intuited space. To the extent that the objects are stretches of time and not only regions of space—and recalling the role of anticipation and retention as necessary, not simply in conception, but in the intuitive grasp of appearances themselves—the suggestion indeed allows us to consider, as genuine objects of consciousness, not simply manifolds of appearances that stand, or that might be conceptualized as standing, in possible relations with other ones, but the very stretches of possible time in which those appearances might be anticipated as stretching into other ones (or in which others might have been anticipated as stretching into them). What I shall argue, then, is that nothing but the essential correlation between these points of view is what assures us that the apprehension of such “syntheses of representations” always constitutes, on a pre-predicative but fully empirical level, the apprehension of possible futures or pasts as “one’s own.” Thereby, it both constitutes the most basic mode of self-consciousness, independently of the determination of oneself as an object, and at the same time provides the necessary basis for that determination.

That the notion in question has a useful bearing on the problem that concerns us I shall try to make more plausible with the following examples. I begin with a case that comes as close as possible to providing an instance of “selfless” absorption in some object or observed scene. If certain philosophers are correct, then there is at least one sense in which no state of consciousness is completely free of some kind of self-relation. According to Sartre, for example, all mental acts include an apprehension of themselves; this in turn is constitutive of a special kind of apprehension of “oneself.”³⁸ So far considered, however, the latter still falls short of a conception of oneself as any kind of object, nor does it involve any kind of predicative ascription in regard to oneself. I want to begin with a case that is at least as (apparently) selfless as that.

It is useful to consider a certain sort of “scene,” for example, a scene of a kind that might be a part of a movie. I choose this sort of example because, in many cases, one manages to get completely absorbed in such scenes, hence to be minimally or even, apparently, not at all self-aware. We may then imagine the effecting of a series of transformations. These will be of a kind that are often employed in such scenes. What I have in mind is any such series as will eventually issue in a case in which, independently of theoretical pre-dispositions, we can all agree that the viewer has become (at least imaginatively) conscious of being actually involved in the scene in question. In other words, whatever a particular theory—for example, Sartre’s, or even Kant’s own—might dictate regarding the viewer’s antecedent condition, we would appear to have a case in which the viewer is at least subsequently “self-conscious.” The relevant consideration will then be this: that, while the transformation that eventuated in the latter condition did indeed eventuate in a state of self-consciousness, it was effected precisely without any loss of one’s complete and (relevantly) selfless absorption in the objects of consciousness.

Take the case of an automobile that is driven, in a scene in some movie, along a road that is at least in part visible to the viewer. Imagine that the camera is focused, for the duration of the scene, through the windshield of the car and onto the stretches of road. One may observe a scene of this sort in an apparently selfless way. One often does so while in fact driving. Now we are also familiar with scenes of a different sort. In these, the camera remains focused on the stretches of upcoming road, but various techniques produce a different effect in the viewing of them. The point is, of course, that they produce the effect of stretches of road that the viewer is actually traveling. That is, they produce an effect whereby upcoming objects, or upcoming stretches of road, are apprehended as successively closer to oneself, and not merely to the character who happens to be portrayed on the screen. (One does not need to imagine, though one may do so, that one is in fact that character.) In viewing a scene of this sort, one has become, in a way, particularly aware of oneself But the point is this: One has not, for all this difference, become any less “selflessly” absorbed in the objects that are around or ahead of oneself: in this case, a movie screen.

Now it admittedly does not strictly follow from the fact that we can obtain this kind of effect, as the upshot of purely cinematographic adjustments, that in the final case one is as purely “object-directed” as before. One might, after all, hypothesize that, at some point in the process, a double state of consciousness had developed: one part of it is still the original, visual consciousness of the upcoming stretches of road as traveled by someone other than oneself (say, by James Bond); the other, somehow together with this, an at least imaginative state of consciousness of oneself as traveling those stretches. In a way, I am not suggesting that there would be anything wrong in putting the case in these terms. Whether we do so or not, the question remains: What is the point of appealing to a double mode of consciousness? A similar query must meet the suggestion, apparently equally truistic, that in such cases one is indeed merely “imagining that” one is traveling a stretch of road. It is clear that this can be said. The problem is that the “imagining,” in such cases, seems indistinguishable from the viewing. (In other sorts of cases, one may be additionally led to the performance of distinct acts of reflection: explicitly taking note of certain feelings in one’s stomach, for example, or wondering why one has come to the movie.)

The following consideration seems to me to constitute a reason for supposing that the apparently two components of such an experience are really inseparable aspects of a single experience. On the one hand, we should note that it is perfectly possible to imagine that one is traveling along the road in question (or along a road just like it), at just that speed, and under just the given circumstances, while viewing the scene in question, but without that scene at all appearing to one as it does in the case. (As to how, specifically, I am in fact supposing it to “appear” in the given case, my own suggestion would obviously be that this involves perceptible aspects—at least of objects as “appearances”—that, unlike one’s speed and the objective road conditions, simply cannot be conceptualized in equally “objective” terms. But the point is that, despite this, we would still be directed toward no more than the appearances themselves.) On the other hand, it seems impossible for the scene in question to be appearing to one as it in fact does, yet without one thereby imagining that one is traveling along that road at just that speed, and so on. This seems to me a reason to suppose that whatever kind of imagining is in question, it does not occur as a series of distinct imaginative acts, merely externally (perhaps causally) connected with a series of acts of a purely perceptual nature.

My suggestion is that the case involves imaginative content, not in the form of acts of imagining alongside purely perceptual ones, but as material through which perceptual acts are constituted in the first place. Of course, there is no reason to suppose such material to be limited to what we might regard as merely passive “anticipation and retention.” In the case at hand, it is reasonable to suppose one’s apprehension to be effected, not simply through a manifold of anticipations and retentions, but through dispositions and tendencies of various sorts. Prominent among them might even be idiosyncratic tendencies, of a kind that becomes manifest when one actually drives a car. In any event, the suggestion is that the imaginative content that Kant takes to be essentially ingredient in any instance of objective apprehension is also what constitutes that special form of empirical self-consciousness—“indeterminate” but in principle determinable as the identification of a particular individual—that on his view is equally essentially embodied in such apprehension. The point of the example was simply to show that such a mode of self-consciousness can indeed be constituted by means of the mere ingredience of imaginative material in an experience, hence without any concept of oneself as subject of experience. (The content in question might even include, in this way, the anticipation and retention of a variety of upshots of “one’s own” possible actions, without presupposing any actual concept of oneself as an agent.)³⁹

Sometimes, Kant himself appears to go so far as to connect the mere “sensations” that are ingredient in perceptual states with a certain kind of self-consciousness as well, namely, with a consciousness of oneself as a being possessed of sense organs. Insofar as it functions as material in intuition, sensation is merely material “through which” an appearance is apprehended. Thereby, it is also reflected in any appearance as such. But even when it is ingredient in ordinary acts of perception, sensation, Kant holds, makes us aware of our own sense organs as well as of something in appearances.⁴⁰ It is hard to know how seriously to take this. Kant himself is not clear as to what is really involved in this allegedly dual character of mere sensations. In particular, it is not clear whether he does not shift between two different notions of “sensation.” With one of them—the one that seems prominent in the Aesthetic—it would not be sensations as such that make us aware of our sense organs. It would rather be certain dispositions (or at least anticipations and retentions—which is the closest Kant himself ever comes to “dispositions”) that are typically “associated” with sensations in that sense. On the other hand, those same dispositions might also be regarded as constitutive of “sensations,” in a different sense of the term. Such a view is at least independently defensible. It is worth digressing to reflect on it.

Take the case of physical pains. No sensory states seem more apt than these to make a subject self-conscious. It might even seem to be “constitutive” of pains that they do so. Yet it has been argued that the supposedly essentially painful character of pains is not at all a function of their purely sensational aspect.⁴¹ It is rather a function of dispositions of various sorts. This might be defended without any explicit appeal to my own analogy between the apprehension of perceptible qualities through sensations and their apprehension through whatever dispositions are in question. It is nevertheless useful for our purposes to extend the analogy as far as possible. So we may suppose, for a start, that the apprehension of a pain in a limb that actually feels painful does indeed involve apprehending that limb through particular pain-sensations of one’s own; as a correlate of this, the limb (at least qua appearance) comes to exhibit a corresponding and phenomenologically unique quality.⁴² Conceding this, the question will then be whether the bare apprehension of this quality—despite the fact that it is the objective correlate of one’s own pain-sensations—needs to amount to the apprehension of what we would normally call painful. As suggested, the contrary is arguable.

Now in some sense, it seems, nothing can make one more self-aware, and necessarily so, than the awareness of something that is painful in a part of one’s body. In general, however, it is not the case that, in becoming conscious of a part of one’s body, one is thereby rendered self-conscious. In many cases, one can easily forget that some limb is in fact one’s own, and selflessly contemplate the way that it happens to appear. Now why should we suppose that the case of pain is different? Why should a perceptible quality, just because it is perceptible only through pain-sensations of one’s own, also be inevitably experienced as painful to oneself? Empirically, it does not seem to be the case that it is.⁴³

The resolution of the apparent paradox may lie in the fact that there are really two qualities in question, each of them as immediately available to intuition as the other, and both of which one might be inclined to regard as the quality of phenomenal painfulness. One of them might be regarded as the correlate of mere pain-“sensations.” But if it is, then there can be no a priori reason for supposing that the apprehension of that quality is ever in fact painful. That is, there can be no a priori reason for supposing that the apprehension of that quality is painful to oneself (or, for that matter, to anyone else). For the latter, what would need to be added? It is implausible to suppose that what needs to be added is just certain sorts of judgmental ascriptions in regard to oneself. That would be incompatible with the fact that the immediate experience of oneself as pained is as immediate and unreflective as the apprehension of any sensory quality in intuition.

It is this immediate and pre-predicative aspect of the painfulness of pains that suggests that their apprehension involves no more than the ingredience of a particular type of sensation in intuition. Were we prepared to broaden the notion of sensation to the purpose, we might even concede this point. But our question concerns what such a broadening would need to involve. Given the immediate and intuitive nature of the case, only one alternative seems open. One’s “self’ can have entered the picture in the first place, in the mere apprehension of a part of one’s body, only to the extent that the quality of painfulness is apprehended in that part, not through mere sensations of pain at all, but through specific dispositions and tendencies as well—or, if through “sensations,” then only to the extent that this notion has already been broadened to the purpose. What we usually regard as painfulness, as an immediately perceptible quality, would then be the correlate of this type of apprehension. And it is only as such a correlate that it could possibly be truistic to say that the apprehension of immediately perceived pain is the apprehension of something as painful to oneself.

Whether he should have done so or not, it seems clear that Kant himself did not regard sensations, insofar as they are supposed to convey a consciousness of one’s own sense organs, as involving anything more than “sensations” in the sense that is more generally in question in the Aesthetic—where Kant merely considers them as corresponding to some “matter” in appearances. In particular, he does not appear to regard them as involving anything like the imaginative “dispositions” that need eventually to serve as material for conception. Nor do I have any reason for supposing that Kant considered extending to pains his more general approach to appearances and their qualities in the first place: regarding them, not as modes of consciousness, but as the intentional correlates of modes of consciousness. In any case, I chose the example of pain because it appears to involve, and essentially so, a consciousness of oneself as a specifically embodied being, but it also appears to involve a self-consciousness that is immediate and pre-predicative.

The supposition that, to the extent that pain does indeed involve an instance of self-consciousness of this kind, it must involve the apprehension of qualities “through” something more than sensations—or if only the latter, then only in a suitably broadened sense—is perfectly compatible with the immediate and pre-predicative aspect of such self-consciousness. For my purposes, then, the upshot is this: that just because, when so construed, the apprehension of pains is essentially self-conscious, we have an additional reason for concluding that the mere apprehension of objects through the right sort of material, and apart from the presence of concepts, is of itself constitutive of a basic and irreducible mode of self-consciousness. It is constitutive, that is, of a consciousness of oneself precisely in one’s consciousness of certain perceptible dimensions of appearances (in this case, of parts of one’s body as appearance). In the case of pains, we have a name for the quality in question, though we have seen it to be ambiguous. In other sorts of cases—for example, in the case of the “subjective” qualities of upcoming stretches of road, or in the case of Sartre’s “streetcar-having-to-be-overtaken” ⁴⁴—language may come closer to failing us.⁴⁵

It will be easier to appreciate this point, and in general to appreciate the significance of insisting, in the present context, on the notion of an object of awareness as an intentional correlate of the awareness of it, if we bear in mind my earlier suggestion concerning the general relation between our awareness of appearances and our awareness of space and time. Kant’s notion (A22/B37) that time is the form of inner sense, while space is the form of outer sense, may generate the impression that we apprehend perceptible “material” only as filling regions of intuitively apprehended space, or portions thereof, and never as filling or even partially filling intuitively apprehended stretches of time. This is because we may think that what one immediately apprehends in intuitively apprehended time, as such, is never outer appearances—that is, never what we are free to take, or on occasion to refrain from taking, as actually existent matter in space—but only states or ingredients of states of the apprehending subject. But Kant maintains that it is outer appearances that constitute the material of inner sense (B67; cf. B154-6, B163, B220, B292; I have argued elsewhere against taking this claim as expressing the subjectivistic doctrine that outer appearances are made out of our own subjective states). We can recognize this point, while at the same time accommodating Kant’s own distinction between the two “senses.” The point will simply be that while outer appearances are perceptually apprehended as filling or as partially filling time as well as space, the temporal relations among them, construed as actual existences in space, are never themselves an object of perceptual apprehension; they are at most an object of judgment. (Strictly, what immediately “appears” in an intuitively apprehended stretch of time may of course then amount only to a temporal portion of the real material object judged to be in it.)

This finally frees us to recognize that, in the apprehension of outer appearances, one is indeed always apprehensive of a manifold of stretches of one’s own possible future and past. For that future and past will always be presented (or at least “appresented”) as a kind of intentional object in its own right. (More exactly, it will be presented as a possible future or past portion of the total intentional “object” apprehended.)

If one is likely to find a specific role for “anticipation and retention,” in Kant’s doctrine of synthesis, what one of course generally assumes to be anticipated or retained, in connection with a given experience, is just possible future or past experiences, or the correlative appearances, appropriately “connectible” with the given one. In turn, this might be thought to require the ability to conceive of oneself as a being capable of stretching into the future or past, in such a way as to serve as the possible subject of those experiences, or as the possible perceiver of those appearances. It might be thought to require this self-conception, because there is simply no purely intuitive awareness of stretchable selves, over and above the awareness of whatever experiences (or appearances) are in principle ascribable to (or perceivably by) them. But our new framework provides for a very different picture. For it accounts for a sense in which one is intuitively aware of a stretchable self precisely by providing a sense in which one is intuitively aware, not simply of given experiences, and of possible future and past experiences co-ascribable with them—but of the very (albeit possible) stretching of this or that experience (or, correlatively, of this or that “appearance”) itself into the future and past.

Surely, to be immediately aware of this or that experience as stretching, or as having stretched, is a fundamental form of the awareness of oneself as stretching, or as having stretched. But it is not a form that needs to rest on any conception of a self to which experiences are ascribable. It needs to rest on no more than the awareness of experiences themselves. This intuitive character of such a self-experience is not, of course, incompatible with one’s ability to conceptualize its object in various ways, for example, as a possible stretch of one’s own experience. The point is simply that its intuitive character does not presuppose one’s ability thus to conceptualize its object.

It should be clear what role is played in this account by the broadening of what may have been our initial grasp of Kant’s theory of temporal intuition. It is crucial to be clear, as well, what role is played by our theory regarding the apprehension of intuited objects through imaginative anticipations and retentions. That one does not merely experience anticipations and retentions that are in some way externally connected with the apprehension of given appearances, but that given appearances are themselves apprehended through anticipations and retentions, is what allows us to say that, in the apprehension of those very appearances, one is in a special sense self-aware. That the appearances in question are apprehended through imaginative anticipations and retentions necessarily gives them a perceptible dimension that they would otherwise fail to exhibit. In particular, it gives them a perceptible dimension that is a correlate of one’s own future and past (at least qua “anticipated” and “retained”). Ordinary language may lack the terms for a clear expression of this dimension of appearances (apart from those terms that already involve an explicit reference to “oneself’ as anticipating and retaining in their regard). We have already seen that Kant had dealt with at least a part of this problem in his treatment of “transcendental affinity.” What we are now encountering is another aspect of the same problem.

Thus Kant’s point, I suggest, is simply that, however we conceptually express the relevant dimension, the spatiotemporal dimension that is correlative with the ingredience of the appropriate sort of anticipation and retention in intuition—as opposed to their external connection with intuitions—is constitutive of an original and transcendental self-awareness: one is aware of “one’s own” experiences, as possibly stretching into the future and past, precisely in the sense that one is aware, not simply of given appearances, but of given appearances through a manifold of one’s own anticipations and retentions. This mode of self-awareness is “transcendental,” not simply because it is a necessary condition of the possiblity of knowledge, but because it is a form of self-awareness in which no self appears as an object. It is “original,” because it is the necessary foundation for eventual determination of oneself as an object.

Kant may in fact have no “argument” that is meant to lead to this conclusion in a rigorous way, over and above his observation that nothing else is available, as material in intuition, for the original constitution of a self-concept.⁴⁶ In any event, once one reflects on the significance of the fact that, already below the level of conceptualization as such, appearances are not only apprehended through mere “sensations,” but through manifolds of one’s own anticipations and retentions, the claim in question may not appear to be far from the truth. To recall our previous examples, it at least seems to provide a plausible account of how, in merely traveling a road in the absence of self-reflection, or in merely experiencing a pain in some part of a body, one may ipso facto be aware of oneself as traveling that road, or as in pain. To account for this apparent identity of the consciousness of certain objects and the consciousness of oneself, we do not need to introduce any conceptual factors or criteria that are not already involved in the apprehension of the objects themselves: the road and some bodily part. By virtue of its own internal structure, the apprehension of those objects is at the same time self-apprehension.

Now apart from a genuine self-concept, one will, of course, be unable to judge of any intuitively apprehended future, in a way that is truly judgmental in Kant’s sense, that it is indeed a stretch of “one’s own” possible future. But we have only been attempting to account, so far, for a mode of self-apprehension that is the necessary basis for any self-determination—as of any object-determination— through concepts. Beyond this point, it is likely that Kant had few details in mind. But the basic view that is to be gathered from the materials now at hand seems to me to amount to a plausible view concerning the relation between original self-consciousness as Kant construes it—that is, the consciousness of oneself, not as an “object” of some description, but merely as a kind of “synthesis of representations”—and a self-ascriptively determinate consciousness of one’s identity.

Kant’s point may then simply be the following: In general, one’s imaginative anticipations and retentions regarding appearances stamp the world of appearances with a unique and immediately experienced “orientation”; only relative to and derivatively from the intuitive apprehension of the latter is it then possible to form any genuine concepts of the various “points of view,” or of the various possible alternative “centers” of orientation, from which a particular consciousness might be presumed to be active in the first place. The formation of a determinate self-concept may in turn be held to rest on a concept of just this sort, therefore to rest on precisely the kind of anticipational/retentional structure that is essential to the conceptualization of objects.⁴⁷

Obviously, one may form a purely conceptual notion of various centers of orientation, in terms of purely general features determinative of those centers. The point of view from here or there may be defined in terms of such features. But our question concerns the possibility of regarding oneself as occupying such a center. Clearly, I do not simply mean by myself “whatever being is defined by such-and-such point of view.” For, however that point of view is described, both temporally and spatially, I may always find it significant that I in particular happen to occupy it.⁴⁸ What more is needed? The suggestion is this: that one must not merely be able to locate the “center” in question, but that one constantly live it on a pre-conceptual level. The only available, non-circular, account of such “living,” in turn, seems to me to be precisely what we have already discovered, in the notion of the sensible/imaginative material through which objects are apprehended. To apprehend the world from any point of view in the first place can only be to apprehend it through an appropriate body of material. As we have already seen, the world, qua appearance, necessarily reflects the material through which it is originally apprehended. That was the point of the account earlier proposed of transcendental affinity. It is now the same point that finally accounts for that special mode of apprehension that is involved in apprehending some particular point of view on that world as “one’s own. ”⁴⁹ (In considering the plausibility of the suggestion, we need to bear in mind that the full range of imaginative material, ingredient as material in intuition, will stamp the world with an “orientation” that is much more than a matter of visual perspective. The orientation will also reflect, for example, one s anticipations concerning the possibility of altering any given perspective.)

However the notion is elaborated, it is not, after all, implausible to suggest that one’s fundamental self-concept is simply that of whatever subject is in the appropriate way defined by some spatiotemporal “point of view” on the world. The point is simply to see that such concepts amount to “self”-concepts only because they originally derive from a more primitive and original mode of self-consciousness involved in the constitution of such points of view in the first place. The latter does not determine a subject through spatiotemporal concepts (though it does, of course, “determine” it in some other sense). Nonetheless, original self-consciousness guarantees that the subject is so determinable. At least, it guarantees that some subject is so determinable, namely, the subject precisely qua correlate of the “unity of objective experience.”

I have added the final qualification, because all of this can only concern the so-called “empirical” subject. But Kant himself seems to take it to follow, from the fact that the conditions for the determination of any empirical subject are derivative from the original constitution of conceptual consciousness itself, that the empirical subject is not as such (i.e., qua empirical) the ultimate subject at all:

The subject of the categories cannot by thinking the categories acquire a concept of itself as an object of the categories. For in order to think them, its pure self-consciousness, which is what was to be explained, must itself be presupposed. Similarly, the subject, in which the representation of time has its original ground, cannot thereby determine its own existence in time. (B422; cf. A401-2)

The empirical subject is a genuine “subject of consciousness.” But it is, as we have seen, a subject of consciousness only so far as it is constituted as an empirical subject by the flow of consciousness itself, in the course of its own constitution of the experience of objects. Are we therefore required to conclude that the originally constituting—yet obviously empirical—activities of the consciousness in question are themselves ascribable to the empirical subject only to the extent that they are more truly ascribable, in some other sense, to some “other” subject as well (that is, to a noumenal one)? Is there, that is, one consciousness but two subjects of it?

Dies macht nun ein Doppeltes ich aus, aber nicht des Bewustseyns (ich erscheine mir selbst, ich bin mir auch in diesem empirischen Bewustseyn doch der Beobachtete und zugleich Beobachter, der . . . [Kant’s note breaks off at this point].⁵⁰

This is a question that I propose to leave outside our present reflections.⁵¹

Despite all the foregoing, it should be clear that I have not attempted to construct a valid argument that in fact succeeds in linking self-consciousness with the consciousness of objects. I have at most presented the rudiments of a theory of the former that seems to me to link it with the latter in a way that best fits the overall “argument” of the Aesthetic and the Deduction. In so doing, my primary aim has been to improve on accounts that merely concern the conditions for the ascription of representations to oneself as a subject. The problem with those accounts is that they involve no serious reflection on what it is to be aware of oneself as a subject in the first place. As we have seen, it is Kant himself who seems to want to tell us something as to what self-consciousness really is, not merely as to certain of its necessary conditions. And it is in this context that he tells us that self-consciousness, in its transcendental structure, can be nothing more than the consciousness of a “synthesis of representations.” At the same time, he also tells us, it must be a consciousness that is in some way one with a consciousness of objects. Commentators cannot, of course, fail to be sensitive to the first of these points. I have tried to be sensitive to both of them.

But there is a more serious shortcoming in the analysis as I have proposed it. Even if we accept the proposed theory of self-consciousness, it would at most be the case that object-consciousness entails transcendental self-consciousness (thus at least the possibility of a more determinate mode of self-consciousness). It would not be the case that the latter entails the former. For if the central idea is that transcendental self-consciousness is constituted through the ingredience of manifolds of anticipations and retentions in intuition, then it is not clear why it should require full-blown conceptualization of appearances. Presumably, there is nothing in principle preventing the ingredience of anticipations and retentions in intuition, independently of their serving as material for conceptions as well. In the absence of forms of the appropriately intellectual sort, the mere form of intuition might suffice for the ingredience in question. (I have already, at least, rejected the apparent difficulty that, as merely ingredient in intuition, the material in question could not constitute a kind of self-consciousness, simply because it could not amount to any “consciousness” whatsoever.)

It may be that Kant overlooked the possibility of the actual ingredience of anticipation and retention in intuition, apart from their serving as material for conception. As we have seen, he appears to have been sufficiently perplexed by their nature in the first place as to have been tempted, at points, even to regard their purely “imaginative” syntheses as, at bottom, intellectual. (But then, despite the mystery of it, Kant did know better.)⁵² Quite apart from the obscurity of their intrinsic nature, Kant may also have been struck by the following fact: that there is only one function to be served by the ingredience of anticipation and retention in intuitions in the first place, namely, the making possible of the conceptualization of the latter.⁵³ Apart from that, anticipations and retentions might as well be merely externally connected with intuitions. Perhaps these factors led Kant to be insufficiently critical concerning the possibility of self-consciousness, apart from the conceptualization of appearances. But even if they did, they would of course still not explain Kant’s even stronger conclusion, namely, that the constitution of orginal self-consciousness presupposes, not simply the conceptualization of appearances, but their conceptualization precisely as involving (what we would normally regard as) concrete objects.

It is in the “Refutation of Idealism” that Kant himself explicitly connects self-consciousness with the consciousness of concrete external existences. But for two reasons that text fails to help us. First, and without explanation, the so-called proof of the thesis supplements the perfectly reasonable claim—on which I shall have more to say in the next chapter—that self-consciousness presupposes consciousness of some kind of “permanence,” with the hardly evident claim that self-consciousness presupposes the consciousness of something concretely existing as a permanent in intuition (B275-6). (Since no such permanent is available in “inner” intuition, Kant then simply concludes that it must involve matter in space.) Second, Kant in any case presents the issue as one that concerns the possibility of “empirically determined” self-consciousness from the start, whereas our question, in the Deduction, is one that concerns the possibility of a self-consciousness that is merely determinable. Indeed, Kant presents it as a question that does not concern mere self-“consciousness” at all, but rather full-blown cognition of oneself (B277).

It is not easy to be sure just what Kant has in mind in the “Refutation.” One might easily presume that he is specifically concerned with the possibility of a conceptually determinate mode of self-identification. This, at least, would draw a distinction between the “Refutation” and the Deduction. But on the other hand, it is also arguable that Kant is concerned with the same question in both places. It may simply be that, in the later passage, he is implicitly admitting that the original Deduction (in both of its editions) has in fact failed to meet the objections in question. That at least would explain why Kant feels that he can conclude, in the “Refutation,” that “empirically determinate” self-consciousness does not simply presuppose some knowledge concerning objects existing in space, but presupposes the “immediate consciousness” of them (B276, B277n). Still, it is difficult to see why the empirical determinable consciousness of oneself— not merely fullblown self-knowledge—needs to presuppose such consciousness.

I have only the following suggestion. It is a suggestion that relates to that element in any consciousness of concrete objects on which Kant himself places special emphasis in the first-edition Deduction, namely, the consciousness of some kind of “necessity” in the apprehension of appearances (A104ff). (Indeed, necessity seems to be the only one of the categories to which Kant gives particular attention in the first-edition Deduction. I shall have more to say about it in the next chapter.) On some level of clarity or other, Kant may have reasoned in the following way:

(1) Original self-consciousness is constituted by the ingredience, in intuition, of anticipations and retentions concerning the course of possible experience. (Argument: There just isn’t anything else available in intuition for the purpose.)

(2) But original self-consciousness, if it is to be determinable as the consciousness of some identifiable individual, cannot merely involve the consciousness of possible experiences. (Argument: In a sense, anything is legitimately imaginable as possible. So if original self-consciousness merely involved the anticipation or retention of certain courses of experience as possible, it would fail to bear on the self-consciousness of any particular individual. So far as would be compatible with such consciousness, the “I” of which one was thereby conscious could be anyone: one could be imagining somebody else’s, or even nobody’s, possible future or past.)

(3) Therefore, original self-consciousness must be constituted by the anticipation and retention of possible experiences, not simply as possible extensions of a given experience (or as experiences possibly having stretched to become a given experience), but as extensions of a given experience that are necessary, at least relative to certain antecedent conditions (or as experiences whose stretching to become a given experience was necessary, relative to certain antecedent conditions).

(4) But the representation of experience in such terms as these simply is its representation in terms of objective concepts. (Argument: A104ff, concerning the concept of an “object.”)

(5) Therefore, original self-consciousness is impossible apart from the objective conceptualization of appearances.

The argument is, no doubt, far from conclusive. But on the basis of the materials that are to be found in the text itself, it seems to me likely to reflect Kant’s own thinking. In any case, it seems to me better directed than attempts to formulate conditions for the possibility of self-consciousness simply in terms of the possibility of the self-ascription of predicates. Before we can see what is needed for the latter, purely judgmental, mode of self-consciousness, it is surely reasonable to suggest that we first need to have reflected on the latter’s more basic ground in an original self-consciousness.