I. Introduction

THIS CHAPTER contains a preliminary defense of my reading of the theory of consciousness and concepts developed in the Deduction. It is both broader and narrower in focus than some of the others. In the first respect, I do not limit my attention to Kant’s claims and arguments as they are presented in the Deduction. I give equal attention to some sections of the Logic that are preliminary to the Deduction, devote a section to the Schematism chapter, and make relatively free use of Kant’s unpublished notes and notes derived from his lectures on logic. As to topics that may seem to be slighted, there is first of all the problem of the relationship between the first and second versions of the Deduction. In the present chapter, I make little use of passages specific to the second-edition Deduction. The second area of neglect concerns the concept of consciousness.

Although I am centrally concerned with Kant’s theory of consciousness, I often formulate questions in this and the following two chapters in terms that do not explicitly invoke the concept of consciousness, for example, in terms that involve such notions as those of synthesis and imagination, or of understanding or (synonymously) intellect or conception. One might suppose that there are in fact both conscious and unconscious forms of all these things. In one respect, there is no need to question this. The supposition might be accommodated very easily. It might be accommodated simply by regarding all of these things “dispositionally,” or as “faculties.” I have no objection to this, so long as it is clear that it is not the same as saying that we need to distinguish conscious and unconscious forms of some single kind of activity or process.

Whatever we say about understanding as a mere disposition or faculty, the activities or processes that are its paradigmatic manifestations are essentially and uniquely conscious for Kant. They are so in a twofold sense. First, they embody a consciousness of objects distinguishable from subjects of experience. Second, they embody a form of consciousness of themselves and of their own subjects as well. In at least these respects, the workings of the imagination are not essentially conscious for Kant. Kant suggests that at least some of them are essentially unconscious (A78/B104, A141/B180-1). One may of course wonder whether he is clear as to what is thereby supposed to be essentially unconscious. For example, just how clear is he regarding the relationship between faculties and their workings in the first place? In any event, we shall be concerned with something called “imagination” only to the extent that it provides an essential ingredient in what is paradigmatically conscious in Kantian terms. As I develop the relevant notion of “ingredience,” the question is largely semantical whether, viewed in themselves, the workings of the imagination are essentially unconscious or not (and indeed even whether or not they may be regarded as somehow “judgmental”). What remains the case is that denying their status as full-blown consciousness does not entail that they are then essentially hidden with respect to the ordinary course of conscious experience. While all of these issues are raised by Kant’s arguments, and while more might be said to anticipate them, considerable preliminaries are in order before we can formulate our questions with regard to the relevant notion of consciousness. In particular, we shall need to await Chapter Six for a discussion of the problem of self-consciousness.¹

It is obvious that most of the difficulty posed by any reading of the Deduction stems from Kant’s attempts to get clear about the connection between imagination and intellect. At the least, its first-edition version appears to suffer from what Jonathan Bennett has called “neurotically inept exposition.”² Whether this is due, as Bennett suggests, to “sheer terminological confusion,” to outright self-contradiction, or to some other cause is uncertain. Apart from specific problems, which I postpone to the next chapter, regarding Kant’s attempts to distinguish between “reproductive” and “productive” aspects of imagination—and to draw a corresponding distinction between mere “associations” and a genuine grasp of “affinities” involving the objects of consciousness themselves—the primary source of obscurity lies in Kant’s unclarity as to which of the two, or whether both of the faculties, should be regarded as responsible for one or more modes of a necessary “synthesis of the manifold of intuition.”

Kant sometimes attributes such synthesis to a pre-conceptual faculty:

Synthesis in general, as we shall hereafter see, is the mere result of the power of imagination, a blind but indispensable function of the soul, without which we should have no cognition whatsoever, but of which we are scarcely ever conscious. However, to bring this synthesis to concepts is a function which belongs to the understanding, and it is through this function of the understanding that we first obtain cognition properly so called. (A78/B103)

The intellectual component in cognition, Kant says, lies in the addition of an intellectual element to what would otherwise remain a purely imaginative synthesis. Apparently, it lies in the contribution of a special kind of unity to that synthesis (A79/B104, A94, A118, A130, A155/B194, A158/B197), or at least of some kind of consciousness of such unity (A103-6, A115-6). This may puzzle us not only in its concession of autonomously synthetic powers to imagination. It may puzzle us in a respect that is generally neglected but that I plan to take seriously. Understanding is needed, Kant says, in order to bring syntheses “to concepts.” As we shall see, an adequate interpretation of this, and of the correlative notion of a consciousness of “unity,” will eventually require an appeal to something much deeper than our ordinary notions of “applying” or “predicating” or, in general, “employing” concepts. The functions of unity in question will be needed to make sense of the notion of the very forming of concepts in the first place.

Elsewhere, Kant appears to attribute synthesis directly to intellect. Perhaps most notably, in his personal copy of the Critique, he inserted a correction to the passage from which I just quoted. As emended, it simply says that synthesis in general is “a function of the understanding.”³ In any case, whether or not there is a pre-conceptual faculty of synthesis, and whether or not it has an essential role in conceptualizing intuitions, Kant also attributes some kind of synthesis to concepts in their own right, or at least he insists that any synthesis that is ascribable to imagination also needs to be in “accordance” with, and apparently to be guided by, the understanding (A78/B103, A97, A110-13).⁴ The second edition might be supposed to resolve the apparent contradiction. It may appear to do so by maintaining that the relevant sorts of conceptual and imaginative syntheses are really just the same thing considered in two respects:

It is one and the same spontaneity, which in the one case, under the title of imagination, and in the other case, under the title of understanding, brings combination into the manifold of intuitions. (B162n)

This does not require denying that synthesis is a function of imagination. On the other hand, it is apparently compatible with very different views of the relationship between imagination and understanding. One would reduce the former to the latter from the start; the other, the latter to the former.

To most commentators, though not to all,⁵ it seems unlikely that Kant construed understanding as a mere function of imagination. However, the opposite might seem to be the case, at least with respect to the particular sort of imaginative synthesis in question, that is, with respect to whatever sort is presumed essential to the conceptualization of intuition. In the second-edition Deduction, Kant may seem at last to be clear: “This synthesis is an action of the understanding on the sensibility” (B152). Perhaps the same claim can be found, if less explicitly formulated, in the first edition as well: “The unity of apperception in relation to the synthesis of imagination is the understanding” (A119; Kant’s emphases). But it is in fact implausible to suppose that Kant’s efforts are reductive in either of these fashions: “Imagination and understanding are two friends,” he tells us, “who cannot do without one another”; to be sure, Kant adds, they are friends “who also cannot abide one another.”⁶

It is no doubt the tension that is generated by such cognitive pull and push, rather than indecision or confusion, that accounts for much of the difficulty in the Deduction. In any event, I want to defend a reading that is reductive in yet a third way. It is not reductive of either of the “faculties” as such. Both are essentially involved in the “application” of concepts to intuitions. On the other hand, the reading might be regarded as reductive with respect to the very notion of applying concepts in the first place. More exactly, it might be said to be reductive with respect to the notion of applying empirical⁷ concepts. As the second edition itself seem to affirm, such actions are to be reduced in favor of those of a “pure” intellectual faculty.⁸

As is well known, Heidegger has argued that Kant shifts in his view between the two editions: “In the second edition of the Critique of Pure Reason the transcendental imagination . . . is thrust aside and transformed—to the benefit of the understanding. . . . While in the first edition, all synthesis, i.e., synthesis as such, arises from the imagination as a faculty not reducible either to sensiblity or understanding, in the second edition the understanding alone assumes the role of origin for all synthesis.”⁹ My own reading will agree and disagree with some aspects of Heidegger’s. It will disagree with the view that the second edition reverses the position of the first. As regards the first edition, I follow Heidegger in seeing a radical dependence of understanding on imagination. It remains a question as to how specifically to construe that dependence.

II. The Transformation of Intuition

Kant’s introduction to the Logic begins with reflection on what empirical concepts “contain.” In this regard, he makes what may appear an odd suggestion. First, he repeats the familiar point that intuition is what accounts for the fact that thoughts have a “content,” that is, that they are thoughts about “objects.” Then he draws an expected distinction between pure and empirical forms of both intuition and conception. The apparently odd suggestion concerns this distinction. What Kant seems to say is that empirical intuitions and concepts are those that contain sensations (wenn Empfindung . . . darin enthalten ist). Sensation, in such a case, is said to provide the matter or material (Materie) of the cognitions in question (A50/B74). At A95, similarly, Kant distinguishes empirical from a priori concepts in virtue of the fact that the former consist of elements of a possible experience (aus Elementen einer möglichen Erfahrung besteht). Apart from that, a would-be concept would only amount to the “logical form of [zu] a concept, not the concept itself through which something is thought.”

It is not clear how concepts might be supposed to contain sensations as, or even as part of, their material. One suggestion is that Kant is speaking loosely when he says this. Shortly after the passage in question, he says that earlier, though he doesn’t say where, he had used the term Materie for the content (Inhalt) of cognition (A59/B83). Now the latter term, in the discussion immediately preceding this observation, seems to be used only in order to designate either the objects of cognition themselves or cognitive relations with those objects.¹⁰ It is not used for anything like the sensations which, in the Aesthetic and elsewhere, are said to provide material for cognition. So by calling sensation the matter of (empirical) concepts, not just the matter of intuitions or empirical cognitions, Kant might be speaking loosely. He might only mean that the ability to experience sensations is an essential element in one’s general capacity for relating empirical concepts to their objects.

On the other hand, in the discussion that had immediately preceded A59/B83, Kant seems to have been speaking of Inhalt in a rather different sense from the way in which he speaks of Materie in the passage that we are considering from A50/B74. (Nor does he use the latter term in the later passage.) In the discussion immediately preceding A59/B83, one could as well have spoken of space and time as part of the content, hence of the “matter,” of cognition. The point, in that context, was that a concern with spatiotemporal representation as such (whether pure or empirical) is not a concern that abstracts from “content,” since it does not abstract from spatiotemporality; because of this, it is not appropriate for consideration in General Logic. In this sense, space and time are a part of cognitive content. At A50/B74, by contrast, the point is to exclude the representations of space and time themselves from the class of representations comprising “material” for cognition. This is because, as in the Aesthetic, the representations of space and time are purely formal. So in the sense relevant in the context, something else is needed, namely, sensations, in order to provide material or content for cognition.

What is striking about the passage at A50/B74 seems to be that Kant is now extending the point about the need for ingredient material beyond the need for a distinction between pure and empirical intuitions. It is now extended so as also to apply to the need for a distinction between pure and empirical concepts:

Both may be either pure or empirical. Empirical, if sensation (which presupposes the actual presence of the object) is contained in it [darin enthalten]; but pure, if no sensation is ingredient in it [beigemischt]. (A50/B74)

Thus it may be possible to speak in two different senses of the Materie of concepts. In one sense, the matter of a concept is the range of objects, or of possible objects, that fall under it. By extension, it may be whatever is transcendentally necessary (including the pure form of intuition) for establishing cognitive relations with objects. In a second sense, the matter of a concept may be something whose actual ingredience in a concept is a condition of the establishment of cognitive relations with objects. We already know that, in the second of these senses, or at least in an analogous sense, sensation is the matter of empirical intuitions. The question is whether Kant now supposes that sensations, or something analogous to them, must be regarded as similarly incorporable into empirical conceptions.

If this is Kant’s suggestion, then he is suggesting that intuitions, not just judgments, must be capable of receiving judgmental form as a part of their internal structure. It is difficult to see what the upshot of an action of this kind could be, if not the very conversion of intuitions into conceptions. Now where, one might ask, would empirical concepts fit into such “conceptions”?¹¹ One is inclined to suppose that concepts are what we need to apply to intuitions in order to yield a certain sort of cognitive state as an upshot, namely, in order to yield a perceptual judgment as a composite of the two. But Kant now seems to be proposing something different. What he seems to have in mind is the generation of perceptual judgments, not by means of the application of concepts to intuitions, but rather by means of the application of the mere forms of judgment to them. Intuitions, by this means, are turned into perceptual judgments; what we call “application” of concepts to the former is just whatever act of understanding is involved in their conversion to the latter. If this is Kant’s view, then we can at least see one point in the suggestion that sensations need to serve as matter in empirical “conceptions.” Certainly, they could at least be said to serve as matter in intuitional judgments. And it is Kant himself who insists that “the only use which the understanding can make of these concepts is to judge by means of them” (A68/B93). Still, what of the suggestion that sensations are also matter in empirical concepts, not merely in empirical judgments?

Kant speaks in a number of places of intuitive representations becoming concepts. At A76/B102, he refers to a process by which representations are “transformed into concepts”: in Begriffe zu verwandeln. At A147/B187, he speaks of the categories as “functions of the understanding for [zu: my emphasis] concepts. At A350, similarly, he says that “consciousness” is that which “makes all representations into [zu: my emphasis] thoughts”: was alle Vorstellungen zu Gedanken macht. He does the same in various lectures and Reflections on Logic and elsewhere. Throughout, he takes the central question to be “how representations become concepts . . . thus how it comes about that repraesentatio singularis becomes communis.”¹² We also need to be serious about Kant’s claim that “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” (in einer Anschauung: A79/B104-5). The function is, of course, that of the “pure understanding,” through its forms of judgment or the categories. (Perhaps it would be better to say: through its forms of judgment in their role as the categories.)

Kant attributes to this function of understanding a unity of synthesis that does not merely bear on a manifold of several intuitions, but rather on a manifold contained in a single intuition: in einer Anschauung. This is what we ought to expect, if his view is that any act of empirical conceptualization involves the elevation or transformation of intuitions, not by means of a distinct act of “applying” concepts to them, but rather by means of the activation of pure intellectual functions in regard to those intuitions. Were the former what is in question, one might suppose that all determinate content, or at least determinate conceptual content, needs to be “brought” into intuitions by the concepts that are to be predicated of them. In a truistic sense, we may say that this is so on any account. But if the application of concepts is to be constituted by an “application” of pure intellectual functions in the first place, then all determinate conceptual content presupposes a body of material that is autonomously (if not also antecedently) ingredient in intuitions in its own right. This is because the forms of understanding do not differ from case to case.

Another way to put the point is this. If the conceptualization of an intuition, and so predicating something of its object, involves the “application” of something intellectual to intuitions, then it could not do so merely by means of a function that is exercised in regard, as it were, to intuitions as units, apart from the question of what those units literally contain. This could not possibly work, so long as we were insisting, with Kant, that the addition in question is of a purely formal sort. If it is of a purely formal sort, then what is bestowed on an intuition, in conceptualizing it, is exactly what is to be bestowed on any other one. All “material” similarity and difference would have to come from the intuitions themselves. But if variation in conceptual content were the upshot of operations on intuitions as units, then it would follow that conceptual content always co-varies with variation in those units. That would of course be absurd. In whatever way our account is “reductive” with respect to predication, we want to be able to suppose that qualitatively different intuitions might be conceptualized by means of the very “same concepts,’’ and that a given intuition might have been conceptualized by any number of different concepts. Clearly, the sameness and difference in question could not stem merely from the application of pure intellectual forms (though of course some kind of sameness—a purely formal one—does stem from this). There is only one possible alternative. Intellectual form must be able to yield concrete predication only so far as it somehow applies (but in a different sense from what one usually calls “predication”) to a body of material in the latter—that is, to a body of material that may remain constant or vary within the formal unity of a given intuition. Such “application,” one must suppose, in some way forms a full-blown concept (or at least a “concept in use”) precisely out of that body of material.

I do not propose that sensations, in the usual Kantian sense, could ever provide such material for formation into concepts. The reason for this is simple: intuitions that contain different sensations, in the usual Kantian sense, remain subject to the very same empirical predications; by the same token, the purely sensational “content” of an intuition does nothing to determine the way in which that intuition is to be conceptualized. Nevertheless, it is implicit in my argument so far that if sensations cannot be formed into empirical concepts, then something importantly analogous to them must be able to be so formed. This latter must be something analogous in the sense that, like mere sensation, it also serves as matter within intuitions themselves.

It may be this that Kant somewhat hesitantly acknowledges in his suggestion that sensations need to serve as matter in empirical concepts. In any case, as we have seen, Kant is at least not hesitant in suggesting that, in some sense or other, empirical conceptualization is an operation by which empirical intuitions are themselves converted, if not literally into concepts, then into determinate empirical conceptions. It is difficult to see how to make sense of this, apart from something like the proposal in question.

It may be with respect to their focus on such issues as these, not with respect to their doctrines or even their degree of mere clarity, that the first and the second versions of the Deduction ultimately differ. Certainly, the first appears to inquire more explicitly as to what concepts are in the first place.¹³ In a dispositional sense, one may of course say that concepts are nothing other than capacities for judgment, that is, for forming conceptions or cognitions. But Kant himself could not ignore the question, nor should we: What is it to form a conception or a cognition? It seems clear, in the light of what I have argued so far, that Kant presumed that an account of this needs to recognize that, in a generally neglected sense, intuitions themselves must provide the very “material” for such formation. In comparison, everything else would be mere “form” in a sense more literally formative than generally recognized:

Experience contains two very dissimilar elements, namely, the matter for cognition from the senses, and a certain form for the ordering of this matter, from the inner source of pure intuition and thought, which, on occasion of the former, are first brought into action and yield concepts. (A86/B118; final emphasis added)

This unity of consciousness (of the connection of our representations) is as much a priori in us as the foundation of all concepts as the form of appearance is in us as the foundation of intuitions.¹⁴

Here a concept is considered only subjectively, not with respect to how it determines an object through a Merkmal, but rather with respect to how it can be applied [bezogen] to several objects. (Not with respect to how they arise [entspringen] as representations, but rather with respect to how given representations become concepts in thinking [im Denken zu Begriffe werden]. Only with respect to the form of a concept.)¹⁵

While this aspect of Kant’s theory is not usually noted, Paton, for example, gives it some attention. Unfortunately, he confesses himself not “altogether [to] like the statement that we make concepts (as regards their form) out of given intuitions.” His best attempt to clarify the statement leaves him uneasy. He notes that Kant often identifies concepts with the common characters (Merkmale) of possible individuals. Having done this, Kant then “looks upon what is common as itself an individual intuition which receives its universality by being referred to many different objects as their ‘mark’.”¹⁶

At least frequently, Kant does in fact define the notion of a concept in terms of the fact that concepts represent objects by means of their Merkmale (e.g., A19/B33, A320/B377). Sometimes he goes so far as to say that concepts are themselves Merkmale. Thus somehow, he seems to say, one and the same item is able to serve both as a possible feature of cognizable objects and also as the representational medium through which those very features are cognized.¹⁷

This, together with the assumption that empirical concepts are “derived” from experience, may suggest that Kant adheres to a version of the Thomistic theory of abstraction. It seems to imply that, in order to know them, the features (or would-be features) of objects need literally to become ingredient in the faculty of cognition itself. Kant’s only advance, if it is one, would seem to lie in the conclusion that, for such to be possible, the objects must be “appearances.” But it is Kant himself who gives us pause in regard to any doctrine of abstractionism. He concedes that empirical concepts are always “abstract.” Truistically, they at least need to abstract from many of the features of the objects that are known by their means. But at the same time, they are not themselves abstract-ed from anything, not even “appearances.”¹⁸

Apart from the qualification, we might still make some sense of the notion that (empirical) concepts are “derived” from experience, in particular from intuitions. For example, we might explicate that notion in terms of the formation of concepts out of a body of material available in intuitions. That vould allow us to say that the concepts are “derived” from intuitions, and even to understand Kant’s inclination to put the point in terms of the converting of the “common characters” of appearances into concepts. But it would do nothing to explain an additional tendency to regard them as intuitions, hence particulars. Indeed, it would do nothing to explain an inclination to regard the common features of appearances as entities of any sort. In any case, Kant’s spatiotemporal ontology for the world of “appearances” could surely not welcome the presumption of such a status for common features.

I have argued elsewhere that Kant’s doctrine of sensible form was specifically designed to replace the notion of sensible features as literally ingredient in the cognitive states through which objects (even as appearances) are perceived.¹⁹ It is difficult to avoid supposing that Kant was equally intent on providing an alternative to the corresponding Scholastic approach to “intelligible” form. To do this, he could not escape the question: What is it that one “applies” to an intuition, in order to apprehend the latter’s object through some concept? Can a concept be some item or entity, at one’s disposal for that purpose—even an item or entity that one has oneself formed for the purpose? It seems rather to be the case that (except insofar as we are prepared from the start to identify “concepts,” irrelevantly, with capacities or dispositions for certain kinds of activities) concepts must be indistinguishable from, or at best mere aspects of, certain kinds of activities, namely, the very activities that one normally calls “applying” concepts in the first place.²⁰ What gets “applied,” in the most basic sense, is simply the activities themselves. Despite my own suggestion that concepts are made out of something, thus despite the apparent concession that, once made, they are entities in their own right, the latter involves a view that we need to avoid.

We cannot abandon the idea that empirical concepts function as a kind of predicate in intuition, and that a concept is therefore in some sense an item—perhaps even a quasi-linguistic item—attachable to distinct intuitions or “tokenable” by them. But when we do not ignore some of Kant’s own apparently odder suggestions, we obtain a view concerning these notions that is radically different from what the linguistic analogies usually suggest. Again, my own proposal is in a way reductive with respect to concepts. It is reductive with respect to the distinction between concepts and the intuitions that they conceptualize. But it is reductive only in a certain way. It is reductive in the sense that the “predicating” of concepts is an action to be explicated as nothing other than the action of applying something pre-conceptual (though not necessarily pre-intellectual) to certain intuitions (and to a correspondingly ingredient material).

This is obviously not intended to deny that there is such a thing as applying concepts to an intuition. It is only to say that doing so is nothing other than an intuition’s elevation, through the elevation of some material in it, to a specifically conceptualized status. The elevating factor is attributable wholly to something “formal.” But as Kant himself tells us, the formal factor is just consciousness itself, insofar as the latter constitutes “the merely subjective form of all our concepts” (A361). As we shall see more clearly later, thus applied to the problem of the formation of particular cognitions, not merely to the problem of a general unity among manifolds of distinct cognitions, Kant’s pronouncements concerning a corresponding “unity of consciousness” take on a significance that is more radical than typically supposed:

But this unity of possible consciousness also constitutes the form of all cognition of objects, whereby the manifold is thought as belonging to a single object. Thus the 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). (A129)²¹

It is at this point that we need to attend to the role of imagination in conception. Undeniably, it is a role that cannot be played by mere sensations, as Kant himself usually understands the latter. Sensations, in that sense, could never provide material for concepts in anything like the way in which they provide material for intuitions. To suggest that they can, as we have seen Kant does, could at most serve to anticipate a fuller account. It anticipates an account according to which something relevantly like sensations must be material for concepts. Of course, the ground of comparison is crucial. But my suggestion is that whatever is needed to provide material for concepts (that is, for non-“categorial” ones) has got to be comparable to sensations in the respect that Kant himself often takes to be definitive of sensation in the Aesthetic, namely, in the respect that it is a pre-conceptual matter for intuitions, incorporable within the latter by means of intuitional form. Sensation, in the narrow sense, cannot provide matter for concepts, not even for empirical ones. But whatever provides that matter must nevertheless, like sensation, provide it for intuition as well. This is because, as already argued, only in this way can it be anything relevant to the conceptualization of intuition in the first place.²² I shall argue that, in Kant’s view, something “imaginative” needs to play this role. In particular, it is sets of imaginative “anticipations and retentions.”

III. The Manifold of Intuition

In the present section, I do not specifically argue that the pure forms of understanding are modes or operations of consciousness in regard to imaginative anticipations and retentions. But I prepare the way for that enterprise by commenting on the ambiguity of the notion of synthesis, and of the corresponding notion of a manifold of intuition, and by discussing some difficulties that confront other approaches that these notions might seem to allow. In Chapter Five, I discuss some of the same ambiguities and alternatives, but with specific regard to the second-edition Deduction. Aside from passing reference to the second edition, I continue to attend to issues that arise elsewhere.

We may begin by considering Kant’s introduction of the notion of synthesis into the Critique. It may in fact appear to exclude the idea that imaginative anticipations and retentions are what provide material for concepts:

Transcendental logic . . . has lying before it a manifold of sensibility a priori presented by transcendental aesthetic, in order to provide material [Stoff] for the pure concepts of understanding. In the absence of this material those concepts would be without any content, therefore entirely empty. Now space and time contain a manifold of pure intuition a priori, but at the same time are conditions of the receptivity of our mind—conditions under which alone it can receive representations of objects, and which therefore must also always affect the concept of these objects. But the spontaneity of our thought requires that this manifold first be gone through [durchgegangen] in a certain way, taken up [aufgenommen], and connected [verbunden], in order that a cognition be made out of it [um daraus eine Erkenntnis zu machen]. This act [Handlung] I name synthesis.

By synthesis in the most general sense, I understand the act of putting different representations together [zueinander hinzuzutun], and of grasping what is manifold in them in a cognition [ihre Mannigfaltigkeit in einer Erkenntnis zu begreifen]. Such a synthesis is pure, if the manifold is not empirical but is given a priori, as is the manifold in space and time. Before we can analyze our representations, these must themselves be given, and therefore as regards content no concepts can first arise [entspringen] by way of analysis. Synthesis of a manifold (be it given empirically or a priori) is what first gives rise to a cognition [bringt zuerst eine Erkenntnis hervor]. This cognition may indeed, at first, be crude and confused, and therefore in need of analysis. Still the synthesis is that which gathers the elements into cognitions and unites them into a certain content [die Elemente zu Erkenntnissen sammelt, und zu einem gewissen Inhalte vereinigt]. . . . (A76-7/B102-3)

It would appear from this that the manifold supposed to be gathered by means of synthesis is either a multiplicity of the parts of space and time or (or possibly in addition) a multiplicity of the appearances of spatiotemporal objects or possible objects. This may seem to follow from the fact that Kant says that the manifold in question is provided by “transcendental aesthetic.” He in any case does speak of space and time as containing a manifold. On the other hand, we know from our discussion of the Aesthetic that Kant frequently shifts between reflection on cognition as such and reflection on cognition’s intentional correlate. So any talk about pure intuition, or about intuitional form, is bound to be ambiguous.

The form of intuition may be space and time, or the spatiotemporal structure of appearances. But the form of intuition may also be a structure of cognition itself, namely, the structure that accounts, when a cognition is conceptualized, for the latter’s object-directedness. In that case, it would simply remain to be seen what the corresponding “manifold” could be. It is in any case not through the very parts of space and time themselves, as opposed to our representations of them, that intuitional form, in the second of these senses, is able to direct spatiotemporal concepts toward objects. Similarly, the empirical manifold that Kant contrasts with a corresponding pure one may be a manifold of spatiotemporal appearances or possible appearances. But it may also be a manifold through which appearances are to be apprehended, and in that case it presumably ought to comprise, not appearances themselves, but some variety of representation of the latter.

There is a straightforward objection to the idea that conceptual representations are generated by means of synthetic activities directed toward parts of space and time or toward the possible appearances in them. The objection concerns the sense in which the requisite acts of synthesis could possibly manage to get directed toward their presumed targets in the first place, in the absence of already functioning representations able to place those targets before the mind. Assume, for example, that the pure manifold is a manifold of actual and possible regions of space and time. These, we might maintain, are given in a pure intuition, and they are in turn utilized as material for synthesis. But in what sense could one possibly be “given” a manifold of distinct regions in the first place, apart from an at least rudimentary power of discrimination already active? The objection is a fortiori applicable with respect to a manifold of actual appearances, or a hypothetical manifold of images of merely possible ones, insofar as the members of those manifolds are supposed to be distinctly apprehensible items, sufficiently present to mind to be the targets of possible synthesizing.²³

This objection may be thought to reject my own account of the Aesthetic. It may seem to do so because it rejects the idea of a pre-conceptual access to space and time. But it only rejects that idea in a certain sense. My account of the Aesthetic does not entail (nor does it deny) that regions of space and time, or appearances in them, are presentable as objects of intuitional consciousness apart from conceptual functions. What the account entails is only that the application of concepts to space and time, or to appearances in them, requires the ingredience of the concepts in question in irreducibly and uniquely object-directed states. It requires their ingredience in cognitive states whose object-directed character is an internal feature of those states, but is reducible neither to the ingredient concepts or sensations in those states nor to mere relations involving such ingredients. Compatibly with this, one might insist that apprehension of spatiotemporal regions, or of spatiotemporal appearances, as genuinely discriminable items before one’s mind, presupposes the presence of concepts in intuition. If it does, then it would of course be difficult to see how the synthesis of a multiplicity of material in either perceived or imagined space or time could be required, as Kant suggests it is, in order to make conception possible in the first place.

If the objection is compatible with my account of the Aesthetic, it may nevertheless appear to be incompatible with my suggestion that imaginative material is what constitutes the matter of concepts. If imaginative anticipations and retentions are what provide that material, then we seem to acknowledge pre-conceptual states that are genuinely object-directed. We therefore seem to concede that the manifolds of appearances that are the objects of such states may after all be the original targets of synthesis. In the next section, I distinguish two kinds of anticipation and retention. With respect to one of these kinds, it is clear that anticipation and retention presuppose a capacity for forming representations of anticipated and retained objects or appearances, as discriminable intentional objects. But there is also a kind of anticipation and retention that does not presuppose this capacity. A creature that anticipates and retains in this latter way may be said to anticipate, not merely (“de re”) with respect to some real object or situation, but with respect to the merely possible appearances of merely possible objects or situations. (As we shall see in Chapter Six, the case of retention requires special treatment.) Merely possible appearances may then be regarded as the “objects” thus anticipated. But it does not follow that the anticipating subject has any such objects “in mind,” as a manifold of items potentially targetable by mental operations. In particular, it does not follow that the anticipated appearances are, just as such, the objects of intuitional states in Kant’s sense. (Kant claims that imagination is a form of intuition. I consider this difficulty in the next section.)

Furthermore, even if we concede the possibility of purely intuitional “consciousness,” the fact that we may speak of anticipated appearances merely on the ground that some creature is in an anticipational state permits us to speak of “syntheses” of appearances even where the appearances in question are not objects of such consciousness. In that case, one might say that we are dealing with syntheses directed toward “objects” that are, as such, nothing more than the intentional correlates of the very anticipation of them. So construed, I have no objection to the claim that synthesis is a mental operation in regard to a manifold of objects of consciousness. We need only remember, in that case, that the notion of an intentional correlate is different from what I introduced in connection with the Aesthetic. In that context, even if we were unwilling to concede the possibiltiy of truly pre-conceptual consciousness, we could also speak of the intentional correlates of pre-conceptual intuitional states. We could do so simply by referring to the ways in which such states are in principle eventually conceptualizable. What was important, in that context, was to acknowledge the role of intuitional form. Apart from that, the question whether such form alone puts objects before one’s “consciousness” runs the risk of raising an issue that is merely verbal. But if we now also speak of the intentional correlates of anticipation and retention, on the most primitive level of the latter, and if we speak of anticipated and retained appearances as thereby before one’s consciousness, then there is no longer any reason to suppose that we are talking about states with intuitional form in the first place. For we would now be talking about nothing more than states that might, along with sensations, be able to serve as material in intuition.

It may be Kant’s unclarity on this score that accounts for much of the unclarity in the Deduction. This may account in particular for the apparent suggestion that the activities of synthesis are supposed to operate on a body of peculiarly non-objectified “objects” of consciousness. That is, it may account for the suggestion that the body of material for the activities of “pure” synthesis comprises, if not the very parts of space and time themselves, then, even more mysteriously, something that is able to get synthesized into the parts of space and time. Likewise, it may account for the suggestion that the body of material for the activities of empirical synthesis comprises, if not full-blown appearances, or even the images of possible appearances, then something that is at most capable of getting synthesized into appearances. Since imaginative anticipations and retentions, of the sort needed to develop Kant’s theory, cannot plausibly be regarded as intuitions in their own right, and yet since they still appear to have (possible) spaces, times, and appearances as their correlates, Kant must have found it considerably easier to write as if the activities of synthesis operated directly on the latter. Strictly, what he needed to consider was the sense in which the manifold of anticipations and retentions was itself subject to synthesis of some kind. But the status of that manifold in the first place, with its mysteriously pre-conceptual yet non-intuitional “representation,” must have troubled him. On account of this, it may have seemed easier to formulate the doctrine in question in “noematic” rather than “noetic” terms.

Now there are a number of ways in which syntheses might be supposed to function in cognition. Many of them are objectionable, not on account of regarding the former as already bearing on space, time, and appearances as “objects” (which is, in a sense, permissible), but rather because they construe synthesis as already conceptual in nature. If they do that, then they cannot appeal to synthesis to explain how conception is possible in the first place. In the passage that I quoted, the latter is Kant’s aim. He implies that concepts, with respect to their “content,” first arise [entspringen] via synthesis. He also says that synthesis is what “gathers the elements into cognitions and unites them into a certain content [die Elemente zu Erkenntnissen sammelt, und zu einem gewissen Inhalte vereinigt],” and that synthesis is what “first gives rise to [bringt . . . hervor]” cognitions. Finally, that this is in fact Kant’s aim seems confirmed by the passages considered in the previous section. Those passages imply that, analogously to the forms through which sensations are incorporated as matter in intuitions, intellectual forms need to incorporate something as matter in acts of conception. The incorporation takes the form of synthesis. As intellectual as the latter might perhaps be, it would be circular to regard it as specifically conceptual in nature.

As I noted in the first section of this chapter, an account according to which transcendental synthesis is essentially intellectual in nature may consider itself able to regard such synthesis as imaginative as well. The easiest way to develop this notion would be to construe the role of the imagination as in effect a merely potential one. Kant’s point may then be taken to be that concepts provide a special awareness of imaginable possibilities. One can imagine all sorts of possibilities, but concepts are what, by their very nature, dictate that some of those possibilities are more especially in order than others: for example, the possibility, upon apprehending a given appearance, of proceeding via a certain route to the apprehension of another. One can imagine a situation, for example, in which a sow’s ear, or the appearance of one, gets transformed into a purse, or at least into the appearance of a purse. But the concept of a sow’s ear, one might say, conveys no special anticipation of this particular possibility. In general, we may suggest, any (empirical) concept’s identity is constituted in terms of its provision of such rules for imaginative possibilities. To this extent, a concept “synthesizes” the latter.²⁴

Obviously, these notions might also be thought adequate to account for the centrality of “rules” in Kant’s theory of synthesis: “a concept is always, as regards its form, something universal which serves as a rule” (A 106). I shall have more to say about the problem of rules later. For now, we may reflect a bit further on the positions available as to the relationship between rules and synthesis. One might suppose, for example, that synthesis is nothing more than whatever it is that rules qua rules “do,” namely, dictate or prescribe norms with respect to certain possibilities. Alternatively, but equally intellectualistically, one might take synthesis to be whatever is involved in one’s more or less explicit intellectual consciousness or conception of such rules. Finally, a less intellectualistic position might attempt to ascribe synthesis directly to imagination itself, not either to rules as such or to a specifically intellectual consciousness of them. It might even attempt to construe the rules as grounded in imagination from the start:

Imagination by itself could give us only pictures or images, but because one and the same understanding is able to conceive the various principles [n.b.] upon which the imagination synthetises [sic] the given manifold, the concepts of the understanding are able to give us knowledge of objects.²⁵

Though it is unclear what Paton intends by the principles upon which imagination synthesizes, what it is for imagination to synthesize “upon” principles in the first place, and therefore what synthesis itself is supposed to be, it is clear that Paton takes concepts to be, at least in the first instance, concepts of rules.²⁶ At the theoretical extremities, this leaves us with two positions as to the sense in which imagination might be relevant to the rules in question. It might be presumed to be relevant in the fairly minimal way that I have already indicated, namely, in the sense that the rules in question must have a prescriptive bearing on whatever is in principle imaginable in certain contexts. However, imagination might also be presumed to be relevant in a more active or operational way. It might be supposed, as Paton suggests, that imagination already operates in a rulelike manner and that concepts merely express an intellectual consciousness of that operation. Paton himself seems to waver between these extremes. After all, if imagination really provided only pictures or images, it would be difficult to see what could be involved in its also synthesizing the latter, and in its operating “upon” principles in doing so. By contrast with Paton, in any case, others straightforwardly identify concepts with rules from the start, and they consistently distinguish concepts themselves from one’s consciousness or conception of rules.²⁷

Strawson has tried to do more justice than this to imagination as an actively ingredient element in its own right. Kant’s point, he suggests, is that certain imaginable possibilities are in some way actually brought to mind in conceptualizing appearances. They are brought to mind in such a way as to enter the very apprehension of the latter. Consequently, the manifold of appearances in question is not merely a potentially present one:

[T]he actual occurrent perception of an enduring object as an object of a certain kind, or as a particular object of that kind, is, as it were, soaked with, or animated by, or infused with . . . the thought of other past or possible perceptions of the same object. . . . Nonactual perceptions are in a sense represented in, alive in, the present perception; just as they are represented, by images, in the image-producing activity of imagination.²⁸

This approach comes closer to the sort of view that I want to propose. But despite its attempt to do justice to imagination as an actively ingredient element in experience, it is at bottom intellectualistic. This is because the imaginative content of a given experience is still something that a concept alone is supposed to bring to experience. The only reason for regarding that content as imaginative is that, in an unexplained way, the “thought of other past or possible perceptions” is supposed actually to enter into the given experience, thus not to remain a thought that is merely externally attached to it. I am not objecting that it remains unexplained in what way such thoughts might become ingredient in experiences. I am objecting to the suggestion that the imaginative component is in the first place imaginative because of the operation of purely intellectual factors on a manifold of sensibility. Kant himself, as I have noted and as we shall see more clearly later, seems to hold that imagination contributes, not as a mere upshot of conceptualization, but as something that helps to make the latter what it is from the start.

A similar point may apply to Allison’s defense of the claim that, even in the second edition, Kant continues to ascribe an essential function to a non-conceptual capacity for imagination, precisely with respect to the latter’s ability to project the future and to reproduce the past in any act of conceptualization. So far as I can tell, and despite Allison’s aim to the contrary, the only respect in which the allegedly “imaginative” function in the case seems to differ “from the purely intellectual synthesis that occurs in judgment [is] that it is also conditioned by the form of inner sense.”²⁹ This would seem to be compatible with any projection and reproduction in the case merely being a matter of thoughts or judgments, somehow embedded or implicit in conceptual acts. The only distinctively non-conceptual element, and the only apparent “conditioning” by the non-conceptual form of inner sense, would appear to lie in the fact that what is in question as the objects of that projection and reproduction are non-conceptual items, namely, potentially apprehensible, or already apprehended, times. But what is needed in order to do justice to the role of imagination is rather some reason for saying that the imagining of those items involves more than mere judgments to the effect that there might in fact come to be, or already have been, such things (and also, of course, a sense in which such imagining, though not conceptual, might actually be an ingredient in something that is conceptual). After all, one may judge about the future and past, as well as intuitively apprehend them.

The question of synthesis is, of course, inseparable from that of the nature of the “manifold” supposed to be synthesized. Whatever their differences, all of the accounts on which I have commented seem to agree on one point in this regard. They appear to presume that, if a manifold of items is in question as a target for synthesis, then it is a manifold of objects, or of potential objects, of consciousness. In this sense, the accounts all deal with a manifold of items that could not themselves be ingredients in cognitive states. At most, they are items that might be apprehended or represented by means of cognitive states, or by what is ingredient in such states. I have agreed that it is possible to speak in this way of the manifold of intuition. In fact, Kant very often seems to intend a manifold of spatiotemporal appearances, when he speaks of the manifold to be synthesized, although on other occasions he seems to mean something else. But one question still needs to be faced, if we regard the synthesized manifold as a manifold of appearances.

We need to face the question as to how those appearances might be present to consciousness in the first place, in order to be targets of synthesis. Now one answer, again, is easy: we might simply suppose that they are present to consciousness by virtue of being conceived of. The problem is that the answer precludes any appeal to synthesis in an account of what conception itself is. At most it could serve in an account of the logical implications of certain conceptions. The alternative is to be serious in what might seem a paradoxical supposition, namely, that while a manifold of appearances is necessarily apprehended in any (empirical) conception, its apprehension must be accomplished precisely through some non-conceptual mode of representation. Given this, it is not implausible to speculate that any conceptual apprehension of possible appearances, in the conceptualization of a given appearance, must be accomplished through the ingredience in cognition of some other sort of manifold altogether, namely, through the ingredience in it of manifolds of non-conceptually imaginative anticipations and retentions.³⁰

Insofar as a manifold of anticipations and retentions is ingredient in intuition, we may speak of a manifold of objects or appearances as an “intentional correlate” of that ingredience as well. In other words, the notion of synthesis is essentially Janus-faced.³¹ It is comprehensible only to the extent that we are able to grasp the correlation between its “noetic” and “noematic” dimensions. This, again, may account for much of the ambiguity in Kant’s own treatment of that notion. What Kant describes as our capacity for synthesis may sometimes be nothing other than our capacity for ordinary conceptual dealings with the objects of experience (including spaces and times) and the possible appearances of them. If so, then we need to be concerned with synthesis in some other sense as well. We need to be concerned with synthesis in a sense that allows that function to serve in the constitution of conceptual representation in the first place. The point is to see that this can be accomplished only if an intuitional state can contain, and apprehend objects through, a set of sub-states that satisfy at least two conditions: (1) they must be states of an anticipative/retentive nature, in a sense yet to be clarified; (2) while conceptualization, in a sense yet to be clarified, must consist in the elevation of them (though not necessarily by means of a temporally subsequent act) to a higher level of consciousness, they must be in themselves, in the sense in question, non-conceptual in nature.

A number of passages seem to indicate that the primary notion of the manifold of intuition is that of a manifold actually present in a given intuition, not merely of a manifold (and thereby a merely potential manifold) of distinct intuitions. Consider the following passage:

The unity of apperception is thus the transcendental ground of the necessary conformity to law of all appearances in one experience. This same unity of apperception in respect to a manifold of representations (determining it out of a single one [es nämlich aus einer einzigen zu bestimmen]) acts as the rule, and the faculty of these rules is the understanding. (A 127; emphasis added)

As it happens, Kemp Smith translates the passage so as to conceal the point in question. He replaces “determining it out of a single one” with “determining it out of a unity.” It is difficult to see what the latter could mean. In any event, while Kant most often speaks only of a “manifold of intuition,” and sometimes directly of a manifold of distinct intuitions or even appearances, several passages also carry a different emphasis:

. . . die mannigfaltigen Vorstellungen, die in einer gewissen Anschauung gegeben werden. . . .(B132) . . . das MannigfaltigeeinergegebenenAnschauung. . . .(B137) . . . Vorstellungen in irgendeiner gegebenen Anschauung. . . . (B138) . . . alles in einer Anschauung gegebenen Mannigfaltige in einem Begriff vom Object vereinigt. . . . (B139) . . . Das mannifgaltige in einer sinnlichen Anschauung Gegebene . . . alles Mannigfaltige, sofern es in Einer [sic] empirischen Anschauung gegeben ist. . . . (B143) . . . das Mannigfaltige in einer gegebenen Anschauung. . . . (B143) . . . Ein Mannigfaltiges, das in einer Anschauung, die ich die meinige nenne. . . . (B144)

In addition, a number of other passages speak of a manifold in, rather than of, “an intuition,” though they perhaps do not so strongly suggest that what is in question is a particular instance of intuition, not simply the totality of, or perhaps a subject’s mere capacity for, intuition in general.

Thus two different sorts of manifold might be regarded as providing “material” for synthesis and therefore material for the concepts that one applies to an intuition: a manifold of ingredient sub-states in the intuition itself, and a corresponding manifold of anticipated or retained appearances, as objects or as potential objects of consciousness. Kant himself seems to be operating with two different notions of material in the passage quoted above (A76-7/B102-3). He concludes the passage by saying that it is synthesis that is first required in order for cognition to have any content (Inhalt) to begin with. But without any “content,” it is difficult to see what sort of material could be available as a target for that synthesis to work on. Now it seems unlikely that Kant only means, by the claim that synthesis is necessary for cognition’s possession of content, that synthesis is involved in conception. That is obviously his view. But if it were all that he means, then the synthesized manifold could be regarded merely as the correlate of the act of conception as such. It seems unlikely that this is all that Kant means by saying that synthesis is required for content. For in the same passage he also distinguishes between the original “material” for conceptual representations and their contents: in the absence of the manifold of intuition, Kant says, the requisite material (Stoff) would be lacking for concepts, and so “those concepts would be without any content (Inhalt), therefore entirely empty.” The last point presumably means that they would be without any object.

Kemp Smith’s translation of the preceding sentence may encourage the supposition that the Stoff in question already amounts to “content” in the relevant sense, hence at least to a rudimentary sort of object of consciousness. As he translates the passage, the pre-conceptual manifold of intuition is already “presented . . . as material for the concepts of pure understanding,” which is necessary lest those concepts lack content. What Kant says is only that the manifold needs to be available in order that those concepts have a content. A similar distinction seems to be operative at A95: If an empirical concept did not consist [besteht] of “elements of a possible experience,” then [er würde alsdann] it would have no Inhalt, “since no intuition corresponds to it; and intuitions in general, through which objects can be given to us, constitute the field, the whole object, of possible experience.” There is no suggestion, in either of these passages, that the manifold of material to be synthesized in a concept, or the elements of experience of which a concept “consists,” must be in some way part of that concept’s very “content.” At least there is no such suggestion, insofar as the content of a concept is regarded as either spaces or times or as any sort of object capable of appearing in the spaces or times to which concepts are applicable.

One might, of course, distinguish notions of content, as well as notions of an “object of consciousness.” Surely Kant does not deny that, in some sense or other, non-human animals, and human infants, are exposed in intuition to, and are thereby “conscious” of, a variety of contents, hence a variety of “objects,” in at least a minimal sense.³² So one might always maintain that, by means of synthetic activities in question operating upon purely animal/infantile contents, hence upon purely animal/infantile “objects,” the latter are somehow converted into distinctively human ones, thus into objects of distinctively human consciousness. But whatever other difficulties stand in the way of it, reading the passage this way requires supposing that Kant used the term Inhalt, in the course of a single page on which that term is crucial, in two radically different ways: at the end of the passage, for the full-blown objects of human understanding; at the beginning, for some rudimentary stuff somehow supposed to have originally become those objects, at least with respect to one’s own conceptualized consciousness of them. That sort of shift, in Kant’s case, might seem notoriously possible. In any event, it would make more sense to suppose that, whatever rudimentary stuff is in question, it is not supposed to be formed into the objects that one eventually discriminates in phenomenal space and time, but rather into instances of one’s distinctively human discriminations of such objects. That is, the stuff in question is merely certain of the subject’s own states, ingredient in a given intuition and precisely through which conceptual discrimination needs to be effected.

There is a second point to be drawn from the passage on behalf of this reading. It is that Kant takes pains to distinguish between the fact that space and time themselves may be said to contain a manifold of some sort (presumably, a manifold of sub-regions) and a very different sort of fact about space and time, or rather about our representations of the latter. Space and time contain a manifold, Kant says, “but at the same time [aber gleichwohl] belong to the conditions of the receptivity of our mind—conditions under which alone it can receive representations of objects, and which therefore must always affect [affizieren] the concept of these objects.” Here Kant seems to be distinguishing between what pertains to space and time as such as objects of possible consciousness, on the one hand, and what pertains specifically to our representation of space and time, on the other hand. (We shall see the same distinction later, in the second-edition Deduction, when Kant distinguishes between space “as an object” and space taken in some other, but closely related, way: B16on.) He then seems to be saying that it is in the latter of these respects that we are to consider the manifold of pure intuition that space and time “contain,” when we consider that manifold in relation to the constitution of conceptual contents. The point would seem to be that some special connection needs to be effected with respect to one’s spatiotemporal representations, in order that we may eventually be in possession of concepts applicable to spatiotemporal regions (and of course also applicable to the appearances therein). This might be what explains the fact that Kant employs the two notions of (a) some sort of Stoff for a concept, on the one hand, and (b) a concept’s Inhalt, on the other. Possession of the former seems not to be identified with possession of the latter; it is rather a necessary condition of it.

IV. Kinds of Anticipation and Retention

It is time to attend more specifically to what I have been calling imaginative “anticipation and retention.” The next chapter will develop the notion in relation to Kant’s distinction between productive and reproductive imagination. For the present, I limit my discussion to some preliminary clarification and distinctions. Perhaps the first point to notice is that I have chosen to speak of anticipation and retention rather than to use the English term more immediately suggested by Kant’s own single term, Reproduktion. This is because Kant does not in fact view “reproduction” in primarily past-oriented terms. Far from it, he means to include all of our ordinary empirical anticipations. This is evident from the fact that his primary effort in regard to that notion is to establish a correlation between “the synthesis of reproduction in imagination” and a natural, anticipable, order of things in the world of appearances:

It is a merely empirical law, that representations which have often followed or accompanied one another finally become associated, and so are set in a relation whereby, even in the absence of the object, one of these representations can, in accordance with a fixed rule, bring about a transition of the mind to the other. But this law of reproduction presupposes that appearances are themselves actually subject to such a rule, and that in the manifold of their representations a coexistence or sequence takes place in conformity with certain rules.³³ (A100)

We shall have to postpone examination of whatever reasoning might lie behind the apparent inference in this passage: the inference from one’s mere capacity for associative reproduction to the apprehension of an objectively anticipable order among objects. But it is in any event clear that the “transition of the mind” that is here intended is not simply a calling forth, or the bringing before one’s mind, of something previously experienced and still retained. It is more a matter of the anticipation of a course of experience that one expects to continue (and to continue to change) at least on the whole as it has.

Kant, I presume, had no very clear theoretical idea as to what the difference might be between a mere calling forth of some possible appearance and the actual anticipation of it (relative to certain equally anticipable circumstances). He seems to suggest that the capacity for genuine anticipation stems mainly from a conditioning process on the basis of past experience. That may be why he speaks of “reproduction.” Obviously, not all anticipations arise in this way. Some of them are the upshot of more or less conscious inferences, or in any case take the form of perfectly reasonable judgments. But then judgment and inference are themselves ambiguous terms. A dog that has been conditioned to anticipate the presence of water or food under certain circumstances may be said to judge (or at least to believe) that water or food is forthcoming under those circumstances, and even to have drawn corresponding inferences. It is difficult to be sure what needs to be involved in such cases. Kant himself would perhaps prefer say that what the dog anticipates is merely sets of possible sensations, or perhaps intuitions, but never genuine objects such as water or food. But the question remains as to what it could be to do even this much. It seems unlikely, in any case, that Kant would seriously have thought that it involved nothing more than the “reproduction” of past sensations or past intuitions.

Some type of dispositional analysis, or a narrowly functionalist account, may seem appealing. If we attempted to pursue them, we would at least need to be clear about one thing. We would need to be clear about the distinction between purely dispositional or functionalist accounts of anticipation and retention, of the sort that needs to be in question here, and such accounts of conceptual acts themselves. We may tend to ignore the distinction, because it is permissible to regard even the most primitive anticipation and retention as constitutive of (correspondingly primitive) “belief’ or “judgment.” But steering clear of purely verbal issues, we need to remember that the relevant capacity for anticipation/retention is merely supposed to provide “material” for acts that are truly conceptual in Kant’s sense. The substantive thesis lies in the claim that, if any sorts of beliefs or judgments can be construed in purely animal/infantile terms, then they at most provide the material for the sort of function that in fact distinguishes us from (other) animals and infants.

As noted, Kant himself suggests a still more primitive view of animal/infantile anticipation and retention. He seems to suggest that it amounts to no more than the reproduction of past sensory states. If he supposed this, then he was wrong. But if he was wrong in this way, it is important to remember that there is nothing in his theory that requires him to be so. And there is no reason why his erring in this way need diminish his main claim: that whatever such primitive doings are, they need to be incorporable in intuitions, not merely externally or mechanically associated with them. They need to do this precisely so that they might in turn (though not necessarily subsequently) be relevant to the conceptualization of intuitions.

Kant himself is prepared to grant that animals are in fact capable of a kind of judgment. They are even capable of a kind of “reflective” judgment, at least with respect to their own intuitions. Now in human beings, the capacity for reflective judgment means the ability to reflect on an intuition, with respect to its susceptibility to some particular mode of conceptualization or other: zum Behuf eines dadurch möglichen Begriffs. This is a claim that is essential to Kant’s aesthetic theory, as well as to the general connection of that theory with his conception of purposiveness in nature. In particular, the claim involves the notion of a certain sort of suitability or affinity between the work of the mere imagination and that of an at least potential understanding. I return to this issue in the next chapter. For now, the point is simply that a certain sort of reflective judgment also takes place in animals. But:

however only instinctually, it is not in reference to a concept that is thereby to be attained, but rather to an inclination that is in a way thereby to be determined [nicht in Beziehung auf einen dadurch zu erlangenden Begriff, sondern eine etwa dadurch zu bestimmende Neigung]³⁴

Now while purely animal anticipations and retentions are not true judgments, and involve no conceptual form, in what Kant takes to be a privileged sense of these terms, there is of course no reason to deny that human beings often anticipate and retain in an animal way. In itself, the concession leaves two possibilities. One is that instances of animal anticipation and retention simply occur as a part of a person’s condition on an occasion, externally interacting with other subjective states of that person. No doubt this is constantly the case. But something else may also be the case. The other possibility is that instances of animal anticipation and retention actually enter in a more direct way into other subjective states, namely, into concepts and judgments of the sort that are distinctively human. The possibilities are not exclusive. At a given moment, one may anticipate and retain in both sorts of ways. What is important is just that, in another sense, this does not involve a distinction between two “ways” of anticipating and retaining at all. In both cases, purely animal anticipations and retentions may be in question. That is, we need not be talking, in either case, about anticipation and retention in the form of explicitly anticipative or retentive judgments. In the second of the cases, we might be considering anticipations and retentions merely insofar as they function as material in truly conceptual and judgmental states of some sort. (Even the latter states need not themselves amount to conceptually anticipative or retentive judgments. They might, according to the theory, be ordinary judgments about what is presently the case. The fact that anticipations and retentions are what constitute their material no more entails that they are anticipations and retentions in their own right than the fact that certain intuitions incorporate sensations within themselves entails that they are themselves sensations.)

I have said that Kant often appears to regard anticipation and retention as unduly primitive in nature. Other times, he may seem to do the opposite. This is because he sometimes attributes a transcendental dimension to “reproductive imagination.” But we need to note an ambiguity. I said that, in a strict sense, merely animal imagination, even when it functions as the matter of conceptual acts, does not in itself involve a distinctively conceptual “way” of anticipating and retaining. It is by hypothesis still purely animal anticipation and retention. The only difference is that it is now functioning as the basis for distinctively human mentality. In another sense, however, we are of course concerned with a very different and a distinctively human “way” of anticipating and retaining. For, while the anticipations and retentions are in themselves purely animal, they are no longer functioning on a purely animal level. So if we suppose that Kant is regarding the sort of imagination now in question with particular concern for this function, then it is understandable that he would tend to regard it as itself a “transcendental” capacity. In this sense, Kant’s position would be “functionalistic” with respect to the more primitive capacity. The problem remains how the function is to be explicated. As we shall see more clearly later, Kant himself appeals to a primitive notion of distinctively intellectual consciousness: imaginative anticipations and retentions serve as material, not merely in intuitions, but in the concepts that one “applies” to intuitions, to the extent that an appropriate level of one’s consciousness of that material is embodied in those same intuitions.

The “functionalist” dimension may explain Kant’s apparent persistence in regarding imagination as in all cases, implausibly, a type of intuition.³⁵ In some cases, anticipations and retentions are indeed types of intuition. In these cases, one may actually picture or in some way relive (or fore-live) an anticipated or retained situation. In other cases, it does not seem reasonable to adopt this line. In particular, this does not seem reasonable in the case of those anticipations and retentions that are merely supposed to provide the material for the employment of concepts. In any event, we might at least recognize that the notion of imagining “in intuition” (B151) is ambiguous from the start. In some cases, one may be anticipating and retaining, therefore imagining, “in” intuition, simply in the sense that one’s anticipating and retaining serves as material in intuition. Assuming that Kant’s own interest in anticipation and retention is limited to this function, and to what presupposes this function, it is perhaps no wonder that he tended to regard it as intuitional in its own right.

We might also notice this. While it is implausible to suppose that all the possibilities one imaginatively anticipates or retains need actually be present to one’s intuitional consciousness, it is perhaps at least arguable that it is only insofar as such material serves as material in conceptual acts, hence in intuitions themselves, that one is ever able to call forth the corresponding imaginative possibilities as explicit objects of intuitional consciousness. Once again, it may remain a mystery how animal imagination manages to relate to the future or past in the first place, apart from already constituted concepts to guide it. Perhaps for that reason, Kant declares it to be “a blind but indispensable function of the soul . . . of which we are scarcely ever conscious” (A78/B103), and a function “whose real modes of activity nature is hardly likely ever to allow us to discover” (A141/B181-2). At the same time, though apparently in contradiction to this, it is also understandable that Kant took the function to be one of “intuition.”

If we do not adopt some such view as this, regarding the type of anticipation and retention that Kant has in mind in connection with his doctrine of synthesis, then we would have something much more mysterious than what Kant is content to leave us. To see this, we need simply ask what we are in fact to make of the anticipating and retaining that uncontroversially lies within the capacity of lower life. It is difficult to believe that dogs and cats, in anticipating the arrival of water or food, are anticipating these in imaginative “intuition,” in any sense that involves the occurrence of distinct acts of imagining what is forthcoming. But having conceded that there is in fact a pervasive type of anticipation and retention, as part of our animal heritage, then it is equally difficult to suppose that these functions have nothing essential to do with human conceptual abilities. Obviously, the latter are not—certainly not for Kant—reducible to the former. (For the functionalist in the narrower, causal sense, if they are not so reducible, then they presumably differ at most in degree of complexity.) But it would be plainly incredible if the one did not at least have something essential to do with the other. Presumably, we still are capable of purely animal anticipation and retention. What would be incredible is the suggestion that such a capacity might stand idly alongside our truly conceptual abilities. Nothing would seem more reasonable than that the more primitive capacity should somehow function as material for the more sophisticated. In any event, if we intend to take seriously Kant’s own attempts to apply the notions of matter and form to the faculty of understanding, we need to suppose that, however haltingly, he is aiming at just this view of things.

We need to complicate matters a bit further. I have noted that Kant’s own notion of “reproduction” includes both retention and anticipation, and that the latter is what is really given priority in his thinking. At least this is the case with Kant’s tendency to couple that notion with “association,” and to implicate the latter, in turn, in our capacity for the representation of a reliable order of nature. (At A115, Kant even explicitly distinguishes between mere Reproduktion and association.) But Kant may seem to be confused on these matters. For in the context in which he couples reproduction and association, with the latter’s primary orientation toward the future, he also ascribes something altogether different to the former. It is something apparently much more primitive and much more obviously oriented toward the past. In particular, Kant seems to regard one and the same “reproductive” capacity as responsible both for anticipation, with respect to a general order of nature, and also for the much more primitive ability simply for the holding in mind of what is immediately past: “the first parts of the line, the antecedent parts of the time period, or the units in the order represented” (A 102).

Here as well Kant might of course have noted corresponding future-orientations: for example, the anticipation of immediately upcoming parts of the line in question, or of successive parts of a time period, or of the units in some other sort of progressively unfolding order. All the same, there seems to be a significant difference in the cases. Both might be said to involve anticipating “more of the same,” and to do so on the basis of some kind of retention of the way things have already been. But an obvious difference is this: what Kant himself regards as distinctively associative anticipation and retention has nothing in particular to do with what one is in fact anticipating or retaining as immediately upcoming or receding. It has rather to do with things that one is able to anticipate or retain either as able to occur or not to occur (or as able to have occurred or not to have occurred) under certain sets of merely conditional circumstances. To put it another way, the ability to anticipate “more of the same” that is involved in the case of mere repetition, while no doubt also necessary for one’s ability to apply concepts to objects in experience, seems not to provide an appropriate comparison for the anticipative capacity that Kant suggests is more intrinsically ingredient in the very concepts one applies to objects. The point, of course, is that different things (or things that one will eventually be able, conceptually, to regard as different) continue to repeat their behavior, and to behave “in the same way,” by behaving in all sorts of different ways. We are only able to experience such similarities and differences, as instances of sameness or repetition, once we have already developed appropriate concepts. Inasmuch as Kant’s endeavor is to elucidate the very functioning of the latter, the comparison does not seem appropriate. As for the mere ability to “retain” parts of a sequence, which is what Kant himself calls to our attention, the comparison seems even less appropriate.³⁶

The approach that I am suggesting may have the advantage of explaining why Kant appears to lump together two such different kinds of retention and anticipation: one of them seeming to have to do with what is immediately arising or receding, and thereby to have only the most general bearing on the problem of conceptualization, the other exhibiting the opposite characteristic. We can see the common ground more clearly once we have noticed a special feature of the more immediate kind of anticipation and retention. It is a feature that may not seem at all evident in the case of anticipation and retention of the less immediately and more purely “associative” kind. But the point is to see that the cases need to be regarded as agreeing with respect to just this feature. At least they need to do so, so long as the associative capacity is supposed to have a bearing on one’s ability to conceptualize intuitions. It is obvious that the cases agree in something. They agree in that both involve anticipation and retention, in connection with ongoing and receding intuition. Kant’s attention may have been on something less obvious.

To put it a bit paradoxically, what is noteworthy about the more immediate kind of anticipation and retention is that it could not possibly be regarded as a case of (what we may ordinarily consider) “association” at all. That is, it could not possibly be regarded as a case in which anticipations and retentions manage to get associatively connected to an occurrent intuition. It does not matter how “immediately” one may suppose certain anticipations and retentions to have gotten connected to an intuition. The point is that they will be able to contribute to the temporal structure of an intuition only to the extent that they actually enter into it. If they did not enter the intuition in question, then we could at most say that, while experiencing the latter and very closely connected with it, one also managed to retain something of the immediately receding past, and to anticipate something of the immediately arising future. That would be to say something significant about oneself. But it says nothing about the particular intuition whose structure was presumed in question. It therefore fails to do justice to the fact that an intuition must, in some special way, anticipate and retain in its own right. It fails to do justice to the fact that we do not simply experience special triples of intuitions: one that relates to the immediate past, one that relates to the immediate future, and the two of them joined to a third that, taken on its own, bears on an infinitesimally immediate present. It is an alternative to this sort of approach that Kant appears to be seeking from the very opening of the Deduction.

Consider the account of the synthesis of “apprehension in intuition”:

Every intuition contains in itself a manifold which can be represented as a manifold only in so far as the mind distinguishes the time in the sequence of one impression upon another; for each representation, in so far as it is contained in a single moment, can never be anything but absolute unity. In order that unity of intuition may arise out of this manifold (as for example in the representation of space) it must first be run through, and held together. This act I name the synthesis of apprehension. (A99)

In this first of several, and apparently conflicting, comments on “apprehension,” Kant seems to attribute the latter to intuition as opposed to imagination. He then proceeds to the “synthesis of reproduction,” which he finally attributes to imagination as such. It is in the latter context, as we have seen, that he couples reflection on “associative” anticipation and retention with reflection on the phenomenon that he had apparently already included under the heading of “apprehension.” At most he now adds the hint of an orientation toward the future as well as the past:

But if I were always to drop out of thought the preceding representations (the first parts of the line, the antecedent parts of the time period, or the units in the order represented), and did not reproduce them while advancing to those that follow, a complete representation would never be obtained. (A 102)

After this, Kant concludes that “the synthesis of apprehension is thus inseparably bound up with the synthesis of reproduction.” One might rather wonder what could possibly have distinguished them. I shall return to this point in the next chapter, when I also consider Kant’s reformulation at A120-1. Right now our question is whether the kind of anticipation and retention involved in the two sorts of cases is not meant to exhibit more commonality than what is already superficially evident. I have suggested that more is at stake. Kant’s repetition of the point that he had already covered, in the treatment of apprehension, must hold the key to it.

What joins the cases of immediate and less apparently immediate anticipation and retention is not simply that they both involve an ability to retain and anticipate. What they need to be seen to share, at least in the context of a concern with the conceptualization of intuitions, is what is most prominent in the apparently more immediate of the cases, but which might seem to be altogether absent from the other. If my suggestions so far have been correct, the point of comparison should be clear. The conceptualization of an intuition, with respect to a given object, does not merely relate that intuition to a manifold of anticipations and retentions concerning that object. To the contrary, we would not in the first place have been conceptualizing a particular intuition unless a manifold of anticipations and retentions were, within the structure of that very act, an internal component of the intuition itself.

When that condition is not satisfied, we might still bring something forth that has a regular, perhaps a causal, connection with a genuine conceptual act. For example, we might externally attach, in a law-or a rule-like way, a name or a word or some other symbol (perhaps an “image”) to a particular intuition. The name or word or symbol will no doubt be one that we “associate” with the intuition in question, by virtue of certain anticipations and retentions. To this extent, it will have a real connection with some concept, predicable of the given intuition. But the attachment in question would still not be an instance of conceptualizing that intuition. It could at most be a convenient and mechanical substitute for it. This, we may suppose, is the real point of comparison between the apparently merely superficially similar phenomena Kant considers: the internal structuring that is more obviously required for one of them (that is, for the apparently more immediate kind of anticipation and retention, in regard to what is immediately arising and receding) needs to serve as a model for the structure of the other as well (that is, for the case of apparently merely “associative” anticipation and retention).

V. Synthesis and Rules

We have already noted that many of the notions central to Kant’s argument are ambiguous. The notions of synthesis and of the manifold of intuition are examples. I discuss them further in Chapter Five, in connection with the second-edition Deduction. In the next chapter, I consider some problems in regard to associative imagination, for which I have prepared the way in the preceding section. In the present section, I return to a notion whose ambiguity was touched on earlier. It raises issues that are more prominent in the first-edition Deduction, but also in a chapter that is common to both editions, “The Schematism of the Pure Concepts of the Understanding.” Our question concerns the claim that concepts express “rules,” and perhaps even that they are themselves rules. One of the advantages of the proposed approach is that it can help make sense of some ambiguities surrounding this point as well.

Kant says that “a concept is always, as regards its form, something universal, which serves as a rule” (A106). In fact, understanding may even be defined as the “faculty of rules” (A 126): “Sensibility gives us forms (of intuition), but understanding gives us rules” (A 126). Fascination with analogies between mental and linguistic representation has generated considerable interest in this aspect of Kant’s thought. But it is not easy to see what Kant’s point really is. Kant insists that concepts are essentially “predicates of possible judgments” (A69/B94). We might therefore regard concepts as analogous to linguistic predicates. Accordingly, we might take Kant’s position to be that systems of rules are what confer an original meaning on such predicates as predicates. If so, then the rules presumably prescribe, relative to possible sets of circumstances, the various possible intuitions to which those predicates paradigmatically (though of course not exclusively) attach. The privileged set of intuitions would be the manifold of intuitions that are synthesized by a concept: a concept “can be a rule for intuitions only insofar as it represents in any given appearances the necessary reproduction of their manifold, and thereby the synthetic unity in our consciousness of them. The concept of body, in the perception of something outside us, necessitates the representation of extension, and therewith representations of impenetrability, shape, etc.” (A106). But there is an obvious problem in all this. It contradicts the presumption from which we began, namely, that concepts are, or serve as, rules of some sort. The view that we end up with is, quite to the contrary, that concepts are merely governed by, or in some way essentially “subject” to, rules.

Kant’s position appears to be contradictory. It demands that concepts be predicates of possible judgments, but it also demands that they be the sources of rules that confer meaning on those predicates in the first place. We might try to avoid the difficulty by distinguishing two senses of concept. But that postpones the difficulty. The question remains as to what, outside of a concept qua predicate, could possibly confer a specific predicative status upon a term. To say that what does it is also a “concept,” but this time a concept qua rule for predication, rather than a predicate itself, only names the problem. To say that what does it is a quasilinguistic conceptual “system” does no better. In any event, any source of the rules, outside of the concepts for which they are presumed to be rules, would seem to threaten Kant’s own insistence on the autonomy and spontaneity of understanding as a faculty. As I argued in Chapter Two, what may be the main alternative, namely, appeal to a causal system of some kind, is particularly deficient in this regard.

Kant may be taken to speak to this problem in the eleven-page chapter that, according to Heidegger, “forms the heart of the whole work.³⁷ Heidegger’s enthusiasm is not universally shared. The main points of difficulty seem to be two. First, there is the difficulty that, by raising the question, Kant must be supposing, absurdly, that a concept can be something apart from the rules for its use or application in the first place, the question then being: What in fact constitutes those rules?³⁸ The second difficulty concerns Kant’s attempt to reply to the difficulty. (Kant does not put it, in the case of empirical concepts, in terms of reply to a difficulty; his main concern in the chapter is with a corresponding question concerning the categories.) The reply appeals to a relationship between concepts and their “schemata,” a relationship that is somehow supposed to mediate the former’s application. The objection is simple. Kant’s description of the schemata of empirical concepts (and of any non-categorial ones, hence including pure mathematical concepts) seems to equate schemata with rules for the use of concepts in the first place. Thus the purported solution either restates the problem or, by suggesting that schemata are a “third thing” altogether, generates an infinite regress.³⁹

Pippin sums up the general problem this way:

Some way (usually attributed mysteriously to the imagination) must be found for thinking of a rule in terms of the content of experience. We don’t, that is, simply encounter [apart from conceptualization] lists of Merkmale in experience, calling for or warranting the application of this rule and not that. Neither do we, especially in such empirical cases, merely impose a conceptual unity on a manifold. We have to understand the use of a rule in terms of the content of empirical judgment, and this poses a problem, since the rule is not a representative [conceptual?] entity. It merely stipulates universal conditions for the recognition and connection of sensible intuitions. But—and now the problem—in applying such a rule, we must formulate a “schematic” representation of all these markers as a whole in order to use the rule to distinguish sensible contents. The Merkmale of the concept dog are “imagined” together “universally” (allgemein) in order to use the concept, or to “connect” it with sensible particulars, “images.”⁴⁰

The difficulty, I take it, is this: If the rules for concepts are not themselves conceptions (presumably, conceptions regarding the paradigmatically possible application of concepts to intuitions), or if they are not at least “representative entities” of some kind, then they are unable to provide guidance for the application of the concepts for which they are supposed to provide rules in the first place. But, of course, the rules in question could not already be, without circularity, conceptions, and it is difficult to see what other sorts of representations they could be. To put the problem in still other terms: For a case of real conceptualization, as opposed to a purely causally grounded “response” to intuitions, it must be precisely one’s grasp of the intuitions in question that accounts for the specific modes in which one conceptualizes the latter. But that seems impossible according to Kant’s approach, since it seems to make rules for the use of concepts superfluous; one’s immediate “grasp” should already suffice.⁴¹

Despite his apparent identification of non-categorial concepts with rules of some kind in the Deduction, Kant does appear to regard the “schemata” of these concepts as rules as well. In particular, he appears to regard them as rules for the application of the concepts in question. Of a pure mathematical concept, for example, the schema—“in itself always a product of imagination”—is the “representation of a universal procedure of imagination in providing an image for a concept” (A 140/B179-80). It signifies (bedeutet: Kemp Smith says it is) “a rule of synthesis of imagination” (A141/B180). Regarding an empirical concept, similarly, Kant says both that the schema to which the former stands “in immediate relation” is a rule, and also that it “signifies [bedeutet] a rule” (A141/B180). In either case, schematism is “an art concealed in the depths of the human soul, whose real modes of activity nature is hardly likely ever to allow us to discover, and to have open to our gaze” (A141/B181).⁴²

Kant appears to proceed from here to attribute sensible schematization in general to “pure a priori imagination”:

[T]he schema of sensible concepts, such as of figures in space, is a product and, as it were, a monogram, of pure a priori imagination, through which, and in accordance with which, images themselves first become possible. These images can be connected with the concept only by means of the schema to which they belong [welches sie bezeichnen]. (A141-2/B181)

This may seem to contradict what Kant had said earlier. He had earlier said, at least regarding the schemata of pure sensible concepts, that they “can exist no where but in thought [in Gedanken]” (A141/B180). Now he seems to say that they exist in the imagination. (The later passage is in any case not perfectly clear. Kant’s example involves concepts “such as of figures in space.” Is it meant to apply only to pure sensible concepts, not to sensible concepts generally, including all empirical ones?)⁴³

All of this, according to the suggested approach, makes perfectly good sense. In particular, it is understandable why Kant should identify concepts with rules of some sort, and yet also why he should see the need to distinguish them from rules. And it is understandable why Kant should say that schemata are to be found only in thought, and yet also why he attributes them to the imagination. The apparent contradiction reflects what might be regarded as the paradoxical nature of concepts themselves. In a way, concepts can be nothing other than the rules for their own application. The air of paradox diminishes when we elaborate the notion in the terms that I have suggested. The point would then be this: that what concepts are made out of is nothing other than what serves to indicate, relative to circumstance, the manifold of the intuitions to which those concepts are (paradigmatically) applicable. As a multiply instantiable aspect of possible cognitions, a given concept is paradigmatically applicable to many intuitions. So we need to ask: What specifies the relevant intuitions, relative to particular circumstances? What provides one with the needed “grasp” of those intuitions in the first place? The simple answer is that what does this is our grasp of the rules for a concept. A deeper suggestion, and one that avoids the problems so far encountered, is that this is accomplished precisely by means of the anticipation/retention structure of conception itself.

This is compatible with conceding that it is only as an ingredient in some conceptual structure that a manifold of anticipations and retentions can be dignified with the title of a set of “rules” in the first place. In itself, such a manifold could at most provide the anticipation and retention of possible intuitions. As such, it might therefore succeed in indicating a manifold of intuitions to which some particular concept is paradigmatically applicable. But it cannot indicate those intuitions precisely as intuitions to which some concept is applicable. In this view, the anticipation and retention of a set of intuitions, as those to which a single concept is applicable, is rather constituted by the incorporation of those anticipations and retentions in a higher form of consciousness. What sort of item or entity is the concept thereby in question? In the proposed view, it is not an item or entity at all. Nor is it what may therefore appear to be the most likely alternative: a mere disposition of some kind. It is simply a mode of consciousness itself, or at least of possible consciousness. That is, it is simply an aspect of consciousness, regarded as an aspect potentially instantiable in any number of states of consciousness. The anticipations and retentions in question indicate this manifold (or at least a paradigmatically representative portion of it).

An additional complication is needed to elaborate this view. As we shall see more clearly later, it does not suffice, for the formation of conceptual contents, that just any set of anticipations and retentions get embodied as material in some structure of higher order. In order for such a structure to qualify as conceptual in form, it must include anticipations and retentions of a special sort. It must include anticipations and retentions that are themselves of a higher order. In particular, it must include not only anticipations and retentions of possible intuitions, relative to various circumstances, but also anticipations and retentions of intuitions to which those anticipations and retentions are paradigmatically appropriate.

I return to this complication in Chapter Six. For now, we should at least have a sense in which (non-categorial) concepts can be both nothing other than, and yet also more than, systems of “rules.” They are systems of rules in the sense that they are wholly made out of nothing other than what are in effect rules for their own application. But by the same token, they are also more than rules. They are the very “predicates” on whose application those rules are supposed to bear. What distinguishes a concept from the rules for its “use” is, of course, just the form of consciousness that distinguishes it from its own matter. According to alternative accounts, as we have seen, the situation is very different. There, the systems of rules that are presumed to constitute the significance of predicates qua predicates (or whatever sorts of systems are presumed to do this) are by their nature external to those predicates themselves. Such, for example, are the causes and effects to which functionalists need to appeal. Even if, on the functionalist’s account, the typicality of certain causes and effects is internal to some term’s status as predicative in nature, the very fact of such typicality is not part of the very being of that term in the first place. It is an external fact about it. According to the present account, if predication is to be possible, then a certain set of a subject’s anticipations and retentions needs to be incorporated into the very “term” (that is, into the very aspect of consciousness) on whose predicative status that set is supposed to bear.

On at least three occasions, Kemp Smith blurs this point. That is, he obscures Kant’s suggestion that something within a conceptual state qua conceptual, but not unambiguously identifiable with a concept itself, pre-conceptually represents or indicates the manifold of possible intuitions to which a concept is supposed to be applicable. Kant says, for example, that a sensible concept is necessarily connected, in an a priori and imaginative manner, with the manifold of “images” to which that concept applies “only by means of the schema that indicates” those images (nur immer vermittelst des Schema, welches sie bezeichnet). Kemp Smith instead translates: “only by means of the schema to which they belong” (A142/B181). This substitutes the vaguer notion of images “belonging” to a schema for Kant’s own notion that those images need to be “indicated,” thus in some way present to mind, by means of a concept’s schema. The substitution, of course, avoids any question as to what could in fact provide such “indication,” short of concepts themselves.

Two related passages in the first-edition Deduction are also mishandled. In both of them, Kant calls explicit attention to the anticipational structure of consciousness. In the first of them, Kant is moving from his treatment of the synthesis of imaginative “reproduction” to his treatment of the synthesis of “recognition in a concept”:

If we were not conscious that what we think is the same as what we thought a moment before . . . in its present state a new representation . . . would not in any way belong to the act whereby it was to have been gradually generated [wodurch sie nach und nach hat erzeugt werden sollen]. (A 103)

Kemp Smith translates: the new representation “would not in any way belong to the act whereby it was to be [my emphasis] gradually generated.” Kant’s own formulation seems designed to suggest that the possible “generation” of additional representations, from a given one that was originally subject to some mode of predication, must be regarded as already foreshadowed in the given representation, independently of (though perhaps not antecedently to) the act of predication as such.⁴⁴ A similar alteration occurs at A105. Kemp Smith substitutes “if the intuition cannot be generated in accordance with a rule” for “if the intuition cannot have been [my emphasis] generated in accordance with a rule” (nach einer Regel hat hervorgebracht werden können).

All of this has been meant to apply to non-categorial concepts. But it is worth noting how the approach might extend to fit Kant’s treatment of the relationship between categories and their schemata. I have argued elsewhere for an account of the latter and will not repeat it here. Others have also argued for one or the other of the following propositions: either that the schemata of the categories are pure intuitions or else that they are pure intuitions “determined” with respect to one or more of the pure judgmental or intellectual forms. In any case, the most plausible view, consonant with the approach that I have proposed, is both that the schemata of categories are pure intuitions and that the categories themselves, to the extent they are concepts, not mere forms, can be nothing but those very schemata. That is, they can be nothing but those schemata—considered precisely insofar as the latter have been appropriately informed (“determined”) by the forms in question. Thus categorial concepts, like any others, can only be characterized by a way in which some intuitional “material” is elevated through embodiment in a state whose form is that of the forms of judgment. The difference is in the material. The material of the categories can only be pure intuition as such, that is, intuition as a whole considered with respect to its form qua intuition. By contrast, the material of empirical concepts is not intuition as such, but manifolds of the sorts of anticipations and retentions that are embodiable in their own turn in intuition. In any event, on account of the common structure, and despite the difference in “matter,” we may agree with Heidegger: “All conceptual representation is essentially schematism.”⁴⁵

The point can be clarified by relating it to some views similar to my own, at least regarding the categories and their schemata. Gram, for example, also takes transcendental schemata to be various pure intuitions. On the other hand, he sees those schemata as at most relating to the categories by way of “falling under” them.⁴⁶ In this respect, it might seem that my view ought to more in accord with Allison’s. Allison criticizes Gram for failing to recognize that pure intuitions are themselves conceptually “determinable.” Because of the latter fact, transcendental schemata are to be regarded not simply as intuitions able to “fall under” concepts, but rather as intellectually “determined” intuitions in their own right.⁴⁷

Allison’s approach seems to be like the one that I have proposed. This is because according to that approach, and at least in the categorial case, conceptual “determinations” need to enter into intuitions themselves and not to remain merely externally connectable with them. Thus Allison appears to presuppose a more phenomenological approach, while Gram’s is more purely logical: “Although Kant, of course, begins with the radical separation of sensibility and understanding, intuition and concept, the very heart of his account of knowledge consists in the claim that any cognition of an object involves both elements.”⁴⁸ But it is not clear to what extent my own point is in fact comparable to Allison’s. We can see the difficulty by putting the issue in explicitly noetic terms—that is, by focusing explicitly on intuitions and conceptions as “representational states.”⁴⁹ In these terms, if we are really regarding a category’s schema as both intuitional and conceptual—if we are really regarding it as a pure intuition “determined” by conceptual form—then it is difficult to see how a category’s schema is supposed to be distinguished from a category itself. So “determined,” the schema in question would seem to be a pure mode of representation, both intuitional and conceptual in nature. So it would seem to be identical at least with a “schematized” category.⁵⁰ But Allison explicitly distinguishes the schemata of the categories from the corresponding categories themselves. The latter, “as rules for the transcendental synthesis of the imagination, serve to determine time. . . . [T]he schema is in each case the product of such determination.”⁵¹ In contrast to both these approaches, I take the schemata to be relevantly “determined” by the pure forms of judgment, and the corresponding categorial concepts simply to be the resultant schemata/intuitions, viewed as so determined.⁵²

From the point of view of their status as already formed categories—that is, as concepts—it will still make perfectly good sense to regard the categories as concepts that, in their own turn, are also able to “determine” intuitions. Indeed, to assure that this makes sense is just the point of Kant’s concern with schematism in the first place, as the actual “procedure of understanding with these schemata [das Verfahren des Verstandes mit diesen Schematen]” (A140/B179). His concern is to show how the categories, though in a sense nothing more than the pure forms of judgment, are yet not mere forms for the relating of terms internal to judgments. Rather, they are able to “subsume” objects, to be “homogenous” with concrete representations of the latter, and consequently really to “contain something which is represented” in the latter (A137/B176). My own interest is not in the specifically categorial dimension of the subsumptions in question. It is rather in the Verfahren des Verstandes by which any conceptual content needs to be constituted or formed. But so far as I can see, the relevant Verfahren, in the case of categories, must be a procedure by which categorial concepts, like any others, are constituted out of a certain type of material.

Kant’s own primary interest in the Schematism chapter is, of course, to show how, by virtue of being formed out of pure intuition, a category is formed out of something that is in its own turn embodiable, and necessarily so, in the intuitive representations of sensible objects. According to that account, any act of empirical conceptualization is necessarily an act of categorial determination. For it is always an act in which intellectual form is “applied” to a state whose form is already that of pure intuition. In the case of empirical concepts, by contrast, the corresponding question had already been answered in the Deduction: the fact that their material is simply anticipations and retentions already explains why empirical concepts are embodiable in sensible intuition.

In order to appreciate these points, we simply need to recall throughout that the “formation” of a concept in any particular case is nothing distinct from a particular act of conceptualization. Thus concepts, in the relevant sense, are not items or entities that might, for example, be the mere products of any kind of action, nor any kind of “possession” somehow internal to a subject. If they were, then we would indeed have a problem explaining their connection with the subject’s occurrent grasp, in any act of conceptualization, of the actually operative “rules” for the use of the item on that occasion. (Of course, the question of rules, for obvious reasons, only arises in the case of empirical concepts in the first place.) To be sure, there is nothing in principle wrong in speaking of “concepts” as a kind of possession. As we have seen, perhaps we may even identify them with certain dispositions of a person. But if we do, then we will at most be speaking of dispositions in regard to the very formation of concepts—in the other sense of the term. The issue, it should be clear, is not a merely terminological one.