Two Perspectives on the Object of Knowledge

 

by Stephen Palmquist (stevepq@hkbu.edu.hk)

 

 

 

Reason proceeds by one path in its empirical perspective, and by yet another path in its transcendental perspective [Kt1:591].

 

1. Kant-s Six -Object-Terms-

 

      Kant-s use of the word -object- (Objekt or Gegenstand) is a potential source of much confusion and ambiguity. Sometimes he uses it in a broad sense, either nontechnically to refer to an ordinary -thing- encountered in im­mediate experience, or technically to refer to anything which stands in some potential, actual or necessary relation to the knowing subject. At other times he uses it in a narrower sense to refer to an object in general as it is viewed at one of several stages in the knowing process. Consequently, its meaning is not always evident when Kant uses it without a qualifying adjective [cf. B20: 76 and G12:778]. The first step to coping with this situation is to rec­ognize that he explains the role of the object in his theory of knowledge (i.e., in systemt) primarily by implementing six technical -object-terms- (as I shall call them). A clear understanding of these special terms and of their perspec­ti­val interrelationships will provide an interpretive framework for understand­ing Kant-s use of the word -object- when it appears on its own.

 

      Throughout Kt1 Kant makes frequent use of three object-terms: -thing in itself-, -transcendental object- and -appearance-. Because he does not describe their relationships univocally, his inter­preters and critics have made numerous proposals as to how these terms should be treated. Oftentimes they are de­fined in conjunction with one or more of three other, less used object-terms (viz., -phenomenon-, -negative noumenon- and -positive noumenon-), which Kant formally intro­duces towards the end of the Analytic.[1] These six terms, to which I will limit my attention in this chapter, compose one of the most ob­scure, yet most fundamental, conceptual networks in Kant-s entire System. Of course, the Copernican Perspective, upon which Critical philosophy is based, emphasizes the way expe­rience is constructed a priori by the subject. But by establishing here a framework for understanding Kant-s conception of the object, a great deal of complexity can be avoided in Part Three, so that we will be fully prepared to concentrate there on the various conditions imposed on experience by the subject in each of Kant-s three Critical systems.

 

      Most interpreters take two or more of Kant-s six object-terms to be syn­onymous. Certainly the most commonly equated terms are -appearance- with -phenomenon- and -thing in itself- with -noumenon-; but other syno­nymies have also been suggested.[2] Among those who admit the possibility of dis­tin­guishing between one or another of these pairs, many would nevertheless agree with Weldon that the difference -is not of any great importance- [W16: 107]. Such tendencies reflect a widespread neglect of what should be re­garded as an important fact when interpreting Kant-s usage: that he does not make any significant use of the second set of object-terms until his elabora­tion of the main tenets of systemt (using only the first set) is nearly complete [see note VI.1]. In light of the considerable effort Kant put into constructing his System according to an architectonic plan, in relation both to the particu­lar terms in which it is expressed and to its general outline [see Kt1:xxxvii-xxxviii and III.3-4], it seems likely that such a sudden change of terminology could serve as a -clue- of some sort to help us understand his theory. The in­troduction of a new set of object-terms in the middle of Kt1 would be redun­dant unless their meanings were designed to be distinguished from those of the original set.

 

      Although some commentators do distinguish between all or most of Kant-s object-terms [see Ap. VI], I have been unable to find any who try to account for the gap (which is rarely even acknowledged to exist) be­tween his introduction of the two sets.[3] This neglect, though unfortunate, is only a minor problem, however, since the task of determining what Kant means by each term is fundamental, and is not necessarily dependent on an understand­ing of why he introduces them in the order he does. Indeed, the main purpose of this chapter will be to determine the extent to which each term can be given a distinctive meaning without misrepresenting Kant-s in­tentions; the explanation of the late introduction of the second set of terms will be signifi­cant only insofar as it aids in this task.

 

      The conceptual key which can, as I have argued in Chapters II-V, unlock many of the complexities, ambiguities and apparent inconsis­ten­cies in Kant-s System is the -principle of perspective-. By recognizing its thoroughgoing in­fluence, especially as applied to the fundamental distinction between the tran­scendental and empirical perspectives in systemt (to which the object-terms have their primary application), it should be possible to render his the­ory of the object of knowledge at least intelligible, if not to establish its va­lidity. As explained in IV.3, adopting an empirical perspective in one-s search for knowledge involves examining particular experiences in order to determine what -is true- (synthetic a posteriori) about them, without necessarily distin­guishing between the subject and object of knowledge. But adopting a tran­scendental perspective involves carefully distinguishing the subject from the object and examining experience in general in order to determine what -must be true- (synthetic a priori) about it-i.e., in order to determine the conditions of knowledge which make experience (i.e., the -empirical knowledge-) possi­ble [Kt1:28-9]. Since Kant-s theory is couched in these -radically epistemo­logical- terms [A10:44], his discussion of -different objects- should be inter­preted as referring not so much to ontological distinctions [cf. P14:74-5] as to -different perspectives- on one object, encountered in ordinary (immediate) ex­perience. Thus, as Allison puts it, -the general distinction between the tran­scendental and the empirical...is not intended as a distinction between two kinds of being, but between two perspectives or manners in which we can consider one and the same thing- [A6:194; s.a. B20:29].

 

      Throughout Kt1 Kant frequently discusses the implications of the tran­scendental and the empirical perspectives together in the same passage. This gives the impression that all such passages are equally concerned with estab­lishing the elements applicable to both sides of this perspectival distinction. In one sense this is quite true. For it is possible to pinpoint transitions from one perspective to the other in specific passages, such as Bird does in B20:14 with regard to Kt1:A121. Moreover, as we shall see in Chapter VII, Kant does introduce both em­pirical and transcendental elements at virtually every step in systemt. Nevertheless, as we saw in III.4, each of Kant-s systems is divided into four -stages-, each of which is defined by a specific perspective [see Figure IV.2], and the shifts between these stages can be correlated quite closely with four of the major divisions of the Transcendental Doctrine of Elements [Kt1:33-732].

 

      Kant starts by focussing in the Aesthetic on the transcendental perspec­tive. For, although he asserts that objects are both -ideal- when viewed from the transcendental perspective and -real- when viewed from the empirical perspective, the primary goal of his exposition is to elaborate the basis and implications of the former claim. Thus, as Bird hints in B20:47-51, the Aesthe­tic more or less assumes the empirical reality of space, time and ob­jects, but argues for their transcendental ideality. The first shift occurs in the transition from the Aesthetic to the Analytic of Concepts. In the latter sec­tion Kant takes a step back from the perspectives concerned with transcenden­tal ideality and empirical reality and adopts a logical perspective, from which ob­jects are viewed as relating not to sensibi­lity-not to time and space-but to the understanding and its concepts. Thus in this section the transcendental and em­pirical perspectives are more or less of equal im­portance, both being sub­ordi­nate to the main goal of estab­lishing the status of the categories as -pure con­cepts of understanding-.

 

      In turning from the Analytic of Concepts to the Analytic of Princi­ples, Kant is modulating from the logical to the empirical perspective. This is ev­i­denced first by the reintroduction of time in the Schematism chapter and then by the reintroduction of space in the chapter on the Principles (especially in the second edition, where it includes the revised Refutation of Idealism). The Analytic of Principles focuses primarily on the empirical perspective in the sense that its primary target is the description of the synthetic a priori princi­ples which determine objects in the ordinary physical world to be empirical objects (a task quite different from that of the Aesthetic, which seeks to estab­lish the synthetic a priori -form of appearances-). The reader of Kt1 who fails to realize that Kant intends the -objects- dealt with in the Refutation of Ideal­ism and throughout the Analytic of Principles to be regarded as empirical objects [cf. E3:107n] is bound to misunderstand the significance of Kant-s arguments in these sections.

 

      Finally, having completed both the empirical and the transcendental sides of systemt, Kant modulates in the Dialectic from the empirical to the hypo­thetical perspective, where he considers the illusions resulting from the specu­lative attempt to adopt a quasi-logical perspective, which is actually a confla­tion of the transcendental and empirical perspec­tives. Both the dogmatic acceptance and the skeptical rejection of numerous speculative knowledge-claims are shown to result from an improper confusion of these two key perspec­tives. The alternative is to view such notions as God, freedom and im­mortal­ity as practical ideas which serve the end of unifying knowledge, rather than as legitimate transcendental and/or empirical objects of knowledge. Thus systemt ends by giving equal weight once again to the transcendental and em­piri­cal perspectives, for both are subordinated to the hypothetical perspective, in preparation for adopting the practical standpoint in systemp.

 

      One way of assessing the plausibility of this overview as an ac­count of the perspectival organization of Kt1 is to compare the relative frequency with which Kant uses the terms -transcendental-, -logical-, -empirical-, -hypotheti­cal- and -practical- in the Aesthetic, Analytic of Concepts, Analytic of Princi­ples and Dialectic of Kt1. Table VI.1 uses Pq10 to make just such a compari­son. The results, predictably, are not altogether conclusive, because of the above-mentioned fact that Kant often develops the transcendental and em­piri­cal im­plications of his theories side by side. Moreover, the five key words are not al­ways used in their strict, perspectival senses, so the most we can expect out of such a statistical analysis is a rough idea of Kant-s tendencies. Neverthe­less, the ten­dencies revealed by the table confirm that the correlations I have made are at least possible, and in some cases even probable.

 

      The following conclusions (in order of probability) can be drawn from the table. (1) The Transcendental Doctrine of Elements undoubtedly em­phasizes the transcendental-empirical distinction more than the logical-hypotheti­cal (or -practical) distinction: the former pair of terms is used over seven (or six) times more frequently than the latter. (2) If -hypothetical- and -practi­cal- are to be specially associated with the perspective assumed in any one part of the Doctrine of Elements, it must be with the Dialectic; for all but seven of the 63 occurrences of these two words come in this section-by far the highest percentage of occurrences on the entire table. (3) The -logical- cannot be associated with the Aesthetic, but is used significantly in the other three sections. Of these, it seems to be associated most close­ly with the Analyt­ic of Concepts, where it is used half again as many times per page as in the two subsequent sections. Moreover, it has a higher percentage of use in the Analyt­ic of Concepts than any of the other terms in that column. (4) Be­cause of the ubiquity of -transcenden­tal- and -empirical- (each occurring an aver­age of more than once per page), they are the terms most difficult to 

   Table VI.1: Frequency of Perspectival Labels in the

Transcendental Doctrine of Elements (TDE) of Kt1a

NOTES:

aThe boxes with thick borders indicate the set of figures which should be most significant for each column and row, in order for Kant-s usage of these key terms to support my overview. (Page numbers refer to Kemp Smith-s translation of Kt1.)

bThe average numbers per page provide an accurate means of assessing the relative importance of a given word in different sections (as listed in the rows), because the dif­ferences in the number of pages in each section are in this way taken into ac­count. However, these figures are not as useful for comparing figures in a column, because the five terms differ widely in their total numbers of occurrences.

cThe percentages provide an accurate means of assessing the relative importance of all five words in a given section (as listed in the columns), because the differ­ences in the overall usage of each word are in this way taken into account. However, these figures are not as useful for comparing figures in a row, because the four sections differ widely in the number of pages they occupy.


asso­ciate with any one sec­tion in particular. The most sig­nificant difference between their fre­quencies of use comes in the Analytic of Principles, where -empirical- is used nearly half again as often as -transcendental-. The per­cent­age of the former is also higher, though the per­centage of -logical- in this sec­tion sur­passes both-a reflection of the close connection between the two parts of the Analytic [cf. note VII.8]. (5) Finally, the frequency of -transcen­dental- and -empirical- in the Aesthetic is almost identical. But be­cause the former is in general used by Kant slightly less frequently than the latter, its percentage of total usage ends up being slightly higher (5.1% as compared with 4.7%). Thus Table VI.1 supports all four of the proposed cor­relations between per­spectives and sections of text-though admittedly with varying degrees of probability.

 

      In Chapter VII I shall examine in detail the extent to which the theories elaborated in Kt1 support this interpretation of the perspectival modulations in systemt. If it is accurate, then it would seem reasonable for Kant to intro­duce a new set of object-terms at some point in the Analytic of Principles. This would help the reader to know whether the -object- referred to at any given point is meant to be viewed transcendentally or empirically. As it turns out, this is precisely what he does: -thing in itself-, -transcendental ob­ject- and -appearance-, I propose, refer to various ways of considering the ob­ject from the transcendental perspective, while -phenomenon-, -negative nou­men­on- and -positive noumenon- refer to parallel ways of considering the object from the empirical perspective. The latter set is introduced at the end of the Ana­lytic of Principles, once the synthetic a priori rules governing the empiri­cal perspective have been fully established. Moreover, the Appendix [Kt1: 316-49] added after the chapter on Phenomena and Noumena emphasizes the perspectival nature of his theory as it is worked out to that point [cf. note VI.3]. What more appropriate place could there be for these sections than just before the Dialectic, where Kant always assumes the two perspectives are fully compatible?

 

      This perspectival interpretation of the progression of Kant-s argument in Kt1 seems far more plausible, and accords to him a great deal more integrity as a philosopher, than the alternative interpreta­tions which tend to depict him as unknowingly propounding contradictory and incompatible theories in the Aesthetic, the Analytic and the Dialectic. For the purposes of this chapter, I shall therefore adopt it as a tentatively valid hypothesis. In accordance with its description of the general movement of Kant-s thought from the transcen­dental to the empirical, I will examine the terms applicable to the former perspective in VI.2, and those applicable to the latter in VI.3.

 

2. Kant-s Transcendental Object-Terms

 

      Kant-s transcendental perspective in systemt is concerned not with partic­u­­lar objects but with -objects in general-. When an empirical act of knowing is viewed from the transcendental perspective, it is there­fore described in terms not of specific facts, but of a knowing subject -repre­senting- to itself an unknown thing in the form of a -representa­tion-. This unknown thing, Kant ar­gues, cannot be represented as it is in itself, because its ori­ginal representa­tion must stand in some rela­tion to the person perceiv­ing it; that is, a thing in general, regarded from the transcendental perspective, must become an ob­ject of experience in order for it to be known by a subject.[4] The original rep­resentation of an unknown thing, which in this form is nothing but a com­pletely indeterminate -something in general=x- [e.g., Kt1:A104], is called -the transcendental object- [e.g., 236,A109]. Bird observes that -the un­known- as a descrip­tion of the transcendental object -is not a reference to any mys­te­rious realm of intelligible objects- [B20:79]; rather it re­fers to the general epistemo­logical requirement of -having experience-. Thus it presents -a philosophical task to solve- [79]: viz., how does the unknown -x- (an object of possible ex­perience) come to have the status of a real object of empirical knowledge? The steps from the former to the latter point, and the rules governing them, will be examined in full in VII.2-3. For now it is enough to point out that to view a subject-s experience of a real empirical object from the transcenden­tal perspec­tive is to -posit- the object as existing -in itself, beyond the [mere] thought of it- [Kt1:667t.b.]-i.e., to view it as a transcen­den­tal ob­ject which is believed to be in some way based on the unknowable -thing in itself-. The latter is Kant-s term for an object as considered in its original, unrepresented, and so -transcendent-, form: -The true correlate of sensibility, the thing in it­self, is not known, and cannot be known, through [any] representations; and in experience no question is ever asked in regard to it.-[5]

 

      The problem which inevitably arises for the interpreter of Kant is that, if the thing in itself is unknown and unknowable, it would seem to be difficult, if not impossible, to defend Kant-s use of it as the transcendent starting point for systemt. One reaction to this diffi­culty is to conclude that -Kant was sim­ply mistaken- [W3:3], and that the thing in itself is knowable after all; but, as I demonstrate in Appendix V, this approach is radically untenable, both in itself and as an interpretation of Kant. Another reaction is to drop the thing in itself altogether, regarding it as superfluous [see e.g., S17:38-42]. However, this phenomenalist approach, though a coherent alternative, does not solve the problem, but merely ignores it or denies its existence [see Ap. VI]. As an alternative to these approaches, I have argued in Chapter V that faith is an adequate justification for our initial assumption of the thing in itself; but in order fully to understand its function (and hence, its legitimacy), it is still necessary to present a clear and coherent interpretation of the System to which it gives rise. We can now begin this task by demonstrating that Kant-s theory of the object of knowledge is self-consistent, even when it comes to his assumption of the thing in itself.

 

      Much of the unwillingness of Kant-s interpreters to accept his starting point results from a common misunderstanding about what the unknowability of the thing in itself entails. When Kant denies the knowability of the thing in itself, he is not denying empirical knowl­edge, for the thing in itself does not directly -affect- the senses: the af­fection of a subject-s sensibility by an object occurs only in immediate experience.[6] (Hence empirical knowledge of the thing in itself would be a contradiction in terms.) In fact, from an empir­i­cal perspective the objects of everyday experience (of which we certainly do have knowledge) can be regarded (nontechnically) as things in themselves [see VI.3]. But in its technical sense, the concept of the thing in itself is implied by a limitation which arises once a person reflects on affection from a tran­scendental perspective. This limita­tion is that transcendental knowledge of the conditions and sources of empirical knowledge is not possible unless one re­gards the object as an object of experience. Accordingly, it is the possibility of gaining transcendental knowledge of the thing in itself which Kant denies. Empirical knowledge is based on our experience of real objects, but transcen­dental knowledge is based on our reflection on the transcen­dental object (which is itself empiri­cally unknowable [Kt1:641-2]).[7] Considered in its transcendent state, however, as a thing in itself, even its existence is open to legitimate doubt. (This, I believe, is Kant-s concession to the skeptic [see V.3].) But once it is accepted as a valid assumption, its unknowa­bility is self-evident and unproblematic [see A9:319-20]. For the unknowability of the thing in itself simply means nothing can be known about empirical ob­jects if they are considered apart from all relation to the forms by which a knowing subject experiences them.

 

      Unfortunately, Kant-s doctrine of the transcendental object is no less ob­scure than his doctrine of the thing in itself. The former is certainly closely related to the latter-indeed, so closely that the two are sometimes wrongly in­terpreted as merely synonymous terms. Ewing, for example, follows Vaihinger and Kemp Smith in claiming that -it is hardly possible to doubt that the transcendental object is identified with the thing-in-itself- [E5:101; cf. W2:107]. Nevertheless, others interpret it as related more closely to the de­terminate forms of the object-s represen­tation in experience [see Ap. VI]. For example, Paton suggests (with some uncertainty [P2:2.449]) that the tran­scen­dental object should be identified with the concepts produced by the unity of apperception [443f]. But Paton-s dubious interpretation of Kt1:A250 ig­nores the fact that Kant says the transcendental object is the -correlate of the unity of apperception- [e.a.], and that it -cannot be separated from the sensible data- [A250]. Kant-s point, then, is that the transcendental object does for the object what the unity of apperception does for the subject: each provides in its own way the basis for the unity of experience.

 

      If it can be identified neither with the thing in itself nor with any of the determinate forms of an object-s representation, then the transcendental object must be regarded as a preconceptual, nonsubjective (but none the less tran­scendental) element in systemt. Kant implies this in a reference to reason-s pursuit of an unconditioned object, which can also be applied to the relation between the thing in itself and the transcendental object:

 

reason relentlessly seeks the unconditionally necessary and sees itself com­pelled to assume it [cf. the thing in itself], though it has no means by which to make it comprehensible and is happy enough if it can only discover the concept [viz., the idea; cf. the concept of the transcendental object] which is consistent with this presup­position.... [Even though] we do not indeed com­prehend [it,]...yet we do comprehend its incomprehensibility...[8]

 

Along these lines Buchdahl argues in B27:77 that the transcendental object -is not the sort of thing which as such has or lacks a constitu­tion but is always the aspect of the object still awaiting acquisition of such a constitution via a process of realization [i.e., via intuition and conception].-

 

      Findlay also puts forward a plausible account of these object-terms when he says Kant

 

mainly uses the term -Transcendental Object- when he conceives of such ob­jects...as being what we have to conceive as being the underlying, unknown ground of appearance and experience; while the term -Thing-in-itself- is mainly employed when he conceives of them as existing independently of whatever we may conceive or believe. [F3:3; cf. Kt1:A190-1]

 

Similarly, Allison says Kant-s purpose in introducing the transcendental object is to show that the thing in itself -for us can be nothing more than a mere something=x- [A10:60n]. Indeed, only when the thing in itself is regarded as a transcendental object (i.e., as an element of systemt, given in im­mediate expe­rience) does its -existence- take on a literal meaning for us [cf. Kt26:(68)]. Yet in spite of these clear distinc­tions, both Findlay and Allison tend to use the terms indiscriminately [e.g., F3:16-24; A10:69; s.a. Ap. VI].

 

      A metaphor which can help us avoid such inconsistency is to picture the mediating function of the transcendental object as a doorway between the thing in itself and the subject: although it is directly related to both, it can­not be wholly identified with either. As such, it refers to an object, consid­ered transcendentally, as it would be just at the point of being apprehended by a subject (hence, not to the thing in itself), but without taking into account the subject-s determining influence. Only this object is -given-, though not as a determinate object of -experience- [Kt1:522-3,642]. Thus Kant equates -an ob­ject in general- (i.e., a transcendental object) with -an object as it may be given in intuition- [411ne.a.]- an option denied to the thing in itself. Since the transcendental object bridges the gap between the indetermin­able thing in itself and the determinate forms of a represented object, it is not surprising that Kant regards it as similar at times to the thing in itself and at other times to the determinate object.[9]

 

      Once we have a clear understanding of Kant-s initial assumption that the thing in itself can be represented to a knowing subject only as a transcenden­tal object, it becomes easier to understand his elaborate theory of how the transcendental object comes to be represented in a determinate form through intuition and conception. Most of the details of this theory will not concern us until Chapter VII; but Kant does make use of yet another object-term, which we must examine if our analysis of his transcendental perspective on the object is to be complete. That term is -appearance-.

 

      In Kant-s primary transcendental sense appearances are objects intuited in time and/or space by a subject [Kt1:A384]. Since -the transcendental object of our...sensible intuition gives [such] intuition- [585], the resulting appear­ances must in some way -conform to- the former [522-3]: indeed, -the tran­scendental object [lies] at the basis of appearances- [641; s.a. A104-10], which in turn reflect -the way in which our senses are affected by this un­known something- [Kt2:314-5; s.a. B20:78]. Inasmuch as the -transcendental appear­ance- (as I shall call it) is -grounded- in an unknown something which exists independently of the subject [Kt2:314,354; cf. B20:4-5,76], the tran­scendental object (and so also the thing in itself) can be regarded as -the cause of appear­ance-.[10]

 

      Notwithstanding their transcendental relationship with the thing in itself, such appearances -constitute an object which is merely within us- transcen­den­tally [Kt1:A129]: from the transcendental perspective these transcendental appearances -do not exist in themselves but only relatively to the subject- [164] as -possible perceptions- [246; cf. A250,A375n,527]. Kant defines an appearance as an -undetermined object of an empirical intuition- [34], and ex­plains that we -have no knowledge of any object as thing in itself, but only in so far as it is an object of sensible intuition, that is, an appearance-.[11] The result is that the one aspect of an object viewed from the transcendental per­spective which we do encounter in experience-viz.,the transcendental ap­pear­­ance-is -ideal- rather than -real-(hence the name -transcendental ideal­ism-[12]).

 

      Kant employs the term -appearance- in several other ways as well, which I will expound briefly in hopes of dispelling any tendency to regard them as incompatible. Applying the principle of perspective to his use of this term will reveal that its meaning is not -incoherent- or -absurd-, nor is it -non­sense- arising out of a -verbal gymnastic-, as some commentators have sug­gested [S8:43; P14:75-6; H8:19; s.a. E3:103], but is merely a reflection of the complexities in­herent in a rigorous investigation of the epistemo­logi­cal un­derpinnings of human experience. The word-s primary sense, as por­trayed above, is as a tran­scendental term employed in the context of the tran­scenden­tal perspec­tive. Another equally straightforward sense is as an empiri­cal term employed in the context of an empirical perspective. Kant clearly dis­tin­guishes in Kt1:69-71 between an ordinary empirical appearance and his special transcendental type: the former is a partial or even illusory perception, whereas the latter is that aspect of a real empirical object which, being -inseparable from the [act of] representation of the [empirical] object, is not to be met with in the object [=thing] in itself, but always in its [perspectival] relation to the subject- [70n; cf. B20:51,192]. Likewise, in Kt69:269 he dis­tin­guishes between a thing-s -physical appearance- (Apparenz) and appearances as -Phenomena- [see VI.3]. But Kant uses the term -appearance- in the former, empirical sense only on rare occasions [e.g., Kt1:428].

 

      The trap into which the interpreter must be careful not to fall is to think these two senses of -appearance- provide an exhaustive account of Kant-s usage. On the contrary, two other, more subtle senses are discernible. First, an empirical appearance can be regarded from a tran­scendental perspective. The appearance would then be nugatory, since illusions play no part in con­stitut­ing empirical knowledge (as transcen­dental appearances do). Consequently, Kant mentions this possibility only in order to pass it off as insignificant, or to warn against adopt­ing an incorrect epistemology.[13] But the second subtle use plays a cru­cial role in systemt: a transcendental appear­ance can be regarded from an empirical perspective. In this case appear­ances are the things and ob­jects which make up the natural world of our everyday experience.[14] Only when this transcendental term is viewed from the empir­ical perspec­tive, then, is Grabau correct in suggesting that -for Kant appear­ance in experience is the mode in which we grasp reality- [G12:771]; for when it is viewed from the transcendental perspective it implies our inability to grasp reality. Likewise, only from the empirical perspective is it proper to say -the domain beyond the sphere of appearances is not full of things in themselves, but is empty- [775; cf. Kt1:310; Kt2:361; see V.2]; for the tran­scendental appearance, viewed from the empirical perspective, is a -real- rather than an -ideal- constituent of knowledge (hence the equally valid name, -empirical realism-[15]).

 

      Neglecting the difference between the two perspectives from which a transcendental appearance can be viewed will almost always render an in­ter­preter-s account of Kant-s theory inadequate. Barker, for instance, sees the two ways Kant has of talking about appearances, but, ignoring their perspectival relationship, regards them as -by definition mutually exclusive- [B2:281]. He accuses Kant of unknowingly contradicting him­self [289], inasmuch as he -decisively embraced neither the [empirical] theory of appearing nor the [transcendental] theory of appearance, but oscillated between them without recognizing any need for making a choice- [283]. But when the principle of perspective is taken into consideration, we can see that no -choice- has to be made, since both theories have their own valid sphere of application. This in­terpretive key enables us to reconcile such claims of Kant-s as that appear­ances are -in us- (i.e., transcendentally) even though they exist as objects of perception -outside us- (i.e., empirically) [cf. Kt1:59 and A373].

 

      Kant-s varied use of such an important term inevitably gives rise to some confusion. The foregoing account of the four ways in which -appearance- can be used is an attempt to dispel some of this confusion. It may be helpful at this point, therefore, to summarize the results:[16]

 

(1)  A transcendental appearance viewed from the transcendental perspec­tive is a representation in the knowing subject which, though unde­termined, is ready to be assimilated into a perceptual experience. This is an appear­ance in Kant-s primary technical sense.

(2)  An empirical appearance viewed from the empirical perspective is an or­dinary perceptual illusion. Kant uses the word in this nontech­nical sense only occasionally.

(3)  To view empirical appearances from the transcendental perspective would be to attempt to include perceptual illusions in an explanation of the possibility of experience. Kant explicitly warns against this use of the word.

(4)  A transcendental appearance viewed from the empirical perspective is a real object of empirical knowledge. This is an appearance in Kant-s sec­ondary technical sense.

 

Fortunately, Kant helps to reduce the chances of confusion by introduc­ing, at the end of his discussion of the conditions and rules governing the empirical perspective, a new object-term denoting an appearance in the fourth sense. This new term, -phenomenon-, is the first member of the empirical set of object-terms to be examined in VI.3; but before turning to this task, I will attempt to clear up some possible misconceptions re­garding the thing in itself and its relation to the other transcendental object-terms.

 

      Many interpreters who ignore the perspectival relationship between Kant-s object-terms interpret them as referring to things which are somehow actually different objects [e.g., K3:217-8; P14:75-6; S17:90-1]. Such an in­terpretation finds support in Kant-s occasionally careless use of the word -object- [see VI.1]. For he sometimes calls the thing in itself an -object- [e.g., Kt1:70n] even though-inasmuch as it is -in itself- before it is repre­sented to a subject-it is strictly speaking no more than a possible object (i.e., a thing which might become an object through representation [cf. G12:778]). But in­terpreting the thing in itself as an object which transcends the object given in experience, yet stands in some sort of -causal- relationship to it, creates the insurmountable problem of explaining what this means without assuming a transcendent employment of the categories, which would contradict Kant-s own Critical principles [see VII.2.B and B20:75-6].

 

      Fortunately, the perspectival interpretation of Kant-s System re­veals a more coherent picture. The various stages in the determination of an object of knowledge all result from the taking up of some perspec­tive by the sub­ject; the thing in itself, however, is independent of any perspective, because it is not actually viewed by the subject at all [S8:42]. Kant has this in mind when he speaks of the thing in itself and its representational determinations as two equally legitimate ways of referring -to the very same thing- [Kt2:344; s.a. Kt1:xxvii and G12:771n]. In relation to the -appearance-, -the thing in itself...is not itself a strange [absonderlicher] object, but only a special [besondere] perspective (respectus) for constituting [the thing in] itself as an object- [Kt9: 22.44; cf. A3:653; B27:97; H11:37-9]. In this sense we can say that -the thing in itself is...given in its appearance- [G2:471]. The latter is -a thing considered in a certain relation, in virtue of which it falls under a certain description-, and the former is -the same thing, considered in abstraction from this relation, and therefore as not falling under this description- [A10:54]. Those who ignore or overlook Kant-s principle of perspective, and continue to refer to the thing in itself as an object, often end up using the term in a way closer to Kant-s use of -phenomenon-, in which case (as we shall see in VI.3) the object under consideration is knowable in principle.

 

      When the thing in itself is interpreted from the transcendental perspective as referring to an empirical object insofar as it is not an object of any sub­ject-s knowledge, it becomes a more reasonable (indeed almost a trivial) notion to maintain; for, whatever else might be required for the possibility of empirical knowledge, we obviously cannot know a thing empirically without encountering it as a represented object. -To state that we know only appear­ances and not things in themselves is-, as Schrader rightly asserts, -to state an obvious tautology, namely that objects are known only as they are known- [S4:173]. Moreover, if the thing in itself were interpreted in this way and yet rejected, it would be impossible to give any plausible account of the source of -the empirical differences in shapes and sizes- of the objects of everyday expe­rience [P2:1.139; cf. E5:191-2]. Only what is common to all representations is, for Kant, supplied a priori by the subject in the act of representing. So the thing in itself must be posited and assumed to determine in some sense the raw material for any possible object of knowledge.[17] In other words, a thing other than us must be represented to us in a perceivable form, otherwise we could never become conscious of it; and this, Kant assures us, -would practi­cally amount to the admission of [its] non-existence- [Kt1:A117n].

 

      In accordance with this interpretation, Kant-s claim that things repre­sented in the form of transcendental appearances or of the tran­scendental object also exist apart from the subject in the unrepresented form of the thing in it­self can be understood (and hence accepted) as an argument against Berkeleyian idealism. Berkeley-s view-sometimes attributed to Kant, usually by those who reject the thing in itself [e.g., S17:242,246]-is that the existence of a thing is due only to the nature of the knowing subject (possibly with some help from God). But Kant-s view is that a subject-s knowledge of a thing, and so also the representations which constitute that knowledge, exist only in the subject. Whereas for Kant the existence of a real object is ultimately derived from the thing in itself which exists independently of any subject, for Berkeley objects have no nature at all -in themselves- [Kt1:69-71; cf. V.3]. Berkeley-s error results from his failure to see the difference between regarding an object from an empirical perspective and regarding it as a thing in itself. Kant-s modulation to the empirical perspective in the Analytic of Principles is intended at least in part to point up this important difference; hence it is to his often neglected empirical object-terms that we shall now direct our atten­tion.

 

3. Kant-s Empirical Object-Terms

 

      Why does Kant wait until the last chapter of the Analytic of Kt1 to in­troduce his empirical set of object-terms? As suggested in VI.1, he has a very specific reason: empirical terms for his primary transcen­dental distinctions are now required because the same object which he began by viewing from the transcendental perspective is now being viewed from the empirical perspec­tive. The empirical status of the object under consideration in the Analytic of Principles, and so also the function of this section in systemt, could perhaps have been conveyed more clearly by introducing these new terms at the be­ginning of this section, rather than postponing their introduction until after the principles governing the empirical perspective had been fully elaborated. However, as we shall see in VII.3.A, the position of the chapter on Phenomena and Noumena does have a basis in the architectonic structure of systemt.

 

      Kant hints rather early on that, whereas an appearance (i.e., a transcenden­tal appearance regarded from the transcendental perspective) is -an empirical in­tuition...which becomes experience- (i.e., becomes empirical knowledge) through -the concept of understanding arising from it- [Kt66:142(22)], a -phe­nomenon- is an appearance which is already -in agreement with the cate­gory- [Kt1:186]. In his secondary writings he uses -phenomenon- in much the same sense: it refers to -an object of experience- [Kt23:401n(179n)] or -...of the senses- [391n(165n)], the detailed knowledge of which -can be confirmed by experience- [Kt55:318(83); s.a. Kt53:76(156)]. Although in these works he prefers this term to the term -appearance- [see e.g., Kt46:435(101),440(109), 453-5(130-3); Kt55:318(83); Kt57:234(45)], he does explain the differ­ence be­tween them in Kt19:394: -in phenomena that which precedes the logi­cal use of the intellect is called appearance, and the reflective cognition which arises when several appearances are compared by the intellect is called experi­ence.... [Thus] the objects of experience are called phenom­ena...- Kant-s first proper definition of -phenomena- in Kt1 agrees with his usage elsewhere: phenomena are -appearances, so far as they [i.e., the appear­ances] are thought as objects according to the unity of the categories- [A248-9].

 

      An appearance, then, is the undetermined or -nondescript- form of what is destined to be regarded as an empirical object, while a pheno­menon is such an object -into which appearances may be discriminated by means of the under­standing- [B20:56-8,53; cf. Kt1:34]. When the two terms are interpreted as names for -the same things, only viewed in different ways- [B20:54], it be­comes evident that phenomena are always transcendental appearances viewed from an empirical perspective (i.e., -categorised appearances- [55]), but that appearances viewed from the transcendental perspective are not phenomena be­cause they are regarded as if they are not yet categorized (and thus, not yet known empirical­ly). Therefore, just as the transcendental object is trans­formed into an appearance only when the transcendental forms of intuition are added to it, so also the appearance is transformed into a phenomenon only when it has been synthesized with concepts in such a way as to become a real, determinate object of empirical knowledge.

 

      This interpretation reveals that Kant does not intend -phenomenon- to be -pejorative- merely because he contrasts it with -noume­non-, nor to -imply that the reality we elaborate in judgment is somehow defec­tive-.[18] On the con­trary, of all Kant-s object-terms, it is the only one which refers straight­for­wardly to -objects- proper [Kt1:137]-i.e., to the objects of everyday expe­ri­ence with which science is concerned, which are -immediately perceived- in sensation [A371]. But it does so in such a way as to remind the reader that if an empirical phenomenon were to be viewed from a transcendental perspec­tive, then it would have to be re­garded (but not -condemned- [F3:20]) as an appearance. Accord­ingly, it would be doubly misleading to claim with Weldon that the thing in itself (as such) is -the necessarily unattainable goal of empirical investi­gation- [W16:140e.a.]; for Kant maintains not only that the goal of empirical inves­tigation is always and only the phenomenon, but also that this goal is adequately attainable [see B20:24]. Although -the com­plete de­termination of an individual [object] is an infinite task- [W25:41e.a.; cf. Kt7:418], Kant allows that for empirical purposes -the cognition of [phe­nomena] is most veridical.-[19] Furthermore, although we could say with Ewing that phenomena -should be treated in science as if they were things-in-themselves relatively to us as empirical beings-,[20] this empirical use of the term must be clearly distinguished from its technical use, according to which phenomena viewed from the transcendental perspective must be regarded mere­ly as appearances. In other words, -empirically con­strued, an object "in itself" just denotes the correlate (Phänomena) of empirical judgments- [P7:375; cf. A10:67]. Of course, this is the case only because the objects we think of em­pirically as independent­ly exist­ing -things- are the same entities as those which must be called -ap­pearances of a thing in itself- from the transcen­den­tal perspective.[21]

 

      If the phenomenon is the final goal of empirical investigation, then why does Kant also introduce the -noumenon- at this point [Kt1:306]? His ex­pressed reason is not to stress the importance of the change from the transcen­dental to the empirical perspective, which these terms signify; this he does in the Appendix to the Analytic [316f]. Rather, he introduces the -noumenon- in order to emphasize that the transcenden­tal elements in his account of the synthetic a priori construction of empirical knowledge can be applied only to ob­jects of possible experi­ence. To elucidate this limitation, he needs a specifi­cally empirical object-term not only for the transcendental appearance, but also for the object in its unknowable state. The former need is clearly fulfilled by the word -phenomenon-; but it is not so clear in what sense the word -noumenon- can be an empirical equivalent of either the transcendental object or the (transcendent) thing in itself.

 

      As we saw in VI.2, the thing in itself is the reality which is knowable as represented to us in the form of a transcendental appear­ance, but which is by definition unknowable in its unrepresented form. Once the unknown thing is considered to be a transcendental object standing in some relation to a tran­scendental subject, it can be deter­mined through intuition to be a transcenden­tal appearance, and through conception and synthesis to be an empirical object [see VII.2-3]. With this progression in mind, it seems natural to infer in em­pirical reflec­tion that the empirical object, which we can now call a phe­nomenon, must be related -to something, the immediate re­presentation of which...must be something in itself, that is, an object independent of sensi­bility- [Kt1:A252e.a.]; hence it must be re­lated to -an object determinable through mere concepts [341], or to -an object of a nonsen­sible intuition- [A249]. In other words, -something which is not in itself an appearance must correspond to [the phenomenon]. For appear­ance can be nothing by itself, outside our mode of representation- [A251].

 

      To regard a phenomenon as released from the subjective conditions of ap­pearance which particularized it (viz., from -our mode of intuition- [Kt1:307]) -i.e., to regard a particular object of knowledge as if it could be an object with­out being represented-is to form the concept of what Kant calls a nou­menon. This -entirely indeterminate concept of an intelligible entity- [307] accompanies a phenomenon [A251] whenever an object is judged empir­ically to have -objective reality-;[22] for this judgment consists precisely in de­ter­mining that the phenomenon has such an independent (and thereby in itself unknowable and indeterminate) na­ture. -The understanding, when it entitles an object from an [empirical] perspective mere phenomenon, at the same time forms, apart from that perspective, a representation of an object in itself--i.e., a noumenon [Kt1:306; s.a. B20:74]. Hence-much as in the case of -phenomenon- and -appearance--a noumenon is always a thing (i.e., an ob­ject) in itself, but a thing in itself is not a noumenon unless its representation is considered to be particularized as an object in a subject-s phenomenal expe­ri­ence. In both cases the new terms are brought in as empirical equivalents for the original, transcendental terms. This explains why Kant never uses -noumenon- until the third chapter of the Analytic of Principles, and then oc­casionally equates it with -thing in itself- in subsequent sections: in these sec­tions Kant is exposing the mistake of viewing the thing in itself as an ob­ject -or vice versa-so he naturally treats -thing in itself- as synonymous with -noumenon-.

 

      Kant is careful to warn, however, that the concept of -noumenon- a­rises out of an -ambiguity- which -may occasion serious misapprehen­sion- [Kt1: 306; cf. B20:71-3]. For although -even the wisest of men- has a ten­dency to speculate on this basis [Kt1:397], it is a mistake to conclude that the noume­nal aspect of objectively real phenomenal objects opens up to us -a world which is thought as it were in the spirit (or perhaps intuited), and which would therefore be for the understanding a far nobler...object of con­templa­tion- [Kt1:A250]. Such a conclusion would require us to view the nou­menon in a literal way as -an intelligible object- [311e.a.] or an -object of pure thought- [393e.a.]; but when viewed from mankind-s limited, empirical per­spec­tive, -noumenon- is -not the concept of an object, but is a problem unavoid­­ably bound up with the limitation of our sensibility- [344]. As phe­nomenal be­ings, we can acquire empi­ri­cal knowledge only by means of sen­sible intu­ition: -our kind of intui­tion does not extend to all things, but only toobjectsofoursenses,...consequentlyitsobjectivevalidityislimited-[342-3]; yet we can infer from this that -a place therefore remains open for some other kind of [non-human] intuition, and so for things as its objects- [343].

 

      Our ignorance of that -space- in which we could experience the nou­menon as an empirical object raises yet another problem. Such ignorance is re­quired by systemt, because from the transcendental perspective a thing can be an ob­ject of knowledge only if it is represented intuitive­ly to our senses as an ap­pearance, and not directly to the understanding as a thing in itself [Kt1: 307]. Yet something like noumenal knowledge seems to be necessary in order for us to judge an empirical object to be independent from the empirical subject, and thus to be an objectively real phenomenon. Rather than eliminat­ing this difficulty by retracting one of his claims, Kant uses our natural but unattain­able desire to know the phenomenon as a noumenon to uncover a number of long-standing philo­sophical illusions in the Dialectic. But before doing so, he refines his terminology by proposing one further distinction, in order to clarify how it is possible to know the objective reality of phenome­nal objects in spite of our necessary ignorance of their noumenal nature.

 

      The meaning of the general distinction between phenomenon and nou­menon, as w