The claim put forward in this chapter is that Kant’s theory of the process of the constitution of an object of cognition relies heavily on his belief that productive or transcendental imagination plays a central role, but in the end, he does not succeed in giving a convincing account of what this role consists in. In order to substantiate this claim, I will discuss some basic features of his epistemology under the perspective of how imagination can fit in.
A familiar way to characterize the gist of Kant’s epistemological message is to start with his formulation of the supreme principle on which all synthetic judgments are founded (i.e., on which the possibility of the objective validity or the truth of a synthetic judgment is founded), and according to which “Every object stands under the necessary conditions of the synthetic unity of the manifold of intuition in a possible experience” (KrV A158/B197). The claim, obviously, is that objects as objects of cognition (not as objects, e.g., of thought or imagination or as formed matter) depend on the conditions under which they can be experienced, and this leads to another formulation of the same principle which says: “The conditions of the possibility of experience in general are at the same time conditions of the possibility of the objects of experience” (KrV B197). Because Kant defines experience as empirical cognition (KrV B147) and thinks of cognitions in terms of judgments, this statement can be translated into the claim that an object of cognition, i.e., an object about which I can make an empirical judgment, has to be such that it conforms to the conditions of empirical judgment, i.e., it has to be such that it can be addressed by a judgment.
What, then, are the conditions of an empirical judgment? First, one has to notice that the phrase “conditions of an empirical judgment” is shorthand for the longer formulation "conditions of the objective validity of an empirical judgment.” Thus, the question is: What are the conditions of the objective validity of an empirical judgment? Kant’s answer is well known: Whatever else might be involved in the formation of an empirical judgment, there are two conditions that are the most basic: (1) there has to be something or other that is or at least could be “given” through the senses and that conforms to the requirements Kant puts forward for what he calls an “intuition” and (2) there have to be concepts available that can capture what is “given” as an intuition. These two demands for Kant are both eminently plausible and totally uncontroversial because they express what he rightly takes to be implied in our normal understanding of the term “empirical judgment”: Every empirical judgment has to have a non-conceptual content, i.e., a “given” object it is about (otherwise a judgment is just either a mathematical or a conceptual truth or a Kantian principle of the understanding) and a conceptual form, i.e., it must connect concepts in a way that yields a judgment. Given that this understanding of what is required for an empirical judgment is right (and Kant never doubted this), then it makes sense to think with Kant of such a judgment as the result of a process of transformation of non-conceptual content into conceptual form. If one is inclined to identify the realm of intuitions with the domain of “the non-conceptual,” and if one thinks of “the non-conceptual” and “the conceptual” as a complete alternative (vollständige Disjunktion), then the question as to the conditions of the objective validity of empirical judgments seems to require an answer to the different and systematically prior question as to the details of the transformation process of non-conceptual “stuff” into conceptual representations. And indeed, Kant himself points in the direction of this question when he declares that the problem one has to solve in any attempt to analyze the conditions of an empirical judgment consists in explaining “how subjective conditions of thinking [i.e., concepts] should have objective validity, i.e., yield conditions of the possibility of all cognition of objects [i.e., non-conceptually “given” intuitions]” (KrV A89 f./B122). Now, it seems that Kant himself can and does establish a link between the non-conceptual and the conceptual in a very prominent passage without relying on imagination at all (but only on sensibility and understanding). The passage I have in mind is what Kant, in the second edition of the Critique, refers to by the name of a metaphysical deduction.1 Here in the metaphysical deduction Kant can be seen as offering in two famous sentences a very straightforward answer as to how it comes about that non-conceptual content can get conceptual form by invoking just two faculties, i.e., sensibility and understanding. Kant can be understood to say that what brings about the transformation of what is rooted in non-conceptual sensibility into the conceptual form of a judgment is the understanding alone.
However, even if one agrees with the view that the metaphysical deduction successfully establishes a link between the conceptual and the non-conceptual elements involved in the formation of an empirical judgment by insisting on the necessary contribution of (conceptual) rules of the understanding in the act of object constitution about which (empirical) judgments can be rendered, one cannot avoid the impression that this success is based on our willingness to accept a premise that is argued for by Kant neither in the metaphysical deduction itself nor elsewhere in the “clue”-chapter. This premise concerns the organization of the matter that is unified by the activity of the understanding into representations of objects, i.e., it concerns the manner in which, according to Kant, sensible representations have to be present to the mind in order to count as representations out of which the understanding can form representations of objects. Quite obviously the metaphysical deduction implicitly presupposes that the sensible representations the understanding operates synthetically on are such that they can be unified into object representations. This again presupposes that a manifold of sensible representations can be unified into intuitions because the sensible representations the understanding is working with are intuitions. But why and how is this presupposition justified? Why is it that we have to think of sensible representations as unifiable into intuitions, and how does it come that – if they are unifiable this way – they can lead to representations of objects? These questions have to be taken seriously, not so much because of worries concerning the soundness of the argument of the metaphysical deduction, but for a substantial reason. Kant has to avoid the impression that for him objects of cognition are based on random collections of sensible representations the understanding happens to single out as intuitions arbitrarily in the endeavor to unify them according to its unifying rules. This is so because otherwise he would have to give up on his very reasonable claim that it is not enough for an empirical object of which we can have knowledge to be grounded in whatever sensible representations happen to be around, but that there has to be an inherent affinity between those sensible representations that end up as intuitions out of which an object of cognition can be formed. In other words, there must be more to sensible representations than just the fact of their occurrence in the mind if they are to qualify for representations out of which “given” intuitions can arise that can be united into the representation of an object. This means that Kant has to concentrate more closely on the details of what is happening on the sensibility side of object formation up to the point where intuitions start to play a role. And this is exactly what Kant is doing by invoking imagination in both versions of the “deduction” chapter (though the first version is much more detailed with respect to imagination than the second and follows a different strategy).
However, when it comes to the role imagination is supposed to play in his account, a lot remains obscure. The first thing to worry about is why and how Kant wants to draw a distinction between imagination and the understanding at all. The background to this worry is the following: Kant obviously thinks of imagination as a faculty whose synthesizing activity in the process of the constitution of an object of cognition is subject to and guided by rules that in the end reflect the demands which flow from the unity-providing faculty of apperception, and which are founded in the necessary unity of apperception, i.e., the unity of (self-) consciousness.2 These rules are the categories and the synthesizing activity is directed toward items that are given by sensibility. At the same time, Kant thinks of the understanding as a faculty that by exercising a synthetic activity on what is given by sensibility has to employ the very same categorical rules as imagination does in transforming intuitions into units of reference for concepts, i.e., into objects of cognition. If imagination and understanding perform the same tasks under the same regulatory constraints on the same material provided by sensibility, why and how are these faculties to be distinguished?
Does this mean that in the end one is better off to follow Kant’s own practice in the years after publication of the first edition of the Critique and to downplay the role imagination plays in the process of object cognition? Although such a move is quite tempting and presumably will not find much resistance from many current Kant scholars, it might well go with a price Kant himself, for very good reasons, has not been prepared to pay. This price consists in the inability on Kant’s part to explain how it comes that already on the receptive level of sensibility material is available that is fit for the synthesizing activities of the understanding if there were no different synthesizing faculty, i.e., his so-called “imagination,” involved. If this is right, then there must be a difference between the performance attributed to imagination in producing intuitions on the one hand, and representations of objects on the other. In order to see why this is so, and to show that Kant has been very well aware of this problem lurking on the side of the organization of sensible material, one has to go back and look again at his analysis of how the representation of an object of cognition comes about.
The view attributed here to Kant concerning the constitution of an object of cognition is that of a process which leads from sense impressions or sensations to the representation of a full-blown cognitive object, i.e., an object about which objectively valid judgments can be made. This process contains two distinguishable stages. The first goes from sense impressions to representations of individual items (undetermined objects), i.e., intuitions; the second from intuitions to conceptual representations of (determinate) objects. Both classes of representations, intuitions and concepts of objects, are constituted in that they are products of some activities of the representing subject. From what was discussed earlier, it is quite likely that the second stage, from intuitions to concepts, is not really of interest for someone who attempts to secure imagination as a self-standing and independent status within Kant’s conception of object formation. This is so because what is going on at this stage is definitely dominated by the understanding in that the understanding provides the rules (rooted in the transcendental unity of apperception), without which the necessary unity of an object of cognition would not be possible. At this stage, imagination has the thankless task to do whatever is necessary to support the understanding in the endeavor to “bring a transcendental content,” as Kant calls it in the metaphysical deduction (KrV A79/B104), into a manifold of intuitions, thereby giving, so to say, the spatio-temporal stability and determinateness necessary for an object that can be referred to by a concept to an otherwise rather instable conglomerate of perceptions in an intuition. The understanding achieves this determination by somehow forcing imagination, which is used to deal with material given in sensibility under space-time conditions, to act synthetically in an object-constituting way, i.e., to perform a “transcendental action” (KrV A102, B154) or to fulfill its “transcendental function” (KrV A123).
This second stage of the process, from intuition to representation of an object Kant is picturing, is based, in any case, on two assumptions: (1) The synthetic activity of the understanding is restricted to object constitution; the understanding cannot do anything else other than bring together representations (whether they are intuitions or concepts) into representations of objects, i.e., into representations that are determined by the (categorical) rules necessary for thinking a given manifold as objectively unified or for taking a suitable manifold as being united in the representation of an object. (2) The synthetic activity of imagination is not restricted to object constitution, i.e., imagination can bring together sensible representations in ways that result in complex representations of as yet undetermined objects which do not qualify as representations of cognitive objects. These two assumptions, together with the explicit goal of the second stage to provide the details of the process of cognitive object constitution, lead to the quite sensible assessment that at the second stage imagination is of no interest for its own sake but has to be addressed only in its function as the servant of the understanding.3 And this might have been a reason for Kant to downplay the role of imagination in contexts where he is primarily interested in pointing out the achievements of the understanding.
However, although assumption (2) already shows that within the two-stage scenario, imagination at the second stage might have a lot more to do than just synthesizing items under the spell of the understanding, it is primarily the first stage, from sense impressions to intuitions, which is in need of imagination, as exercising an activity that is different from that of the understanding and not restricted to synthesizing alone. Why is that so? The short answer to this question is: Because there is a difference between the emergence of a representation that has the status of an intuition and the formation of a representation of a cognitive object, with the result that not every intuition has to be a representation of an object. Before this assertion can be substantiated, one has to look in more detail at what is going on at the first stage of object constitution, i.e., on the way from sense impressions or sensations to intuitions.
According to Kant, sensations are the ultimate building blocks of representations of objects, i.e., for Kant the phenomenology of object constitution has to start with sensations (KrV A19 f./B34). Sensations are taken by him to be affections of what he calls “sensibility” (Sinnlichkeit), and because he knows of two kinds of sensibility, i.e., outer and inner sense, he distinguishes between affections of the outer and the inner sense. Although it is not quite clear how to interpret this distinction,4 it is at any rate clear that he wants sensations to be affections of the organs of the (five) senses, and thus physiological events. Sensations as physiological events are, as Kant rightly points out, “for us nothing” (KrV A120) if they are not conscious states of a subject or “connected with consciousness” (KrV A120). As conscious states, sensations have a content, i.e., they represent qualities like colors, shapes, sounds, etc., and occur sequentially in time. These content-filled representations Kant calls “perceptions” (KrV, B147). They are the material out of which we form intuitions, and imagination (and not the understanding) is supposed to be active in this forming process.5 Given that there are indeed these three elements, i.e., sensation, perception, and intuition involved in this process, and that these elements form a constitutive sequence such that sensations make possible perceptions and perceptions ground intuitions, the process seems to be divided into two phases, the first leading from sensations to perceptions, the second from perceptions to intuitions. The question as to the role of imagination in this process, then, is: In which of these phases and in what manner is it involved? Is imagination operative in both phases, and, if so, in the same or in a different manner? Or is it at work in only one of these phases? The following remarks intend to tentatively reconstruct aspects of what Kant takes to be the phenomenology of the constitution of intuitions (not of cognitive objects!) guided by the ambition to uncover a couple of features unique to imagination.
One could be tempted to expect already, during the first phase, i.e., at the very beginning of the way from sensation to perception, a genuine function for imagination as a subjective synthesizing faculty. This is so because there has to be some explanation as to how the process goes on that transforms the purely physiological event of a sensation into a content-full conscious mental representation, i.e., a perception. This process has to involve an activity of the representing subject if one does not want to maintain either of two (within the Kantian framework) rather implausible hypotheses concerning the essential characteristics of sensations: The first would be that there is no need for a process because bare sensations already have representational content, the other that bare sensations are causally efficacious in that they are the causes of perceptions. While the first hypothesis would be hard to reconcile with Kant’s convictions as to the non-representational character of sensations, the second would violate his view as to the passivity of sensibility. If the transformation process cannot be explained by relying on peculiarities of bare sensations (like having representational content or being causally efficacious) then, it seems, what remains is to have recourse to some mental activity of the subject. So, why not think of imagination doing this job of transforming sensations into perceptions? Imagination within the Kantian framework is, after all, intimately connected with non-conceptual activities at the level of sensibility and is even said to be “a necessary ingredient of perception itself” (KrV A120fn.).
However, it might be objected that this suggestion does not look very promising as long as one tends to think of imagination as a faculty whose exclusive task consists in synthesizing individual items into more complex wholes. If imagination could indeed do nothing else other than synthesizing, then surely enough it would be a very implausible candidate for playing a role in the transformation process from sensation to perception. This is so because what is needed here is not a faculty whose activity consists in synthesizing but rather a kind of discerning faculty. What is required here is something like an interpreting faculty that can somehow manage to individuate sensations which are stipulated as featureless items by giving them discriminable representational content, thereby elevating sensations to perceptions. Such a faculty would provide qualitatively distinguishable individual perceptions which, according to Kant, “by themselves are encountered dispersed and separate in the mind” (KrV A120). This seems to imply that such a faculty has not much to do with a connecting activity. And if the exclusive function of imagination is to provide connections then, it seems, it is not fit for this transformational task from sensation to perception because there is no connecting, no synthetic activity involved.6
There are two ways to counter this objection. The first is to doubt that it really is the case that for Kant imagination is restricted to its connecting function, i.e., that it can do nothing other than putting given elements, perceptual or otherwise, together. The second is not to doubt that imagination is restricted to acts of connection but at the same time to claim that Kant conceives of synthesis of imagination not as a simple and uniform, but as a complex activity.
I take this second way to be unconvincing because of what Kant explicitly says about synthesis (e.g., KrV A77 ff./B102 ff., A97, A101, A116, B130). However, there is no textual obstacle to proceed on the assumption – underlying the first way suggested earlier to overcome the objection – that the activities of imagination are not restricted to the performance of acts of synthesizing, and hence that imagination can be involved in non-synthetic activities. If imagination can play a role in the transformation of sensations into perceptions, then it does have an autonomous occupation that distinguishes it from the understanding in the process of creating representations of objects based on intuitions because this occupation is not meant to be a synthetic one. It might also have a genuine function within Kant’s two-stage model of object formation in the domain in which we form intuitions out of perceptions, i.e., in the second phase of the first stage. In terms of the two-stage model imputed here to Kant, this means: If imagination is to have a genuine function and if it cannot figure as a self-standing activity at the second stage (because at this stage it would exercise a synthetic activity under the spell of the understanding), then it has to get assigned a genuine role in either one or both phases of the first stage, i.e., in the transition from sensations to intuitions. Because this transition, or at least a part of it, is the main achievement of what Kant calls “apprehension,” it is especially with respect to its apprehending function that imagination can be expected to lead a life of its own, independent of any direct interference from rules of the understanding.
What, then, is happening on the level of apprehension? The apprehending mind, we are told, “is to bring the manifold of intuition into an image” (KrV A120). In order to do this, it must “antecedently take up the impressions into its activity, i.e., apprehend them” (KrV A120), or it has – to quote another formulation – “first to run through and then to take together” (KrV A99) a manifold of perceptions. However, this process of taking up impressions (which I take to be perceptions), though quite understandably conceived by Kant as taking place antecedently to the acts of running through them and taking them together, cannot be the beginning of the whole process from sensation to an image because there still is the step from sensation to perception which is unaccounted for if apprehending starts with perception. How is one to conceive of this step (the first phase of the first stage previously mentioned) or, in Kant’s words, how do sensations become “modifications of the mind in intuition” (KrV A97)? Maybe it helps to discuss this question by looking a bit more closely at a concrete example. Assume that I, by chance, fall into a swimming pool filled with cold and rather dirty water. What is my situation as to the sensations with which I am confronted? I guess I will have millions of sensations, i.e., physiological occurrences that result from affections of all my five senses. These are, as purely physiological events, no modifications of the mind, though they are definitely changes in my bodily state. In order to make them modifications of the mind I have to transform (some of) them into individual episodes of which I am conscious, i.e., into perceptions. Thus, immediately after falling into the pool I cannot but become aware, in a very rapid succession in time, of a lot of things, each of which is a disparate item. I have somehow to notice that there is that feeling of coldness on my skin, that I have the optical impression of darkness, that I have the taste sensation of brackish water, that there is a muffled sound around, etc. How am I to account for this transformation from an all-encompassing non-individuated and unstructured physiological event into single episodes which I can distinguish from one another, thereby making them items which can be “encountered dispersed and separate in the mind” (KrV A120)? Obviously, it is more than likely that one will introduce an activity that has to accomplish this transition. An activity in Kant’s map of the mental is based in a faculty. Because this activity at work in the transition is supposed to be part of the cognitive process, it must be related to one of the cognitive faculties. These are sense, imagination, and apperception (KrV A94, A115). Sense by definition is not the right candidate for a transitional activity because it can only passively receive data; it is the faculty of receptivity. Apperception is no good candidate either because just by itself it only provides the form of unity to whatever is such that it can be connected. It might be objected that already the very concept of an individual item that can be distinguished from other items presupposes the possibility to view this item as a unit, thus making it dependent on the apperceptive activity in its categorizing function. Though this is right, it does not make apperception the material source of perceptions; it just accounts for their singularity.7 Hence the only candidate left for the transforming activity within the Kantian taxonomy of cognitive faculties is imagination.
Admittedly, there is not that much in Kant’s texts by virtue of which one can claim that a reasoning along the lines pointed out here is part of his considered view as to the transition from physiological states to perceptions of individual sensations that can count as a manifold out of which intuitions are formed. However, granted that one were allowed to attribute to him such a reasoning, he would be in the position – to use a rather daunting English metaphor – to kill two birds with one stone. On the one hand, he would have gained the means to address a problem and to point to its solution, which inevitably arises for every causal theory of perception that starts with physiological episodes in the shape of sense impressions as the ultimate building blocks of perception, i.e., the problem of individuating these impressions. On the other hand, he would be able to give us an argument as to why he wants to think of imagination as a self-standing and irreducible faculty without which cognition would not be possible. A stance on imagination, as suggested by the previous reasoning, would imply that the task of imagination – at least in its productive function8 – does not solely consist in synthesizing items into wholes. It would hold the activity of imagination responsible for providing a representational material which first of all makes synthesis possible by creating individual and discernible units (perceptions).
Although a reasoning like the one sketched a moment ago – irrespective of whether Kant really would subscribe to it or not – might be a way to secure the faculty of imagination an albeit rather fragile autonomy against the faculties of sense and apperception, it could be objected that it is by no means sufficient to demonstrate the independence of imagination from the understanding. It could well be that imagination does the job of transforming physiological states into individual perceptions, i.e., content-full unities of which I am conscious in a way that somehow involves the rules of the understanding. After all, perceptions are meant to have characteristics that make them suitable for playing the role of a manifold out of which intuitions can be formed, which in turn can function as the material on which the understanding performs its synthetic activity which leads to the representation of an object. Thus, the very fact that they are necessary elements in the constitution of object representations seems to submit them to the very same rules that are operative on the level of intuitions. If my imaginative processing of sensations would lead to conscious episodes that in no way could fit together, e.g., if I had sometimes an acoustic perception of colors or a tactile perception of smells, whereas at other times my acoustic perceptions would contain feelings of pressure and my optical perceptions would shift between having as their content noises and tastes, then these perceptions could not become elements of an intuition, i.e., a manifold that can be brought together in an intuition. Hence, imagination must somehow transform physiological events “in the right way.” And, is it not the case that “the right way” has to be defined in terms of object-constituting synthetic rules, which would mean that in the end imagination becomes again dependent on the understanding?
However, this objection is not really convincing because it appears to be self-defeating. Already the notion of “the right way” obviously makes sense only in contrast to other ways. And if there are other ways in which imagination can perform the transformational task at hand, then this seems to imply that just the opposite from being dependent is the case: The very fact that imagination could do differently, i.e., could act in such a way that there would be no chance to get the resulting perceptions into an intuition shows that it is a self-standing and independent faculty that (presumably within certain limits) is free to do whatever it wants. It is not that difficult to come up with examples that make it likely that imagination indeed acts in ways that do not aim at providing materials for the formation of intuitions, i.e., that do not result in providing items which then can function as elements out of which intuitions can be formed. Think of sexual episodes: Obviously, in these episodes there are present many or “a manifold” of sensations “accompanied by consciousness,” i.e., perceptions which, however, are so diverse that they cannot be unified into an intuition which could become the basis for the representation of an object. Viewed from this perspective, that there are options of acting available to imagination, the fact that imagination can and does, but by no means has to create perceptions that meet the standards of the understanding, indicates that it is not constrained by the rules of the understanding. Rather, it seems, it is up to imagination to “decide” whether it conforms to the demands connected with what the understanding needs in order to form representations of objects. The only rule it has to obey, as Kant points out quite explicitly (KrV A99), is the rule of sensibility, i.e., that it has to let perceptions form a sequence in time.9
If one agrees that it is acceptable to attribute to Kant the idea that imagination has some sort of “freedom” from the rules of the understanding, at least at the initial phase of the first stage of gathering elements for creating a representation of an object, a quite natural thing to ask is whether this “freedom” extends to the second phase of this stage as well, i.e., to the phase from perception to intuition. Remember that both phases of this stage are conceived of as involving a mental activity (namely imagination) in its capacity to operate exclusively on the level of sensibility, which is meant to imply that this activity is not assumed to be subject to conceptual rules but is supposed to operate in a non-conceptual way. Here, in the second phase, the task no longer consists in arriving at perceptions but to create out of perceptions intuitions. Kantian intuitions have to be understood as unified collections of perceptual data of which I am conscious. Therefore, the task imagination has to perform now in this second phase consists in actively going through (durchlaufen) the perceptions at hand in order to select those that qualify for bringing about an intuition of an as-yet-undetermined object.
How does this selection process work? According to the reconstructive picture suggested here, the following is the case: Imagination finds itself confronted with a large number of perceptions, only some of which are such that they can be connected into the unity of an intuition. Thus, to turn again to phenomenology, in any given situation I will have some sound perceptions, some color and shape perceptions, and different smell and taste perceptions, all of them ordered in time without having any reason, to use Kant’s words, “to summon to the subsequent [perceptions] a perception from which the mind did move on to another [eine Wahrnehmung, von welcher das Gemüt zu einer anderen übergegangen, zu den nachfolgenden herüberzurufen]” (KrV A121). In order to bring some object-representation-enabling structure into this ordered sequence of inhomogeneous perceptions, imagination has to pick out those that are of the right kind, i.e., those that fit together into the unity of an intuition. If, e.g., I were conscious of a colored sound perception accompanied by a tactile smell perception and followed by an optical perception of a taste, there would be no way for imagination to make them fit together into an intuition; and even if my perceptions were not that strangely “non-objectifiable” as these synesthetic perceptions are but would comply with the “normal” sense distribution (colors are perceived by the optical sense, tastes by the gustatory sense, etc.), it could still very well be the case that they cannot mesh with one another into a single intuition. Just think of the perception of a color in front of you and the almost simultaneous perception of a sound in the far distance. Actually, the very possibility of becoming conscious of different sensations as individual events depends on the inherent independence of perceptions from intuitions. A perception is in its own right more than just an ingredient of an intuition, or, in other words, I can have a perception without thereby automatically having a (maybe incomplete) intuition. Hence, imagination has to seek out the right kind of perceptions, i.e., those that can be used in the process of constituting a unitary intuition.
Imagination acting in this selective capacity is guided by just a single criterion: It is guided by whether a perception can be integrated into a unity of an intuition, i.e., a unity that is compatible with what Kant calls (in section 18 of the B-Deduction) the “objective unity of apperception.” The phenomenologically sound assumption behind this move on Kant’s part is the idea that only those collections of perceptions can have the unity of an intuition that do not interfere with my being able to think of myself as an identical subject. If any collection of perceptions would in principle qualify for the status of a unitary intuition then, as Kant rightly points out, “I would have as multicolored, diverse a self as I have representations of which I am conscious” (KrV B134). This shows that the demand for unity in connection with an intuition is not immediately related to the demand for the unity of an object based on the category “unity” in the table of categories. That an intuition has to have a specific unity is due directly to the demands connected with the possibility of a unified and identical self. Therefore, the unity of an intuition is not owed to an object-constituting concept and is no conceptual ingredient provided by the understanding.10
If imagination brings about intuitions by starting with perceptions and uniting (some of) them in the manner just sketched, can it be called “free” or “independent” or “autonomous” in the same way as it was in the process of transforming physiological events into perceptions? Obviously not. Whereas in the latter case imagination can be said to be “independent” or “autonomous” because there appears to be no interference at all from either apperception or the rules of the understanding in the transformational process, and it can be termed “free” because it is free to follow either no rule at all or rules that are exclusively its own, imagination in the former case is definitely restricted in its activities by the demands, though not of the understanding but of apperception concerning unity. Though being subject to the conditions under which the unity of the self, i.e., apperception is possible is without doubt a constraint on imagination’s “freedom,” it is a much less rigorous constraint than the restraints that come with being subject to the requirements of the understanding. If one were to call the freedom imagination enjoys in the context of the production of perceptions “absolute” freedom, one can think of imagination’s freedom in the process of forming intuitions as a “relative” freedom. But even in this state of relative freedom, imagination would be completely independent of the operations of the understanding, and in this sense, autonomous.
That there is a difference between being dependent on apperception alone and being dependent on the rules of the understanding (perhaps together with apperception) shows not just in the different degrees of freedom connected with these dependencies. It also shows in the results of the activities of imagination. Whereas, in the process of uniting a manifold of perceptions into an intuition, all that imagination has to achieve is intuitive unity; in contrast, it has to accomplish objective unity or the unity of an object when acting in the service of the understanding. Kant very sensibly makes it quite clear – at least in the A-deduction – that there is a fundamental difference between an intuition and the representation of a cognitive object (KrV A124). An intuition being just a unified collection of perceptions which somehow fit together (can be brought under an apperceptive unity) has to be such that the understanding might be able to use it as material for creating a general representation (i.e., a concept) of an object. Or, to put it a little bit metaphorically, an intuition has to contain the promise of an object. In order to become the representation of an object an intuition has to be treatable or, so to say, manageable by the understanding, that is to say it has to be accessible to categorization. This accessibility condition is not fulfilled by a single intuition but presupposes what Kant somewhat vaguely calls “a manifold of intuition.”11 Otherwise one would never get to a general representation of an object (of cognition). In other words: Kant, by insisting on the difference between an intuition and the representation of a (cognitive) object, does account for an attitude which indeed has a solid basis in the phenomenology of perception, i.e., the attitude that many, if not most of the intuitions we have, do not end up in representations of objects. It is just those intuitions that can give rise to reproductive and recognitive activities that make a representation of an object possible. These activities, though they involve imagination “in its transcendental function” as well, are subject to the categorical rules of the understanding and start their work on the next, the second stage of the long way from sensation to the representation of an object whose concept can function as a predicate in a judgment. Here in the second stage, which begins with intuitions, conceptual elements in the shape of the categories have their debut, and imagination loses both freedom and autonomy. Fortunately, Kant does not abandon imagination to this sad fate. According to him, imagination happily regains far away from cognitive processes a new and authentic life in aesthetic contemplation. It took Kant a couple of years before he found the right means for this resurrection of imagination in the shape of a theory of taste.
Let me highlight at the end what I take to be the most interesting positive points that result from ascribing to Kant the model of cognitive object formation outlined so far. These are mainly two, the first one relates to an aspect essential for every systematic account of object formation in cognition, the second relates to Kant’s special presentation of such an account. The systematically interesting point is the following: Every theory of cognitive object formation, as far as it bases object formation on perception and thinks of perceptions as having semantic content, has to somehow answer the question as to how this semantic content comes about. There are many answers around, both from the philosophy and the psychology of perception, which, rather vaguely, connect the origin of semantic content to brain activities and neuronal processes and try to establish in this way a direct link between physiological events (brain activities, neuronal processes) and semantic content. What these theories fail to provide is an account of the conditions that have to obtain for an item to be able to have a semantic content. At the least, one has to expect that such an item can be distinguished from other items and that it has a specific singularity that is such that there is a one-to-one correspondence between the item and its semantic content. Thus, the systematically interesting lesson to be learned from Kant’s approach is that a theory of perception based on sensory input first of all has to account for the distinctiveness and singularity of what psychologists nowadays call “percepts” before it can be of any help for a theory of cognitive object formation.
This leads directly to the second point that can be seen as a positive result of attributing to Kant the model of object formation outlined here. In the beginning, it was said that there always have been and still are serious worries about whether and how the faculty of imagination can be acknowledged as a constitutive and self-standing factor in Kant’s attempt to somehow bring together conceptual and non-conceptual elements in his conception of an object of cognition, as outlined in the first Critique. This, however, seems to be an ill-founded worry if one is prepared to follow the suggestions put forward here. Rather, it shows that imagination has in a certain sense to carry the main burden in the laborious process to make, out of amorphous and unstructured physiological sense-impressions, representations of full-blown cognitive objects. In this process, imagination in the initial sensory stages plays a surprisingly autonomous role and reveals an admirable and almost unrestrained range of creative activities whose rules (if there are any) are unknown to us. It is only in the later, conceptual stages of this process that imagination has to succumb to foreign demands and to follow the categorical rules of the understanding. All this confirms quite nicely Kant’s assessment of imagination according to which it is “a blind, though indispensable function of the soul without which we would have nowhere any cognition at all” (KrV A78/B103). So, ultimately Kant was right to insist on the essential role imagination has to play, even in purely epistemological contexts, and why this is so can be explained by going back to his view on cognitive object formation elaborated here. It cannot be denied that he himself, in the course of time, became somewhat reluctant to highlight this role as clearly as he did in the first edition of the Critique. Yet, this reluctance seems to be rooted not so much in doubts about imagination’s function as in the complexities connected with imagination’s activities. It surely does not indicate that Kant changed his general view in a fundamental way.12
1 There is a lot of guesswork over more than 200 years as to why the metaphysical deduction is supposed to be a deduction, what makes it metaphysical, and even where exactly it is located. Kant is of no help because he nowhere cares to give even a hint to an explanation of what he means with this term. Though there has never been a general agreement as to what the metaphysical deduction is about, it is safe to say that nobody would object to Guyer’s formulation according to which “the so-called metaphysical deduction is meant to establish that the categories are the conditions of the possibility of cognition of objects” (Guyer 2001:318).
2 Cf. e.g., KrV A124.
3 This does not mean, at all, that this second stage has no relevance for the success of Kant’s epistemological enterprise. On the contrary, questions concerning the details of his account as to how we are supposed to transform intuitions into conceptual representations of objects have been raised ever since the publication of the first Critique.
4 This is due to the obscurity connected with the conception of self-affection discussed by Kant himself extensively in sections 24 and 25 of the B-Deduction.
5 In both versions of the deduction, it is not quite clear how exactly Kant wants us to understand the relation between sensations, perceptions, and intuitions. Sometimes he writes as if an intuition is just a collection of different perceptions, which in turn are conscious sensations (see earlier references), and sometimes one has the impression that he wants perceptions to be the intentional correlate only of intuitions, which would mean that intuitions are the constitutive basis of perceptions or that they make perceptions possible (KrV B160, B164). This confusing ambiguity extends into later parts of the first Critique as well. Cf. the Anticipations of Perception, (KrV B207 ff.).
6 Unfortunately, as far as I can see, Kant nowhere tells us how he would like to explain the transformation of sensations into perceptions. Maybe he would give an interpretation that relies on the conditions an item has to fulfill in order to become included (aufgenommen) into the unity of consciousness. One could read his remarks on KrV A108 ff. and A121 fn. as pointing in this direction.
7 This means that apperception in its so-called original (ursprünglich) state is not confined to, in the end, conceptual, i.e., categorizing operations, but has initially only the much more basic function to provide numerical identity or “unity in the time-relation of all perceptions” (this term is used here in its first edition meaning as referring to conscious impressions!). This is nicely confirmed in a passage from the beginning of the Analogies of Experience in the first Critique which states: “In the original apperception all this manifold, according to its time relations, is to be unified; because this says its [original apperception] transcendental unity a priori under which everything stands what is to belong to my (i.e., my unified) cognition and thus can become an object for me” (KrV A177/B221).
8 Whether Kant thinks of reproductive imagination as a capacity that can do anything other than synthesizing is not quite clear. On the one hand he takes it to be a rule-guided synthetic activity (e.g., KrV A100), and on the other he wants it to be active in providing representations of objects without their presence (KrV A100).
9 To think of (productive) imagination as having some degree of “freedom” while performing in what was called earlier “the first phase of the first stage,” the act of transforming physiological events into perceptions helps to connect Kant’s considerations regarding imagination in the first Critique to what he says about imagination in the third Critique.
10 The distinction between unity as an object-constituting concept (a category) and unity as an achievement and characteristic of apperception is pointed out nicely by Kant in KrV (B131).
11 There is a well-known ambiguity in Kant’s use of the term “manifold of intuition”: On the one hand, this term refers to the manifold of perceptions out of which a single intuition is formed. On the other hand, he seems to refer, with this expression as well, to the manifold aspects a single object can have when given “in intuition.” In this second use the term “intuition” designates a mode of awareness of an individual object, in the first, a distinct entity.
12 This chapter is a very condensed version of a much longer piece that is scheduled to appear in 2018 under the title Kant’s Power of Imagination, with Cambridge University Press. The version printed here owes many thanks to Dina Emundts, Eckart Förster, Gary Hatfield, and Sally Sedgwick for helpful comments.