IN THIS chapter I try to develop a general framework that will be fruitful for a study of Kant, but also independently appealing in a study of the mind. In the first of these respects, we must take account of several distinctions that Kant himself draws. But the language in which he draws them is ambiguous. The ambiguity reflects, at least in part, Kant’s desire to emphasize a correlation between those aspects of subjective activity through which things are apprehended and those things themselves, and their various aspects, qua apprehended (or apprehensible).
This sort of correlation is most likely to be seen by those concerned with the role of “conceptual schemes” in the constitution or determination of objects of cognition. But there is a perfectly defensible, and distinctively Kantian, approach to the problem of “constitution” that goes beyond a conceptual idealism of this sort. To understand it we need to clarify the distinction between matter and form in intuition. Some readings of that distinction are too subjectivistic. This is so, for example, when the immediate objects of intuition are taken to be, or to be “formed” out of, items such as sensations, sense impressions, or “sense data.”1 Other readings are insufficiently subjectivistic. Such, for example, are those for which the relevant distinction is simply between the set of one’s most basic conceptual formations and some appropriately binding or limiting set of non-conceptual limitations upon such formations.2 Our need is to formulate a sense in which intuitional form is in its own right an irreducible mode of subjective directedness, without regressing to the first type of reading. As we will see, the corresponding “matter” can then only be a kind of subjective material that is not merely somehow binding or limiting, with respect to conceptual acts, but much more internally material through which those acts are able to function, and much more literally ingredient in them than is generally acknowledged.
We need first to relate the notion of intuition to some more general notions in Kant. It is supposed to be a species of Vorstellung (A320/B376-7)—as I shall call it, “representation.” But Kant uses both terms in a number of ways. Sometimes, the objects of ordinary perception are said to be representations (A30/B45, A114, B147, A492/B520) or intuitions (A 163/ B204, B207, A370). And sometimes Kant uses the latter term for a particular aspect, the purely spatiotemporal aspect, of the objects of ordinary perception (B202, B207). In addition, he speaks of space and time as intuitions (A25/B39, A27/B43, B207). But this may all be regarded as derivative from a usage that more directly concerns cognitive states themselves.
In the primary sense of the term, we may say that an ordinary sense perception,3 not its object or any part or aspect of its object, is an instance of intuition or of intuitional representation. With qualification to be added presently, another example will be any instance of ordinary imagination. These are paradigmatic examples of “representation” for Kant, because (with qualification to be added here too) they are ways of having one’s awareness directed toward possible “objects.” As such, they represent (or in some way “present”) those objects in a state of consciousness. However, other sorts of Vorstellungen are not, even with qualification, concrete ways of being directed toward possible objects. Sensations and concepts are examples of this. Obviously, sensations and concepts are not intuitions for Kant. But they are nonetheless representations. The primary reason for this is that, while not themselves ways of being directed toward objects, they are nonetheless able to contribute, in an internal way, to ways of being directed toward objects.
A question remains concerning the notion of a purely “internal” contribution to representation. With regard to concepts, that is the main concern of this study. In any event, in addition to eventually clarifying a sense in which sensations and concepts are representations, though neither alone a way in which consciousness is directed toward objects, we may avoid supposing that representations are peculiarly subjective objects of consciousness. Indeed, we cannot even suppose that, in every case, they are items or entities in consciousness. Intuitions, in the primary sense, and (with some qualification) sensations, are “states” of a subject, or at least of a potential subject, of consciousness. But concepts are neither states of a possible subject nor peculiar objects. Our task will be to develop a sense in which concepts are nontheless (possible) aspects of consciousness. They will in that case be, as suggested, “representations” in virtue of contributing, in an appropriately internal way, to the latter’s object-directedness.
There may seem to be a number of problems in the proposal to regard intuitions, in the primary sense, as those states of a subject that are its truly “object-directed” states. The least of them may be that Kant himself defines intuition in terms of the notions of representations that are “singular” and “immediate.” I have argued elsewhere that both of these notions amount to nothing other than the internal capacity of certain states to direct a subject, in a sense to be clarified, toward objects.4 There may seem to be other problems. As we shall see more clearly in the next section, the proposal requires taking seriously the notion of being intuitionally directed toward merely possible objects. It does this because intuitional “form” is an intrinsic aspect of intuitional states. As such, intuitional form does not require the obtaining of any factual, e.g., causal, relations with actual objects.
A further complication concerns abstract thoughts. As distinguished from such genuinely intuitional states as perceptions (and, with qualification, imaginings), these do not always seem to involve a way of being directed toward objects. Or it may seem that they do so at most in a way that is different from what is in question in the case of intuitions. It is hardly clear, in any case, that we could do justice to abstract thoughts by merely relegating them to the class of those things that are at most able to contribute, in an albeit internal manner, to ways in which other states might direct one toward objects.
The same might appear to go for instances of mere “belief” or “knowledge.” Many people find it useful to regard these as modes of “representation.” That Kant himself does so may seem to be entailed by his use of the term Erkenntnis. Kemp Smith generally translates it as knowledge (in the plural: modes of knowledge), when in many cases some other term is preferable, e.g., judgment or cognition.5 In any event, one might suppose that belief and knowledge are constituted out of concepts, and concepts are of course representations. Now Kant places particular weight on the connection between concepts and “rules.” In some sense he seems virtually to identify the two. But rules are something that one can learn or master. Having done so, one has apparently entered into a more or less enduring subjective “state” that may be said to contain or embody, in a purely internal way, the particular conceptual abilities that were in question. These states might then be regarded as representational states. We do not need to pursue this issue now. Later I shall argue that Kant’s conception of concepts as rules, or as embodying rules, does not imply anything like this sort of view. It does not involve this sort of dissociation of concepts from the concrete course of ongoing consciousness. In fact, concepts are, in a sense, merely forms or aspects of consciousness itself. If this is so, then we should not presume that mere states of believing and knowing are ever representations in the Kantian sense (although they may involve or presuppose representations in that sense, and may also be representations in some other sense).
The question remains concerning occurrently conscious thoughts and judgments in general. Even apart from the question of abstraction, and from that of merely “dispositional” belief or knowledge, it may seem necessary to extend our general approach in order to accommodate them. For example, one might extend the notion of representation so as to include, not simply particular ways of being directed toward objects (perception and imagination), nor merely whatever elements are capable of contributing in an internal way to ways of being thus directed (sensations and concepts), but also modes of consciousness that are in their own turn constituted out of representations in either of the former senses. It may seem obvious that thoughts and judgments, so long as they are occurrent states of consciousness, not mere dispositions or knowledge merely “possessed,” must be constituted out of one or the other (or some combination) of intuitions, sensations, and concepts. But in fact, I shall suggest, Kantian “judgments” are themselves intuitions. They are intuitions, that is, that have been converted into judgments (i.e., into Erkenntnisse). The main obstacle must of course lie in the challenge of purely abstract cognition.6 As it happens, Kant himself does not seem clear in this regard:
Since no representation, save when it is an intuition, is in immediate relation to an object, no concept is ever related to an object immediately, but to some other representation of it, be that other representation an intuition, or itself a concept. Judgment is therefore the mediate cognition of an object, that is, the representation of a representation of it. In every judgment there is a concept which holds [gilt] of many representations, and among these also grasps [begreift] a given representation that is immediately related to an object. Thus in the judgment, “all bodies are divisible,” the concept of the divisible is related to various other concepts; but among these it is here especially related to the concept of body, and this concept again to certain appearances that present themselves to us. These objects, therefore, are mediately represented through the concept of divisibility. Accordingly, all judgments are functions of unity among our representations, since instead of an immediate representation, a higher representation, which grasps [unter sich begreift] this one and several others, is used for the cognition of the object, and thereby many possible cognitions are drawn together into one. (A68-9/B93-4)
It is not clear what point Kant is making. But we might suppose him to be suggesting that even the most abstract judgment still demands the embodiment of concepts in intuitional material, at least if it is to be a judgment in any way about “objects.” If so, it remains unclear what the form of embodiment needs to be. As we shall see later, the only kind of embodiment that constitutes a genuinely cognitive “response” to an intuition, thus to an object qua object of that intuition, is an operation by virtue of which the latter is itself converted into the very judgment in question. It must be an operation by virtue of which the intuition is made an Erkenntnis in its own right. Perhaps Kant is suggesting that we take this as our model for judgment generally. (That all judgments are intuitional has in fact been argued by others; I comment on such a case in Chapter Two. The apparent absurdity may in part be removed by recalling that, with qualification, intuition includes “imagination.” In addition, nothing prevents the admission of purely abstract “judgments” in some sense at most related to the proper Kantian one. I comment on this point in the Conclusion.)
Consider the judgment that all bodies are divisible, as a judgment actually “formed” on some occasion. Kant may be supposing that this judgment involves the imaginative presentation of some possible body or bodies. If it does, then one is of course not merely conceptualizing the latter as some possible body or bodies. One must be conceptualizing it as all bodies. (There need be no question of error, in virtue of having, in that case, mistaken some for all bodies. Taken “in themselves,” the objects need really be neither some nor all bodies. It is not, after all, a matter of imagining actual bodies. It is more like imagining possible bodies and taking them as “representative” of all. The question of “possible objects” will occupy us in the next section.) The full judgment would consist, then, in the “application” of the concept divisible, not simply to the concept body (or even to the concept all bodies), but precisely (and in an appropriately modified sense of the term) to the conceptualized intuition itself. Of course, this would raise problems for Kant’s claim that some judgments are purely analytic, hence do not relate concepts to objects at all.7 In any event, such is the theory of judgment that I shall argue to be implicit in the Transcendental Deduction.
There is an additional ambiguity concerning the notion of consciousness, reflected in Kant’s inclination to deny that intuition, apart from concepts, is genuine consciousness at all: it is “blind” without them (A52/B75); without them, “nothing to us” (A120; cf. A112, Aug, B132). But while there is an apparently strong inclination to this effect in the Critique, the inclination is also contradicted in that work and elsewhere in Kant. It seems clear that unconceptualized intuitions are in some sense still instances of “consciousness,” yet is also clear that Kant sees a sense in which consciousness depends on concepts.
If Kant is otherwise right, we could always stipulate a sense in which intuitions, even apart from concepts, are modes of consciousness. We could say that they are modes of consciousness, for example, in that they contribute a unique sort of object-directedness to any state in which they are appropriately ingredient. If it is the intuitional dimension in any conscious state that is responsible for its object-directedness as such, then that may be reason enough for conceding that intuitions are themselves instances of “consciousness.” Their conceptualization, we might then say, is simply their elevation to a higher level of consciousness. In any event, Kant has no particular theory about the kind of consciousness peculiar to intuition as such, over and above his views concerning its role in the constitution of object-directedness.
Kant of course distinguishes between pure and empirical intuition. The latter, I have suggested, includes ordinary sense perception. But ordinary sense perception involves concepts. This may incline us to say that it cannot be an example of empirical intuition. But that would be a mistake. The mistake would rest on supposing that the conceptualization of an intuition involves the attachment of a concept or a judgment to it, or perhaps the imposition of some external structure upon it (or upon it together with others). Again, one of the points of this study is to propose an alternative to this view. Conceptualized intuitions are still intuitions. They are conceptualized intuitions by virtue of some internal alteration or elevation, not by virtue of something that is attached to or imposed on them.
Sometimes, what Kant calls “pure” intuition is just a type of imagination, namely, the type that plays a special role in mathematics. But Kant also uses the term in other senses. He uses it for space and time, for example, and for the purely spatiotemporal aspects of objects of intuition. Most crucial for our purposes, he also uses it for a certain structure in our representations of any object of intuition, that is, for the pure “form” that is in them.
As a variety of imagination, pure intuition is a kind of imagination that is in a sense doubly pure. It is pure because pure intuitions are devoid of sensations (or if they contain sensation, then only inessentially). And they are pure because the concepts in them need to be “pure” (any others occurring inessentially). Like any instance of merely (“purely”) imagining something, a pure intuition does not involve sensation in the way that a perception does. In this respect, any instance of it is like an instance of merely imagining a cup of tea, instead of actually seeing one. But there is also a feature of pure intuition that distinguishes it from imagining in general. It may be tempting, from what Kant says, to suppose that it lies in the fact that pure intuition does not involve any concepts at all. (It is of course difficult to see what imagining a cup of tea could be, apart from the employment of concepts.) But this would be wrong. Whatever he might mean by it, Kant sometimes speaks of mathematical concepts as “pure” concepts (A140/B180; cf. A719/B747). What Kant sometimes calls pure intuition is simply that special kind of imagination in which such concepts are “constructed” or represented for the purpose of mathematical cognition A713ff/B741ff).
A more difficult issue concerns the connection between pure intuition and what Kant calls the “categories.” In fact, Kant sometimes seems to distinguish between two sorts of categories. One sort appears to be so pure as to contain no reference to the conditions of spatiotemporal representation at all, hence to be even more pure than pure intuition: the so-called unschematized categories. Whatever we say about them, there is in any case a problem concerning the schematized categories.
For those who think of concepts as “rules,” the relation between categories and pure intuition may seem simple. The pure intuitions in question might just be space or time themselves, or determinate aspects of these; the connection with the categories would simply be that the latter are rules that involve an essential reference to the former. But suppose that concepts, as in our own view, are forms or modes of consciousness, not merely rules (though non-categorial concepts must of course involve rules in a way yet to be analyzed). Then how shall we conceive of the relationship between pure intuition and the categories? One suggestion, which I have elaborated elsewhere, is that the schematized categories, if they are concepts at all—not mere forms for the forming of concepts out of material in intuition—must be, in a sense, pure intuitional imaginings in their own right, or at least some formal aspect of the latter, over and above the pure “form of intuition” that these contain.8 But it is not clear what sense can ultimately be made of this view. In any event, I shall be concerned with the Kantian “categories” in this study only in their role as forms for the “formation” of concepts.9
We are finally ready to return to the notions of matter and form in intuition more generally. The first main point is to be sure that we do not fall prey to a common misinterpretation. It is one that finds some support in Kant’s text, but also much to confute it. It is a form of the subjectivistic interpretation that I mentioned earlier. In considering it, it is necessary to comment on another possible source of ambiguity. The ambiguity concerns the notion of sensation (Empfindung). It is likely that Kant uses this term in more than one way. In its primary employment, it is used to designate certain sorts of states of the perceiving subject. (The latter notion may of course prove ambiguous too: as between the perceiver as a “thing in itself’ and as mere “appearance.” In that case, the notion of sensation will be doubly ambiguous.) The crux of the subjectivistic reading is this: that sensations, as states of the subject, are the “matter” or the material of the “appearances” that are the immediate objects of sensory intuition.
Proponents of such a reading might put the point differently. They might put it simply by saying that sensations are the material of sensory intuitions. This is something Kant himself says (e.g., A20-2/B34-5). The question is what he means. As noted, Kant sometimes means by “intuition” what might more properly be said to be the objects of intuition. It does not follow from this that when he speaks of sensation as the material of intuitions he is adopting a form of the subjectivistic view. He might, on those occasions, be speaking of intuitions more properly, namely, as instances (at least when conceptualized) of ordinary sensory experience. In that case, he would only be saying that sensations are the material of ordinary sensory experiences. On the other hand, he might be using the term peculiarly. He might be regarding sensation as the material of objects of consciousness (appearances), but yet not any longer saying that our own subjective states are the material of those objects. The question is additionally complicated by ambiguities in the notion of “appearance.”
In one sense of the term, appearances might indeed be regarded as composed of subjective material. This must involve a merely empirical, as opposed to the “transcendental,” notion of appearance. In the transcendental sense, even material objects are appearances for Kant. The way these objects in turn appear may be said to involve “appearances” in a merely empirical sense (A29-30/B45). Now there is a perfectly legitimate sense in which Kant may speak of such appearances as formed out of sensations. So understood, they would presumably be states, or aspects of states, through which subjects apprehend objects or possible objects. As Kant says, such appearances would only be “changes in the subject” (A29/B45). Unfortunately, he also seems to say something that makes little sense. He gives, as an example of such an appearance, the colors that one perceives in perceiving some object. It is difficult to make sense of the notion that some of the states of a perceiving subject actually are colors. At most, Kant ought to say, they are states through which colors are apprehended.
We may suggest an explanation of such formulations. Two very different things might be meant by appearances, even in the empirical sense, and by the “ways in which” material objects appear. The distinction is a difficult one, and it is not unlikely that Kant occasionally slipped on it. First, appearances in the empirical sense might simply be the intentional correlates of sensory apprehension, considered precisely as such correlates, and hence without regard to their reality as material objects. For the sake of our present discussion, we may call them correlate-appearances. Considered just as such, correlate-appearances are not formed or made out of anything at all. Hence they are not made out of sensations. Nor are they made out of anything else. Considered as mere correlates of the apprehension of them, such talk is simply misplaced.
Now it is crucial to my reading of Kant that it is precisely correlate-appearances that one ordinarily takes to be materially real objects, and is usually right in doing so. (I try to clarify this notion further in the final section of this chapter.) In that case, one must of course take the appearances in question to be formed or made out of something. What one takes them to be formed or made out of is what Kant himself calls either their Realität (A166/B207ff) or their Materie (A20/B34, B207). Kant thinks of this as whatever it is that really fills some region of space, whenever a material object is filling it. But consider mere correlate-appearances. Even though, just as such, they are not made out of anything, they too can be described in terms of a kind of “filling” of space. They can be described in terms of an at least apparent filling of a region of space, for an at least possible object. We may therefore speak, quite legitimately, of the correlate-material of correlate-appearances. To take the most obvious case, perceived colors would be examples of such material. To make the distinction, we need simply distinguish real colors from correlate-colors. The tendency to do so would be all the more powerful, given that mere correlate-material is in effect what one ordinarily takes, and is usually right in doing so, as the real colors of real material objects.
Thus we can understand Kant’s tendency to regard correlate-appearances as made out of something, even though they are not made out of anything at all. Given that tendency, we can also understand something else. We can understand why, in general, Kant might be inclined to refer to correlate-material as “sensation.” It is, after all, only through sensation that one apprehends correlate-material in the first place.10 And it is difficult to see that any other, except a purely technical term, would do better. It does not follow from this that Kant really supposed that correlate-appearances are formed from subjective states. In any case, correlate-appearances are only one of the two things that might be meant by appearances in the empirical sense of the term. It may be Kant’s tendency to confuse the two that led him to forget that correlate-material is not a kind of material.
But what could be the point of claiming that ordinary experiences have a “material”? To see this, we need to see what Kant contrasts with the latter. We need to see what he means by the “form” of an experience. But here we also encounter ambiguity. Often, what Kant means by the “form of intuition” is not at all an aspect of sensory experiences. Rather, it is an aspect of the objects of sensory experience. Still, what does Kant mean by the form of a sensory intuition, when he does in fact take it to be an aspect of a sensory experience?
In one sense, the form of an experience might simply be identified with its object-directed character. In another sense, it can only be identified with an important part of that character. However we choose to put it, we presumably need to allow for the fact that, in some sense or other, the object-directed character of experience is in part determined by concepts. I do not propose to have Kant deny this.
One might suppose that, for a very simple reason, an experience’s conceptual content cannot completely exhaust its object-directed character. One might suppose that this is so because the object-directed character of an experience is dependent upon sensation. What I am seeing or hearing does not merely depend on what I take myself, conceptually, to be seeing or hearing. To some extent it also depends on what my sensory state really is, and not on what I take it to be. However, this cannot be the reason why Kant regards conceptual content as at most contributory to object-directed character, and intuitional form as more strictly constitutive of it. This is because Kant excludes mere sensations, as much as he does concepts, from the intuitional form of an experience. Getting clear about the reason for this is a necessary preliminary to getting clear about the object-directed character of experience in the first place.
What we need to appreciate from the start is the importance that Kant attaches to the fact that the very same object might be presented either in a sensory or in a non-sensory “experience.” That this is indeed a fact, he takes pains to insist at the earliest possible moment in the Aesthetic:
Thus, if I take away from the representation of a body that which the understanding thinks in regard to it, substance, force, divisibility, etc., and likewise what belongs to sensation, impenetrability, hardness, colour, etc., something still remains over from this empirical intuition, namely, extension and figure. These belong to pure intuition, which, even without any actual object of the senses or of sensation, takes place [stattfindet] in the mind a priori as a mere form of sensibility. (A20-1/B35)
One might suppose that the only sense in which the very “same object” might be presented in a sensory and in a non-sensory manner is one that is purely conceptual: the object of a thought that we think in connection with an experience, or the object of a judgment, might continue to be the object of the very same thought, or of the very same judgment, even after all of the sensations in question have vanished. But Kant makes it clear that his point is different from this. He makes it clear that the “presence” of the very same object is not a purely conceptual matter at all. Even if the same concepts were employed on two occasions, one of them involving sensation and the other not, the fact that the same object (extension and figure) is in question could not possibly be explained by appeal to those concepts. The reason is this: Those concepts merely serve to conceptualize11 whatever object is in question. Therefore they do not by themselves account for the fact that some object is in question in the first place.
Kant’s reasoning turns on a distinction between two kinds of imagination that is independent of his own distinction between pure and empirical intuition. It is the distinction between imagining some (possible) object, on the one hand, and merely imagining or supposing that something is (or might be) the case, on the other. Imagining some possible extension or figure, for example, is clearly different from merely imagining or supposing that there is, or that there might somewhere be, a particular extension or figure, or from imagining or supposing anything else about some extension or figure that is not actually imagined on that occasion. It may be that the latter sort of imagining is a purely conceptual matter. The former cannot be. In the former case, one is not simply employing certain concepts, but employing them precisely in thinking about some imagined object. Now obviously there is much more to be said on the question than this.12 For example, one might attempt to account for the difference between the two sorts of imagining by appealing, not to a primitive intuitional “form” in the more intuitional case, but merely to the presence of special phenomenal qualities, perhaps analogous to sensations, in addition to whatever concepts might be involved. I argue against this suggestion in the next section.
The distinction does not rest on assuming, as Kant may appear to do, that any instance of imagining some extension or figure might actually remain in fullblown consciousness even after the removal of all modes of conceptualization of that extension or figure. Though Kant may be assuming the possibility, his argument does not hinge on the assumption. Kant may very well concede, at this point, that there is no such thing as imagining a triangle or a line of any sort, apart from actually imagining that it is a triangle or a line. (In fact, there is no reason to concede quite this much. One might be said to imagine a triangle or a line, just in case one is imagining something that one would normally be expected to conceptualize as a triangle or as a line. The most one could plausibly claim is that, in order to do the latter, one must be conceptualizing the imagined object in some way or other.) Even if this were conceded, the point would remain that the concepts in question are “object-directed,” in any imagining, in a way that is not itself a purely conceptual matter. I do not have a way of defining what I mean by this notion. But as with Kant, it seems to me appropriate to appeal to it precisely as the point of difference between merely imagining or supposing that something is or might be the case, in regard to some possible extension or figure, and imagining or supposing in regard to some actually imagined extension or figure. The inability to define the difference may simply reflect its primitive character.
Once we appreciate the difference between imagining an object and merely imagining or supposing that something is or might be the case, then we are in a position to focus on that aspect of a sensory intuition that Kant calls its “form.” The latter simply is that non-conceptual aspect of a perceptual experience, even of a hallucinatory one (I won’t bother to repeat the qualification throughout), that it is able to share with purely imaginative but non-hallucinatory intuitions.13 Again, the question does not turn on supposing that it makes sense, apart from the way we actually conceptualize imagined or perceived objects, or may normally be expected to conceptualize them, to talk about the very “same object’’ being presented in a purely imaginative and also in a sensory experience. Even if concepts are involved in both cases, the fact remains that, in the one case, what we conceptualize in some particular way is presented in a sensory manner, and in the other case it is not. Hence sensation no more than conception, nor a combination of the two, is responsible for the “object-directed character’’ of the presentations.
The issue regarding the role of sensation, and its connection with object-directedness, is complicated by a further ambiguity concerning both these notions. I said earlier that, in the primary sense, sensations are “states” of perceiving subjects. But Kant sometimes seems to regard such states as purely internal to perceivers, and other times as essentially involving perceptual relations with actually existing objects.14 In the former sense, but not in the latter, even the sheerest hallucination could be said to involve sensation. In the latter sense, in any event, we would appear to have a sense in which mere sensation is able to determine an “object” for a perceptual state.
It should be clear that this is compatible with what I have said so far. I have not denied that there is a sense in which intuitions are “objected-directed” in virtue, or at least partly in virtue, of purely external relations, or even that sensations may be essential to such directedness. What I have said at most implies that actually taking any object to be an object (imaginary or real), with respect to any intuitional state in fact apprehending it, always involves the (perhaps conceptual) enrichment, or the “conceptualization,” of a different sort of object-directedness than sensation provides. (The notion of “taking,” and of “conceptualization,” will soon become ambiguous as well.) The sort in question is not a matter of one’s actually standing in a perceptual relation with objects. Nor is it a matter of being in a “sensory” state, even in the sense in which hallucinations may be said to be so.
Much more, of course, needs to be said. I have so far only been formulating some distinctions. In the next section, I offer an independent defense of them. A thorough discussion will also require additional concern with “objects,” and even with the question of an object’s “reality.” Regarding the latter, I have argued elsewhere that Kant is committed to a kind of phenomenalism, at least with respect to the reality of objects perceivable in space and time. Thus he is committed to a kind of phenomenalism with respect to perceptual “relations” (at least insofar as they are not presumed to involve some unknowable relation with “things in themselves”). I have also tried to show how this particular kind of phenomenalism differs from more objectionable forms of that doctrine. In the final section of this chapter, I attempt a further contribution to this project. But this can only be done once we have expanded our notion of “material” in a sensory intuition even beyond the two notions of “sensation” so far encountered.
This much not only should give us a preliminary idea of what Kant means by the pure “form of intuition,” in what I propose to regard as the primary sense of that term, and why all instances of imagination, and even hallucination, need to involve it. In addition, it should give an idea of the point of Kant’s claim that sensations, and sensations alone, are the “matter” of sensory intuitions as such. However, we shall also need to see the sense in which, even though sensation is the matter of (sensory) intuitions as such, some other sort of material must be capable of serving as matter in it as well. This, I shall eventually argue, is a kind of material that Kant attributes to “imagination.” But it will be in a sense that we have not yet encountered.
In order further to clarify the notion of sensation itself as “matter,” it will in any case not suffice to say, as I have so far said, that, while the pure form of intuition constitutes a primitive kind of directedness in a subjective state, sensation is what contributes, on a non-conceptual level, to the specific way in which such directedness occurs. We cannot say this, precisely because the same will hold for the “imaginative” material within an intuition. Nor should we say, for example, that sensations constitute that aspect of a sensory intuition whose distinctive “intentional correlate” is the presence of some immediately perceivable quality in perceived objects: say, their immediately perceivable phenomenal color. We should not say this because it is at the least unclear, despite Kant’s own occasional suggestions to the contrary, to what extent such qualities are in fact apprehensible apart from some way in which the subject actually takes those qualities.
One might of course always say that the sensory matter of an intuition is that aspect of it that accounts for the fact that its object is presented in a “sensory” manner, and simply take the latter notion as primitive. But Kant himself often attempts to say more. He sometimes tends to do so, for example, by reference to the sorts of alterations that are normally produced in the state of a subject by means of the stimulation of its sense organs, or that are apt to call attention to a subject’s sense organs, or at least to its own internal condition.15 A related suggestion that is also close to Kant’s own formulations, but perhaps more resistant to counterexample, is this: the distinctively “sensational” matter of any sensory intuition is that aspect of it in virtue of which it is prima facie reasonable to take its object to be real.16 However, in the light of our own conclusions to follow, we might also simply mean this: the specifically “sensational” matter of an intuition is just whatever non-conceptual aspect of it contributes to the way in which its object is presented, but is—unlike any merely “imaginative” material it may contain—unsuited for service as material for the formation of concepts.
However we spell it out, nothing stands in the way of presuming that an intuitional state is itself made or “formed” out of whatever happens to serve as its “matter.”17 Its “form” will then be whatever constitutes it as a state by virtue of which an object is apprehended through that material, as a potential subject for conceptualization. On the other hand, we might also accommodate the supposition—on some level, Kant’s own—that any genuine “apprehension” of objects must always involve at least some degree of conceptualization from the start. The concession, should we make it, does not disturb the general point. In that case, we should say that intuitional “form” is what would constitute a state as one by virtue of which something is “apprehended” through a certain body of material—given that the latter, or some appropriate portion of the latter, should also happen to serve as material for an act of conceptualization.
The approach that I have begun to develop may seem ontologically objectionable. The difficulty stems from the fact that, according to the account in question, ordinary experience always consists of a series of conceptualizations, or at least of some kind of “takings,” of objects that might or might not be, or that might or might not have been, real or actual objects. (A reminder: For convenience and where context assures clarity, I speak both of the conceptualization of intuitions and of the objects that are presented in them.) This is because an intuition’s intrinsic directedness—provided by its “form”—is neutral as between imagining or hallucination, on the one hand, and a concrete apprehension of reality, on the other. Indeed, it is neutral as between mere imagining and the apprehension of what one even so much as takes to be reality. The latter involves a particular way of taking an object, but it does not need to involve a way of taking an object that really exists.
Kant argues, in the “Refutation of Idealism” (B274-9), that a necessary condition of one’s capacity for self-cognition is that one’s intuition always be related to a real world in some way. But in any particular case, what is conceptualized as a concrete, solid, object in space might still be a purely imaginary or hallucinated extension or figure (B278).18 This seems to imply that real objects might have been something other than real. And in any case, it rests on the notion of taking things to be real that, in any particular instance, need not be or have been. We may thus appear to be committed to a realm of purely “intentional” objects whose intrinsic being is other than real being.19
Two strategies may seem to contain the alternatives for avoiding this commitment. Each may be regarded as attempting to favor one of the two parts of an ordinary intuition. The first may appear to be my own. It posits a primitive directional character or quality in consciousness, in order to account for instances of the primitive object-directedness that I have been trying to capture. Those instances would be provided by determinate occurrences of this primitive determinable character. Since the character is supposed to be a purely intrinsic quality, not a relational property, it would not carry any real commitment to objects of consciousness as such. The variety of apparent objects would merely be the phenomenological or intentional counterpart, and so at best a grammatical reflection, of the possible varieties of that primitive directional character. Obviously, I think that there is something right about this position. But there is something wrong as well.
The second approach does not hold that the assignment of objects to states of consciousness, in the neutral mode that I have been trying to capture, rests on the ascription to consciousness of a variety of primitive object-directed characters or qualities. Instead, it may be regarded as no more than the provision of a variety of “translations,” into the language of the ascription in question, of the intentional “contents” of those states. (Independent considerations will call for the distinction between states and their contents.) We need not assume, on this approach, that intentional contents are distinct entities to which intentional states are externally related. We might, for example, regard them as a certain type of instantiable feature of intentional states, or as a variety of types that the latter are capable of “tokening.” In any case, the suggestion is this: to say that a certain state represents (or presents) a possible object, in the relevantly neutral sense, is simply to say, or at least to indicate in an appropriately specific manner, that the state in question is translatable, or has a content that is translatable, by means of an appropriately specific expression in whatever language is in fact employed in the provision of that translation.20
In an obvious sense, the first of these approaches is more ontologically oriented than the second. But they agree in one respect. They agree that it is possible to acknowledge an object-directedness intrinsic to particular subjective states, without having to assume an ontological commitment to the objects of those states, as entities to which they are related. The commitment to relations with objects, as entities of any sort, would enter only with the eventual inclusion of purely extrinsic factors (e.g., with the inclusion of facts about entities that might be, or might be appropriately related to, things that actually affect one’s sense organs). Each of these approaches is inadequate. But each represents a factor that needs to be included in an adequate conception of the kind of directedness that can be intrinsic to subjective states.
The first approach would be, in one way, like a familiar attempt to avoid a commitment to “sense data” in perception. One might, for example, attempt to avoid the commitment by regarding such predicates as sees a green spot, when they are not supposed to involve the real existence of anything seen, as expressing no more that a certain way of being “appeared to” sensorily. That strategy is, of course, consistent with a number of views of perceptual judgment. For example, our judgment about the ways of appearing might be construed as implicitly comparative: one is being appeared to “in the way that (or in one of the ways out of a range of ways that) one is standardly appeared to when. . . .” Or they might simply be taken to rest on acquaintance with a multitude of primitive phenomenal qualities, comparable to one another, but not themselves to be defined in ultimately comparative terms. Perhaps most plausibily of all, a combination of the two is in order: We are acquainted with standard cases of (really) perceiving (real) spots and of perceiving something green. We are also capable of discriminating respects in which such cases agree and differ with respect to aspects of phenomenal quality. That we could not verbally define such respects need be no objection to the supposition that our apparent awareness of a primitive mental “directedness” in sensation is, throughout, merely the awareness of respects in which modes of appearing are both like and unlike certain standard cases of sensory appearing.
These sorts of elaboration might appear to avoid an immediate objection to the strategy as originally proposed. What the strategy appears to require is the construal of primitive directional qualities as, so to speak, indivisible units. For example, one such unit would be involved in apprehending an O as an F, another is apprehending an O as a G. But insofar as we are dealing with indivisible units, not relational properties, the strategy would appear to exclude the possibility of cases in which a single object happens to be apprehended in different ways. So far as we could say, we would simply be dealing with two completely different “objects,” O-as-F and O-as-G; it would make no sense to inquire whether, in addition to this, the O’s are the same or different.21 All we could ask is whether the primitive directional qualities are the same or different. And of course they are different. But as elaborated so as to allow for differential comparisons among phenomenal qualities, or instances of primitive “directedness,” the proposal may appear to avoid at least this difficulty. However, it also seems unlikely that any such approach could be plausibly extended beyond the purely sensory case: for example, to the case of merely imagining a spot, not to mention that of imagining or supposing that there is, or might, somewhere be one.
We might try to extend the comparative approach by continuing to appeal to a standard range of sensory appearings. The purely imaginative cases might then be handled in terms of our ability to discriminate respects in which other cases are qualitatively comparable to them, even though they are not themselves sensory in quality. Once again, that these respects are “primitive” in the sense that we could not verbally define them, in a non-circular way, need be no objection to the supposition that our apparent awareness in either case of primitive mental directedness is merely that of respects in which a non-sensory state is like and unlike possible members of a standard range of sensory appearings.
Whatever additional objection may apply to the suggestion, the suggestion is impossible for the following reason: it implies that our awareness of the similarity between imagining and sensing, apart from whatever corresponds to the specific terms used describe them (e.g., green or spot) is of something purely negative, namely, of the fact that the imaginative cases do not involve any actual mode of sensory appearing. Of course, Hume, notoriously, offered a way out of this difficulty. He did it by proposing that there is indeed a discriminable positive element, shared by imagining and sensing, in addition to whatever corresponds to the particular terms used to describe them. For there is, in his view, some positive degree of “force and vivacity” in each case; it is simply that it is lower in the case of imagining. But except so far as talk about degrees of force and vivacity is merely another way of talking about primitive mental directedness, the suggestion seems crudely ad hoc. In any case, as further discussion will show, the strategy could not possibly explain the essential twofold structure that Kant rightly saw in at least certain mental states, namely, in those in which some content is intentionally directed toward some (possible merely imagined) object.
It will be important to see why the second strategy fails in the same respect. The second strategy does not appeal to irreducible ways of appearing or being appeared to, or to primitive mental directedness. It appeals to a purely functionally defined notion of “content.” The latter is simply a term for the fact that a mental state can be regarded as “saying” something or other. The functional significance of this latter notion might then be supposed to lie wholly in what we acknowledge to be acceptable translations of what is said. As to the question of an underlying ground of translation, should we ask it, the most controversially reductive approach is, of course, one that involves an analysis that is also “functional” in a more narrow sense, namely, in primitive causal terms.22 Kant’s own view might be thought to fall somewhere in between. This is because of his emphasis on concepts, not as states of (nor, apparently, as contents in) minds, but as functions of potential unities with regard to states of mind:
Whereas all intuitions, as sensible, rest on affections, concepts rest on functions. By ‘function’ I mean the unity of the act of bringing various representations under one common representation. (A68/B93)
This suggests at least a general sort of “functionalist” approach. Kant also attempts to be more specific regarding what is needed in order to constitute any commonality of conceptual function. It does not appear to have anything to do with commonality of causal role. It rather involves some kind of commonality of normatively binding “rules” for the employment of any concept: “But a concept is always, as regards its form, something universal, which serves as a rule” (A 106). From this, the Kantian approach might seem to be a particular instance of the translational approach to content. The suggestion would appear to be that a conscious state “says,” for example, that a is F (or at least in some other way “represents” a-as-F) just in case it tokens a type, or instantiates a feature, or in general has a content, whose role is governed by just the same rules as govern that of some corresponding item in a language of translation. I shall return, in Chapter Three, to the question of “rules” in Kant.
On a purely formal level, this sort of approach has value. But it can also be misleading. As we shall see, a closer look at precisely how concepts are able to serve as “functions of unity” among states of consciousness reveals a dimension of representation that needs to be ingredient in states of consciousness in a way that merely instantiated features, or tokened tokens—and in general translatable “contents” as we have so far been understanding them—are not. Kant himself may have a tendency to blur this issue, on account of his tendency to shift between a more formal and a more material way of looking at concepts. On a purely formal level, we may indeed regard (non-categorial) concepts as a kind of quasi-linguistic term potentially tokenable or instantiable by a manifold of distinct states. As such, a concept provides translatable content to possible states of consciousness simply by virtue of being tokenable or instantiable in accordance with appropriate “rules.” When we report the contents of those states—as opposed to theorizing, as we shall be, about what makes them possible in the first place—we are simply not generally interested in what might otherwise be ingredient in them. All the same, conceptual content could not possibly be a matter of merely tokenable or instantiable translatable content. It must rest on something that is more literally ingredient in conceptual states. Conceptualization involves the presence of translatable content, not simply through unifying functions in regard to translatable states as wholes, but through functions that bear on the very material out of which those states are formed in the first place.
The suggestions need to remain vague for a while. We must first return to our distinction between what seemed to be incompatible strategies in regard to the intrinsic directedness of consciousness. I want to argue that what we need is an approach that involves elements of both. To show this, I shall consider the case of a certain sort of imagination, namely, the sort that, as I have already suggested, requires appeal to a primitive “form of intuition”—that is, imagining “an object” as opposed to merely imagining that something is or might be the case. To make the point, one might also appeal to sensory intuition. But while it is obvious that such states would involve more than translatable content, they involve so much more that it is difficult to focus on what is truly relevant in them.
An imaginative intuition presumably always exemplifies some conceptual or descriptive content as well. In a perfectly natural sense, one cannot imagine a bowl of soup without in some sense “conceptualizing” what one imagines as soup (or at least without conceptualizing it in a way that would in turn make it reasonable, perhaps under different circumstances, to conceptualize it as soup). But how could one ever translate the content of an imaginative intuition? A translating sentence seems to be ruled out from the start. In effect, we have already excluded such attempts as “There is, somewhere, a bowl of soup” and “There might, somewhere, be a bowl of soup,” and any more specific variations on such attempts. These sorts of translations might be appropriate to imagining that something is or might be the case regarding a bowl of soup. They are not appropriate to imagining that something is or might be the case regarding a bowl of soup that one happens actually to be imagining. Of course, we could always appeal to something like “This is a bowl of soup, or at least an imaginary one.” But that, for obvious reasons, appears to duplicate the problem. (Or perhaps it does not. I argue below against some recent attempts to explicate the notion of “demonstrative content.”)
Perhaps we need to consider words or phrases, instead of sentences, in order to capture the content of imaginative intuition.23 We might consider, for example: “a bowl of soup.” But the suggestion is unacceptable. The suggestion implies that the state of imagining a bowl of soup, and in general the state of imagining any object, is derivable by mere subtraction from that of imagining that something is the case. Consider the thought that there actually is, or might be, a bowl of soup somewhere. On the suggestion in question, the state of imagining a bowl of soup, at least with respect to its “content,” is obtainable by mere subtraction from this one. The only relevant difference would be that the content of imagining a bowl of soup involves only a part of the content that is involved in a case of imagining that there is, or might be, a bowl of soup somewhere. This seems impossible. The former ought to involve more, not less, than the latter.
It might still be possible to appeal to something other than translatable content and yet avoid a reversion to primitive directional qualities. The most likely candidate may seem to be what some people call “raw feels,” “phenomenal qualities,” or “qualia.”24 Such solutions might introduce a primitive but non-directional feel or quality, in order to account for the additional element in actually imagining some kind of object, as opposed to merely imagining that there is or might be an object of that kind. We might then suppose that the total content of an imaginative state is only a matter of the appropriate (perhaps causal) connection between this phenomenal component and a bit of genuinely translatable mental content. The connection, and neither element alone, would constitute the peculiarly imaginative content in question. The whole of the content might appear to be translatable at first. It might appear to be translatable precisely as “This is a bowl of soup (or an at least imaginary one).” But on the suggestion, what we really have is a connection between a bit of phenomenal quality and a content that is strictly translatable, not as a “statement” to the effect that this is a bowl of soup, or even a merely imaginary one, but at most in terms of a non-sentential phrase: “a bowl of soup, or an at least imaginary one.”
Just as stated, and apart from its own insistence that the relevant “phenomenal qualities” are non-directional, the suggestion might of course be taken to formulate Kant’s own view, and as a way of recognizing primitive directional quality after all, in addition to a mental state’s possession of translatable content. How crude we judge the formulation to be would depend on what we suppose to be involved in the “connection” between translatable content and primitive phenomenal quality. If it simply amounts to whatever it takes to embody, in a conscious state, a content that is translatable by means of a particular expression, then there is really no argument. The connection of that content with a particular raw feel, or with a special phenomenal quality, might be regarded as nothing other than its tokening or instantiation by a Kantian Anschauung. To put it differently, the need for such a “connection” would be just the need to recognize that a state with a given translatable content is not simply a state that has directional quality in addition to that content. Rather, it is a state in which that content itself needs, as it were, to receive the form of directedness in question. This will become clearer later. In any case, the present suggestion is intended very differently. It is intended to rest on a distinction, as well as on an external connection, between two different sorts of subjective states, or at least between two different aspects of a single state. Each is supposed to enjoy an identity of its own and to be in principle describable independently of the other.
The suggestion that two states are involved, or even two distinguishable aspects of a single state, is untenable. This can be seen from the following consideration. If it were really a matter of distinct states, or of distinguishable aspects of a single state, needing to be connected in some way on a particular occasion, then there would be no absurdity in the supposition that any particular translatable content might end up being connected with any primitive phenomenal quality whatsoever, or even with any number of them at once. Consider, for example, the phenomenal quality presumed to be involved in some particular instance of imagining a rose bush. By hypothesis, that quality is what we need to postulate, in order to account for the “plus” enjoyed by actually imagining a rose bush, over and above its merely translatable content. Some presumably rather different quality would need to be postulated in order to account for the “plus” enjoyed in imagining a bowl of soup. And of course we would also need to postulate more specific qualities and differences, in order to account for the different ways there are of imagining such things, so long as these are not explicable by appeal to translatable content alone. But insofar as the postulated qualities are presumed to involve independent states, or independent aspects of states, both with respect to one another and to whatever translatable contents with which they might be connected, there ought to be no absurdity in supposing that those qualities and contents get randomly switched on occasion. We might suppose, for example, that the translatable content that is normally connected with the “feel” of imagining soup is now connected with the feel of imagining a rose bush. But the supposition seems, to the contrary, perfectly absurd.
If we are indeed prepared to countenance “phenomenal qualities” as essential to imagining in the first place, then there of course ought to be some leeway regarding the connection between such qualities and genuinely translatable content. As we might want to put it, although it is also a bit misleading: in anyone’s view, a single mental “image” might get variously taken. The image taken as a bowl of soup (or an at least imaginary one) might instead have been taken as a bowl of colored water. With appropriate background assumptions, it might even have been taken as a rose bush looking like soup. A certain amount of recombination is possible. Nor does conceding this imply, as it might seem to, that “images” are entities that get conceptually taken in various ways in imagining. Certainly, it would seem odd to suppose that phenomenal qualities, as aspects of mental states, are what get so taken.25 It is not clear what we should finally want to say about such matters. In any event, I introduced the notion of a mental “image” only to make a certain point more graphic. The usage conforms to what many people are inclined to say about imagining. However we spell it out, what is supposed to be captured is just the same as what was supposed to be captured by introducing phenomenal quality in imagining in the first place. In either case, we have to deal with the concession that there is something more to imaginative intuition than the tokening or instantiation of descriptive or conceptual content. Whoever is prepared to concede the point must also concede that there is always leeway in regard to “connecting” whatever elements are then in question. My argument need not deny this. (Whether there is comparable leeway regarding the connection of phenomenal quality with translatable content—or ways of “taking”—in general, not simply with conceptual content, remains a point to be reconsidered. I have, for the sake of convenience, so far not distinguished between “taking” and “conceptualizing.”)
When it is reformulated in order to capture this concession, my argument is simply that, while the introduction of phenomenal, but intrinsically nondirectional, quality ought to allow for a degree of play with respect to its actual connection with translatable content in imagining, a construal of such quality and content in terms of distinct states, or in terms of distinct aspects of states, allows for too much leeway. As we have seen, whatever such quality is involved in imagining a bowl of soup might instead have been involved in imagining a rose bush. But appropriate background assumptions were needed to make this intelligible. One needs to imagine, for example, that a rose bush is “appearing” under circumstances in which it happens to look like soup. By contrast, when phenomenal quality and translatable content are merely distinct states, or distinct aspects of states, no such assumptions are any more appropriate than others. (By a construal in terms of distinct states, or distinct aspects of states, I simply mean the supposition, again, that the ingredience of particular quality and the tokening or instantiation of content constitute a single state only by virtue of being connected into that state: for example, by virtue of standing in appropriate causal connection to one another, or perhaps merely by virtue of being instantiated together in that state.)
It will help, in seeing the point, to recall the original motivation of the suggestion. Even though it concedes, in effect, that the translatable content of imagining objects is not in itself a full sentential content, the suggestion may be regarded as resting on a model of content that involves full sentences. What seems to be left out of any purely conceptual content, and seems not to be capturable by our normal means for completing analogous sentences, is what we might call the “referential,” or at least the apparently referential, element in imagining. This element would be left out, for example, of any instance of tokening a content that is merely translatable as “a bowl of soup,” as opposed to “This is (or might be) a bowl of soup.” Though it might be held to capture its strictly conceptual content, the former obviously fails to capture the full content of imagining a bowl of soup. The latter, on the other hand, would seem to capture that content at the cost of re-introducing primitive directional qualities. So to compensate for what appears to be left out, but without running the risk of tolerating primitive directional qualities, one may simply be inclined to postulate the occurrence, among one’s subjective states, of a state (or of an aspect of a state) that is in some way a counterpart of—even though obviously not translatable as—the missing demonstrative term. Like the term in the corresponding sentence, the occurrence of that state at least generates the feeling that one is really referring to something. But unlike a correct use of the demonstrative term, it only needs to involve the feeling that such usage is correct. Unlike whatever genuinely conceptual content is in question, the counterpart of “this” is not translatable, and a fortiori not as this.
Now one might of course reply to the argument so far offered that while any such connection of “terms” (content and quality) is indeed logically possible, we can only in fact manage to effect certain ones, given certain sets of background assumptions. But that would be ad hoc. If phenomenal quality is ever able, together with translatable content, to constitute a single imaginative state, then it ought to be able to do so whenever it is appropriately tied to such content. Consider, again, the linguistic analogy. What we would be talking about, when we are considering such apparently “absurd” connections, would be analogous to the formation of ill-formed sentences. By the same analogy, we ought to expect, on the approach in question, precisely the possibility of generating instances of ill-formed imagining. The fact that ill-formed imagining (at least in the sense of ill-formed imaginative “intuition”) seems in turn to be unimaginable, undercuts the sentential analogy. (Of course, irrelevantly, one may always imagine that various “ill-formed” states of affairs obtain.)
Now as I have already suggested, the concession of a role for “phenomenal quality” in imagining might in fact be nothing other than acknowledgment of the need for a middle position between our two strategies. It might be a way of acknowledging the need for both irreducible directional quality and merely translatable “content” in at least some mental states. But we need to see how to acknowledge the duality without appealing to primitive determinates of a single determinable quality of directedness. To revert again to Kantian terminology, the solution must be to recognize that any kind of translatable content, if it is to be relevant to the directedness of consciousness of the sort that concerns us, must be regarded as the content of a state whose “quality” of directedness is the very form of that state. (This, so far as I shall argue, is compatible with regarding that content as also able to serve as the content of states that are mental, and even “directional,” but yet not so in the way that intuition is.)
Consider again the case of imagining a bowl of soup (as a bowl of soup). Unlike the approach that proceeds in terms of a determinable directional quality, we cannot construe this as a case in which some subject is in a state possessed of the primitive character “imagining-a-bowl-of-soup” (or even of two distinct characters, one to account for the quality of imagining, the other for requisite specificity). Unlike the approach that proceeds in terms of translatable content, neither can we construe the case as one that merely involves a state possessed of translatable content (or even, additionally, of some primitive phenomenal feel or quality). The case must rather be one in which some content—one translatable, for example, as “a bowl (or at least a possible bowl) of soup”—is ingredient in a state of consciousness precisely by virtue of being ingredient in a special sort of state. That is, we must suppose that the content is ingredient in consciousness precisely by virtue of its ingredience in a state possessed of primitive directional quality.
Our appeal must be to a primitive quality of directedness, but not to primitive determinate modes of such directedness. What we need to concede, in other words, is only that there are indeed irreducibly directional states of consciousness. Beyond this, we do not need to acknowledge the existence of a multitude of irreducibly determinate modes of directedness. As I shall argue, in order to account for the specific forms that the directedness in question takes, we need simply appeal, as on the second of our strategies, to the ingredience of purely translatable content in a state of that sort. By virtue of this appeal, any recognition of complexity of “structure” in intentional content ought then to remain as available to us as it did not on the first of our strategies, and it ought to be no less available than it was on the second. In any case, the admission of primitive directional quality continues to avoid a commitment to intentional objects as entities. For it is still regarded as an intrinsic quality, not a relational feature of mental states.
We need to modify the second of our strategies as well, beyond merely combining it with the first. As we left it, that strategy was compatible with a number of views as to the notion of translatable content. For example, it was compatible with a narrowly functionalistic view, according to which translatable content, considered in itself, is solely a matter of causal relations in which a state, or some aspect of a state, stands. But I shall now argue—by what is in effect an extension of the argument concerning primitive “feel” in a state—that a mental state’s translatable content cannot simply be an aspect (or a part) of a state, merely externally connected with its primitive directedness. Rather (adopting once more our Kantian terminology), whatever constitutes translatable content in a state of consciousness must serve as a kind of material in a state whose “form” is precisely that of primitive directedness.
The argument concerns the necessity that translatable “content” in a mental state that is primitively object-directed—that is, object-directed in the way that at least imaginative intuition is—must make some difference in how an object (a possibly non-existent one) is apprehended through that state. I do not mean to overstate the point. In the first place, we are still limiting our concern to mental states in which, as in imaginative intuition, there is indeed some difference between translatable content and primitive directional quality as such. In the second place, the point is not intended to imply that translatable content always makes a difference in what one might call the appearance of some object present to consciousness. In any event, whatever we might say about states possessed of translatable content, but not of primitive directedness, it would seem to be a condition on any theory of translatable content that it at least account for the fact that translatable content needs to make a difference in how one apprehends an object of consciousness, at least when the object in question is regarded as a “correlate” of that very directedness. My argument will be that this cannot be accommodated by views in which translatable content is, as it were, purely “instantiated” or “tokened” content, and in no further sense ingredient in a state whose content it is. We may regard such views as instances of a “formal” approach to content. It should be clear that functionalist approaches, in the narrower sense of the term, will be instances of formal approaches in this sense.
It may seem that a formal approach to content should have no difficulty conceding that translatable content, in a state that is intrinsically object-directed, necessarily affects the way in which one apprehends an object. This may seem to follow from the fact that, even according to a formal approach, content needs to be tokened or instantiated by the very state to which that content is ascribable. However, the fact that content is tokened or instantiated by a state at most implies that the latter exhibits a feature that it would otherwise not have exhibited. This implies that the ascribability of content makes a difference to the particular state to which it is ascribable. It does not imply that it makes a difference in how some object is apprehended through that state. It does not imply the latter, even on the assumption that the state is intrinsically object-directed. Even on that assumption, the question remains: Why should the fact that a state, which happens to be object-directed, comes to exhibit a feature that it might otherwise not have exhibited have any bearing on the way in which an object is apprehended through it? Of course, one wants to say that any translatable content instantiated or tokened by a state is a special feature of that state. It is just the sort of feature that does not merely make a difference in the quality of states as states. Instead, it makes a difference in a state’s cognitive content. But this only restates the problem. How could there be a feature that is special in this way, if it fails to enter into the fabric of an object-directed state, or to rest on something that enters into that state itself, in something more than the sense that it is tokened or instantiated by it?
Apart from something further, what we have so far learned is only that the “content” of a state is a translatable aspect of it. A number of accounts, again, might attempt to explain the ultimate ground of such translatability. But whatever the ultimate ground, the most we are so far offered is this: that a state possessed of primitive directedness is able to come to token, or is able to come to instantiate, an additional feature of a certain special sort, namely, a feature that manages to “say” something that is in some way translatable. In these terms, our question may be posed in the following form: How can the instantiation or tokening of any such feature, even when it is instantiated or tokened by an object-directed state, amount to that state itself “saying” anything at all, and in particular saying something about its own object?
The answer would be easy if one of the features, as with our first strategy, were merely a determinate form of the other one. It would be easy, for example, if the translatable “content” in a directional state were merely a determinate mode of that state’s quality of directedness in the first place. But we have already seen that this is impossible. The primitive directional character of a state is not a determinable quality of which the specific modes of apprehension are determinates. If it were, then all possible modes of apprehending objects would involve equally primitive qualities, and we could not account for the presence of structure in the contents of mental states.
An analogy may be helpful. On the purely formal approach, we may compare the tokening or instantiation of contents by mental states—or at least by those “intuitional” ones to which we are restricting our attention—to the action of inscribing a text on an arrow. An arrow may really happen to point at an object. When it does, then from the fact that something has been inscribed on an arrow, it of course follows that it has been inscribed on a thing that points at an object. But it is clear that this could not explain how the inscribed text might itself manage to point at an object, in anything like the original sense of pointing. Furthermore, even if the inscribed text should happen to point at an object, or to be “about” an object, the mere fact of its inscription on an arrow could do nothing to ensure that it must be about the same object as the one at which the arrow points. This is because the mere inscription of a text on a surface does not, apart from the purely formal fact of co-presence in it, connect that text with any of the other of the features of the surface. It does nothing of the kind, even if, by supposition, pointing at a particular object should also happen to be a feature of that surface. The question still remains: What could possibly constitute the needed connection, between acts of “pointing” at objects and acts of tokening “contentful” texts, short of construing the latter, impossibly, as a determinate mode of the former in the first place?
It is difficult to avoid concluding that the relationship between the contentful “text” in a mental state and the particular instance of primitive directedness in it must indeed be, or at least be explicable in terms of, the relation between “matter” and “form” in a single state. The conceptualization of any object that is intuitively apprehended must be grounded in the presence of something that is ingredient in the apprehending state, not merely tokened or instantiated by it. If it were merely tokened or instantiated, then whatever the ultimate ground of the translatability of the content in question, it would remain at most a feature among the features of that state. In whatever sense it might be said to be an “object-directed” content, it would in no way share in the object-directedness of the state as a whole. For this we need to suppose, not simply a primitive form of directedness in certain states, but a form of directedness that directs them through a body of material of which those very states are composed.
A generalization of this argument seems to me also to show the inadequacy of certain accounts that do not appeal directly to the translatability of content. Consider, for example, the suggestion that the notion of “demonstrative” content is to be explicated in terms of patterns of a subject’s “sensitivity to evidence” for and against the truth of a content.26 Here, what makes a content demonstrative is not its connection with a bit of phenomenal quality. Rather, it is the fact that it is a content, with regard to the instantiation of which the subject has sensitivities that appropriately parallel the pattern of canonical evidence for and against the truth of a (thereby) corresponding possible demonstrative utterance. The problem with this suggestion is that it is perfectly compatible with the instantiation of demonstrative content while unconscious. To rectify it, we would at least need to be more specific about the medium (as it were) in which the content in question is instantiated. In particular, we would need to specify that it is instantiated in a state of consciousness. But then the argument above applies: the presence of demonstrative content ought to make some difference in a state precisely as a state of consciousness. According to the proposal, demonstrative content would at most be a feature of the state in question, in addition to the feature of its being a state of consciousness.27
I conclude the section with two comments on points on which I have already touched. First, I have assumed for the sake of argument that genuinely translatable content must always be “conceptual” in form. And once we have distinguished translatable content from purely “phenomenal quality,” it may in fact appear difficult to defend any other assumption. However we construe the ingredience of translatable content in consciousness, it may seem impossible to draw a non-arbitrary line between those instances of it that are truly conceptual and those that are not. But the account that I propose to develop also accommodates an important distinction. It allows us to recognize an element of translatable content that—aside from the doctrine of “judgments of perception” as presented in the Prolegomena (and on which I comment in detail in Chapter Five)—is below the level of what Kant himself officially regards as conceptual or judgmental. On the other hand, insofar as it is still a question of something more than intuition itself, and the sensations ingredient in it, we might also continue to insist on its judgmental or conceptual nature. What is crucial is simply to see that, so far as we would at most be dealing with some kind of (albeit non-“sensory”) material in intuition, we would by that very fact still at most be dealing with mere material for concepts and judgments, in the official Kantian sense of the latter terms.
We shall find it useful to regard such contents, apart from their actual forming into Kantian concepts, as comprising concepts and judgments of a more purely “animal” sort. Obviously, the limits of toleration are vague as regards the acceptability of translations for contents in contexts that are unambiguously human. Those for translating would-be animal contents might seem to be even more so. This may lead to hesitation to acknowledge that translatable contents are ascribable to animals at all, whether or not we regard those contents as conceptual.28 Alternatively, one may regard the problems of inter-species translation as unusual only in degree with respect to problems involved in the translation of human contents.29 In any event, there is at least nothing in Kant against the supposition that animals, and human beings too, instantiate contents that are not “conceptual,” in the Kantian sense, but are nonetheless translatable. One advantage of the present approach is not simply that it permits the distinction, but that it also provides an account of the relationship between the two sorts of content in human consciousness: the more primitive is simply material for the more sophisticated.
Our distinctions might also be employed to clarify the notion of phenomenal quality. In the case of a genuinely sensory state, one may of course be inclined to suppose that the presence of such quality is precisely what Kant means to account for by appeal to the ingredience of sensations. By supposition, the “material” of a state of consciousness will always make some difference in the way in which it presents an object (or a would-be object), but not a difference that in turn rests on the way in which the object is conceptualized. By extension, we might suppose that the recognition of such quality in mere imagining would need to postulate some analogue of Kantian sensations. According to the present account, however, there will be no reason to deny that conceptualization itself, through the material that is essentially ingredient in it, is perfectly capable of contributing to the “phenomenal quality” or primitive “feel” of a state of consciousness. This is because, again, the material for the formation of concepts “applied” to intuitions must itself be material, as well, in those intuitions. As I shall propose, it is constituted out of a body of pre-conceptual anticipations and retentions concerning the course of possible experience. As “material” in intuitional states, the material in question might of course be regarded as analogous to sensations. But the analogy need extend no further. In the case of purely imaginative consciousness, then, it seems reasonable to conclude that phenomenal quality or primitive feel is simply reducible to the general character of directedness itself, together with whatever such material is in question. There is no need for special mental “images,” nor for imagistic analogues of sensation of any sort.
This of course requires rethinking the suggestion that mere (Kantian) “sensations” are in fact what account for the phenomenal quality of sensory states themselves. At most they could account for that part of phenomenal quality or feel that is peculiar to sensory states—for example, to actually seeing colors as opposed to merely imagining them. But the total phenomenal quality of seeing involves more. Some may insist that it also involves distinctively conceptual elements. Again, the proposed account permits a middle road. It does justice to our inclination to deny that phenomenal quality, in perception, is completely independent of the ways in which perceptions are “taken.” Yet it permits a sense in which conceptualization always involves a conceptual response to what is independently accessible, purely phenomenally, in perception. The latter will still be so in the sense that the “application” of concepts always involves some kind of forming of a body of anticipations and retentions ingredient as material in an intuitional state. (The requisite forming need not follow the latter in order of time. In most cases, an initial body of material may be responded to in a single “act” that is simultaneously introductive of additional material and appropriately formative.) It seems reasonable to assume that the latter ingredience is what provides the phenomenal quality of conceptualization itself.
An obvious consequence follows from this for the problem of leeway regarding the possible variety of “connection” between particular contents and phenomenal qualities in imagining. It should be clear that, in at least one sense, the present account is required to deny that such leeway exists at all. The ingredience of particular anticipations and retentions in an imaginative intuition simply is, according to the present account, the phenomenal quality of that state. Since that ingredience is at the same time constitutive of a kind of translatable content, it follows that, with respect to that kind, no leeway is possible between phenomenal quality and translatable content. Nevertheless, there is still some leeway to be recognized between the presence of phenomenal quality and the ingredience of genuinely conceptual content in an imaginative intuition. Presumably, there ought to be room for recognizing whatever leeway we might have had in mind earlier, in conceding that a given “image” may be conceptualized in various ways (with appropriate background assumptions). The account can recognize this. Any particular set of anticipations and retentions, ingredient in intuitive imagining, may in turn be regarded as potential material for inclusion in a variety of distinct conceptual contents, all regarded as legitimately “applicable” to the intuition in question. All of the points made earlier, in terms of the notion of taking or conceptualizing “images,” can be recognized here as well, but without incurring a commitment to images as peculiar objects. The solution is simply to recognize that the anticipative and retentive material in an intuitional state, whether sensory or imaginative, is in turn available as material for, but no more than as material for, whatever acts of conceptualization are also applicable to it.30
One may finally say something more about the concept of an “object” and of the related notion of an “appearance.” So far, I have attended to object-directedness as an aspect of consciousness itself. Any objects in question were thus considered only with respect to their status as the intentional correlates of the latter. The main endeavor was to see that this did not entail an objectionable onto-logical commitment. This notion of the objects of consciousness, as correlates of the consciousness of them, will prove crucial in some of the later discussions, most notably concerning association and the “affinity of appearances” (Chapter Four), the two stages of the second-edition Deduction (Chapter Five), and the account of self-consciousness and self-knowledge (Chapter Six). But there must also be a sense in which the objects of consciousness are not, for Kant, mere correlates of the consciousness of them. If there were not such a sense, then there would be no way for him to account for a consciousness of objective reality.
I have argued elsewhere on behalf of a kind of phenomenalist reading of Kant, with respect to the reality of objects of possible knowledge. (One may try to be more tolerant as regards “objects” of suitably abstract conception.)31 It should be clear that this need not be incompatible with regarding objectively real objects as something more than the correlates of one’s immediate consciousness of them. The most obvious way to pursue a phenomenalistic line, compatibly with this concession, would be to regard objectively real objects as a kind of correlate, not of particular states of consciousness, but of a suitably generalized possible consciousness. Kant seems to say as much:
In the mere concept of a thing no mark of its existence is to be found. . . . For that the concept precedes the perception signifies the concept’s mere possibility; the perception which supplies the content [emphasis added] to the concept is the sole mark of actuality. (A225/B272-3)
That there may be inhabitants in the moon, although no one has ever perceived them, must certainly be admitted. This, however, only means [aber es bedeutet nur so viel] that in the possible advance of experience we may encounter them. . . . To call an appearance a real thing prior to our perceiving it, either means that in the advance of experience we must meet with such a perception, or it means nothing at all. (A493/B521)
What may seem impossible is the combination of such an approach, in terms of “possible appearances,” and my contention that the apprehension of objective reality involves the apprehension of appearances—i.e., of correlate-appearances—as objective realities (not, for example, merely as components into which such realities are reducible). The difficulty may seem to lie in my denial that regarding appearances as “correlates” in the first place requires granting them an independent ontological status. If, apart from having been actually identified as objectively real, appearances need not be supposed to possess any being of their own (not even some kind of “mental” being), then what sense could it make to suppose that they ever could be so identified? Denying them an independent ontological status seems precisely to remove the point of speaking of them in the first place. If there are no such items to begin with, then how could one possibly take “them” for really existing objects?
To put the point differently, I have argued that talk of appearances merely as candidates for objective regard is really just a way of talking about the intrinsic directedness of intuitional states. It may seem to follow from this that it is impossible ever to make a judgment (at least in intuition) about anything other than intuitional states. That, of course, contradicts the claim that such judgments can really be about perceivable objects.
It should be clear that this difficulty, if it is one, is not peculiar to judgments concerning the objective reality of appearances. If it has any force at all, it should apply to any sort of attempt to conceptualize appearances. For example, even to take some appearance as of a roughly circular (but perhaps merely hallucinated) shape, whether or not we also take it to be objectively real, will have to involve a judgment that is every bit as “directed” toward that particular appearance as a judgment to the effect that it is in fact objectively real.32 And yet, it may seem, both sorts of judgment are impossible on the terms laid down. On those terms, there really “is” no appearance to make such a judgment about in the first place.
This would be a serious objection against a purely formal approach to conceptual content. According to that sort of approach, as I have argued, there is no useful sense in which the conceptual content present in a state of consciousness might be regarded as part of the way in which the immediately apprehended object of that state is apprehended. At most, according to such an approach, an object-directed state of consciousness might, in some more or less abstract way, manage to “say” something about something or other. Or rather: it might manage to instantiate or token a content that manages to say something. But there is no way that such a state, or its tokened or instantiated content, could be construed as saying something about the very object of that particular state of consciousness. At least this would not be possible, so long as the latter is regarded as the intentional correlate of that state’s intrinsic directedness, that is, of its intuitional “form” as such.
One of the main advantages of the present approach is precisely that it does account for our ability to conceptualize the immediate objects of consciousness—whether as objectively real things or in some other terms—without requiring that we abandon the insight that the immediate objects of consciousness are intentional correlates of the consciousness of them. It is able to do this because, according to that account, concepts themselves (at least non-categorial ones) are originally formed out the very material through which one is able to apprehend objects in intuition in the first place. As I shall eventually argue in Chapter Six, what specifically conceptual form adds to any such mode of apprehension is simply a kind of embedding of that same material, by means of the “forms of judgment,” in a higher order consciousness of a special sort.
Insofar as the material in question is not merely externally associated with an intuition, but is as much ingredient in it as sensations ever could be, when one apprehends appearances through them, then we would at least have an explanation of how concepts might be “applied” to immediately apprehended objects, and how those objects might be conceptually “taken” in this way or that, without having to abandon our notion that the immediate objects of consciousness are intentional correlates of the apprehension of them. The account will be phenomenalistic, because, as suggested, the basic material of such conception will be at bottom nothing other than a manifold of anticipations and retentions concerning the course of possible experience. But the account will also be compatible with the recognition of objective reality on the part of at least some intuitional objects. This is so because the corresponding judgments are not about one’s own intuitions, but are every bit as object-directed as the intuitions to which they are applied. (Obviously, the suggestion requires that we rethink what is to be meant by “application” of a concept in the first place.) The anticipations in question are, by supposition, anticipations concerning future possible appearances. (The retentions, which we may ignore for now, might be regarded as “anticipations” concerning past possible appearances, that is, anticipations of what might have been perceived.) But the judgment in question is not reducible to those anticipations. Rather, it is formed out of them. In turn, those anticipations are a part of the “material” through which an appearance is apprehended. That is, they are a part of the material through which the form of intuition originally operates in some case. In virtue of this, the judgment in question is precisely about that apprehended appearance. According to the account in question, that is just what it is for such a judgment to be “about” such an object.33
It should also be clear that, according to the account proposed, the fact that immediately apprehended appearances are, as such, the “intentional correlates” of the apprehension of them is compatible with conceding that—qua identifiable as objectively real—those appearances have qualities that are different from, and even incompatible with, the qualities that they immediately appear to have. We may simply regard this as a case in which the anticipations embodied in the predicated concept are anticipations of sets of future possible appearances (relative to various possible conditions for their apprehension) that are not the sorts of appearances that one would normally be led to anticipate, solely on the basis of the “given” appearance—insofar as the latter is apprehended merely through whatever sensations are in question. (In one sense, of course, the fact that appearances are apprehended “through” certain anticipations must have some kind of counterpart in those appearances as well, qua intentional objects. But all that this shows is that the notion of an intentional object, hence of an intentional “correlate,” needs to be considered with respect to more than one level of analysis.)
As for whether a given judgment is also true of a given appearance, qua objectively real object: according to the proposed account, that would have to be a matter of whether or not the anticipations that the judgment embodies are actually satisfiable (or, perhaps, of whether or not they would be, in some eventual “long run,” justified).34 In any given instance, some of them are likely to be, and others, not. (Of course, Kant argues, in the “Refutation of Idealism,” that at least some of them must be taken to be in fact satisfiable. For at least some given appearances must be taken to be such objectively real objects.) But that need not pose a problem for our account. It simply requires the ability to characterize the sorts of anticipations that are specifically relevant to the minimal reality of an apprehended object. (In fact, our proposal does not even rule out recognizing one or more senses in which one may be said to apprehend really existing objects quite apart from any question as to the anticipations ingredient in apprehension. If one favors a causal theory, for example, one might stipulate that under certain conditions, involving the causal origin of an intuition, that intuition is to count as the perceptual apprehension of a really existing object. But that sort of question is not the one that interests us here.)
There is a final advantage in the proposed account, with respect to the sort of “phenomenalism” that it supports. It allows us to avoid the problem of circularity that may appear to affect varieties of phenomenalism. The apparent difficulty is this. Such views seem to demand a phenomenalistic analysis of judgments regarding ordinary objects. Such an analysis, presumably, will involve judgments to the effect that certain sorts of perceptions are or would have been obtainable, given satisfaction of the appropriate antecedent conditions. But this leads to circularity, because those latter judgments in turn require phenomenalistic analysis, and the only plausible candidates for it will eventually need to mention material conditions among the “appropriate” conditions originally in question. For example, they will need to mention such things as the sorts of perceptions that would be obtained were one’s eyes oriented in a particular direction. Thus we will not have succeeded in reducing judgments about objects to judgments about possible perceptions.
The response to this objection is that the proposed account does not propose the possibility (not even in principle) of reducing judgments about material objects to judgments about possible perceptions. According to the proposed account, judgments about material objects are regarded as formed out of anticipations (I continue, for the time being, to ignore the role of retentions) concerning possible perceptions. In turn, the anticipations in question must be “conditional” ones: anticipations of possible perceptions relative to the satisfaction of equally anticipable conditions. But this does not imply that those anticipations involve judgments, or even conceptions, in regard to those conditions—or at least not that they do so in the relevant sense of those terms. In the relevant sense of the terms, the proposal maintains that the anticipations in question are merely the material out of which judgments and conceptions proper are formed. (Clearly, nothing prevents describing such material as itself involving judgment or conception as well, but in some different sense of the terms.)
It is equally clear that any attempt to give linguistic expression to the material in question will require describing what is anticipated, and thus will require conceptualizing the latter in some particular way. There is no harm in conceding that this, in turn, requires concepts that are in some way already dependent upon one’s conception of material reality. But describing or communicating a set of anticipations, and forming them (or having them formed) into modes of conception in the first place are hardly the same thing. So the required appeal to material conceptions does not support a charge of circularity. (As we shall see in Chapter Six, any complete description of the relevantly anticipated “appearances” would require describing those appearances as appearances eventually perceivable by oneself as well; but this is compatible with maintaining that one’s original self-concept is in turn nothing other than the concept of whatever subject of experience is reflected, as an intentional correlate, precisely in such anticipated appearances, qua anticipated.)