Clarification of Some Ambiguities in Systemt
by Stephen Palmquist (stevepq@hkbu.edu.hk)
A. The Faculty of
Representation
Two ambiguities
arise out of Kant's habit of referring to the 'faculty of representation'. The
first concerns his frequent use of the word 'faculty'. This and many of the
terms used in connection with it are often condemned by modern critics as
reflecting Kant's unphilosophical acceptance of 'the imaginary subject of
transcendental psychology' [S17:97]. The only proper response, they say, is to
'de-psychologize' his theory in order to purify its truly philosophical content
[H5:288]. And this, of course, involves analyzing the arguments in modern terms
in hopes of finding consistent and/or conclusive bits of reasoning-a procedure
the interpretive value of which I have already assessed in I.1. In defense of
Kant's general habit of using such unusual terminology, it should be noted
that, although it appears to the twentieth-century reader as if he is
arbitrarily inventing words at nearly every step, most of these terms were
familiar to philosophers in Kant's own time. 'Indeed', Weldon tells us, 'it is
not too much to say that all Kant's leading conceptions ... and much of his
greatly abused architectonic are clearly foreshadowed in the [works of his
predecessors]' [W16:52; s.a. P2:1.100n and V4:82-8]. This alone, of course,
does not justify our continued use of such terms; but it does suggest they are
meaningful in their own context, so they cannot simply be discarded by the
interpreter without seriously misrepresenting Kant's System.
The specific
reason why we should continue to use Kant's faculty terminology when
interpreting his philosophy is that it is not, in fact, intended to be
psychological. (Kant clearly distinguishes on several occasions between
transcendental philosophy and
empirical psychology [e.g.,
Kt1:152].) Instead, as Vleeschauwer points out, Kant's use of apparently psychological
terminology assumes a 'different perspective' [V4:88]. It is simply the way he
has chosen to refer to the subjective functions
of human knowing in his radically perspective-bound philosophy. Because
philosophical analysis is always limited to one perspective at a time (at
least, if it is to be coherent), the functions can be regarded as distinct even
though each faculty performs its own function on 'one and the same territory of
experience'.[1] Van de Pitte
defends Kant in much the same way when he says:
Kant is unable to define sensibility (except as
receptivity), the categories, or the imagination (except as the agency of
synthesis).... [Yet] this inability does not at all trouble him. For ... it is
not the faculties of the mind [as such] that he is concerned with, but the
functions which must be fulfilled if experience is to be possible. [V3:1024-5]
The second
ambiguity to be considered here is that the term 'representation' can be
used either as a verb to name an a priori function of knowing, or as a noun to
name the state of the object once it is determined by this function. This
means the effect on an object of the faculty
of representation (i.e., of the function of representing) is to determine that it is a representation (i.e., a represented object) [cf. A11:21; F7:xxxi; P8:72-3]. The same ambiguity
arises with many of Kant's other terms as well, and may explain, at least in
part, why, as Wolff points out, 'statements about knowing and statements about
knowledge tend to be interchangeable' for Kant [W21:73n]. Fortunately, as long
as both possible meanings are kept in mind, Kant's intentions can usually be
determined from the context with a fair amount of certainty.
Perhaps in an
attempt to avoid the ambiguity in this example, some translators and
commentators translate Vorstellung as
'idea' rather than 'representation'. Although the word 'idea', when 'taken in
the ordinary English sense' [P3:93n] is admittedly an accurate translation of Vorstellung, it is inappropriate when
interpreting Kant for several reasons: (1) Deliberate ambiguity in the
original should be preserved, as far as possible, in an accurate translation.
Substituting 'idea' in some passages takes away the possibility of
interpreting Vorstellung as referring
to a mental function, since 'idea' does not also mean 'act of having an idea'.
(2) Vorstellung has such a wide
variety of applications for Kant that the rather narrow word 'idea' is too
specific in many contexts. (3) Kant has another technical term (Idee) which must be translated 'idea'.
Since he does not intend to identify Vorstellung
and Idee, it is misleading to use the
same word in translating both. As Kant himself pleads: 'I beseech those who
have the interests of philosophy at heart ... [to] be careful to preserve the
expression "idea" [Idee] in
its original meaning, that it may not become one of those expressions which
are commonly used to indicate any and every species of representation [Vorstellung]' [Kt1:376]. To avoid
confusion, therefore, whenever a text which renders Kant's Vorstellung as 'idea' (or any word other than 'representation') is
quoted in this book, I have substituted the word 'representation' without
further notice.
The
suggestion that Vorstellung should be
translated as 'presentation' [W8:192; H4:xi; P14:27n] would be suitable when
it is used in the context of the original
representation which arises in the first step of systemt; but in all of
its other uses, the 're' connotes a repeated determination of the same object, and should therefore be
preserved. Rotenstreich's rendering of the term as 'image' [R11:134], however,
is wholly unacceptable, except perhaps in steps four and seven, where
representation takes the form of imagination.
B. Placement of
the Transcendental Object in Systemt
Another
source of possible confusion is that Kant discusses his doctrine of the
transcendental object more in the Analytic than in the Aesthetic [see note
VI.8], yet I have placed it in the first stage of systemt [see VII.2.A].
That it belongs in the first stage
should be evident in light of its mediating role between the thing in itself and
the various determinate forms of representation [cf. VI.2]. As Kant says in
Kt6:218n, 'the architect of a system' must begin by positing 'an object as
such' as the basis for all other divisions. Moreover, it would not make sense
to place it after sensibility, since
Kant clearly says in Kt1:522-3 that 'this transcendental object ... is given
in itself prior to all experience.' This original representation differs from
representations which constitute appearances, inasmuch as it remains
'necessarily unknown' [523-4]. Thus, the transcendental object is a
pre-intuitive representation of sensibility in much the same way that a
perception is a pre-conceptual representation of the understanding.
Some interpreters
have recognized, in varying degrees, the connection between the transcendental
object and the first stage of systemt. Gotterbarn
alludes to its primary position in the logical order of Kant's theory when he
says it 'is a preliminary concept which prepares for the conception of
"affinity" and the transcendental deduction of the categories'
[G11:199]. Buchdahl insists it 'is the central notion with which to begin' in
interpreting systemt [B27:52]; its 'essential function is ... as the
starting point of realization procedure which involves, to begin with,
sensibility' [78]. Weldon connects it with sensibility by saying 'the
transcendental object is not itself a possible object of knowledge. It denotes
simply the formal characteristics which render intelligible the manifold of
pure or empirical sensibility, the ground of synthetic unity in phenomena'
[W16:178n]. Likewise, Walsh refers to the transcendental object as 'awaiting
application ... [to] the data of the senses' [W9:162; see Kt1: A394,411n]. All
these comments suggest that there is an important sense in which the
transcendental object lies at the very basis of the Aesthetic, even though it
is not discussed at length therein.
Kant employs
the concept of the transcendental
object in the arguments of his second, third and fourth stages primarily as a
way of referring to the influences of the first stage on the others. That is,
the concept of the transcendental object stands for the assumption that some
Aesthetic element (i.e., some experience in general) is given, without
presupposing anything about its intuitive or sensible character. Thus in
Kt1:A290 Kant says 'the concept of an object in general [is] taken
problematically, without its having been decided whether it is something or nothing.'
The first step, however, is concerned not so much with the concept of the transcendental object as
with the transcendental object itself-i.e.,
with that radically undetermined state of objects (as such), which confirms the reality of their existence.
Thus, Bird is correct when he says Kant devotes more attention to the concept
of the transcendental object than to the transcendental object itself [B20:5];
but this does not necessarily mean the former is more important to the system.
On the contrary, the transcendental object is an element in systemt, whereas its concept is
not.
The
transcendental object, being itself neither a concept nor an intuition, must
therefore lie at the base of both the Aesthetic and the Analytic of Kt1. It
could be called 'the ultimate recipient
of all the predicates we bear', in the sense that it is equal and opposite to
the giver of all predicates (the
transcendental subject). Kant could have made this more clear by naming the
pre-Logic section of Kt1 'Transcendental Sensibility', and including in it first a chapter on 'Transcendental
Object', to be followed by the existing 'Transcendental Aesthetic' as Chapter
Two. This would have shown that the transcendental object is not part of the Aesthetic, any more than it
is part of (i.e., an element in) the
Analytic.[2]
C. Mathematical
Judgments
Kant uses the
assumption that mathematical judgments depend on pure intuition [see note
VII.16] to argue for the intuitive nature of time and space. The inadequacy of
this approach is often emphasized by commentators [see e.g., Pq18:1.3]; but
Parsons points out that this is a 'comparatively accidental feature of Kant's
view' [P1:38], for the validity of his theory of time and space is not necessarily dependent on the validity of
his conception of geometry. Unfortunately, in defending the intuitive nature
of time and space Kant does argue from
the assumption that the judgments of mathematics, as well as those of physics
and metaphysics, have a synthetic a priori status [Kt1:14-8]. His claims have
often been debated, and not infrequently rejected [B5:18-21; R12:145; but cf.
P14:6-7 and H16:376; s.a. W17:17-27]; but inasmuch as they are not constitutive
elements of systemt, I shall merely offer a brief suggestion for a
possible interpretation.
Mathematical
truth-claims could be synthetic when viewed from the empirical perspective of
judgments, as, for example, when
considering how to go about solving a particular problem [cf. M3:26-7], yet
analytic when viewed from the logical perspective of how to prove the validity
of a given proposition. (In
K2:10.528-30(Z1:129-30) Kant draws a similar distinction between viewing
mathematical concepts 'subjectively' as 'a task to be done' or 'objectively'
as a 'proof', though he does not apply this to the analytic-synthetic
distinction.) For instance, given the laws of mathematics, the number '12' does contain in it analytically the
addition of '7' and '5'; nevertheless, if a person thinks of the numbers '7'
and '5', he does not ipso facto think
of the number '12' until the notion of 'addition' and other relevant laws of
mathematics are specified [see Kt1:14-7,740-66; cf. W5:162-3]. Thus, as Beck
asserts, in order to establish the
validity of a mathematical claim, there must be an 'intuitive construction of
the concept' [B5:19; cf. P2:2.129-31]-and this presupposes at least that the
synthetic method has been adopted. I
discuss the nature of mathematical knowledge and its relation to Kant's doctrine
of pure intuition in greater detail in Pq8:14-22, Pq9:279-82 and Pq14.
D. Inner and
Outer Sense
Kant's
reasons for giving priority to the transcendental necessity of time and inner
sense in stage one are not always clear; but in later stages-especially in the
second edition of Kt1-he balances this one-sidedness with a corresponding
emphasis on the empirical necessity of space and outer sense [e.g., Kt1:156].
Strawson regards this balance between the inner and the outer as one of the
crucial aspects of Kant's transcendental idealism, for it insures that 'no
superiority of status, as regards reality or certainty of existence' can be
accorded 'to states of consciousness over physical objects' [S17:21].
Unfortunately,
the precise meaning of these terms, particularly of 'inner sense', is difficult
to determine. Kant does express the 'opinion' in Kt14: 60(94) that 'the secret
faculty ... by which judgment is possible ... is nothing but the power of the
internal sense, that is, the power of making our own representations the
object of our thoughts.' Conceived in these terms, inner sense would seem to be
a reflection of the means by which stage one is related to the concepts which
arise in stage two [but cf. note VII.25]. Similarly, outer sense could be
regarded as the means by which stage one is related to the phenomenal objects
which will arise in stage three. (That stage four is not so closely connected
with stage one is accurately reflected in Figure VII.3 by their correlation
with the opposing ++ and -- components, respectively.)
Much of the
confusion arising out of Wolff's discussion of these terms in W21:191-202 stems
from his own failure to see that 'inner sense' and 'outer sense' are two ways
of considering the contents of sensibility, and that their empirical
manifestation is something quite apart from their function in the
transcendental conditions of knowing. According to Wolff's empirical interpretation,
'inner sense' denotes 'an inner perception of the spatial manifold which
constitutes the contents of consciousness' [199]; likewise, Weldon rejects the
idea that inner sense refers to a kind of 'feeling' of oneself [W16: 152],
equating it instead with 'my own past acts of awareness'-i.e., memory [154-5].
Kant himself refers in Kt30:284(69) to 'inward experience' as involving 'an awareness
of one's psychological state.' But this empirical interpretation is not
incompatible with Kemp Smith's transcendental interpretation: 'The
representations of outer sense are all by their very nature likewise
representations of inner sense. To outer sense is due both their content and
their spatial form; to inner sense they owe only the additional form of time'
[K3:294]. Given the perspectival orientation of Kant's philosophy, these two
positions are quite compatible.
The failure
to recognize the perspectival basis of Kant's System leads many interpreters to
reject his theory of inner and outer sense prematurely, on the grounds that
'time is an abstraction from what is concrete fact or events' [E3:190], and is
therefore 'a condition of there being a finite mind', but not vice versa [186;
s.a. B17:156-9]. In our discussion of stage three (especially of the schematism
in step seven) we saw that Kant is in full agreement with this view of empirical reality. Nevertheless, this
does not preclude the possibility (or reduce the necessity) of viewing time transcendentally as a formal condition
imposed by the subject, as in stage one. In any case, Kant's theory of time and
space is clear enough so that a psychologist recently referred to patients
suffering from 'akinesia' (inability to function physically and/or mentally in
time or space) as having 'a "Kantian" delirium, and ultimate akinesia
as being "aKantia"' [S1:252]!
The arguments
of Swinburne in S19:26-40 and Quinton in Q3:130-46 against Kant's theory that
there can be only one space suffer
from a similar neglect of the principle of perspective. The fact that it is logically possible to imagine in a philosophical
thought-experiment that two unrelated spaces exist (one of which, for example,
could be heaven [S19:40]) is advanced as evidence that Kant is mistaken [28,37;
Q3:143]. But Kant does not deny the logical consistency of such a proposal, nor
even its real possibility; what he denies is that more than one space can be known to be the transcendental ground for the coherent, unified experience of one
rational individual. Neither argument disproves this transcendental claim. On
the contrary, Quinton's thought-experiment posits two coherent, unified, but unrelated (and unrelatable!) sets of experiences;
and Swinburne ends up admitting that 'we have no reliable information' about
any space other than our own [S19:40].
E. Forms of
Imagination
Kant's theory
of imagination plays an important role in both editions of Kt1 [J4:686]; but he
obscures this role by developing the theory differently in each. In the second
edition, for instance, he several times prefers the term 'figurative synthesis'
to denote 'the transcendental synthesis of imagination' [Kt1:151,154], and
suggests that in a sense 'the imagination ... belongs to sensibility' even
though it is a function of understanding [151-2; cf. note IX.1]. What makes
matters worse is that in both
editions of section one of the Deduction [116-29], as Paton contends, Kant's
'doctrine of imagination is undoubtedly inconsistent with the doctrine
expressed elsewhere' [P2:1.53]. But in Y1 Young provides a good analysis of
many of the problems surrounding Kant's view of the imagination.
The
relationship between perception and imagination can help us to distinguish
more clearly between two types of imagination, as well as between imagination
and the other types of synthesis [cf. note VII.38]. Kant sometimes refers to
the transcendental faculty of imagination as the 'productive imagination' in
order to clarify the difference between it and its empirical counterpart, which
he usually calls 'the reproductive faculty of imagination'. The reproductive
imagination 'leads the mind to reinstate a preceding perception alongside the
subsequent perception ..., and so to form whole series of perceptions'
[Kt1:A121]; only in this way can perceptions be experienced as appearances in
time. It is also responsible for the more common use of 'imagination', in which
the reproductive imagination acts as 'the faculty of representing in intuition
an object which is not itself present'
[151; s.a. Kt66:153(32)]. Whereas the productive imagination produces the
original synthetic unity of the manifold and is objectively valid (i.e., a
priori), the reproductive imagination simply reproduces particular images of
a perception, and thus 'has only subjective validity' (i.e., it is a
posteriori) [Kt1:141,152; see note VII.3].
In its most
elementary form the consciousness aroused through the synthetic function of
the productive imagination, when it is 'immediately directed upon
perceptions', is entitled 'apprehension' [Kt1:A120]. This aspect of synthesis,
Kant explains, 'is only a placing together of the manifold of empirical
intuition' [129] by which the imagination prepares the manifold of sensation
for the actual process of synthesis [235]. In fact, the whole first stage of
systemt is implicitly included in apprehension when that stage is viewed in
stage two from the (logical) perspective of the use to be made of it by the
imagination. Although apprehension has an empirical application [see e.g.,
164], it is also transcendental in the sense that it represents the initial
activity of the function of the productive imagination. Kant may have introduced
these extra terms to allow for the logical possibility that imagination's synthesis
might produce something other than apprehended perceptions, or that there might
be a manifold which is not synthesizable [E5:110-1]. But if empirical knowledge is to result, it is
necessary for the synthesis of productive imagination to give rise to
apprehended perceptions [Kt1:246].
This
necessary epistemological connection may be what persuaded Kant to propose in
the first edition of Kt1 his rather ambiguous theory of a threefold synthesis,
made up of intuitive apprehension, imaginative reproduction and conceptual
recognition [A98-111]. Although he claims at times that these terms can denote
transcendental syntheses [A100,102], he admits elsewhere that the distinctions
are 'merely empirical elements of experience' [A125]. In either case the first
and third terms are apparently intended to describe the influence of synthesis
on what precedes (viz., intuitions) and follows (viz., concepts) the actual
function of synthesis in imagination (i.e., in step four). Thus, when its close
connection with the synopsis (i.e., sensation) given in stage one is in view,
the phrase 'synthesis of apprehension in intuition' [A98] simply indicates
that synthesis presupposes the apprehension of a sensible object in step three
[162n,237]. Likewise, with respect to the conceptual goal towards which step
four heads, 'synthesis of recognition in a concept' [A103] simply indicates
that synthesis prepares the way for conception in step five [cf. H11:183].
(Paton agrees that Kant should have made a point of clarifying that
'recognition involves understanding or apperception' [P2:1.379; but cf.
Kt1:A115]. Elsewhere he makes the interesting suggestion that the first three
categories correspond to this threefold synthesis [P2:2.56-7n].) And the second
member of this threesome, 'the synthesis of reproduction in imagination'
[Kt1:A100], has already been accounted for as the empirical counterpart of
productive imagination. In the second edition of Kt1 Kant, apparently realizing
the redundancy of this threefold synthesis, concentrates more on the simple
fact that synthesis as such is a necessary condition of knowing.
F. Categories and
Conceptual Schemes
Before
arguing in the Transcendental Deduction for the necessity of categorial
conception as the fifth condition of knowing, Kant lists and briefly describes
each of his twelves choices for categories in the passage known as the
Metaphysical Deduction [Kt1:91-116]. He asserts that 'the function of thought
in judgment can be brought under four heads': quantity, quality, relation and
modality [95], each of which in turn has 'three moments' [see Table III.7]. The
table in which Kant lists the resulting twelve functions is sometimes called Kant's
'Table of Judgments' [see e.g., S21:3-19], even though Kant himself never uses
this phrase in Kt1. This practise stems, no doubt, from the fact that Kemp
Smith's translation uses this phrase as a page heading on two pages. Yet this
label is highly misleading, since the initial table does not list the forms of judgment (which would belong in stage three), but 'the mere form of
understanding' as abstracted from
judgment [95]. Kant himself refers to the table as 'the table of logical
functions' [112], so this more accurate name is the one interpreters ought to
use.
Kant then
proceeds to list in a 'Table of Categories' the twelve 'pure concepts of the
understanding' which correspond to the twelve 'logical functions' [Kt1:105-6].
This new table differs from the first in being slightly less abstract: whereas
the first lists only a set of bare 'logical functions', the second specifies
the pure concepts through which each
function can be applied to appearances in order to produce empirical knowledge
[143; s.a. E2:178]. But for both Kant claims a kind of certainty and
completeness [see e.g., Kt2:322-6] which has seemed rather presumptuous to most
interpreters. He decrees, for example, that the logical functions listed in the
first table 'specify the understanding completely, and yield an exhaustive
inventory of its powers' [Kt1: 105]. Hence, there need never be 'any addition
to the transcendental table of categories, as if it were in any respect
imperfect' [115]. Not only can this table 'reveal and make noticeable every gap
in the understanding' [265], but it will also, he predicts, supply 'the complete plan of a whole science' [109].
Despite the
ostentatious nature of such assertions, Kant devotes little effort in Kt1 to
the defense of his twelve categories or to the clarification of their
individual meanings [but see Kt10:101-9(106-16)]. Instead he boasts: 'In this
treatise, I purposely omit the definitions of the categories' [Kt1:108], though
he later confesses that 'we are unable to define them even if we wished'
[A241]. (Some commentators have attempted to defend and/or refine Kant's
choices [see e.g., E2:148-94; E5:137-40; P2:1.295-6]-a worthy cause which we
need not join here.) The fact that he never even explains the nature of their
twelvefold structure-a task I have attempted in III.3-has given rise to the
commonly held opinion that Kant 'was stupid to the point of perversity in the
detail of his doctrine of categories' [M2:213; cf. 244-5]. Yet the justification
for such neglect is that the categories only obtain 'reality' when their
'application' is taken into account [see Kt7:385]; and Kant does eventually
support his initial dogmatic account by examining in the third stage of systemt the principles
through which each category is applied in producing empirical knowledge [see
VII.3.A].
Although
critics are often quick to point out the absurdity of Kant's claim that his own
particular choice of twelve categories is universally valid for all time, they
are usually slow to suggest any alternatives. Strawson, for example, takes it
for granted that our 'schemes of thought ... are not static schemes, but allow
of ... indefinite refinement', just as in all sciences or in the 'development
of social forms' [S17:44]. Martin goes so far as to call Kant's view 'pitiful'
[M3:82], suggesting that the one point of agreement between all interpreters
of Kant has been 'that this rigid structure has got to be made dynamic and
living.' He suggests this could be done by emphasizing that, in spite of the
'static' form of his tables, 'Kant's aim [was] to free himself from a static
conception' by regarding the categories as 'fundamental acts of pure thought'
[M3:81]. Walker, by contrast, claims Kant's choices are 'not so arbitrary and
artificial as is sometimes made out' [W2:26]-a claim supported by our findings
in III.3, which suggest it may be more important to insist on the necessity of
there being twelve categories, than
to insist on Kant's choices being necessarily correct for everyone everywhere.
Premature rejection of Kant's table of categories can be avoided by
recognizing they are not intended to function like a scheme of 'social forms',
or of any empirical science, but merely as a bare, architectonic structure
defining the logical relations
implicit in all thinking.
Kant does
actually explain quite clearly that the certainty of his table rests on its formal structure. Thus, for example, he
says 'the third category in each class always arises from the combination of
the second category with the first' [Kt1:110; s.a. Kt2:325n;
K2:10.344(Z1:112)]. We have already discussed at length some of its other
formal features in III.3. Once its formal structure is recognized as the source
of its certainty and completeness, we can allow that other types of 'conceptual
schemes' (as they are usually called nowadays) might be able to produce
empirical knowledge (as it is understood in other cultures or societies), but
that such alternative schemes will be complete only when they are described in
terms of this same twelvefold form. A Kantian response to such a proposal would
be that modern talk of conceptual schemes nearly always assumes the Empirical
Perspective, whereas his Table of Categories assumes the Transcendental
Perspective, inasmuch as it defines the set of conditions which must be met by any viable conceptual scheme (whether it
be animistic, astrological, religious, Newtonian, relativistic or any other).
The main
purpose of Kant's arguments in the Analytic of Concepts is to establish the
formal status of categories as a
necessary condition for empirical knowledge, so he focuses in the
Transcendental Deduction only on their general
function. As a result, many interpreters have pointed out that, as Körner puts
it in K9:57, 'even if his proof of the complete list of Categories was not
successful he may still claim to have established that we do apply Categories
in making objective empirical judgments.' Although Körner is right to distinguish
the validity of the Transcendental Deduction from that of the Metaphysical
Deduction, he unfortunately assumes Kant was trying to prove his categories in the Metaphysical Deduction; yet this is
open to considerable doubt. Why would Kant wish to prove their validity, when
the pure categories as such are, in
Paton's words, 'of no use at all', being 'merely the pure form of the use of
understanding' [P2:2.438; cf. Kt2:324]? Kant's proofs for his choice of categories really come only in the
Analytic of Principles.
Because Kant
does not actually present conclusive proofs concerning the nature of each
category until the section on the Principles, Wolff suggests: 'The principles
are not applications of the
categories ...; they are the categories'
[W21:204]. But this completely ignores the difference in perspective between
the second and third stages. The categories function as the condition through
which knowledge relating to the logical perspective is synthetically realized,
while the principles function as the condition which determines the form of
knowledge relating to the empirical perspective. Given this difference, we
should expect to find Kant defending each in a different manner.
Another
problem is that Kant does not always clearly explain just how he intends us to
think of these categories as applying to appearances. For instance, do they all apply to every appearance, or does
each appearance reveal only one (or several) of the categories? Kant sometimes
seems to suggest the latter, as when he says 'All the manifold ... is determined in respect of one of the logical functions of
judgment' [Kt1:143e.a.]. But Paton argues that all twelve 'forms of thought
must be present in every judgment' [P2:1.226n; s.a. 212-5; E2:194; E5:147-8].
In contrast to both of these extreme views, I believe Kant wishes us to think of
each judgment as being determined by one
of the three 'moments' or logical functions listed under each of the four headings, so that each judgment will have
precisely four categorial
characteristics determining its form.
G. Varieties of
Deduction
One of the
most complicated aspects of Kant's exposition concerns the various sorts of
'deduction' to which he refers. The difference between the Metaphysical and
Transcendental deductions of the categories was mentioned in Appendix VII.F
[s.a. note III.17]. Two further complications are also worth discussing: the
difference between the first and section versions of the Transcendental
Deduction; and the difference between Kant's so-called 'objective' and
'subjective' deductions.
I have, for
the most part, played down the many differences between the first and second
editions of the Transcendental Deduction. This is not because the differences
are of no significance, but rather, because the most significant points Kant wants to get across are the arguments
which actually constitute the three steps in stage two; and these must be
present in both versions, since Kant claims in Kt1:xxxvii that 'in the form and
completeness of the [architectonic] plan, I have found nothing to alter' in
the second edition. Nevertheless, the second edition is a slight improvement
over the first in the sense that it reveals these three steps more clearly,
with fewer digressions into side issues [see e.g., note VII.34 and Ap. VII.E].
In the
opening pages of Kt1 [Axvi-xvii] Kant distinguishes 'two sides' of his inquiry.
The 'subjective deduction' investigates the nature and possibility of 'the
cognitive faculties' of the understanding; it is of secondary importance and
gives the appearance of being 'somewhat hypothetical in character'. But the
'objective deduction', which 'is intended to expound and render intelligible
the objective validity of [the] a priori
concepts' of the understanding, he regards as the 'essential' purpose of his
inquiry. Because Kant himself never specifies exactly how or where he develops
these two sides in his exposition, Weldon is probably right to say 'there is no
point in attempting such a distinction' in an interpretation [W16:173n].
Nevertheless, interpreters have sometimes made use of it in various ways.
Strawson, for example, identifies the 'subjective deduction' with Kant's
'transcendental psychology', and attempts to 'disentangle' it from the truly
philosophical 'objective deduction' in order to discard the former [S17:16; cf.
H11:171]. In sharp contrast to this approach, Wolff equates Section 2
[Kt1:A95-114] with the subjective and Section 3 [A115-130] with the objective
deduction [W21:85,100,164], and contends that the former 'is the key to the interpretation of the entire Critique' [80]. (See also E3:97;
O3:149-67; P4:175-8; V4:84-5.)
From the
little Kant does say about this distinction, it could be argued, in contrast to
these interpretations, that the subjective deduction is the argument of the
entire second stage and that the objective deduction, though alluded to at
various points in the Analytic of Concepts, is in fact the argument of the
third stage [cf. H13:75]. This alternate interpretation is supported by Kant's
stress on 'objective validity' in step seven, as well as in the remainder of
the third stage. Such an interpretation can help us to see how the two deductions,
as 'two sides to the same argument' [W2:83], are actually mutually dependent
upon each other [cf. E5:70,85 and P2:1.529].
H. Schematism
Some
commentators interpret Kant's statements concerning the reciprocity between
intuition and conception to mean 'there is no third operation besides thinking
and sensing, all knowledge being for [Kant] covered by these two' [E5:36; s.a.
E3:98]. Schematism can then be either rejected as superfluous, or regarded as
nothing but a return to intuition by the concept. Allison follows Gram's
arguments to this effect by equating pure intuition with schemata
[A7:86-88,104; cf. H11:96-7], though he rejects Gram's claim that this is
incompatible with Kant's notion of the schema as a hybrid 'third thing'
[A7:88]. Bennett, by contrast, offers a rather different sort of reduction:
'The schema of any empirical concept is just the concept itself' [B17:151].
The fact that
Kant has already established a close connection between the first two stages
has led some commentators (such as Adickes and Kemp Smith) to regard the whole
Schematism chapter as repetitive and unnecessary. To this Paton rightly
rebuts: 'It is vital to Kant's argument to specify the transcendental schema
which is contained in each category' [P2:2.68]; indeed, it is 'absolutely
essential' [76]. For, whether or not there really is a third thing such as the schema, as Kant
himself may have believed, is not as important as the fact that there must be
some third function. The functions of
intuition and conception in isolation
are not sufficient to explain how empirical knowledge arises: for that, a
third function, that of combining
these two functions in a whole (i.e., schematism) is necessary. As long as we
recognize the new and different use in stage three of that which first arises
in stage two (viz., synthesis and concepts), we can agree with Deleuze when he
claims that, just as understanding judges
by means of its concepts, so also imagination schematizes by means of its syntheses [D2:18]. Synthesis is
pre-conceptual in step four, but post-conceptual in step seven.
Vaihinger and
others have assumed the activity of schematism must be an example of a
so-called 'noumenal act', inasmuch as the function of reintroducing time
cannot itself be an act in time. Paton rejects this interpretation, arguing
instead that such a synthetic process must take time [P2:2.393]. Perhaps a
better solution would be to say that this whole question is incoherent. As a transcendental function, its place in
the chronological sequence of events is not at issue; it is not an 'act' in any
literal sense (phenomenal or noumenal), but a logically necessary function. As temporal functions, intuition,
conception and schematism are all mixed up in a 'bundle', not unlike Hume's,
and cannot ordinarily be isolated as discrete acts.
After noting
its inscrutability, Chipman speculates that the schema may be what physicalists
are looking for when they investigate neural mechanisms [C7:115]. But
Dahlstrom, recognizing that empirical schemata may well be obscure, insists the
schematism of pure concepts, with which Kant is primarily concerned, 'must be
both scrutable and necessary' [D1:220]. He argues against those who believe
that, because the condition for knowing a concept is being able to apply it
correctly, concept and schema are always equivalent (e.g., Bennett
[B17:142-51], Warnock [W11:77-82], and Wilkerson [W20:94-5]), by pointing out
that 'a schema provides a condition for cognition [viz., time] by means of a
concept, not supplied by the concept itself' [D1:219]. He adds that such
critics 'too often demand either the sort of knowledge Kant reserves for fundamental
principles (Grundsätze) or the
knowledge Kant attributes to God.'
I. Principles
The scholarly
discussion of Kant's principles is so massive as to fill a small library all on
its own. There have been many attempts to revise or reconstruct some or all of
his arguments in various ways. Strawson, for instance, attempts to reconstruct
Kant's 'analogies of experience' by regarding permanence, cause, and
interrelatedness not as 'absolutely indispensable elements in terms of which
we see the world if we are to see an objective world at all', but rather as
'natural hopes' which ordinarily
characterize our experience [S17:146]. Of course, much hinges in such a
revision on just what the word 'ordinarily' is taken to mean. This is
especially true because Kant himself claims the principles determine not, as Strawson's wording seems to suggest,
the way we see the world, but only
the way we attain ordinary theoretical
knowledge about the world. Kant would argue, no doubt, that Strawson's
attempt to transfer the principles to the judicial standpoint of hope [see
IX.1] would render the possibility of real theoretical knowledge hopeless!
Interestingly,
Kant defends something not unlike his three analogies in one of his first
philosophical essays, where he argues that 'any substance which is cut off from connection
with others is incapable of any change',
which is an 'absurdity' [Kt11:409(243)e.a.]. To this 'principle of determining
reason' he adds 'the principle of succession' [410-2(245-8)] and 'the principle
of co-existence' [412-6(248-529], thus insuring that 'all things are found conjoined
in a universal mutual connection' [413(248)]. But this essay differs from his
mature position in at least two respects: first, he refers to these as
'principles of abstract metaphysical knowledge' [416(252)] instead of as necessary
conditions for the possibility of experience; and secondly, he 'secures' his
argument for 'mutual reciprocity' by appealing to 'the divine idea' [413 (248)]
rather than by constructing a transcendental argument.
Among Kant's
principles, the 'postulates of empirical thought' are perhaps the most
difficult to pin down, since 'modality
... contributes nothing to the content of the judgment' [Kt1:99-100]. Paton
suggests that possibility, actuality and necessity correspond directly to the
other three categories (viz., quantity, quality and relation, respectively)
[P2:2.57]; in both cases the third member acts as the synthesis of the previous
two [see 2.338n,341]. In K2:11.45(Z1:147-8) Kant gives one of his most concise
accounts of what is entailed by the postulates, from which we can construct the
following table:
Table Ap. VII.1: The Postulates as Forms of Judgment
A common mistake among interpreters is to ignore the perspectival difference between the categories and the principles. Because he ignores this difference, Rosenberg fails to see how Kant subsumes individual empirical instances under such general concepts [R9:349-60]. Yet the entire third stage of systemt is intended to provide just such an explanation. Likewise, England claims that because the categories (a term he uses to cover the principles as well) are not 'constitutive factors in the
object, but notions
employed in the act of apprehension' [E3:103], 'their significance lies in the
realm of knowledge', not in 'the
realm of existent fact' [110]. Yet this ignores the fact that, as far as
empirical objects are concerned, Kant believes the realm of fact must be
located within the the realm of
knowledge. Thus, England's complaint that Kant gives 'no justification
whatsoever ... for supposing that the categories enter into [the] very
structure' of 'empirical things' [109] is unjustified for two reasons. First,
the categories are merely logical forms, so they do not as such constitute objects. And secondly, the supposition that principles constitute objects [see note
VII.49] does not require a separate justification, since it is part of the
general Copernican Perspective upon which the entire Critical philosophy is
built! The justification, then, is in part [see V.1] the coherence of the
System to which this assumption gives rise.
J. Ideas
In some
respects, Kant's treatment of the ideas of reason resembles his treatment of
the thing in itself (a similarity which will be treated more fully in Pq20 [cf.
Ch. X]). The latter denotes the unknowable substratum (0) of the phenomenal
world, whereas the former denotes the way we conceive of this reality (+++)
as it impinges on our empirical judgments. For example, concerning the
principle of substance, Kant says: 'The simplicity and other properties of
substance are intended to be only schema of this regulative principle and are
not presupposed as being the actual ground of the properties of the soul'
[Kt1:711]; to assume the latter would be to confuse the idea with the thing in
itself.
Kant
criticizes Plato for conflating the roles of the idea and the thing in itself
[Kt1:370-5]. Plato regarded ideas as 'the original causes of things' [374],
whereas Kant believes they should be regarded only as 'ends' [375]. Rosenberg
is therefore unfair to Kant when he says 'Kant misunderstood the nature of the
Platonic Ideas and took them to be conceptual forms of knowledge when in fact
they are absolutely non-conceptual' [R8:285]. Kant is well aware of Plato's
sense of 'idea', and regards it as, for the most part, worthy of 'respect and
imitation' [Kt1:375]; but he intentionally replaces
Plato's use of the term with his own, narrower use. Kant's main complaint is
against anyone who, like Plato, tends to use the word to include too much
[376], especially when it is regarded not just as a tool to regulate the way
we think about an unknowable mystery,
but also as itself constituting that
reality and/or our knowledge of it.
