A Commentary On Kant's Critique of Practical Reason (UCP 1960) - L.W. Beck
A Commentary On Kant's Critique of Practical Reason (UCP 1960) - L.W. Beck
A Commentary On Kant's Critique of Practical Reason (UCP 1960) - L.W. Beck
A Commentary
on Kant's
Critique of Practical
Reason
CHICAGO" LONDON
The Cniversity of Chicago Press, Chicago 60637
The university of Chicago Press, Ltd., London
ISBN: 0-226-04076-3
LCl": 60-5464
Foreword
I Introduction 19
2 The Problem of the Critique of Pure Reason 10
3 The "Copernican Revolution" 11
4 Theoretical and Speculative Reason 13
5 The Third Antinomy 14
6 Transition from the Problematic to the Assenoric Judgment
of Freedom 26
III THOUGHT, ACTION, AND PRACTICAL REASON
I
1
Two Aspects of Conduct
Conative and Cognitive Elements in Action
•
3 Practical Reason and Will
4 Pure Practical Reason
IV NAME, PURPOSE, AND STRUCTURE OF THE "CRITIQUE"; CoMMEN-
TARY ON PREFACE AND INTRODUCTION 41
I Introduction 42
1 Title of the Book 41
3 The Purposes of the Book 45
4 The Unity of Theoretical and Practical Reason 47
5 How Is a Synthetic a Priori Practical Proposition Possible? 51
6 Relation to the Metaphysics of Morals 53
7 The Structure of the Book 55
8 The Polemics of the Preface 56
PART II
Xl
""
Xll Table of Contents
3 The Problems of the Analytic 68
4 Division of the Commentary on the Analytic 69
5 Summary of Chapter I, §§ 1-8 70
VI THE ANALYTIC OF EMPIRICAL PRACTICAL REASON. I. FORMAL CON-
SIDERATIONS; COMMENTARY ON § 1 75
1 Empirical Practical Reason 75
2. Practical Principles 76
3 Rules 79
4 Maxims 80
PART III
BIBLIOGRAPHY
INDEXES
§ I. INTRODUCTION
Few substantial philosophical treatises have been written in the haste in
which the Critique of Practical Reason was composed. The book, how-
ever, shows few of the signs of hurry that marred some of Kant's other
works. Indeed, the short time he devoted to thinking it out as a single
self-contained book and to its actual writing may be responsible for a
virtue singularly lacking in the Critique of Pure Reason, which he med-
itated on for twelve years and "brought to completion in some four or
five months,"l presumably by making more or less judicious use of
manuscripts composed at various times during the twelve-year period.
For, unlike the first Critique, the second adheres to one single straight
line of argument, developed without exploring blind alleys and with-
out getting almost lost on tangents. In it, we do not, as it were, listen to
Kant thinking aloud or watch him trying first this and then that way
of making himself understood. The book has the magisterial tone and
direct style that come to a man who has thought out all he wants to say
before he puts one word on paper.
By Kantian standards, at least, it is an exceedingly well-written book.
To many, this may seem damning with the faintest praise possible. Cer-
tainly Kant's style is not to everyone's liking, not even to his own. But
few men have had juster estimates of their own style than Kant had of
his, and his decision not to try to write in a popular way was justified
by the nature of the materials he was dealing with and by the demands
he rightly made of him who ventures to read philosophy.2 And it can
1 To Garve, August 7, 1783; to Mendelssohn, August 16, 1783 (X, 338,345). The
last pan of this sentence does not commit one to the "patchwork theory" (cf. H. J.
Paton, Kant's Metaphysics of Ezperience [London, 1936], I, 41).
2 For an explicit caveat, see Prolegomena, end of Introduction. In an undated
note <504"), he says: "If, like Hume, I had the power to embellish my work, I would
hesitate to use it. It is true that some readers will be frightened away by its dryness.
But is it not necessary to frighten some of them away, with whom the matter would
come into bad hands?"
3
4 Tbe lVriting of the "Critique of Practical Reason"
be excused, if an excuse is needed, by the urgency which marked his
literary activity in his old age. The fastidious reader should remember
that all the best-known works of Kant were written in excusable haste
by an old man who had first discovered his message to the world late
in life and who was never robust but carried (by modern university
standards) a crushing load of academic duties.
Given the circumstances, what is surprising is how well the book is
written. Style cannot be considered apart from content, and the
strongly masculine, sachlich, and encyclopedic style of Kant's thinking
permits and perhaps requires3 the often condemned Kantian style of
writing, which Schopenhauer called "brilliantly dry." It does not ex-
clude occasional eloquence and poignancy of expression, and it never
allows the tone to be less than elevated. Unfortunately, however, these
high virtues do not entirely exclude the vices of pedantic distinctions
followed by carelessness in their use, overcondensation 4 alternating
with repetitiousness, and lack of the art of perspicuous paragraphing.
These weaknesses cannot be denied, but they are frequently exag-
gerated by philosophers who write no better. It is often remarked by
Kant-scholars that when, after dogged effort, they come to the root of
some perplexing sentence or paragraph, they find that the complexities
of the writing cannot always be attributed to the ineptitude of the
writer, but more often to the demands of the thought expressed. They
often realize that it would be very difficult, if not impossible, to say
better exactly what Kant was saying and that any simplification of style
would almost inevitably involve oversimplification of the thought. At-
tempts at simplification have usually produced only faint and distorted
echoes of Kant's meaning, and most of them, if they succeed in making
him seem simple, do so only by making his views seem silly.
Except in a few passages, simplification of style is not needed in the
Critique of Practical Reason. There are few passages in which the at-
tentive reader will experience any real difficulty in discerning what
Kant means. Diagramming of sentences, often said in the legends of
3 Cf. Critique of Pure Reason, A xviii-xix.
4 Kant could appreciate a joke about his own style, at least this aspect of it. Zeiter
told the following anecdote to Goethe and said that Kant had been amused by the
event reported. Wlomer, a banker, told Kant that he had read some of his works
and would have read more if he had had more fingers. "How so?" inquired the
philosopher. "Well, my friend," replied the banker, "your way of writing is so rich
in parentheses and conditional clauses which I have to keep my eye on that I put
one finger on this word, a second on that, and so on for the third and fourth. Before
I get to the end of the page all my fingers are used up" (Vorlander, Immanuel
Kant: Der Mann und das Werk, II, 99).
The Writing of the "Critique of Practical Reason" 5
scholarship to be prerequisite to understanding the Critique of Pure
Reason, is not called for here.
We do not know much about the actual writing of the book. But in
the light of evidence about to be adduced, it is probable that the book
was begun in the spring of 1787, brought nearly to completion by June,
and finished by September. I speak only of the actual writing; most of
the contents of the book had been clear in Kant's mind at least as early
as 1785. But the plan of writing a Critique of Practical Reason as a
single, integral work did not slowly mature. The book was not long
anticipated, and Kant came suddenly, and presumably with reluctance,
to the decision to write it.
In order to understand this decision, it is necessary to go back to-
ward the beginning of Kant's literary career to see what the book was
that he did want and plan to write.
§ 2. THE DEFERRED PLAN TO WRITE A METAPHYSICS OF MORALS
Kant's writings in the 1750'S and 1760's were almost all in the field of
the natural sciences. But his interest in them was genuinely philosophi-
cal, and he was concerned most of all with questions of the method and
scope of the sciences. Even in these works, however, when occasion
offered, he commented freely upon ethical and religious questions
which had become involved in the cosmological speculations of the
day. This direction of his thinking is especially clear in his most impor-
tant scientific work, the Universal Natural History and Theory of the
Heavens (1755)'
The so-called Prize Essay: An Inquiry into the Evidence of the Prin-
ciples of Natural Theology and Morals, published by the Berlin Acad-
emy in 1764, is the first of his works which deals more than inciden-
tally with questions of moral philosophy. Even here, as is seen from its
title, it was the foundations and methods of ethics that were in the cen-
ter of his interest. This essay, often interpreted as manifesting a com-
mitment to the moral-sense theory of Shaftesbury and Hutcheson, and
hence to empiricism in ethics, ends with a query the answer to which
gives direction to all his subsequent work in moral philosophy. He says
there that "it is still a question to be settled whether it is simply the
cognitive faculty or whether it is feeling ... which decides the basic
principles of practical philosophy."1S
Metaphysics is regarded in the Prize Essay as a "philosophy of the
ultimate grounds of our knowledge," as philosophy "applied to the
more universal rational insights." While rejecting the alleged identity
15 Prize Essay, 11,300 (Beck, 28 s).
6 The Writing of the "Critique of Practical Reason"
of the methods of mathematics and those of metaphysics-an identity
almost definitive of the dominant metaphysics of the time-Kant be-
lieved that metaphysics, following its own method of analyzing experi-
ence, could find indemonstrable propositions as certain as those of
mathematics. 6 The basic indemonstrable propositions of metaphysics
would be self-evident rational principles, though not definitions or in-
tuitively certain axioms as in mathematics. He thought metaphysics to
be possible as a rigorous science, and already to exist in natural theol-
ogy. The metaphysical foundations of morals, however, were yet to be
found: "The primary grounds of morals are not yet, in their present
state, capable of all requisite evidence."7
Kant must soon have felt capable of supplying this lack. In a letter to
Lambert the next year, he announced a fateful plan that, in one form
or another, was to attract, even almost to haunt, him throughout his
mature life. It was the plan to write a work "on the metaphysical foun-
dations of practical philosophy.us The title of this work and a report on
his progress with it were given in a letter to Herder in 1768, in which
he said that he was then at work on a "Metaphysics of Morals" and that
he hoped to complete it within a year.'
We can only speculate on what would have been the contents and
structure of this work, had it been completed at that time. But we do
not need to make any conjectures about it for our present purposes.
61t was in this spirit that he favorably compared Rousseau to Newton, for both
discovered "the hidden law the observation of which justifies providence" (Bemerk-
ungen zu den Beobllchtungen, XX, 58-59). Similarly, he esteemed the work of
Shaftesbury, Hutcheson, and Hlime as containing the method by which "the abiding
nature of man" can be discerned in the variety of his empirical appearances. This,
he said, is "an excellent discovery of our days," and accordingly he promised to
evaluate historically and philosophically what happens before showing what ougbt
to happen (Nllcbricbt von der Einricbtung seiner Vorlesungen in dem Winterbalb-
jabr von 176$-1766, II, 311). Each of these passages indicated the analytical method
he was to use, beginning with observations and rising by abstraction from empirical
to metaphysical concepts. The passage in the announcement of his lectures, however,
does not indicate that Kant at this time thought a descriptive, merely empirical,
ethics to be possible, though it has been interpreted as evidence for this by some
writers.
7 Prize Essay, n, z9B (Beck, z8z).
16 Literally, "Critique of Pure Reason," though probably these words were not
meant then as a title of the book.
17 To Herz, end of 1773 (X, 144).
The Writing of the "Critique of Practical Reason" 9
purely intelligible world; but there was not yet any sign of his having
to give up the hope and expectation of being able to establish specula-
tive metaphysics on a solid ground. After the letter of 1772, however,
Kant's "recollection of Hume" awakened him from his "dogmatic
slumber," and he gave up forever the dream of a speculative metaphys-
ics of the intelligible world. Still, the transcendental philosophy or
critical examination of pure reason, as elaborated in the Critique of
pure Reason, did provide the prerequisites, in Kant's opinion, for a
"Metaphysics of Morals" in two senses: a systematic presentation of the
a priori laws of morality, and a practical rather than a speculative
answer to traditional metaphysical questions.
So in the Critique of Pure Reason, as it was at last published in 1781,
Kant still spoke of the metaphysics of morals. He distinguished two
parts of philosophy: critique, or propaedeutic philosophy, which "in-
vestigates the faculty of reason in all its pure a priori knowledge," and
metaphysics, or the "system of pure reason" which "exhibits in sys-
tematic connection the whole body . . . of philosophical knowledge
arising out of pure reason." "M:etaphysics," he continued, "is divided
into that of the speculative and that of the practical employment of
pure reason, and is therefore either metaphysics of nature or metaphys-
ics of morals."l8 The entire faculty of reason, including the practical,
was subjected to critique; that is, the Critique of Pure Reason was
meant to be propaedeutic to metaphysics in both its divisions. 19
In spite of mentioning the metaphysics of morals as a division of pure
philosophy, in the first edition of the Critique of Pure Reason Kant did
not promise to write such a work; he projected only a "Metaphysics
of Nature."20 Transcendental philosophy, which is the system of all
principles of reason so far as it concerns knowledge of objects, con-
tains only pure a priori concepts and principles, and thus it excludes
ethics. For though the highest principles of ethics were said even then
to be a priori and intellectual, "in the construction of a system of pure
morality empirical concepts must necessarily be brought into the con-
cept of duty."21
18 Critique of Pure Reason, A 841 = B 869.
19 Ibid., A xii, note.
20 I bid., A xxi.
21 Ibid., B 29. The Critique of Pure Reason excludes moral philosophy from
transcendental philosophy, not because the former is not pure, but because the
latter is concerned solely with the cognitive (cEo A 801 = B 829). Kant tacitly
widened the concept of transcendental philosophy (as he narrowed that of meta-
physics) until it is hardly distinguishable from critique itself; and though he never
explicitly included moral philosophy in transcendental philosophy, we shall see the
10 The Writing of the "Critique of Practical Reason"
Instead of a "metaphysics of morals," therefore, we have in the Cri-
tique of Pure Reason a "Canon of Pure Reason," i.e., a statement of the
a priori principles of the correct employment of pure reason. This em-
ployment is entirely and exclusively practical, not speculative. The
Canon does not answer the practical problem as such, viz., "What ought
I to do?" but the problem he calls both theoretical and practical, to wit,
"What may I hope if I do my duty?" Thus was introduced the discus-
sion of the immortality of the soul and the existence of God, which are
Ideas of pure theoretical reason, dialectical and empty for theoretical
cognition. It is only in relation to man's will that these Ideas have any
necessary use, as guides or regulative principles for the pursuit of hap-
piness through becoming worthy of it. 22
Again, Kant did not long remain satisfied with a mere Canon of Pure
Reason. He soon returned to the long-deferred plan, not mentioned in
the Critique itself, to write a systematic work on ethics, presumably to
include a metaphysics of morals, for he referred in 1783 to being at
full apparatus of transcendental philosophy in discussions included in the Critique
of Practical Reason.
The tacit inclusion is due not merely to a widening of the scope of transcendental
philosophy, however; it is also to be attributed to a deepening of the level of
moral analysis itself. Thus the passage JUSt quoted says that in a system of moral
philosophy empirical concepts will have to be brought into the concept of duty.
whereas in A 15, the corresponding passage in the first edition, he had said, with
less caution, that "the concepts of pleasure and pain, and of desires and inclinations,
etc. will have to be presupposed." In both editions he distinguished between the
doctrine of virtue, or applied ethics, and pure ethics, the fonner of which alone is
in part dependent on empirical or psychological principles (A 54-55 = B 79). The
Critique of Practical Reason and the Foundations of the Metaphysics of Morals are
meant, of course, to be pure ethics in the sense of this passage, even though in the
Critique he makes use of definitions of psychological concepts needed in the concept
of duty and in depicting the relation of man to the moral law, which requires no
psychological concepts for its fonnulation. But in actual composition of the second
Critique and even of the Metaphysics of Morals, as we shall see (below, p. 53),
Kant made no very consistent effort to separate discussion of pure from applied
ethics, or metaphysics from critique, or either from system.
We shall see that the ascribing of something very like, if not identical with,
transcendental status to practical principles was one of the factors leading to the
decision to publish the second Critique on the same level with the first and not
subordinate to it, as a metaphysics of morals would have been subordinate to it.
Later Kant came to regard the first Critique as not a sufficient propaedeutic to all
metaphysics, and he indicates this implicitly in constantly referring to the first
Critique not by its correct title but by such expressions as "the critique of specu-
lative reason," as in the first paragraph of the second CTitique (3[87]).
22 Critique of Pure Reason, A 806 = B 834. It is not quite accurate to say that this
is their only use; for they are also regulative of inquiry, but they contribute nothing
substantively to knowledge.
The Writing of the "Critique of Practical Reason" I I
§ I. INTRODUCfION
More than half the Preface of the Critique of Practical Reason is con-
cerned with issues that had been raised and, in Kant's view, settled in
the Critique of Pure Reason or deals with objections that had been
made by readers who detected what they regarded as inconsistencies
between the first Critique and the Foundations of the Metaphysics of
Morals. Since the Critique of Practical Reason was to repeat and elab-
orate some of the doctrines of the Foundations, Kant regarded it as es-
sential to his purposes to show that his doctrines here and in the F oun-
dations were consistent with those of the first Critique and, indeed, that
they gave additional support to his findings there. While the full proof
of consistency and mutual support could be given only in the body of
the Critique, there are many assertions in the Preface that can hardly
be appreciated unless the main conclusions of the earlier Critique are
kept in mind. In writing this book, Kant certainly assumed familiarity
with the former work. Yet he explains the fact that "concepts and
principles of the pure speculative reason are now and again reexamined
in this work"; only in this way, he says, can the "old and new," i.e., the
theoretical and the practical, uses of the "concepts of reason" be com-
pared and connected, so that the "new path" of the Critique of Practi-
cal Reason can be "clearly distinguished from the previous one."1
In this chapter I shall attempt to give a summary statement of the
argument and conclusions of the first Critique, sufficient only to make
the Preface easily comprehensible. From time to time in later parts of
this commentary it will be necessary to study specific parts of the ear-
lier work in order to provide the proper background for understand-
ing specific parts of this one. The reader who is generally familiar with
the Critique of Pure Reason will not need to go through this entire
chapter but is advised to go directly to § 6, which is more directly con-
cerned with the text of the Preface.
1 Critique of Practical Reason, 7 (91); d. Reflexionen 5019, 5036.
20 The Limits of Theoretical Reason
§ 2. THE PROBLEM OF THE "CRITIQUE OF PURE REASON"
The Critique of Pure Reason is a sustained effort to determine sys-
tematically the answer to the question "What can I know?" The an-
swer is "I can know the truths of mathematics and of the sciences of
nature, but 1 cannot know the objects of traditional speculative meta-
physics."
This answer is reached by examining the presuppositions of knowl-
edge. In this examination, it is found that the ground which makes it
possible for us to have knowledge in mathematics and science does not
extend to the alleged knowledge of metaphysical objects. This ground
is the givenness of the objects of knowledge to our sensibility, or to
what Kant calls "intuition."
All our knowledge and alleged knowledge is expressed in one of
three kinds of judgments.
i) A judgment may be analytic, such as "Every red object is col-
ored" or "Every body is extended in space." Such a judgment is called
"analytic" because its predicate is found by analysis of the concept of
the subject. Such judgments are certain, and they are important in or-
ganizing and articulating our knowledge. But an analytic judgment
does not extend our knowledge; such a judgment does not tell us more
than we already know, at least implicitly, in entertaining the concept
of the subject. Moreover, it does not tell us anything whatsoever about
the existence of the subject it mentions. "A triangle has three angles"
is absolutely true and would be just as true if no triangle had ever ex-
isted in the world; "God is a perfect being" is such a judgment, and we
do not need to know whether God exists in order to know that the
concept of God contains the concept of perfection as one of its predi-
cates.
ii) A judgment is synthetic if the predicate is not logically included
in the concept of the subject. Thus "This girl is young" and "All men
are less than two hundred years old" are synthetic judgments. Whether
they are true or false must be found out by experience; they are based
upon experience and refer to experience if they are true. Even if true,
however, such a judgment is not necessarily true; it is quite conceivable
that the girl is not as young as she looks and that human life might
extend beyond two centuries. Such synthetic judgments, whose truth
can be judged only in the light of experienced facts~ are called "a pos-
teriori," in distinction to the analytic judgments, which are a priori.
iii) Kant believed, however, that there are some synthetic judgments
which are necessarily true and therefore can be based neither upon ob-
servation and induction from experience nor upon a merely logical
analysis of the concepts they contain. These are synthetic a priori judg-
The Limits of Theoretical Reason 21
Note that only the possibility of another kind of causality with its
own law is established by the resolution of the third antinomy. Noth-
ing is said (except incidentally) in the Critique of Pure Reason to show
that freedom as a mode of causality is actual or that there is an a priori
law for such causality. If there is not such a causality, however, the at-
8 Critique of Pure Reason, B xxx.
7 This transition is fully discussed in chap. x.
The Limits of Theoretical Reason 17
tempt of theoretical reason to establish in principle a complete system
of causes is condemned to failure; even theoretical reason needs such a
concept for its own completion, but cannot establish it. Without such
a conception, however, the very being of theoretical reason is endan-
gered, and its lack may "plunge it into an abyss of skepticism" (3
[88]).
Such a concept can be established only by showing that it alone can
do for some realm of experience what the principle of natural causation
does for the sciences of nature. That is, there must be some realm of
experience which, upon analysis, shows the necessity of some a priori
synthetic judgment which is possible only if free causes are asserted
actually to exist. The Foundations of the Metaphysics of Morals and
the Critique of Practical Reason show that there is an unconditional
necessity in the moral law. The moral law is an a priori synthetic prac-
tical proposition, and these two works show that it is possible. It is pos-
sible if and only if the will is a free cause. "There really is freedom,
for this Idea is revealed by the moral law." The moral law is the ratio
cognoscendi of freedom, and freedom is the ratio essendi of the moral
law (4 n. [88 n.]).
Nevertheless, we thereby have no knowledge of freedom. A cate-
gory, that of causality, is applied to a supersensible object, viz., our-
selves as noumena. We think ourselves free, though in another context
(nature) we know ourselves as phenomena under the law of nature.
The contradiction between what we must think and what we know is
resolved just as the third antinomy was resolved: we distinguish our
reality from our phenomenal appearance. We thereby gain no knowl-
edge which has been interdicted by the Critique of Pure Reason, but
likewise, if we properly understand the meaning of practical reason, 8
we involve ourselves in no contradiction.
In the Dialectic of the first Critique, Kant considers two other Ideas,
viz., that of the soul as substance characterized by immortality and that
of God as a perfect being. He refutes arguments for each, but not to
prove that the soul is not immortal or that God does not exist. He
proves only that theoretical proof of each is impossible. Each is a neces-
sary object of thought, playing a regulative role in the guidance of our
search for completeness in theory; but neither is an object of knowl-
edge.
The Critique of Practical Reason converts the problematic judg-
ments of the Critique of Pure Reason (such as "The soul may, for all
8 Critique of Practical Reason, 6 (90): "This must have seemed an inconsistency so
long as the practical use of reason was known only by name," i.e., so long as it was
thought that practical reason was only a special kind of cognitive faculty.
28 The Limits of Theoretical Reason
we know, be immortal and God may exist") into assertoric judgments.
They were "mere Ideas, . . . unsupported by anything in speculative
reason," but they now "attach themselves to the concept of freedom
and gain, with it and through it, stability and objective reality" (4
[88]). This they do by being shown to be conditions of a necessary
object that must be possible if the law is not to be vain and delusive,
namely, the summum bonum (4 [89]).
But neither do these assertoric judgments, employing the categories
beyond the limits of experience, express or add to our knowledge.
They are postulates of a practical but rational faith, necessary because,
without them. moral experience could not be made fully intelligible.
There is nothing in them of a subjective will to believe or an emotional
need.9 They are rational, but not cognitive.
9 In the Preface, Kant is mainly concerned to deny the accusation that he has
treated these judgments as if they gave knowledge. Later (143 n. [242 n.]) he deals
with the opposite criticism that, since they are not cognitive, they must be arbitrary
and subjective.
III
was concerned only with "pure" in this sense ('iJber den Gebrauth teleologischer
Prinzipien in der Philosophie, VIII, 183-84). When used in reference to a faculty,
"pure" indicates that the faculty is a priori legislative (Critique of Judgment, V, 179
[.s)). It is imponant not to confuse these two meanings, though they are closely
related. The moral law is pure in senses a and b; the concept of duty is pure only
in sense a; practical reason is pure, or may be pure, in the sense that it is an a priori
legislating faculty, giving the moral law.
21 De anima 433 a 23.
22 Treatise of Human Nature, Book II, Part III, sec. iii (Selby-Bigge ed., p. 415).
Kant was not the first in his time to state that reason is both the necessary and the
sufficient condition of action; in this he was anticipated by a critic of Hutcheson's,
namely, John Balguy (The Foundation of Moral Goodness [1718] [Selby-Bigge,
British Moralists, II, 91-<}3]). and by Price (Review of the Principal Questions of
Morals [1758], chap. viii). There is no evidence that Kant knew of Balguy or Price.
23 For this distinction d. above, p. ,.
IV
§ I. INTRODUCTION
In this section I shall deal with the over-all structure of the Critique.
Details of the various subdivisions will be discussed as they are reached
in the course of the commentary.
Kant's architectonic has been treated sometimes with amusement,
often with contempt, and almost always with impatience. It has been
argued that, had he not held to the rigors of his architectonic schemes,
his works could have been organized in a less forbiddingly scholastic
pattern and could have been made shorter, since his ideal of complete-
ness of articulation led him to write whole chapters (e.g., at the end
of the first Critique) in which he had little interest and little to
contribute.
The Critique of Practical Reason is almost wholly free from exces-
sive architectonic zeal; only two parts may seem to owe their origin
to a forced parallelism with the first Critique, and in the course of
this commentary we shall find reason to believe that they are inte-
grally necessary parts of the organic structure of this work.
The organization is easily grasped but presents some interesting
features that call for discussion. There are two statements of the plan
of the work (16, 89-91 [102, 183-84]). In the latter statement there
is a comparison of the structure of the two Critique's; but, in making
this comparison, Kant incorrectly recalled the division of the first
Critique. This error is itself instructive. The accompanying tabulation
presents the structure of the relevant parts of the first Critique in fact,
their structure as Kant recalled it, and the structure of the correspond-
ing parts of the second Critique.
35 Tiibinger gelehrte Anzeigen, May 13, 1786. Kant returned to deal with Flatt,
one of the critics of whom he was most contemptuous, in Metaphysik der Sitten, VI,
1°7·
36 Critique of Practical Reason, 8 (94). This "truth-loving critic" is even more
fulsomely praised in Opus posttmtum, XXI, 416. Contrast with the contempt for
Flatt in Critique, 14 (100).
lIT Pistorius' review must have been known to Kant when he wrote the Preface to
the second edition of the Critique Of Pure Reason. But he said there (B, xli) that he
made no changes after the tirst chapter of the Dialectic "because I have not found
among competent and impartial critics any misapprehension of the remaining sec-
tions." Yet Pistorius' questions did concern the status of the Ideas and the noumena-
phenomena distinction. I do not know what may have been Kant's intentions with
respect to the Pistorius criticism in April, 1787. Perhaps the review by Pistorius was
in part responsible for his rewriting the section on the Paralogisms, though credit
for this is generally given to the Garve-Feder review, with its accusations of ideal-
ism. Whatever the cause, only a few weeks later he found that the understanding of
the Dialectic as a whole was so inadequate that it had to be explained again in the
second Critique. It may be worth noting that Kant praises Pistorius both times (ef.
n. 36 above) in the context of criticism (c) instead of (a).
88 Kant himself had mentioned the apparent circle in the argument (Foundations
of the Metaphysics of Morals 450 [6c))), and he explained it again to Kiesewetter
(letter of April 10, 1790 [XI, ISS]). But in the Foundations he pointed out a some-
What different route of escape. There he emphasized that there is in the experience of
thought itself an awareness of free spontaneity, and this is independent of the ques-
tion of the validity of the moral law. Even if one denied the reality of freedom in
words, he would be expressing his freedom as the precondition for his argument's
having any legitimate claim to validity. This argument is repeated in the review of
Schulz's Sittenlehre (VIII, 13) but is absent from the second Critique. The thought
was an enduring one, however, and the term "autonomy" is extended tinally to cover
theoretical as well as practical reason in Opus postu'I7rUm, XXI, 93, [00.
60 Name, Purpose, and Structure of the "Critique"
portion of Section I of chapter i of the Analytic, "Of the Deduction
of the Fundamental Principles of Pure Practical Reason."
c) The error, alleged again by Pistorius, that "the concept of the
good was not established before the moral principle."39 This is admit-
ted, and it is defended in chapter ii of the Analytic.
d) The alleged discovery that "there can be no a priori knowledge
at all" (I2 [97 D. This was the view of Feder,40 who had revised the
review submitted by Garve of the Critique of Pure Reason. By this
time, of course, Kant knew to what extent Feder was guilty and Garve
was innocent of the worst features of that review, to the stimulus of
which we owe so much of the Prolegomena. So Kant allowed himself
to indulge in bitterness and irony in dealing with this alleged dis-
covery.41
In reply to this, Kant takes the following course. He first shows that
it is inconsistent, for what we know by reason is a priori, and to prove
that we have no a priori knowledge would be to prove by reason that
reason is not valid. Next he criticizes the Humean conception':'! of cau-
sation, on the ground that its substitution of common human consent
for the a priori principle is inexplicable if the concept is "fundamen-
tally false and a mere delusion of thought," which he believes is the
correct inference to be drawn from Hume's premises. But, third, he
exculpates Hume from the absurdity of the doctrine of attempting to
prove by reason that there is no reason, by pointing out that Hume,
unlike the unnamed critic Feder, did not make his empiricism so uni-
versal as to include mathematics in his strictures. Mathematics escaped,
however, only because of Hume's error in regarding mathematical
39 Critique of Ptactical Reason, 8--<) (94--<)5) i cf. 62-64 (154-55).
'0 Uber Raum und Caussalitiit [sic] zur Prufung der Kantischen Philo sophie
(Gottingen, 1787). Kant identifies Feder as his target in the letter to Schutz, June 25,
1787 (X, 4(0).
41 With his characteristic and unvarying admiration for Hume, Kant first dis-
tinguished between the genius who made the original discovery and the epigone
who reduced it, in his opinion, to an absurdity, and he resen'ed his scorn for the
latter. The note at the end of the passage (13 n. [100 n.]) is a defense against mis-
labeling Hume a "genuine empiricist"; but in this defense Kant could not resist
making one more indirect rejoinder to Feder's having labeled him an idealist (cf.
the comparable footnote in Prolegomena, IV, 375n. [Beck, 124n.]). The allusion to
William Cheselden (J688-J752) refers to the physician who cured a blind man and
whose report on his adjustment to seeing was repeated by Herder (Vierter Wald
[J769], Werke [Suphan ed.]. IV. 49) and thus, probably, passed on to Kant.
42 I see no reason to think that Tetens was the opponent whom Kant had in mind
here, as asserted by Cay von Brockdorff, Die deutsche Aufklarungsphilosopbie
(Munich, 1916), p. 105.
N tr111e, Purpose, and Structure of tbe "Critique" 61
judgments as analytic. But though Hume was wrong in holding them
to be analytic,43 what is important here is that he correctly held them
to be a priori and thus maintained an internal standard of criticism for
other forms of knowledge. Genuine empiricism is not just skeptical of
philosophical speculation but would destroy both mathematics and
morality, and on these two counts it is not, according to Kant, to be
taken seriously, since here a philosophical theory, itself speculative,
directly collides with the "highest evidence."
A lesser man might have spent his limited time in extended debate
with his critics or (at his own expense) have ignored their objections
and misunderstandings. But it is one of the marks of Kant's greatness
that he was able to deal systematically, if not sympathetically, with
other men's errors and misunderstandings and out of the bits and pieces
of scholarly journalism to take materials to be built into the permanent
edifice of philosophy. It is to this aspect of his talent that we owe much
of the Critique of Practical Reason. The signs of it are scattered
throughout his work, but they are concentrated here in the 'Preface.
43 In Enquiry concerning Human Understanding, Sec. VII. Elsewhere (Critique of
Practical Reason [53 (143) J) Kant says Hume's views lead "inevitably to skepticism"
even with respect to mathematics. This might refer to Hume's actual, though
passing. skepticism of geometry in the Treatise, while the remark in the Preface
refers to the Enquiry. But it is more likely that the judgment does n()t pretend to
state a historical fact, but means only that empiricism, once on the skeptical path,
stops neither at the "chief branches of knowledge" nor at "ordinary reason," which
was the refuge of the common-sense philosophers, who wished to be skeptical of
rationalistic metaphysics while remaining credulous of other kinds of knowledge
that would be destroyed in a genuine and thoroughgoing skepticism.
Part II
v
Survey of the Analytic of Practical Reason
§ I. MEANING OF THE TERM "ANALYTIC"
Throughout the Critique, Kant uses the word "analytic" as though it
were entirely familiar to his readers. But we must refer back to the first
Critique if we are to understand precisely what he means by this word
and why he chose this name.
General logic, he there tells us, contains an analysis or resolution of
the fonnal procedure of understanding and reason into their elements
and exhibits them as principles of a logical criticism of knowledge; as
general logic, it entirely abstracts from all content of our judgments
and concepts and deals exclusively with their fonnal relations.! Tran-
scendental logic, on the other hand, does not disregard the important
distinction between a priori and a posteriori knowledge and is con-
cerned with the grounding and structure of concepts and judgments
known a priori. Transcendental logic, therefore, contains only rules of
the a priori thought of objects, and hence it excludes other concepts
and judgments which are fonnally identical with them as respects their
logical interrelations but which, as regards their cognitive content, are
drawn from experience.2 The analytical part of transcendental logic
resolves the a priori or "transcendental" operations of understanding
and reason into their elements, and thus it serves as the standard of
criticism for all our knowledge of objects, including even empiTical
knowledge. The transcendental analytic, therefore, is a "logic of truth"
and not, like the analytical part of general logic, a logic of mere formal
consistency. Transcendental analytic consists in the "dissection of our
a priori knowledge that pure understanding of itself yields"; the con-
cepts and principles which it exhibits must be pure and not empirical,
intellectual and not intuitive, simple and not composite, fundamental
and not derivative, systematically complete and not rhapsodically ag-
gregated.a
1 Critique of Pure Reason, A 60 = B 84-
2 Ibid., A 56 =B 60. a Ibid., A ~ =B 8<}.
65
66 Survey of the Analytic of Practical Reason
Except that it is practical reason and not understanding or theoretical
reason that is here under examniation, the term "analytic" has the same
meaning in the present Critique. The analytic is the "rule of truth" in
practical philosophy, and its most important tasks are the differentia-
tion between the (empirical) doctrine of happiness and the (pure) doc-
trine of morality (92 [186]) and the systematic exposition of the a
priori principles and concepts of the latter.
7S
76 The Analytic of Empirical Practical Reason
tween them, one can still discern a great difference in the connotations
of the two words. "Will" suggests more directly the dynamic impulse
involved in action; "practical reason" suggests something cold and de-
liberate and without any "push." Such difference in connotation must
have been present also in Kant's mind, for we have seen how he often
speaks of reason as determining the will, as if there were two factors
here in opposition to each other.
However little he is justified by his own definitions in speaking in
this way, it is easy to understand why he did so. In acts of will, there
are two distinguishable factors: a want, which I have called the "dy-
namic" or "conative" factor, and what we recognize as that which we
ought to do, the recognition of which I have called the "cognitive"
factor. What I want appears as an incentive, and it speaks to my in-
clination, impulse, drive, and propensity. What I ought to do in order
to achieve what I want is discerned by reason. Reason may discover a
rule for what ought to be done if we are to obtain what we want; in
this respect, Hume was right in calling reason the slave of the passions;
but it is an intelligent slave, which, while serving its master, neverthe-
less guides, disciplines, and in part controls the master for the master's
own best interests. When the conative factor is the condition which
makes a specific rule of action relevant to choice, the reasoning in-
volved in formulating or picking out and applying the rule to the spe-
cific situation is based on empirically given data of two kinds: the
present data (feeling), which indicate the existence of a particular wish
or want, and the facts of past experience, which indicate how this wish
or want can be satisfied.
Reasoning functioning in this way is called "empirical practical rea-
son." The analytic of empirical practical reason is the study of the fun-
damental principles and rules of the conduct of reasoning in this con-
currence with desire, whereby blind pushing desire is converted into
interest. Such a study to be undertaken here has two parts: a study of
the formal aspects of such reasoning (in the present chapter) and a
study of the dynamic factors and the specific kinds of volitions result-
ing from them (in the following chapter).
§ 2. PRACTICAL PRINCIPLES
Practical propositions "assert [aussagen] an action, through which,
as a necessary condition, an object becomes possible."2 The term
"practical proposition" therefore refers to all sorts of rules of action,
2 Vorlesungen uber Logik, S 31.
The Analytic of Empirical Practical Reason 77
such as directions for a construction in geometry, a recipe, a counsel,
or any statement of a means to an end. Any proposition which is
effective through being entertained in deliberation on action is a
practical proposition, even though its content may be the same as that
of a theoretical proposition expressing mere knowledge without direc-
tion for use. 3 The familar theoretical proposition, "Alcohol causes
drunkenness," becomes, when related to a want, a practical proposi-
tion, "Drink alcohol" or "Don't drink alcohol," depending upon the
direction of the specific want that makes theoretical knowledge of
the effect of alcohol practically relevant. Not every theoretical prop-
osition, of course, has a practical counterpart; a practical proposition
concerns only what is possible through will, and it states what, in some
sense, ought to be done instead of what is done.
The first definition in § I is that of a species of practical propositions,
viz., principles (GTU17dsatze). "Practical principles," we are told, "are
propositions which contain a general determination of the will, having
under it several practical rules." Practical propositions are thus of
different levels of generality, from a specific command like "Shut the
door now" or a general precept like "Do not take unnecessary chances
when driving an automobile" to a general policy of life such as "Do
not permit any offense to go unavenged." Kant wishes to preserve the
name "principle" for those on the third level of generality. This is in
keeping with the requirement, implicit in the word GTU17dsatz, that
principles must be basic, not derived from more fundamental proposi-
tions. In the Lectures on Logic/, he restricts the word "principle" even
further to propositions which are a priori synthetic. At this stage of
his exposition in the Critique of Practical Reason, however, he is not
prepared to assert that there are any practical principles in this strict
sense;5 here he means by "principle" only a proposition expressing a
3 Practical propositiOns differ from theoretical propositions only in their V orstel-
lungsRrt, not in their Inhalt, with the exception of practical propositions which
concern freedom under laws (Erne Einleitung in die Kritik der Urteilskraft, XX,
1915). The statement above, therefore, is true of "technically practical propositions"
but not, according to Kant, of "morally practical propositions." This is not quite
correct, however. If we maintain, more successfully than Kant did, the distinction
between imperative and law, a morally practical imperative (i.e., the categorical
imperative) will correspond to the moral law, which is a theoretical statement about
the freedom of rational beings in general.
f Vorlesungen iiber Logik, § 34 (IX, IlO); cf. also Critique of Pure Reason,
A J49 = B J88.
• They would be law,s, and he is not yet in a position to assert that there are any
practical laws.
78 The Analytic of Empirical Practical Reason
determination of the will which is not regarded as being taken for the
sake of some prior commitment or policy. 6
Fundamental practical propositions, then, are propositions which
contain a general determination (Bestimmung) of the will, having
under it several practical rules. What is a general determination of the
will? The word Bestimmung is one of Kant's favorite words, and he
overuses it. Among the English words needed to translate it are "defini-
tion," "determination," "predicate," "decision," and "motive." Here it
seems to have the following meanings: a determining cause of an
action, in which sense it means "motive" (Bestimmungsgrund), and a
decision that is taken, 1 by virtue of whkh the will is not indefinite
volition but a specific or determinate volition with a specific direction
and goal. The essential point in "general determination of the will,"
therefore, is that the practical fundamental principle must contain or
express a lasting policy or settled disposition of will, not a capricious
resolve or a variable rule of thumb.
Also in the first sentence the word "contain" requires brief comment.
Enthalten is a curious word here; one might say that a proposition
"contains" terms, but hardly that it "contains" a determination of the
will. One would be tempted to translate enthalten as "asserts," but this
would disjoin the proposition too much from the determination of the
will and make it merely theoretical. I think Kant means this: the
practical proposition formulates what the determination of the will is,
gives expression to it, and does not merely refer to it. It is a proposition
the knowledge or entertainment of which is itself a factor in the ~ill's
decision.
The question for an analytic of empirical practical reason is this:
What principle or principles exist which do indeed contain general
determinations of the will, so far as these determinations can be found
in experience? The question for an analytic of practical reason in
general is this: Are there any fundamental principles in the strict sense
of a priori synthetic propositions which contain a determination of the
will; and, if so, what are they?
6 Having formally introduced the word Grundsatz, Kant in characteristically
regrettable fashion frequently ignores it. (It has been said of Kant that he succeeded
in being technical without being precise.) Almost invariably in its stead he uses the
word Prinzip, which. though declared synonymous with Grundsatz (VorJesungen
fiber Logik, § 34). is much looser in meaning. Thus a rule or precept (V orscbrift)
is often caned a Prinzip, while a genuine Grundsatz is also called Prinzip (e.g., in
§ 4. p. 17 [ I1 4}). Where necessary for the sake of clarity, I shall refer to Grundsatz
in the strict sense as a "fundamental principle," even though this is strictly pleonastic.
T For example. Critique of Practical Reason, 90. four lines from end of page (184.
1.18).
The Analytic of Empirical Practical Reason 79
§ 3. RULES
Kant gives no formal definition of "rule," and his usage is highly
variable. Sometimes it means "law,"s sometimes "imperative,"9 and
sometimes merely a precept or common maxim which guides us in
dubious cases-the kind of rule that serves as a "go-cart for the stupid"
who do not trust themselves to act wisely in the complexity of par-
ticular cases and who therefore fall back upon some simple standard
commonplace. 1o
In the Remark on § I, however, he shows us that he means a rule to
be concerned with the empirical contingent character of action in spe-
cific situations, given the general determination expressed in the prin-
ciples. Rules are thus distinguished from fundamental principles which
express an actual policy of life (maxims) and from fundamental princi-
ples (if there are any) which are universally vaIid.l1 Rules express what
is, on the average, right under a general principle, but they do not hold
necessarily and without exception. They require healthy common sense
and good judgment for their wise application, and they cannot be ap-
plied automatically by appeal to a general policy.12
Kant does not say that the rules fall directly under the principles, as
if they were logically contained in them. They are contained, he says,
in the general determination of the will, not in the principle. Thus rules
may be identical when principles are different, and different when prin-
ciples are identical; two men can aim at the same end and achieve it by
following different rules. It is important to note that Kant must insist
upon this loose relationship between rule and principle; otherwise he
could not distinguish, within a single type of action, those cases of it
that are moral from those that are merely legal. Precisely this impor-
tant loose relationship is what is overlooked in some of the most com-
mon criticisms of Kant's ethics, those which insist that uniformity in
moral actions is overemphasized by Kant.
Kant is not asserting the universal applicability of any rules. The uni-
8/bid., 31 (119). They are confused also in the first Critique; cf. Paton, Ktmt's
Metaphysic of Experience, I, 493.
9 Critique of Practical Reason, lO, second paragraph (106).
10 Reflexioncn 5235-8; cf. Critique of Pure Reason, A 134 = B 173.
11 Foundations of the Metaphysics of Morals, 389 (4), distinguishes rule from law
by reference to the empirical conditions involved in the former. But the proper
dichotomy is between rule and principle, and under principles it is necessary to
distinguish laws from mere maxims; then rules fall under both maxims and laws.
12 The same is true of rules of theoretical judgment (d. Critique of Pure Reason,
A 133 = B (72).
80 The Analytic of Empirical Practical Reason
versality he insist!l upon is not that of a rule but that of a principle,
which permits latitude in the choice of rule. But his failure to abide by
his terminological distinction between "rule" and "principle" and his
tendency to regulate his own life by "maxims" (really rules, as de-
scribed by his acquaintances who wrote biographies of him) perhaps
led him into the error of sometimes thinking of rules as valid without
regard to circumstances, as in On an Alleged Right To Lie from Altru-
istic Motives written in reply to Benjamin Constant. I3
A rule is always a "product of reason," for only reason (as the fac-
ulty of thought in general, not as one of the three cognitive faculties)
can give us the knowledge of the relation of the means to an end which
is expressed in a rule. Because a rule has a cognitive component that is
invariant with respect to the purposes which may be served by using
the knowledge it expresses, Kant calls a rule for the use of a means to
an arbitrary end a "technical principle" (Prinzip; should, of course, be
Satz) and not a practical principle. 14 A technical rule is objectively
valid as a statement of what ought to be done by a rational being who
has the general determination under which the rule falls. But it is sub-
jectively practical, i.e., actually relevant in the determination of be-
havior only for a being who has the general determination in question.
We may therefore properly call rules which prescribe means to ends
"conditional rules," even though they embody knowledge independent
of the subjective condition of its use.
We cannot answer the general question, "What are the conditional
rules of an empirical practical reason?" The answer would be too com-
pendious to state: it would include all practical art, all applied science,
and all knowledge of the ways of the world. But we must ask, in an
analytic of practical reason generally, "Are there any unconditional
practical rules which follow directly from unconditional principles?"
And the answer to this question would have to be given in an analytic
of pure practical reason.
§ 4. MAXIMS
Practical principles are then classified as maxims or as laws, as sub-
jective or objective. Kant's fonnulation of this division is peculiar, how-
13 Recent studies have convincingly shown the relevance of circumstances to
rules in Kant's ethics; cf. especially H. J. Paton, "An Alleged Right To Lie: A
Problem in Kannan Ethics," Ktmt-Studien, XLV (1954), IC)O-203, and "Kant on
Friendship," Proceedings of the British Academy, XLII (1956), 45--66; Marcus G.
Singer, "The Categorical Imperative," Philosophical Review, LXIII (1954), 577-91;
and W.I. Matson, "Kant as Casuist," Journal of Philosophy, LI (1954),855-<)0.
14 Critique of Practical Reason, 26 n. (113 n.). In Critique of Judgment, Introduc-
tion, I. they are called "technically practical" as distinct from "morally practical."
The Analytic of Empirical Practical Reason 8.
ever, because the principium divisionis is not whether the principle
itself is regarded as objective or subjective but whether the condition
is valid for the individual only, or for every rational being. And, fur-
thermore, it is a logically faulty division, because "maxim" is broader
than "law" and, in fact, includes "law" as one of its species. We must
examine each of these peculiarities.
What is meant by "condition" (Bedingung)? Most simply, we can
say that it means the same as "determination" (Besti11lmung) in the
sense of a general setting of the wilf.1!S But why, then, in this paragraph
setting forth the formal definitions, where precision and elegance are
called for, did not Kant repeat the word "determination"? The reason
lies in his logical terminology. It was his practice to call the major
premise in a syllogism of the first figure a prinfiple (sometimes a rule),
and its middle term (the subject of the principle) he called the condi-
tion. 16 In the light of this usage, we can say that "condition" means
that which is involved in the maxim as the general determination of the
will, and thus it becomes the middle term in a practical syllogism:
To avenge a wrong is always my purpose-Maxim or principle.
To tell this lie would avenge a wrong-Rule.
Therefore, I purpose to tell this lie-Decision.17
HTo avenge a wrong" is the condition, and it may be held only by my-
self or ascribed to all rational beings.
I turn now to the relation between maxim and law. The term
"maxim," like "condition," is also taken from logic, deriving from senten-
tia maxima, the name of the first major premise in a polysyllogism. Logi-
cally, then, "maxim" and "principle" have the same meaning, and law
is only a species of them. In fact, however, Kant seems to distinguish
sharply between them; in the Foundations,18 he says that a maxim states
how we do behave and a law prescribes how we ought to behave. In-
stead of this dichotomy, however, he is interested in establishing the fol-
lowing possibilities: (a) a rational man acting according to some maxim
while holding it to be valid only for him, and for him only because its
condition is the actual state of his own motives; (b) a rational man rec-
ognizing a condition valid for, though not necessarily effective in, all
15 Indeed, Kaubler (Der Begriff der Triebfeder in Kants Ethik, p. 41) suggests
that Bedingung is a typographical error for Bestimmung; and BertiTnmung is used
in this sense in Metaphysik der Sitten, Einleitung (VI, 111).
16 Vorlesungen uber Logik, ~ 57 (IX, 110); Critique of Pure Reason, A 311 =
B 378, A 300 = B 357; cf. Critique of Practical Reason, 90 ( 184).
17 Critique of Practical Reason, 90 (184); Metapbysik der Sitten, VI, 437-38.
18 Foundations of the Metaphysics of Morals, 410D. (38 n.).
82 The Analytic of Empirical Practical Reason
rational beings; (c) a rational man recognizing a condition as present
and effective in all rational beings as such and therefore as valid for and
applicable to himself. There is thus a trichotomy, not a dichotomy, to
wit: (a) mere maxim, (b) law, (c) law which is also a maxim. Ill
The questions requiring an answer are these: (1) What is the su-
preme maxim or maxims of an empirically practical reason? (z) Are
they laws, or mere maxims? (3) Can a law itself be a maxim? The first
of these questions must be answered in the light of empirical facts abOllt
man and will concern us in the next chapter. The third question belongs
in the analytic of pure practical reason. The formal question in an ana-
lytic of empirical practical reason is the second.
§ 5. ARE THERE EMPIRICAL PRACTICAL LAWS?
Kant answers that they are mere maxims, and he supports this answer
with many arguments. In this section, however, we shall be concerned
only with the answer insofar as it is based upon purely formal or epis-
temological considerations; the other arguments properly fall within the
scope of the next chapter.
A mere maxim must depend upon conditions which reflect individ-
ual differences among rational beings. These differences are differences
in their desires, in the conative and not in the cognitive components of
their volitions. A law, on the other hand, might seem to depend upon
what all men do, in fact, have in common, and this need not necessarily
be something common to their cognitive faculties but might be some-
thing lying in their conative makeup. In fact, Kant believes that all men
do have a common desire, namely, the desire for happiness. Why, then,
are we not, in Kant's opinion, able to state a practical law based upon
this putative empirical fact-or, indeed, upon any other alleged fact,
if we are able to find some other common conative component in all
men?
A law cannot be based upon any actual features inductively discov-
ered in the generality of mankind. For let us suppose that one desire,
upon examination of a fair sample of mankind, turns out to be present
in them. We might then make the statement, "All men have desire D,"
in which D would be a general determination of their wills. But, even
so, D would not issue in a law that all men should act so as to satisfy D.
For let us see what would happen if we found a man in whom D was
absent, as we might always possibly find one by continuing our induc-
tion. The fact that everyone else had D would not constitute the slight-
est reason why he should do actions called for by D, or should feel any
shame for lacking D; it would, in fact, provide a ground for rejecting
19 Cf. Critique of Pure Reason, A 812 =B 840.
The Analytic of Empirical Practical Reason 83
the generalization. The generalization is a posteriori, but a law, says
Kant, must be universal and necessary, that is, a priori. This, he says, is
true of laws in natural science as well. 20
Few would agree with Kant in this sweeping statement. There is an
element of brute fact in what seem to us to be the supreme "laws of
nature," and they might, for all we know, have been quite different. It
is just a brute fact about our universe that bodies attract each other; yet
we call the description of this attraction a "law."
It is nevertheless clear that Kant is here pointing out an important
distinction in morals which depends upon our different responses to an
exception to an alleged law. There is no possible logical inference from
how men do, in fact, behave to how they ought to behave. No empiri-
cal generalization can do anything but surrender in the face of a well-
authenticated exception. An a priori principle or law, on the other
hand, can "condemn" an apparent exception, and it does so even in
natural science. A body that does not fall according to Galileo's law-
and none does-is simply not a freely falling body. We do not revise
the law when we find an "exception" to it; we reclassify the object. The
law which can be held in spite of apparent exceptions is accordingly
not a mere summary of experience; it is prescriptive, not merely de-
scriptive, even though experience may have suggested its form and may.
in fact, corroborate it. There are such laws in science; there seem to be
such laws for men's practical conduct.
A law is prescriptive, whether it is a theoretical or a practical law.
(Perhaps it would be better to say that a law is prescriptive but that it
may prescribe for theory-construction, in which case we call it "theo-
retical" even though it represents a practical decision, or it may pre-
scribe for conduct, in which case we call it "practical.") Kant, therefore,
in warning against allowing any empirical principle to be called a law
is calling attention to "the most important distinction which can be
considered in practical investigations" (26 [114]), to wit, that between
factual generalizations and practical!!1 prescriptions. No logical ir.fer-
ence between them is possible. There are, therefore. no empirical laws
of practical reaJon, but only empirical maxims (21 [108]). If there are
any laws, they are not empirical.
§ I. INTRODUCTlON
Empirical practical reason is reasoning used in the guidance of behav-
ior for the sake of satisfying some desire. In chapter vi we examined
the formal aspects of this guidance as exercised through principles,
maxims, rules, and imperatives. In this chapter we shall be concerned
with the moving forces in such conduct, with its efficient and final
causes instead of its formal causes. This will require us to try to under-
stand Kant's views of and terminology in psychology or, as he would
say, anthropology. We must understand what he means by desire, pleas-
ure, pain, interest, and happiness in their empirical, psychological set-
ting. We shall then follow Kant's argument that these concepts and the
principles employing them are not competent to fill the office ascribed
to moral concepts and principles.
§ 2. DESIRE
Desire is the faculty of a being, through its ideas (V orstellzmgen,
representations or "ideas" in the Lockean sense), to cause the reality of
the objects of these ideas. 1 This succinctly indicates the two factors in
desire that we have already distinguished: the cognitive factor, or idea,
and the dynamic or conative factor. The latter is sometimes called "de-
sire proper" (Begierde) j when it is of a settled dispositional character,
Kant calls it "inclination" (Neigung)j often the name "impulse"
(Antrieb) or "incentive" (Triebfeder) is applied to it. 2
1 Critique of Practical Reason, 9 n. (94 n.) j Metaphysik der Sitten, VI, ZJ I. Fol-
lowing Wolff and Baumgarten, the faculty of desire (Begehrungsvermogen) js
distinguished from that of aversion (Abscbeuungrvermogen) at Critique of Prac-
tical Reason, 58 (148), but no further use is made of this formal distinction, and
nothing is lost in Abbott's not speaking of faculties at all in this passage.
2 The usage of Neigung to refer to settled disposition is not always observed. It
generally means any inclination and is approximately equivalent to Triebfeder ex-
The Analytic of Empirical Practical Reason 91
Interest is that by which reason determines the will. 3 But since prac-
tical reason is identical with will, it is more accurate to say that interest
"indicates an incentive of the will as it is presented by reason" (79
[172]). In a dependent will, i.e., one not wholly rational, there is always
an interest which is expressed in a maxim; unless it is so expressed, and
thus subject to intelligent inspection, we do not have a will but only
blind impulse. Will differs from mere desire in that in the latter there is
an image of an object which is the target of behavior; while in willing,
on the other hand, there is also guidance by knowledge of a law or
principle which relates the action objectively to what it is that is de-
sired. Thus animals have desires, but only rational beings can have a
wiII.
The objects of the ideas-the objects being the final causes of action,
the ideas being among the efficient causes-are also called "objects of
interest or of inclination."4 The objects are the purposes of the action,
since purpose is defined as the concept of an object considered as the
cause of the reality of the object:~ The idea is one of the efficient causes
of the object by virtue of being one of the factors which determine the
cept when Triebfeder refers to the moral incentive or motive. Abbott translates
Triebfeder as "motive" or "spring." "Spring" follows a usage going back to the
early seventeenth century, but not common now. There is good etymological justi-
fication for it, since Feder refers, e.g., to the mainspring of a watch. "Motive" is a
less fortunate choice, because Kant carefully distinguished between Triebfeder and
BewegungsgTUntJ (= "motive") in the Foundations of the Metaphysics of Morals,
427 (45). Yet the fault is Kant's, not Abbott'S, for in this chapter of the Critique
he is using Triebfeder in the sense in which the Foundations defined Bewegungs-
grund. Picavet translates it as mobile, Capra as movente, Born (following a Kantian
parenthesis) as elater. I have followed the translation suggested by Greene and
Hudson in their rendition of Religion within the Limits of Reason Alone. But I
have done so with growing dissatisfaction, for the meaning of Triebfeder is obvious
to a Gennan, while incentive must be explained to a reader of English. It does
not seem possible to find an entirely suitable English equivalent, and I suspect that
the reason for this is that Kant himself did not use the word univocaJIy. On the
tenninology and its variations see Kaubler, Der Begriff der Triebfeder in Kants
Ethik (1917).
3 Foundations of the Metaphysics of Morals, 459 (80 n.); Metaphysik der Sitten,
VI, 212-13; Critique of Judgment,
§§ 2,3'
4i "Object of inclination" and "object of interest" must be distinguished from
"object of practical reason." It is only the latter that is defined as the good or evil.
"Object of choice [WiUkurJ" is used by Kant in the legal sense, referring to that
which I may dispose of as I choose, i.e., property (Metaphyrik der Sitten, VI, 246)
as well as in the sense of "that which I choose" (e.g., Critique of Practical Reason,
36 [125]).
GCritique of Judgmtnt, V, 180 (17).
92 The Analytic of Empirical Practical Reason
person's action which, in turn, will produce the object. An object is a
purpose if we ascribe its realization to a chain of causes one member of
which is an idea or conception of the object which determines the per-
son's behavior in actually producing the object.
The word "object" denotes two quite different things. It may mean
an actual state of affairs, a physical thing and its psychological effects
that can be brought into existence by action. The production of such
an object requires empirical knowledge of its causes and skill in apply-
ing this knowledge. It is in this sense only that the word "object" is to
be used in analyzing an empirical practical reason. But the word has an-
other, quite unusual, meaning. It may refer to an internal setting of the
will, to an act of decision itself without regard to the causality of will
in bringing its object (in the first sense) into existence. This is the
meaning that the word will have in the analysis of pure practical
reason. 6
§ 3. PLEASURE
Kant's theory of desire is hedonism. "To wish for something and to
have satisfaction in its existence, i.e., to take an interest in it, are iden-
tical."1 Desire is always directed to something-its object-the existence
of which is expected to give pleasure.
The converse, however, does not hold. While his theory of desire is
hedonistic, his theory of pleasure is not exclusively conative. For it is
possible to experience pleasure without antecedent desire. Pleasure can
arise from the mere contemplation of an object or from the experience
of an idea in imagination. Such pleasure is called "contemplative pleas-
ure" and is best seen in the enjoyment of beauty, where the pleasure is
characterized as being disinterested. Pleasure which is the goal and re-
8 Critique of Practical Reason, 15 (101). Similarly, Foundations of the Meta-
physics of Morals, 431, 1. u. (Abbott [50, four lines from bottom] paraphrases and
misses the word "object.") It is essential for Kant to use the word "object" in this
unusual way, for otherwise moral action as such would have to be declared without
object, i.e., without interest or purpose. 'Vhen he says that there is an interest present
in all action but that interest need not be assumed as the determining cause of all ac-
tion, he is referring to interest in the existence of the object, which is the first sense.
But there can be an interest in making the right decision whether the decision can be
rendered effective in the world or not. The latter is the moral interest, and its object
is a certain setting of the will to act from a certain motive (d. Foundationr of the
Metaphysics of Morals, 413 n. £30 n.], and Critique of Practical Reason, H [Il2]).
This is the sense of "object" in chap. ii of the Analytic, when Kant is discussing the
object of pure practical reason, and it will be discussed in chapter ix of this com-
mentary.
1 Critique of Judgment, ~ 4 (V, 20<) [43]).
The Analytic of Empirical Practical Reason 93
ward of action, however, is called "practical pleasure"-practical pleas-
ure is an "interest of inclination."8
Both kinds of pleasure may be defined by reference to the subjective
state of the person. Pleasure is the consciousness of the causality of an
idea to keep the subject in the state in which he is having this experi-
ence. This is true even of disinterested aesthetic pleasure. Pain is, by
corollary, our consciousness of an effect that an idea has in making us
try to, or wish to, change our subjective state.9 Pleasure is present when
there is harmony or facilitation of function; Kant calls it the idea of the
agreement of an object or action with the subjective condition of the
person (desire)1° or with the work of our perceptual and imaginative
capacities (as in the disinterested pleasure of art). It would perhaps be
clearer to say that pleasure is the feeling produced by such agreement.
Pain is the spur to activity to bring about such agreement and is thus
connected with desire. l l
The faculty or capacity for experiencing pleasure and pain is feeling.
Feeling is one species of the general affection of the sensibility (sensa-
tio ), the other species being sense proper (Sinn) .12 If the content of
our sensory experience is of such a kind that it can be related to an ob-
ject of knowledge and thus, notwithstanding its dependence upon the
subject's psychological constitution, can become a component in knowl-
edge of the object, the content is called "sensation." Receptivity to it
is sense proper. The color green, for instance, is a sensation and is sub-
jective, since it depends upon the person; yet it is related to the object
in such a way (by means of an a priori synthesis under the category of
subsistence and inherence) that one can correctly say that the object is
green. This objective attribution cannot take place with feelings, and
hence they are subjective in a double sense.
There are only two elementary feelings-pleasure and pain. All other
feelings, such as the feeling of the sublime, the beautiful, and respect
are defined by the accompaniments, contexts, causes, or "objects" of
the pleasure or pain we feel. Whether the origin of the pleasure lies in
some physical stimulation, the physical fulfilment of a desire, or some
8 Metaphysik der Shten, VI, 212.
II Anthropologie in pragmatiscber Hinsicht, § 60.
10 Critique of Practical Reason, 9 n. (94 n.), in the definitions drawn from psy-
Chology.
11 Antbrop%gie, § 60.
12 Metaphysik der Sit ten, Einleitung, Ii Critique of Practical Reason, 23 (109) j
Erne Einleitung in die Kritik der Urteilskraft, XX, 226. Pleasure is referred to
"inner sense." This is not the inner sense of the Critique of Pure Reason but the
inner sense or capacity of feeling as described in Antbropologie, § 15.
94 The Analytic of Empirical Practical Reason
idea held in contemplation, the feeling is always an effect upon our
sensibility. In a hedonistic calculus, there is no place for qualitative
differences (23 [I 10]).
A feeling of pleasure or pain, we have seen, cannot be ascribed as a
property to an object. It cannot even be related to the existence of an
object by a necessary rule. For whether a feeling of pleasure or pain
will arise in the presence of an object can be learned only in the actual
experience. All our action aimed at the satisfaction of desire, i.e., at
pleasure in the existence of an object, is therefore based upon experi-
ence; all knowledge of the pleasurableness of any existing state of affairs
is, therefore, a posteriori. From this it follows that no law, i.e., no nec-
essary principle valid for all rational beings, can be based upon any
maxim for the realization of practical pleasure.
§ 4- THE LOWER FACULTY OF DESIRE
The distinction between the lower and the higher faculties of desire
( 22 [109]) is the scholastic distinction between appetitus sensitivus and
appetitus rationalis, which, in turn, derives from the classical distinc-
tion between passion and will.1 3 The paragraphs before us are directed
primarily against "the otherwise acute men," the Wolffians,14 who,
while drawing this distinction, did not sharply distinguish between the
sensitive and the rational faculties in general. In Wolff's doctrine, de-
sire has as its condition the knowledge of a perfection. 15 If this knowl-
edge is obscure or confused, corresponding to the lower faculty of
sense, the desire can mislead us into erroneous or bad conduct; if it is
clear and distinct, coming from the higher cognitive faculty of under~
13 Thus Hutcheson (Essay on the Nature and Conduct of the Passions and Affec-
tions [Selby-Bigge, British Moralists, I, 400 n.] which existed in a German transla-
tion of 1765) quotes the scholastic definition of appetitus rationalis: "A constant
natural Disposition of Soul to desire what the Understanding, or these sublimer
Sensations, represent as Good," and adds, "This many call the Will as distinct from
the Passions." On the concept of "passion" and Kant's views of its relation to feeling
and reason, cf. Karl Bernecker, Kritische DaTStellung des Affekthegriffes 'LIon
Descartes his zur Gegenwart (Diss., Greifswald, 19 1 5).
14 Wolff, Psychologia empirica, § 584; Baumgarten, Metaphysica, § 689; Baumeister,
Philosophia definitiva, §§ 849,852,891.
15 Op. cit., §§ 887-90' Pleasure necessarily accompanies the awareness of a perfec-
tion, according to Wolff. This is denied by Kant (Erste Einleitung in die Kritik der
Urteilskraft, XX, 226). But Wolff is not, at least by his own profession, a hedonist,
for the perfection is desired because it is good and not because its achievement
gives pleasure (d. Verniinftige Gedancken von der Menscben Thun und Lassen,
H 14, 139; Psychologia empirica, §§ 511, 558-59). Many of Wolff's followers did
not observe this nice distinction, which resembles Butler's, and were easy targets
for Kant's criticism here.
The Analytic of Empirical Practical Reason 95
standing, the will is rightly guided to choose a real perfection. Hence
the work of reason or understanding in morals, as elsewhere in the ra-
tionalistic philosophy, is to bring our ideas to clearness and distinctness,
for the difference between a sensuous and a rational concept is a dif-
ference only in clarity and not in kind. Hence the Wolffians, Kant tells
us, should not have distinguished between two faculties of desire. For
them, consistency would require only one faculty, and only he is able
properly to distinguish two. The Wolffians should have admitted:
If the determination of the will rests on the feelings of agreeableness or
disagreeableness which [one] expects from any cause, it is all the same to him
through what kind of notion [whether intellectual or sensible] he is affected.
The only thing [one] considers in making a choice [if Wolff is correct] is
how great, how long lasting, how easily obtained, and how often repeated this
agreeableness is [23 (110)].
But the essential epistemological premise that Kant had held since the
Inaugural Dissertation was that there is a generic difference between
sensibility and understanding. The corresponding ethical thesis here is
that there is a generic difference between any sensuous desire, no mat-
ter how refined, with its maxim, no matter how clear and distinct, and
a higher faculty of desire and its principle, which is not determined by
desire for pleasure at all. In fact, this distinction is even sharper in
Kant's ethical writings than in his epistemology, since knowledge is al-
ways the joint product of sensibility and understanding, whereas mo-
rality depends solely upon the determination of conduct by the rational
faculty. Empirical practical reason is always concerned with the satis-
faction of the lower faculty of desire. If there is a pure practical rea-
son, it must be the higher faculty of desire which is not just a refined
Epicurean version of our empirical and animal nature. The higher fac-
ulty of desire must, of course, have an entirely different structure and
function from the lower faculty.
§ 5. MAXIMS OF THE LOWER FACULTY OF DESIRE
Our next task is to follow Kant's argument that the lower faculty of
desire cannot give rise to laws, but only to mere maxims. If he can show
this and if he can demonstrate that the moral law is a law in the strict
sense of the word, he will thereby show that an ethical system cannot
be based on an empirical foundation in the nature of the human lower
faculty of desire.
Theorem I is as follows: "All practical principles [Prinzipien] which
presuppose an object (material) of the faculty of desire as the deter-
mining ground [Bestimmungsgrund] of the will are, without excep-
tion, empirical and can furnish no practical laws" (2 I [107])' It will
96 The Analytic of Empirical Practical Reason
follow from this that all imperatives which constrain us or require ac-
tion for the satisfaction of the lower faculty of desire are hypothetical
imperatives.
Before studying the proof of this theorem, let us examine for a mo-
ment the word "material." "Material" here seems to be equated with
"object of the lower faculty of desire," but the connotations of the two
words are distinct, and the distinction is of great importance. "Mate-
rial" is also contrasted with "form" and "formal," and Kant means
here: All practical principles which figure as the cognitive component
in volition because of their content, i.e., their reference to an object of
desire, and not because of their form, are empirical. It is of the utmost
importance not to fall into the common misapprehension of this theo-
rem and to conclude that Kant means that the presence of a desire and
hence of a material disqualifies a maxim from being a law. He says ex-
plicitly that there must be an object of desire if there is to be action at
all (34 [123]). The theorem disqualifies only those maxims which are
chosen to guide conduct because of their content, i.e., because of their
reference to an object of desire (material) as the determining factor.
All maxims have material; but only the latter are material maxims. Con-
tent (object of desire) without form is blind impulse; form without
object of desire is practically ineffective-this is as true of Kant's ethics
as the corresponding sentence in the first Critique is of his theory of
knowledge.
I now give Kant's proof of the first part of the theorem, i.e., down
to the last six words. If the desire for an object is the condition for
enunciating the proposition expressing the determination of the will to
realize this object, the practical proposition is empirical. This is because
desire for the object is the ground for determination, desire for the ob-
ject being desire for the pleasure expected from its existence; and
knowledge that the object will give pleasure is merely empirical and
hence, at be~t, only probable. Therefore, the maxim which calls for such
action is valid only under the empirical and therefore uncertain condi-
tion that the object will in fact give pleasure.
Kant's proof can be strengthened, so as to make it independent of the
debatable hedonistic theory of desire, by considering the rules in-
volved in such an action. A rule for the realization of any object can be
learned only empirically. Therefore, a determination of the will and its
train of specific rules guiding action is doubly uncertain: it is uncer-
tain whether the production Of the object will satisfy the desire (what-
ever it is), and it is uncertain whether the object will be produced by
the action initiated and guided by the acceptance of the maxim.
The second paragraph gives the proof of the last part of the theorem,
The Analytic of Empirical Practical Reason 97
to wit, that such material principles can furnish no practical laws. It
would have been better to say. "cannot be practical laws," since his
own wording suggests that laws are related to principles in the way in
which a rule is related to them and thus creates the supposition that
there are a priori rules of morality.
A principle is a law when the condition of the general determina-
tion of the will is correctly regarded as a condition of the will of every
rational being as such. But whether the object will give pleasure is con-
tingent upon the condition of the subject, and this condition is known
only empirically and cannot be known to be necessarily and univer-
sally present in rational beings. To hold as a practical law, the condition
underlying the principle would have to be known to be necessarily
valid for rational beings as such, and this knowledge cannot be got by
any observation. When a theoretical proposition is made into a prac-
tical proposition because of our interest in its content, the practical
proposition is never a law; it is, at best, a mere maxim or a rule prac-
tically relevant on the assumption of the maxim.
With the proof of Theorem I, the stage is set for the analytic of
pure practical reason, which begins at Theorem III. Theorem II and
Remark II of § 8, however, are parts of an analytic of empirical prac-
tical reason, and to them we now turn.
§ 6. HAPPINESS
Theorem II is_"All material practical principles [Prinzipien] are, as
such, of one and the same kind, and belong under the general principle
[Prinzip; should be Grundsatz] of self-love or one's own happiness"
(22 [lOS]). In this section, we shall be concerned with the theorem
only in so far as it expresses the doctrine of hedonism and shall reserve
for later treatment those aspects of it concerned with the ethics of
egoism.
A rational being's consciousness of the agreeableness of life which
without interruption accompanies his whole existence is happiness. That
a being who desires pleasures desires happiness, and conversely, follows
from this. Therefore, any practical proposition which is based on the
desire for pleasure (a material principle) falls under the fundamental
principle of making desire for happiness the supreme motive in action
and choice. The desire for happiness, however, is not merely the ag-
gregate of our desires for pleasures. Happiness, unlike pleasure, is a
concept belonging to understanding, not to feeling; it is not itself a
direct object of anyone impulse. Its pursuit, therefore, is guided by a
maxim of higher order than the rules for the pursuit of a particular
pleasure and is therefore under the regulation of a general fundamental
98 The Analytic of Empirical Practical Reason
principle, while maxims or rules for the pursuit of particular pleasures
can hardly be called "principles" at all. But as happiness is our concept
of a systematic whole of pleasures, the maxims for the pursuit of
pleasures are under the restrictive condition of a general principle
which regulates their relations to each other. In this sense they "belong
to" the general principle (Grundsatz) of happiness.
From this is derived the corollary that all material practical principles
place the motive in the lower faculty of desire and that, if all our
principles were of this kind, there would be nothing worthy of the
name of the higher faculty of desire (22 [109]).
The second Remark in § 3 discusses further the insufficiency of the
desire for happiness as supplying a practical law (25 [I 12]). The desire
for happiness is admitted to be necessary in any rational being endowed
with feeling and to be one~ though not always an effective, determinant
of the being~s interests. But because happiness is merely the "general
name for subjective grounds of determination" and since these grounds
vary from person to person and from time to time within one person,
the principle cannot be a law~ which must be the same for all rational
beings. The diversity of interests summed up under the abstract name
of happiness, even supposing that all men consistently follow the maxim
of trying to achieve happiness as they conceive it, for themselves or
even for others, would lead necessarily to conflicts in conduct. 1s
Kant~s argument at this point leaves much to be desired in cogency.
He is placing a demand upon hedonism that his own theory cannot
meet~ to wit, that identity of principle should lead to identity of rule
and action. While the Critique does not make it as clear as does the
Metaphysics of Morals, there can be a variety of rules of action even
under the moral law itself. To require that the desire for happiness
should issue in identical and universally applicable specific rules and
to conclude that, because it does not do this, it is an inadequate founda-
tion for law is to demand too much of any principle. Kant's own
formulation of the law would not pass this test, though he seems to
have thought, in his oversimplified examples, that it could. The moral
law requires that we should seek the happiness of others; but it does
not require that all benevolent men should do the same action, for the
happiness of others is at least as variable as my own. That my concept
of happiness can issue only in contingent rules, therefore, does not
show that this principle is not a law and does not provide a basis for
destructively criticizing it.
The more valid argument that an empirical principle cannot be or
16 Critique of Practical Reason, 18 (115-16); d. also Versuch den BegTiff nega-
tiver Grossen in die Weltweisbeit einzufiihTe7I., II, t81-81, versus Maupenuis.
The Analytic of Empirical Prl1ctical Reason 99
furnish a law is given in the second paragraph of Remark II under § 3.
This argument, which is formal in character, has already been discussed
in chapter vi, § 5.
Objectionable as the argument in the Remark to § 3 may be from
the standpoint of fairness to his opponents, however, we should not
overlook it in favor of the purely fonnal argument, for in it Kant is
enunciating a richer principle for the moral imperative. He is saying
that a maxim must be universalizable in the sense not only that it must
not be self-contradictory or contradictory to the purposes of the person
who accepts it in a particular instance but that the actions which it
dictates must be compatible with each other. It is not really essential
that the actions be identical, though he implies that it is. It is essential,
however, that the actions which one man undertakes under a maxim
not be incompatible with the actions which I or another man under-
takes under the same maxim; and if they are, then the maxims cannot
be instances of a single law. In a fragment he says this more clearly
than anywhere else: "The regulative principle of freedom: that [the
actions] do not conflict; the constitutive principle: that they reciprocal-
ly promote each other [for the] purpose of happiness."17
If we apply this richer principle to the case in point, I could without
logical contradiction will that each man should strive only for his own
happiness. If I do so, I am willing that each man should follow a self-
consistent maxim but, at the same time, that he should follow a maxim
that will produce inco,!l1patible and mutually frustrating actions. This
maxim cannot be a law, because a law must be a unifying and not a
divisive factor in the world; it must make a system of ends possible. 18
In § 3 Kant is thinking too narrowly of a logical criterion of law,
where logic admits of only identity and difference in terms. But his
theory requires him to think in terms also of real repugnance and
hannony, and not merely of difference and identity. His brief and
pointed statements and examples usually illustrate the latter concern,
even when some of the former concern is presupposed. In the Meta-
physics of Morals, however, the former principle comes into its full
rights, though it has been made explicit in the concept of a realm of
ends in the Foundations and that of a moral commonwealth in Religion
'Within the Limits of Reason Alone.19
The inconsistency between saying that all material maxims are max-
ims of self-love and that some men have a direct and even general
interest in the happiness of others, therefore, does not mean that there
is an inconsistency in Kant's ethical theory. Whether we have or have
not a duty to consider our own happiness or that of others is not af-
fected by a decision as to whether we do, in fact, treasure the happiness
of others only insofar as it contributes to our own. But a serious in-
consistency in his psychological theory and in the analytic of empirical
practical reason is present.
This inconsistency was often present in the writers whom Kant
most admired. Butler, whom Kant presumably did not know as a writer
on ethical subjects,23 had been able to point the way to making it
psychologically intelligible that men could have, and indeed sometimes
do have, a direct sympathetic or benevolent interest in the welfare of
,. others. For philosophers like Butler and his opponent Hobbes,24 who
based their (different) ethical theories upon their (different) concep-
tions of human nature, it was of the utmost importance to know
whether men could possibly be genuinely unselfish. For Kant, on the
other hand, to whom the facts about human nature (whatever they
might be) do not suffice to determine man's duty, the dispute between
the egoists and the altruists lies this side of the ethical question. The
ethical question is independent of questions of human nature. But it
is nonetheless regrettable that Kant did not clarify his own thinking
about human nature ip. this important respect.
Let us not, with some other writers on Kant, exaggerate the im-
portance of this error. It has been said that, in order to break the hold
of pleasure (especially of one's own) on the faculty of desire and to
find maxims which are not selfish and hedonistic, Kant had to leave
the empirical realm altogether and base his ethics upon what amounted
to an actual renunciation of human nature. The conclusion is accord-
ingly drawn that if Kant had had a more adequate psychology, like
that of Butler, he would not have had to reject so completely the con-
ception of human nature as a basis for morals.
23 But he might have found the same thing in Wolff, Verniinftige Gedancken
'/Jon der Menschen Thun und Lassen, § 139, and in Hume, Inquiry concerning the
Principles of Morals, Appendix II.
24 Because Kant distinguished, as we have seen, between practical pleasure and
contemplative pleasure, he is not guilty of the oversimplification of the hedonists
and egoists criticized by Butler. What Butler says of pleasure in general, Kant says
only of pleasure in moral action: there is a pleasure in the action, but the interest in
the moral action is not an interest in the pleasure itself (Critique of Practical Reason,
116,160 [z 12,258]; Metaphysik der Sitte'Tl, VI, 378).
102 The Analytic of Empirical Practical Reason
This, however, is incorrect. Butler's analysis of desire may refute
Hobbes or Mandeville not only in psychology but in ethics putatively
based upon an analysis of human nature; but it does not touch Kant in
any essential point. For Kant's argument against the principle of private
pleasure is not that it is selfish in any morally repugnant sense, and his
argument that we should concern ourselves with the happiness of
others is not predicated upon the alleged fact that we do so concern
ourselves. Precisely the same infirmity in the moral argument for
selfishness is found in the moral argument for altruism, when each is
based upon real or alleged facts of human nature. 25 Neither desire-
that for my own happiness or that for the happiness of others-can
found a duty to seek either.
§ 8. HETERONOMY
All action directed to the satisfaction of an interest in an object of
desire is action according to rules which fall under the material of
some maxim. If constraint is necessary in order that these actions rather
than others take place, the action is commanded by a hypothetical
imperative. In actions under a hypothetical imperative, we have seen
that there are always two laws which are relevant. One of them comes
from reason itself and is the formal principle of hypothetical impera-
tives. The prudent man accepts this principle in the choice of actions
for the satisfaction of any desire. The other is a law of nature, or at
least some more or less trustworthy generalization about facts of ex-
perience, and states a connection between causes and effects. It supplies
the values, as it were, for the variables of the formal principle, so that
the prudent man not only decides to do that which will, in principle,
lead to his goal but decides what it is that will, in fact, lead to it. The
second of these Jaws does not come from reason itself but is borrowed
from experience. 26 In content, it is a theoretical proposition; only the
choice of a specific end by the acting person gives it practical relevance.
In following a hypothetical imperative, reason is thereby acting un-
der one law which it has not itself prescribed; it chooses a law, pre-
scribed as it were by nature to the person who has, as a matter of
empirical fact, the goal which can be achieved by use of this theoretical
knowledge. Kant borrows from politics his name for this mode of
action: he calls it "heteronomy" and distinguishes it from autonomy or
self-legislation. A reason which is the slave of the passions, a will which
follows the promptings of desire and chooses laws of nature as its
23 Foundations of the Metaphysics of Morals, 398 ([4); Critique of Practical
Reason, 34 (113)·
26 Foundations of the Metaphysics of Morais, 435-36 (54).
The Analytic of Empirical Practical Reason 103
matters of morality men differ with their environment, and that the
morality of one locality is not that of another."30 There is, of course,
only a short step from "custom" to "education," especially when ed-
ucation is as fallible and variable (because not guided by an ideal of
universal validity) as Montaigne regarded it.31
The criticism of Epicurus has already been given in Kant's general
condemnation of the hedonistic foundation for ethics, though he ad-
mires the consistency of the Epicurean doctrine, which he holds to be
higher than that of the moral-sense theory (24 [J I I]). In the Inaugural
Dissertation, § 9, Kant had placed Epicurus and Shaftesbury together
in reducing "the criteria of morals to the feeling of pleasure or un-
pleasantness," and for this he was criticized by Mendelssohn, who
rightly saw that there was a difference in saying, with Epicurus, that
pleasure is what makes good things good and in saying, with Shaftes-
bury, that there is a particular pleasure in the contemplation of the
good which serves as a mark of its presence.32 In now separating them,
Kant criticizes the moral-sense theory because, to the degree to which
it tries to distinguish the pleasurableness of the moral feeling from that
of physical feeling,33 it must already presuppose that the concept of
the good is held in mind, conformity to which could give the special
feeling of moral pleasure. Feeling itself has no cognitive value; we have
"no feeling for law as such"; for consciousness of the law is a product
of reason and, without it, conformity to law would bring no pleasure
(3 8 [ 12 9]) .
.But Kant is clearly wrong in this argument. He erroneously thinks
that, by showing the pleasure of the moral sense to be like all other
pleasures, he has a right to say that the moral-sense theory "reduces
everything to the desire for one's own happiness" ( 38 [I 28] ). This
confuses two very different things that the British philosophers had
kept properly separate, viz., the disinterested pleasure we experience
in doing something righteous or in contemplating a righteous action,
whatever it may be that makes it righteous, and the interest we have in
the pleasure accruing to us if we do or contemplate a certain action,
30 Lectures on Ethics, 11.
31 Montaigne characteristically says: ''Whatsoever is told us, and whatever we
learn, ever remember: it is a mortal hand that presents it, and a mortal hand that
receives it" (Essays, Book II, chap. xii), Kant must have known that this did not
represent the whole point of Montaigne's thought on morals, however, for the same
essay gives an extensive criticism of the variability of customs,
32 Mendelssohn to Kant, December 15,1770 (X, 114).
33 But it cannot do so; cf. Critique of Prllctical Reason, 13 (109); Metapbysik der
Sitten, VI, 376--77.
106 The Analytic of Empirical Practical Reason
which may be righteous or unrighteous. 34 If the former pleasure is
disinterested, as Kant might hold when he developed his views on moral
contentment (117 [2 14]), its presence could serve as a mark of ethical
value, comparable to the hedonic mark of beauty, without committing
him to the view that pleasure derived from the moral sense is the goal
of the action itself. If this feeling arises from the immediate apprehen-
sion of the fittingness of actions without the need of an antecedent con-
cept of this fittingness, which would be the concept of virtue, then the
sentence beginning "One must already value the importance of what
we call duty" (38 [128]) begs the question against the theory of the
moral sense. To the assumption that we pursue virtue for the sake of
the accompanying pleasure, both Hutcheson and Price anticipate Kant
in making precisely this objection. 35
In spite of his objections to it, however, Kant prefers the moral~
sense theory to an outright and consistent hedonism because "it pays
virtue the honor of ascribing the satisfaction and esteem for her direct-
ly to morality, and does not, as it were, say to her face that it is not
her beauty but only our advantage that attaches us to her."36
Perfection as an ontological c.oncept is rejected as empty and in-
definite.37 When made specifically moral perfection, it is circular to
argue that perfection is the source of the moral predicates, which must
have been added to it. 38 In the Critique he distinguishes transcendental
and metaphysical perfection from practical perfection, which is "fit-
tingness or sufficiency of any thing to any kinds of ends." This, how-
ever, is insufficient to define more than the highest form of talent and
skill, which are ethically neutral. Thus Wolff's concept of perfection
suffers from the same weakness as his universal practical philosophy,3D
to wit, it is so general that nothing specifically ethical can be derived
from it. Ends must be given, and, if ethical perfection is here in ques-
tion, ethical ends defined by some other principle than mere perfection
or a mere general will must be presupposed. The concept of perfection,
however, is esteemed by Kant more than any of the other heteronomous
84 Cf. Hutcheson, Inquiry concerning the Original of Our Ideas of Virtue or
Moral Good, Introduction (Selby-Bigge, British Moralists, I, 72).
35 Cf. Ibid., sec. II; and Review of the Principal Questions of Morals, chap. i, sec.
iii (Selby-Bigge, British Moralists, I, 92; II, 114).
36 Foundations of the Metaphysics of Morals, 443 (61).
37 As already recognized in the Lectures on Ethics (pp. 24, z6, 39) in his criticism
of Baumgarten, Wolff, and Cumberland that moral rules drawn from the concept
of perfection are "tautologous."
38 Foundations of tbe Metaphysics of Morals, 443 (62).
3D Ibid., 390 (5). Wolff's conception of perfection is given in Ontologia, ~ us.
The Analytic of Empirical Practical ReasO'Tl 107
16 The most notable of those who have fallen into this error is Max Scheler (Der
Formalimrus in der Etbik und die materiale Wertetbik, [Halle, 1916]; also in Jabr-
bucb fur Pbilosophie und phiinomenologische Forschung, Vols. I and II). Scheler's
Kant-criticism is that, for Kane, the a priori must be empty because it is purely
formal. But since ethical principles are also, for Scheler, a priori, there must be an
a priori which is material, i.e., given in an intuition of material value~essences. In
this way, Scheler believed he could avoid the "subjectivity" and "intellectualism"
of Kant's ethics, which are consequences, he held, of Kantian "empty fonnalism."
But this criticism misses the main point of Kant's distinction between the object of
the will, which is always present, and the object of the will as its detennining ground,
which is present only to empirical practical reason. Scheler, in his own construc-
tive work, is left with no rational criterion for material principles, which is what
the fonnal principle suppHes in Kant's theory. In the schema introduced below,
Scheler locates the a priori at a, Kant at h. But while b is purely fonnal, it is not
therefore empty; its content is a.
11 Cf. Reflexion 6633 (ca. 1770): "The supreme principles of moral decision are
rational, but they are only fonnal principles. They determine no purpose, but only
the moral fonn of every purpose. Therefore, according to this fonn in concreto
primary material principles occur [vorkommen]." On the control of one maxim by
another see Paton, The Categoricallmperat;ve, pp. 136-37.
The "Metaphysical Deduction" of Moral Law 119
the other. This is the sense in which the maxim of self-love is regarded
by Kant as the supreme material maxim. But, since even this maxim can
ground only hypothetical imperatives and the explicatum here is un-
conditional obligation, the maxim he seeks cannot be governing by vir-
tue of its material. The supreme maxim must select from among the
material maxims those which should be permitted in our practical syllo-
gisms,Is and it must do so by virtue of being a criterion of maxims with
respect to their form and independent of their content. That is what
Kant means when he says that the principle he is seeking must be
purely formal; that is his obscure way of saying that it must be a prin-
ciple regulating conduct through controlling the form of permissible
maxIms.
In the analysis of moral action, therefore, we must add one more
principle to the two practical principles uncovered in our analysis of
hypothetical imperatives. We must distinguish the following:
a) A material maxim setting forth some general condition of my
will.
b) A principle (law) which determines whether a is subject to mor-
ally permissible execution and which is itself a maxim.
c) The formal principle of the hypothetical imperative, which is the
condition of the choice of acts necessary to attain the end set forth in a,
the actual means being discovered by the cognition of causal connec-
tions between acts and consequences.
F or instance, my purpose (a) may be to promote the happiness of
others. The law (b) authorizes me to act upon this maxim. The prin-
ciple (c) selects from among my empirical cognitions those that make
specific actions incumbent upon me, granting the legitimacy of a. The
law (b) forbids certain types of maxim (a), allows some, and requires
still others.
The easiest case to discern in Kant's discussions is the first, so much
so that it is often taken as the only one. He is often interpreted lD as be-
lieving (i) that the moral law is only a criterion for the admissibility of
material maxims and (ii) that there is always a conflict between natu-
ral, spontaneous material maxims and the law. This interpretation is
only partially correct and must be supplemented in what follows. But
the reason why this half-correct interpretation is so easy lies in Kant's
realistic and unsentimental estimate of human character. Where a mate-
18 This is said in Foundations of the Metaphysics of Morals, 432, 444 (50, 63): a
will which made a material maxim its law would need still another law to give it the
proper universality.
19 If, that is, the objector has given up the still more common error of believing
that Kant thought that b had nothing to do with II at all.
120 The "Metaphysical Deduction" of Moral Law
rial maxim leads to action which does in fact accord with the law, any-
one with our philosopher's sense of hard fact and hatred of cant will
naturally regard the material as the motive and will distrust its embel-
lishment as duty (154 [252])' So that the role of each can be distin-
guished, Kant preferred to deal with the extreme cases in which the
two things, duty and desire, work at cross-purposes. 20 Hence his ethics
seems to be more an ethics of prohibition and of repression of the nat-
ural man than one of the celebration of spontaneous good nature. Un-
doubtedly, thIS semblance is based in part on Kant's own personal habitus
(at least by the time he reached the writing of this book); but even
more it is based upon many readers' failure to remember the polemical
situation in which Kant found himself, one in which he had to separate
and set apart and seem to set in opposition those things that had been
confused and even identified by others.
The most significant case, however, is that in which the moral law
requires me to act on certain maxims, whether they express an ante-
cedent condition of my will or not. If we are to justify the alleged di-
rect interest in moral action, the moral principle must become a maxim
with its own material. This material is not empirical, like happiness, but
is the morally good, which is defined by the principle itself and not de-
sired antecedently to it. Only if the formal principle can itself be taken
as a maxim, so that we strive to instill in ourselves a certain readiness to
act in a given way regardless of whether we have an antecedent dispo-
sition to these actions, can we say that pure reason, the exclusive source
of b, can be practical.
20 Cf. Critique of Practical Reason, 9Z--93 (186), where he compares his procedure
to that of a chemist in separating a compound into its elements. Kant's examples in
the Foundations are of this type. They concern, for instance, the man without sym-
pathy who acts from "practical love." If the purpose Kant had in devising these
examples is remembered, one can see the error of interpreting him as a proponent of
"sour duty." He is simply insisting that a and b be distinguished and that we not
attribute moral value to a alone but to a only when it is instilled in us or acted upon
because of b; what other values a may have is not being discussed here at all, but
the specific moral value is being made salient by contrast, i.e., by choosing cases of
a in which the only prima facie value is the moral one. These are not examples for
imitation but examples for illustration. We shaH return to this matter in chap. xii
when we discuss the relations between Kant and Schiller. It is a pity that the ques-
tion must be discussed so fully, but it is necessary because Kant did not make the
purpose of his examples absolutely clear with repeated statements that could not be
overlooked even by his opponents. Had he done so, the quantity of writing against
his ethics would have been halved and its quality raised immeasurably. I have dis-
cussed the misinterpretation of the examples which arises from neglect of the
polemical context in "Sir David Ross on Duty and Purpose in Kant," Philosophy
md Phenomenological Research, XVI (1955),98-107.
The "Metaphysical Deduction" of Moral Law I ZI
But we have not yet seen what this law is which requires that our
maxims have a certain form and which has an object that cannot be
derived from our natural desires.
§ 5. DEFINITION OF THE PRINCIPLE
The supreme formal principle states that a purely rational being
would act on no material maxim that could not be accepted as itself a
law for all rational beings. All maxims of a rational being as such are
independent of the differences between them, i.e., their desires; the
maxims of a rational being as such are, therefore, laws. The imperative,
which is the formula of this law's constraint on a partially rational be-
ing, is that a maxim should be followed only if it is a maxim that I
could will that all rational beings should follow. The imperative serves
as a criterion for maxims that may function in morally permissible
actions.
Since, however, to be moral and not merely legal, there must be a
direct interest in the action, the principle must not be merely a nega-
tive test for an antecedent maxim, as the previous paragraph makes it
seem. The principle must be something positive, something command-
ing us to establish maxims in our hearts that are recognized as valid for
us as rational beings because they are valid for rational beings generally
or willed to be the maxims that all rational beings should follow. In this
positive, constitutive sense, contrasted with the previous negative and
regulative sense of the principle, knowledge of the principle must itself
b~ a ground for action. We must not merely act on maxims that do pass
the test, but our supreme maxim must be to act upon them because they
do pass it.
ft is very easy to emphasize the negative, regulative aspect and to
neglect the positive, constitutive aspect of the principle and then to
suppose that Kant is refuted by showing that there are maxims which
could be universalized but which do not obligate me. 21 For instance, I
can easily will that everyone should write his name in his books, but
this has no effect on my duty. In the negative, regulative function,
there are, as it were, two volitions: a desire to do something, but only
if I can desire others to do it too. In the positive, constitutive function
there is a more intimate relation between them: J desire (or will) to do
something because I can will it for others too. Only if I will it for my-
self because I can will it for others, does the supreme moral maxim free
21 This error is attributed to Kaat by, among others, C. D. Broad, Five Types of
Ethical Theory (London, 1930), p. uS. For an analysis of the logic of this con-
troversy see Austin Duncan-Jones, "Kant on Universalization," Analysis, XVI
(1955),11-1+ It will be discussed morefully in chap. ix.
111 The "Metaphysical Deduction" of Moral LIl'W
itself from antecedent material maxims. This is, I think, intended in
Kant's statement of the imperative, in the words italicized: "So act that
the maxim of your will could always hold at the slD1Ie time as the
principle of a universallegislation."22
This is stated by Kant to be the sought-for moral law, or supreme
principle of pure practical reason. In fact, however, it is not a law, but
an imperative; and though he has distinguished between law and im-
perative, he nowhere in the Critique states the law in proper form. He
does so only in the Foundations, and there incidentally: "An abso-
lutely good will is one whose maxim can always include itself when re-
garded as a universal law," which is an a priori synthetic proposition.23
§ 6. AUTONOMY
A law is a principle that is universally valid for all rational beings. As
a principle, it must contain a general determination of the will. We
have been attempting to understand what this principle is, considered
as a practical proposition.
In the section immediately following upon his formulation of the
principle, however, Kant begins to use the word "principle" in another
sense. "Principle" has hitherto meant a basic, fundamental synthetic
a priori judgment; now it means the actual prius or condition that is
formulated by this judgment. The principle he now speaks of is not a
judgment but a condition of will or the "general determination" itself.
And he says: "The autonomy of the will is the sole principle of all
moral laws and of the duties conforming to them .... The moral law
expresses nothing else than the autonomy of pure practical reason, i.e.,
freedom" (33 [111]).
In §§ 5 and 6 of the Critique (18-29 [II 6- I 7]) Kant shows that a
formal (synthetic) maxim is sufficient to determine the will only if the
will is not wholly detennined by the material of its maxims. If a ra-
tional being follows a formal (synthetic) principle, he does so because
his will is free from empirical detennination. This is freedom in the
negative sense. And since a will requires some law for its determination,
such a being must follow a law not given to it by nature; this can only
be a law given completely by reason itself and not by reason working
22 Critique of Practical Reason, S 7 (30 [119]). The notion that there are two
volitions-one for my own action and another for the universalization of the muim
of the fanner volition-has not even any apparent verbal justification in later
fonnulas of the categorical imperative in the Foundations. Cf. especially the state-
ment of the "principle of autonomy"; "Never choose except in such a way that the
maxims of the choice are comprehended in the same volition as a universal law"
(Foundations of the Metaphysics of Morals, 440 [59]; italics added).
23 Foundations of the Metaphysics of Morals, 447 (66).
The "Metaphysical Deduction" of Moral Ltl'W 123
on the data of experience. Hence the will must be free also in a positive
sense, i.e., as self-determining.
Thus Kant concludes that a free will, a moral will, and a pure prac-
tical reason are identical. If reason were not capable of prescribing the
law to itself, it would have to borrow the law from nature, and such a
law can support only a heteronomous imperative. Only a law given by
myself as a member of the intelligible world can interest me directly as
a member of the empirical world; all other laws and actions under them
can interest me only indirectly.
Only a self-given law, therefore, can support the phenomena of
morality described above in § 3; and a self-given law of a rational being
must have the formula stated by Kant. With it, the metaphysical de-
duction of the principle is complete, though we shall continue, in the
next chapter, to follow Kant in matters corollary to it.
Before turning to these parts of the exposition of the concept, how-
ever, it is well to consider Kant's conclusions reached thus far, espe-
cially in the light of his philosophy of theoretical law. Some writers
have remarked on what they call the "paradox of Kantian ethics," the
paradox of his having begun with the most complete subjection of self
to law in the history of modern ethics and having ended with the com-
plete subordination of the law to the self. His central moral phenome-
non is the restricting of the self by law; its explanation is the lawgiving
of the self, or autonomy.
But there is no paradox here, and Kant's thought is a development of
an older concept in the light of his previous theoretical development of
the concept of the law in general. Wolff and other rationalists had
made the law of nature sovereign over man and had regarded reason as
man's moral mentor, since only reason could discover nature's law. 24
That obedience to the law of nature would bring men to perfer:tion and
happiness was obvious to them; but happiness was not, they thought,
the reason why men should obey the law of nature. Obligation fol-
lowed simply from an analysis of the concept of perfection. 25
Kant, however, was not, during the critical period, any more taken in
by this concept of natural obligation than by that of natural purposive-
24 Cf. Wolff, Verniinftige Gedancken von der Menschen Thun und Lassen, § 137:
The laws of nature require that we do that which contributes to Our perfection.
Our conscience requires that we act according to the laws of nature. The laws of
nature are the laws of morality and of reason, which discovers them. Similarly,
Crusius, Entwurf der nothwendigen Vernunftwahrheiten, H 133. 137: There is an
inborn idea of natural law, placed in us by God and recognized by conscience,
which is not a theoretical cognition but a Grundtrieb or has as its basis a Trieb of
the will.
25 In this they were followed by Kant in the Prize Essay, II, 199 (Beck, 183).
124 The "Metaphysical Deduction" of Moral Law
ness as a ground for morals. He saw that laws of nature are without
moral import and that they can be guides to action only if some pur-
pose is presupposed. The only purpose he could find which could be
served by our knowledge of natural law was happiness and all its pleas-
urable components. Hence he concluded that Wolff's practical philos-
ophy, though ostensibly an ethics of law, was really an ethics of desire,
unable to ground the authority of the law it celebrated.
Kant was not willing to surrender one jot of the authority of moral
law. His theoretical philosophy provided the clue to his concept of the
source of law. In his theoretical philosophy, reason is the source of the
law of nature; the laws of nature are not passively recorded but are
conditions we place upon experience as criteria of its objective signif-
icance. Only because of their origin can they have the necessity denied
by Hume or explained on the assumption that they are intrinsically and
self-evidently rational or revealed as the laws of God-explanations he
contemptuously rejected.
But if the source of the objective necessity of the theoretical laws of
nature lies in us as rational beings, it is easy to see how Kant could also
locate the source of moral law in autonomous reason and guarantee its
objective necessity at the same time. The only significant difference be-
tween them is that the former laws have a sensuous condition and are
therefore restricted to what is, where "is" means "is for us men, under
the sensuous condition of our experience and in accordance with the
second Postulate of Empirical Thought." The latter, on the other hand,
are not restricted to this sensuous condition and are regarded as laws of
a world we do not know actually to exist, but one which we know
ought, by our actions, to be brought into existence. Reason does not
find its demands for rationality fully satisfied in what is presented as
actual; its demands for complete rationality can be met only in a world
not actually given to us. The discrepancy between what reason de-
mands and what experience presents-this unsatisfied demand of theo-
retical reason-appears as the practical interest in the reign of a rational
law that does not actually reign in nature.
To a being like man, who legitimately uses existence only as a cate-
gory of what is presentable to the senses, such a law prescribes what
ought to be, not what is. Its claim on him lies in the fact that it is not
only a law for him but a law by him. If the law sprang from another
source, be it God himself, its claim on man would be only conditional.
It could be effective only through a system of reward and punishment
working upon his desires. The purity of morals would be destroyed.
Only autonomy can unite the two conceptions which are inherent in
Kant's ethics: that man is independent of everything outside him in the
The "Metaphysical Deduction" of Moral LIlW 125
determination of duty and yet that in the moral struggle we are com-
manded, and we are not "volunteers" (82, 85 [175, 178]).
Kant's ethics, historically, represents a transition between two great
conceptions of the relation of man to the world. Against the eight-
eenth-century position that man is a part of nature and ought to be
subservient to her laws, Kant reacted by inverting the order and mak-
ing nature what she is because of how she appears to us. Then he tran-
scended even this Copernican venture by daring to weigh nature in the
scales of reason and to declare that she is wanting and does not contain
the destiny of man. The practical-what man ought to be and how he
ought to transform his existence-in this conception takes precedence
over what nature is and what she demands of man as part of her order.
Nature produced man but brought him to the stage where he can final-
ly assert his independence of her.26
Continuance of this development of the theory of the creative self is
found in romanticism. For the romanticists, man stands above nature as
her author and judge; but the judge has lost his law. The universality
of law, which is Kant's heritage from the rationalists and naturalists of
the seventeenth and eighteenth centuries, was maintained and even mag-
nified by him, though he anchored it in the abstractly personal, in the
res cogitans. But with Herder and others, this was only a halfway point,
an unstable mixture of rationalism and subjectivism. The development
was to be completed, they thought, only by seeing the concretely per-
sonal, historical man as the source of law.
But, of course, under this dispensation law is no longer rational and
universal, and its necessity is only historical. It is only an expression of
feeling and history, of Zeitgeist, Volksgeist, and Sch'Wiirmerei. To
those who wished to take this last step, Kant addressed his essay on
"Orientation in Thinking": "Men of intellectual power and broad
minds! I honor your talents and love your feeling for humanity. But
have you considered what you do, and where you will end with your
attacks on reason? "27
With acute foresight, Kant told them what would be the end of those
attacks. But the attacks have continued, and we have seen their conse-
quences, and they are as Kant foretold.
26 Cf./dea for a Universal History, Thesis III (VIII, 19).
27 What Is Orientation in Thinking? VIII, 144 (Beck, 3°3).
IX
§ I. INTRODUCTION
(184). Elsewhere (Metaphysik der Sitten, VI,438) Kant ascribes to reason the re-
sponsibility for drawing the conclusion of the practical syllogism, but in this passage
he is obviously referring to the logical, rather than the real, Use of reason. In this
passage he calls practical reason in its real use (in the establishment of the principle
or major premise) the "practical understanding." In Foundations, 412 (29), he says
that, since reason is required for the derivation of actions from laws, will is nothing
else than practical reason. Here also we have to do with reason in its logical use,
and this follows Wolff's tenninology in Psychologia rationalis, §§ 494,528.
Practical Concepts tmd Judg;ment 129
the conditions under which objects can be given in harmony with these
concepts."6 Similarly, in the practic.al syllogism it must decide what
object or act falls under the principle which defines the good and state
the ground of this decision in the form of a rule of judgment, so that
moral judgment and possible facts and acts of experience can be
brought together in such a manner that the latter can be estimated.
But, though formally similar, the two syllogisms are quite different
in their epistemic or transcendental function. The concepts of pure
understanding or categories are applied to intuitions a priori, and the
intuitions are the "third thing" that mediates between concepts so that
an a priori synthesis of them is possible. But they do not apply directly
to intuitions; they apply only to certain formal patterns or structures
of them, which are isomorphous with the intellectual synthesis whose
rule is given in the corresponding category. These formal structures
are the schemata of the concepts. For instance, it is only because
"cause" and "effect" are both referred to intuitions in a specific tem-
poral relationship that we are able to make synthetic a priori judgments
of causality that apply to experiences. But the concept of the moral
good is an Idea that cannot be schematized; we cannot find a structure
of intuition that corresponds to it. The concept of lawgiving, which
is that of freedom in one of its senses, is the concept of a cause that
does not exist in time, and therefore the schema of causation does not
apply to moral decision and its expression in action with the same
epistemic consequences that it has in theoretical knowledge. Still, the
concept of the good, though defined by the principle, is to be applied
to particular acts or objects that can be created by our free choice in
the world of experience. The principle itself must be, as it were, sche-
matized (68 [160]), so that it does not remain empty and inapplicable
in the practical experience of the guidance of choice.
The chapter, accordingly, falls into three clearly distinguishable
parts. The first thirteen paragraphs define the concepts of good and
evil as the concepts of objects of practical reason. The next four para-
graphs develop the categories of practical reason in the light of the
principles. The last section prescribes the condition for the application
of these concepts and principles.
§ 3. THE CONCEPTS OF GOOD AND EVIL
22 As pointed out in Fragment 6, i.e., Reflexionen 7101 and 7104 (XVIII, 178, 184).
Practical Concepts and Judgment 139
The statement "I desire X" is a statement of fact, and at any moment
I may actually desire X and things I know to be incompatible with it.
But "I desire X rationally," i.e., "I will X, under a rule of reason that I
regard as valid not only for myself but for others" means "X is good."
This judgment can be a constraint on my wayward desires for those
things not compatible with X. The judgment "X is good" is a rational
estimate, believed to be objectively valid, and is distinguished from the
sensuous estimate of X, viz., "X would be pleasant" or "X would be
fun." Only a rational being can judge "X is good"-whether, in fact, X
does meet the requirements for being good or not-but an animal can
desire X. The manifold of desires is synthesized by reason according to
practical principles or rules, and their common focus is the object called
"good":
These categories of freedom-for we wish to call them this in contrast to
the theoretical concepts which are categories of nature-have a manifest ad-
vantage over the latter. The latter categories are only forms of thought.
which through universal concepts designate, in an indefinite manner. objects
in general for every intuition possible to us. The categories of freedom, on
the other hand, are elementary practical concepts which determine the free
faculty of choice [einer freien Willkiir]. Though no intuition exactly COrre-
sponding to this determination can be given to us, the free faculty of choice
has as its foundation [or: the determination of the free faculty of choice has
as its foundation] a pure practical law a priori, and this cannot be said for any
of the concepts of the theoretical use of our cognitive faculty [65 (I57)].
Before exploring and identifying this "manifest advantage," we must
pause a moment to consider what Kant has already said. "These cate-
gories of freedom"-earlier called "categories of practical reason" (I I
n. [96 n.])-what are they? The "these" indicates that they are the
"rules" previously mentioned; yet in the table itself it will appear that
Kant has not yet decided whether "rule" or "concept" is the better
name for them. Still, there is nothing wrong or even surprising in some-
times referring to a category as a rule and sometimes as a concept. 23 A
category, though a concept, contains (like any concept) a rule for the
synthesis of representations. (This is true not only of categories but of
every concept: the concept of "cat," for instance, though a posteriori,
is a rule for the synthesis of our perceptual and imaginative representa-
tions which are necessary for us to say, on the occasion of one or more
of them, "There is a car.") The elementary practical concepts are just
the concepts of good and evil in general; and the rules spell out their
Use in the synthesis of desires, observance of these rules being necessary
for the conversion of desire into rational desire, or will.
23 On the general problem of the relation of category to rule, d. Vaihinger, Com-
'lne1ZtilTzu Kantr Kritik der reinen Vermmft, 1,222.
I40 Practical Concepts tmd ]udg;ment
Next, Kant tells us, the categories of freedom are forms of determi-
nation of choice and therefore of the objects of rational choice, i.e., of
the good as object. They are not forms of objects simpliciter, which
would require that intuitions be synthesized; this would constitute
them objects of knowledge, not of practice. This independence from
intuition is the "manifest advantage" of the practical over the theoreti-
cal categories. Of any category that functions in the determination of
an object, it must be possible to exhibit an intuition or pattern of intui-
tions which corresponds to it; otherwise the concept is empty, and only
analytical judgments can be made upon it. The good and evil, how-
ever, have no corresponding intuitions, and no intuition can be given
corresponding to the determination of choice, while the determination
of an object of knowledge occurs through the schematization of the
concept.
What is actually given to keep the concepts of good and evil from
being. empty is not an intuition, the source of empirical factuality, but
a principle which Kant elsewhere 24 calls the "fact of pure reason." That
is to say, knowledge of the principle itself gives content to the practi-
cal concepts, whereas in the theoretical case the concepts must be es-
tablished with respect to intuition before synthetic principles can be
formulated with them. This constitutes the advantage of the practical
over the theoretical categories. Theoretical concepts can be synthesized
only by grace of intuition, which is wholly different from and inde-
pendent of the thought formulated in the categories. 25 Practical con-
24 Critique of Practical Reason, 31,41 (no, 131); cf. below, chap. x, § 3.
25 Thus practical concepts have a wider significance than theoretical concepts in
their cognitive function. See the remarkable statement that can hardly be under-
stood apart from the present context, though it occurs in the Foundations (411-11
[28]): speculative (theoretical) philosophy makes its principles depend upon "the
particular nature of human reason," where "human reason" refers to our entire
cognitive faculty for which intuition is always sensible, not intellectual. All princi-
ples of practical philosophy, on the other hand, depend solely upon pure reason
alone, in the strict sense of "reason." This, however, hardly seems compatible with
the earlier view in the Critique of Pure Reason that excluded practical philosophy
from transcendental philosophy, in spite of the fact that its principles were pure (d.
above, p. 9, n. 21) • But in the second edition, as we have seen, Kant pointed out that
empirical concepts were needed for defining the concept of duty. \Vhat Kant means
can best be stated as follows: The determination of the moral principle is inde-
pendent of the particular constitution of the human mind and is valid for all rational
beings, whether their reason be discursive, like ours, or intuitive and whether their
will be holy or only virtuous. But the concept of duty presupposes that there are
desires, and hence the concept of duty is not independent of our mental constitu-
tion. Concepts of objects used in our cognition are likewise dependent in their cog-
nitive use upon the fact that they are rules for the synthesis of intuitions which,
for us men, are given independently of the concepts.
Practical Concepts and Judgment 141
cept5, on the other hand, are justified by the fact of pure reason, with-
out needing any appeal to empirical or pure intuitions or feelings of
pleasure. Accordingly (omitting one sentence just paraphrased), Kant
continues:
Since in all the precepts of the pure will it is only a question of the deter-
mination of the will and not of the natural conditions (of practical ability)
for achieving its purpose, it thereby happens that practical concepts a priori
in relation to the supreme principle of freedom thereby become cognitions,
not needing to wait upon intuitions in order to acquire a meaning. This
occurs for the noteworthy reason that they themselves produce the reality
of that to which they refer (the intention of the will)-an achievement which
is in no way the business of theoretical concepts [66 (157)].
These sentences explicitly concern only the pure will and are true only
of it. "Practical concepts a priorin are direct cognitions of what ought
to be, because they are consequences of a primordial fact, the law of
pure practical reason that a rational being as such necessarily wills to
act in such and such ways. Their object is directly given (as what
ought to be, not as what is): the concept itself is a cognition, not need-
ing to await the schematization of the concept or its exhibition or con-
struction in intuition.
Here, then, we have the one instance in the whole of knowledge
where thought can create its object directly, and not by being a cause
of a series of events which may terminate in an object. (This object is
respect, which is a necessary corollary of knOWledge of the law.)26 But
the knowledge thus gained is not knowledge that its object exists; that
would be theoretical knowledge requiring intuition. The knowledge
thus gained is knowledge that something ought to be, and the recogni-
tion of this is a necessary condition of its actual existence as an instance
of the good. Kant would go even further and say that the concept cre-
ates the existence of the object,27 since the object is a certain setting of
the will:
Freedom is regarded as a kind of causality (not subject to empirical grounds
of determination) with reference to actions possible through it, which are
appearances in the world of sense, and ... consequently [freedom] is referred
to the categories of the possibility [of actions] in nature, while, however,
each category is taken so universally that the determining ground of the
causality can be placed beyond the world of sense in freedom as the property
of an intelligible being [67 (158)].
After giving his table, Kant says: "I add nothing here to illuminate
the table, for it is sufficiently understandable in itself." No one has ever
granted him this claim. Even Schlitz, one of his most faithful disciples,
felt called upon to suggest to the master a way in which it could be
improved. 29 Since, unlike the table of theoretical categories, almost no
further use was made of it30 in spite of Kant's statement that "such a
division based on principles is very useful in any science, for the sake
of both thoroughness and intelligibility" (67 [158]), we can read back
into it very little that can make it more intelligible. Most commenta-
tors have ignored it; only Mellin has wrestled with its many difficulties,
and, though some of his suggestions are shrewd and valuable, on some
points he was, I think, definitely in error. My own ingenuity is quite
insufficient to interpret all parts of it consistently with the ipsissima
'Verba. Nevertheless, the main features of it are not unintelligible.
Before embarking upon an interpretation of what Kant actually per-
formed here, let us consider what we might legitimately expect him to
do in formulating a table of the categories of freedom.
First, we should expect him to set out in a systematic order the ab-
stract and logically constant features of practical judgments. We should
expect some kind of systematization of various distinctions that he has
29 To Kant, June 23, 1788 (X, HI) ; cf. Ak., XIII, 219.
30 One might have expected the whole Metaphysics of Morals to be organized
with respect to it, as Metaphysical Foundations of tbe Natural Scie71ces 'was
organized according to the table in the first Critique. A few fragments suggest that
the Metaphysics of Morals may have been thought of in this way (d. Vorarhe;ten,
Ak., XXIII, 218, and the [unsent) letter to lung-Stilling written after March I, 1789
[XXIII, 495]). Once in the Foundations (prior to the construction of the table) he
made explicit use of the categories of quantity (436 [55]), and in the Metaphysics of
Morals he later made use of those of relation (Rechtslehre, § 4). Some (rather fancy)
use of the table is involved in Rell'k;on within the Limits of Reason Alone, VI, 101-2
(Greene and Hudson, 93).
Practical Concepts and Judgment [45
already made: laws and maxims, material and formal principles, cate-
gorical and hypothetical imperatives, imperatives of perfect and of im-
perfect obligation, legal and moral judgments, laws and imperatives,
etc.
Second, we should expect him to set up concepts (categories) of
good corresponding to each of the kinds of judgments. We should then
have a table of judgments or rules and a corresponding table of con-
cepts.
But he does not proceed in these ways. He assumes that the impor-
tant formal features of practical judgments are identical with those of
theoretical judgments, and thus he avails himself of the division of cate-
gories arising from the table of judgments in the first Critique. And he
does not clearly distinguish between the judgments and the concepts,
so that in the table we find sometimes one and sometimes the other.
Some of the difficulty is undoubtedly due to the fact that each of the
headings is, as it were, imbedded within the single category of causal-
ity. We are never told precisely what is meant by a "mode of a single
category," and we are left to imagine what is meant by the quantity,
quality, relation, and modality of the single category of causation.
Other difficulties arise from the fact that the categories of practical
reason are not purely formal like those of theoretical understanding,
but always have a content in the peculiar causality (freedom) which is
the ratio essendi of the moral good. The categories of good and evil do
not owe their origin solely to the forms of practical judgment, as the
logical organization of the table suggests, but to the classes of causes or
effects which are specified in various practical judgments. We can ex-
pect to find, therefore, purely logical distinctions mixed up with those
of the practical in general and with distinctions belonging to ethics
proper.
In order to interpret the table, I shall make one change in it. I shall
distinguish principle or rule from category or concept of the good;
where Kant states one, I shall try to supply the other.
§ 6. THE CATEGORIES OF QUANTITY
Practical judgments are either principles or rules. We are here con-
cerned only with principles. Principles are either mere maxims or laws.
Maxims are maxims either of an individual or of the generality of man-
kind. Laws are distinguished from mere maxims of either kind by be-
ing universally valid and by the requirement that wills under law
should coexist in the all-comprehensiveness of a total system of ends.31
We have, then, the following principles and categories:
31 Foundations of the Metaphysics of Morals, 438 (57).
146 Practical Concepts tmd Judgment
[. Mere maxim of individual Subjective, good for the individual
1. Maxim of human beings generally Objective, good for the generality of
(precepts) men
3. Laws as maxims Good for rational beings as such, both
objective and subjective
Three comments are called for by Kant's own listings. First, he er-
roneously contrasts maxim and principle. A maxim is a kind of princi-
ple. Second, he explicitly makes the third a synthesis of the first two,
and thus interprets laws as maxims for individuals. He is able to do this
under the judgment of totality, by virtue of the concept of a realm of
ends as an organic unity of various rational beings. 32 Third, there is a
radical difference between the good in the first two and in the third
categories, a difference which is not adequately shown as a difference
in quantity.
What kind of good is Kant talking about? Perhaps the first two are
das Wohl, the third das Gute. But this is not a quantitative difference.
If we follow Kant in saying that das Gute in a heteronomous system
means that which leads to the end (das Wohl) , then all three are
equally objective as judged by reason, though in Kant's table only the
second and third are called objective.
Here and elsewhere 33 Kant makes it clear that, under the category of
quantity, the transition to a strictly moral concept is reached with the
third category of this group.
§ 7. THE CATEGORIES OF QUALITY
Theoretical judgments, with respect to quality, are either affirmative,
negative, or infinite; the corresponding categories are reality, negation,
and limitation.
Practical judgments are, with respect to quality, either rules (in a
loose sense) of commission or of omission (counsels or commands to do
something or to leave something undone) or permissions to make some
exception to a rule of the former kind. Hence the judgments and cate-
gories of quality are as follows:
I. Rules of commission The good as that which is com-
manded.3'
1. Rules of omission The good as that which is to be
achieved through the avoidance of
certain actions
32 Ibid., 436 (54-55); d. Critique of Practical Reason, %8 (115).
33 Critique of Practical Reason, 67 (159); Foundations Of the Metaphysics of
Morals,43 6 (55),
34 We are not here concerned with the source of the commands or counsels. They
may be arbitrary fiats, civil laws, moral commands, or recipes for making sauces.
Practical Concepts and Judgment 147
3. Rules of exceptions The good as that which is appropri-
ate in obedience to a general pre-
cept
'7 If Mellin is correct, the third appears to be a logical division of duty, but "duty"
must then have a moral or juridical meaning lacking in Mellin's interpretation of the
second category. Bendavid attempted to make the third a synthesis of the tirst two,
but failed completely.
Practical C01lcepts tmd ludlr'11e1lt 151
48 Mellin (op. cit., p. 537) interprets only perfect duty as moral, thus diverging
from Kant's meaning. Hence the order of the categories in his rendition is as follows;
1. Actually dutiful actions, as those conforming to a law actually lying in reason as
such; and actions opposed to this law
3. Duties conforming to a law necessarily lying in reason as such
a) Duties by a necessary law, required by an apodictic imperative, the law being
the moral law (e.g., the duty to tell the truth) = perfect duties
b) Duties under a contingent imperative which corresponds to "a contingent law
(rule) lying in reason as such" (e.g., the duty of a philosopher to teach the
truth) = imperfect duties
The scheme proposed here is:
1. Actions which are, in fact, dutiful or opposed to duty; the legally correct or
incorrect actions
3. Actions which are necessary because they are dutiful; the morally good action
a) Actions or duties of perfect obligation
b) Actions or duties of imperfect obligation
The best evidence for the latter interpretation is that it conforms to the doctrine of
perfect and imperfect obligation in the Metaphyrik der Sitten, where both are
treated as apodictically necessary and differ only in what is commanded (a certain
action, action under a cenain maxim).
49 FoundatiOn! of the Metaphysics of Morals, 411 n. (39 n.); cf. Retlexionen 7114.
71640 7170. The distinction of the schools was that perfect duties could be enforced
by external legislation, while imperfect duties could not. This is only a corollary of
Kant's distinction and appears again in the parallel distinction between Rechtrpflicbt-
en and Tugendpflichten. By internalizing the concept of legislation as autonomy,
the emphasis is shifted from the aspect of enforcibility.
152 Practical Concepts and ludg;ment
maxim is self-contradictory when made into a law. An imperfect duty
is one in which the rule of exceptions may apply, in order to prevent
obedience to one rule from conflicting with another. Moral duties are
always imperfect; legal duties are perfect:
If the law can command only the maxims of the action but not the actions
themselves, this is a sign that it leaves open to obedience a certain latitude for
free choice.... Under a broad [imperfect] duty there is no permit to make
exceptions from a maxim of actions, but only a permit to limit one maxim
of duty by another.... Such imperfect duties are only duties of virtue [and
not duties of law or jurisprudence].GO
Despite all these puzzles, the most puzzling thing in the account that
Kant gives of modality is not in the listing of the specific categories but
in the statement he makes of the relation of modality to the other cate-
gories. "Finally," he says,
the categories of modality initiate the transition, though only in a problemati-
cal way, from practical principles in general to those of morality; and only
later will it be possible to establish the principles of morality in a do wnatic
form through the morallaw [67 (158)].
How can this sentence possibly be reconciled with the one that says
that the transition occurs within each group of categories?51
Perhaps the principles of modality in the first Critique, instead of the
categories of modality in the first Critique, can supply an answer. For
we are here dealing with a question of real possibility, not merely logi-
cal possibility. And in these questions it is principles ot modality that
are important. The second Postulate of Empirical Thought in General
is "That which is bound up with the material conditions of experience,
that is, with sensation, is actual."52 If all our thought of duty and of the
good is not to be merely a logical exercise in the analysis of perhaps
empty thoughts which are only logically possible, then some actual
"material conditions of experience" must be given with which these
concepts can be "bound up." This material condition, of course, can-
not for practical reason be sensation, but it must be something that
could make the same contribution to the practical categories that em-
pirical intuition (sensation) makes to the theoretical. This material con-
dition, or rather this analogon of a material condition, is presented by
the "sole fact of pure reason," the direct consciousness of the moral law
to which Kant appeals in the Deduction. 53
Thus interpreted, the two sentences are not inconsistent. In each
50 Metaphysik der Sitten, VI, 39'>.
51 Cf. above, p. 14Z.
52 =
Critique of Pure Reason, A 218 B 265-66.
53 Critique of Practical Reason, 31 (no); d. below, chap. x, ~ 2.
Practical Concepts and Judgment r 53
family of categories he meant that the transition occurs from the con-
cepts of practical reason in general or empirical practical reason to those
of pure practical reason, though we have seen grounds to doubt that
this actually occurs except in respect to quantity. But there is not yet
given any warrant for saying that any experience requires to be judged
in the light of the categories of pure practical reason. The transition is
made (or attempted) in each of the first three groups; yet this transi-
tion as a logical development is given weight by the categories of mo-
dality which relate the copula in every practical judgment to "thought
in general" without adding any new internal form or content to the
judgment.
Even so, it is not quite correct to say that the categories of modality
initiate this transition problematically. They show what are the three
possible relations of the "ought" to the totality of practical experience,
and in this sense they prepare the way for it. The actual transition,
however, is not made by or within the table, but by our being directed,
by the categories of modality, to the "fact of pure reason" which will
tie all of them to the "material condition" of practical experience. (It
must be remembered that "material condition" here means the fact of
pure reason, not the material of a practical judgment in the sense in
which "material" refers to an actual empirical desire.) By reference to
the experience of choice, all the logical distinctions drawn in the table
become prescriptive of possible meanings of the concept of the good;
by reference to the fact of pure reason, some one or more of each
group comes to be the formula for a moral obligation or the concept
of the moral good.
This ends our tedious examination of the table of categories. It is pre-
sented with hesitation, and it contains more questions than answers,
more conjectures than decisions. We regret that Kant did not work it
out with his accustomed architectonic skill. Something comparable to
it-I make no extravagant claim for my own version-is surely needed
to bring the doctrine of principles and the doctrine of concepts to-
gether. It is perhaps symptomatic of the poor construction of the table,
or at least of its obscurity, that little use was made of it in the book that
it should have served to organize. The cause of the failure may well be
that Kant was here approaching the principles of a philosophy of action
that he never achieved, which we in the twentieth century do not yet
have in the formalized manner that Kant was accustomed to give to all
his deepest work. He failed to achieve it not merely because of the in-
herent difficulty of the task but because he carried over into practical
philosophy a theory of logical fonn that was not adequate to elucidate
all practical judgments. Even today we are still seeking a complete the-
154 Practical Concepts tmd Judgment
ory of the fonns of decision and valuation; interesting work has been
done, but more important work remains to be done, on such notions as
the "quantity" of imperatives.~·
The "metaphysical deduction" of the principle of pure practical rea-
son lacked the elegance of the Metaphysical Deduction in the first
Critique because there was no sound "logical clue" in the table of
judgments suitable to the purposes of the second Critique. The attempt
subsequently made to use this table of judgments in elaborating a the-
ory of practical categories was therefore destined from the beginning
to be a hazardous operation, and we have sufficient evidence to justify
the verdict that in fact the results were somewhat artificial and arbi-
trary, without the completeness and elegance and necessity claimed for
the categories of theoretical reason. All the manifold objections made
to the listing and derivation of the categories of theoretical reason
seem to apply, a fortiori, to those of practical reason. This verdict must
stand, I think, until some Kant-scholar has been able to give a more
convincing exposition of them than I have achieved here.
§ 10. THE TYPIC OF PURE PRACTICAL JUDGMENTIIII
Judgment (Urteilskraft) is the art or faculty of applying a concept
or rule to a particular case.1I6 Kant shrewdly remarks that a weakness
uCf., for example. R. M. Hare, The Language of Morals (Oxford, 1951), pp.
187-<)1. If few effortS have been made by Kant~exegetes to make sense of the table
of categories, even fewer have been made to construct a better one in the spirit of
Kant's works. The only such attempt I have seen is by J. Stilling, "i)ber das Problem
der Freiheit auf Grund von Kants Kategorienlehre," Archiv fUr die Gcscbichtc der
Philosophie. XXI (1908), 518-34. XXII (1909), 1-17. Stilling says that his work
(which leads to sixteen "categories of freedom") is an elaboration of Albrecht
Krause's Die Gesetze des menschlichen Herzens dargestellt als formale Logik des
reinen Gefiihls (Lahr, 1876), a book I have not seen.
115 I take this occasion to call attention to the error in the Chicago edition of my
translation (p. 176), where the section title is incorrecdy given as "Of the Typic of
Pure Practical Reason." Barni made the Same mistake in his French translation.
116 Judgment is also the art of finding a rule for a case presented. This is called
"reflective judgment," in cOntrast to "determinative judgment" (Critique Of Judg~
ment, V, 179 [15]). In the Critique of Practical Reason Kant is concerned excusive1y
with detenninative judgment: the principle being given, find the case to which it
applies. That this is a fundamental insufficiency in his ethical theory is argued by
M. C. Nahm, " 'Sublimity' and the 'Moral Law' in Kant's Philosophy," Kant-Studien.
XLVIII (1957), 502-24. The creative venturesomeness of private moral judgment.
which may lead from new moral insights to new moral principles, is an important
aspect of moral life that does not receive proper attention in the Critique. For
Kant, moral concern lies in the application of known principles to given cases, not
in eliciting new principles from problems of conduct. Note Kant's remark (Critique
of Practical Reason, 8 n. [93 n.]) that no one would want to be the inventor of a
new moral principle.
Practical Concepts and Judgment 155
in this talent, which he ascribes to mother wit, cannot be perfected by
the learning of still other rules, since this talent would again be re-
quired for the proper application of them. Thus a man learned in
medicine may possess sound theoretical principles and know the rules,
but only practice which perfects a native gift can make him a skilful
physician. tl7
When the judgment or rule to be applied is a priori, as in the case
of theoretical categories, the role of judgment is the same. But it is
more difficult to understand how it functions because the concept did
not arise from the kind of experience to which it is to apply. Kant
holds that it is easy to see how one can say, "This plate is circular,"
since the empirical concept of plate includes the geometrical concept
of circle, or at least has been associated with it in experience; and we
can both think and intuit a circle. It is more difficult to see how one
can say, "The sun causes the stone to become warm," since, as Hume
showed, there is no intuition or datum of cause from which the con-
cept has been abstracted and to which it ostensively applies.
Kant's solution is presented in his doctrine of schematism. A schema
is a representation that is homogeneous with both an intuition and a
concept, so that the concept can refer to it directly and thence in-
directly to the specific intuitions. The schema of an empirical concept
is a kind of generic image which contains as its core the defined prop-
erties of the definiendum but covers also a range of variation that
makes it resemble all the members of a class. tl8 There can be no image,
however, for an a priori concept. Its schema is not an image but the
representation of the procedure of the imagination in synthesizing
from possible data of intuition in such a way that this synthesis is
homogeneous with the conceptual synthesis whose rule is the category
itself. The schema is both a rule of this procedure and its product,
which is the required formal structure of intuition to which the
category is applied. Thus the category of causation is the concept of
the dependence of the existence of one thing upon that of another.
But Hume showed that we have no intuition of such dependence. Yet
we do apply the concept to intuition under one condition, to wit, the
regular succession of phenomena according to a (putative) rule. The
schema of cause, therefore, consists not in some (impossible) image
additional to the images of two events but in the "succession of the
manifold [of intuitions] in so far as that succession is subject to a
rule."119 Whether a spe~ific succession is subject to a rule, of course,
is a question for empirical research to decide. But the schema exhibits
117 Critique of Pure Reason, A 133 =B 171.
1i8/bid., A 141 =B 180. lUI/bid., A 144 = B 18J.
1;6 Practical Concepts rmd Judgment
the concept of cause in such a way that it can be applied to these
sequences which ~o fall under the rule and can be withheld from
those which do not. Without the schema, we could not move from
the concept of mere logical dependence, which is analytical, to real
dependence, which is synthetic. 60
An analogous problem arises in judgments which are to subsume
actions under practical rules or concepts, where subsumption may
mean the estimation of a particular action in the light of a general rule
or the decision that a particular action would satisfy the rule and
should, therefore, be performed. To decide whether a particular act
possible in the sensuous world falls under the principles which define
the good requires the power of jUdgment. The problem of practical
judgment, when it is a matter of skill and even of prudence, is not
essentially different from that met with in theoretical judgment, since
all the practical propositions of skill and prudence have theoretical
correlates and since the criteria of a successful technique and of a
satisfactory mode of life are to be found in experience.
With pure practical judgment the problem is more difficult, since
the law is a law of reason, not of understanding, and no intuition can
be adequate to it. We can never be sure, in any experience, whether the
full terms of the moral law have been observed. It therefore seems
absurd, Kant says, to wish to find a case in experience, which stands
under theoretical categories, to which the moral law applies so that
the ideal of moral good can be exhibited in concreto. 61 Similarly, the
60 I do not wish to appear to be falling into the error of thinking that the forms
of judgment which generate the categories are {oons of analytical judgment only.
Certainly I can say "Ghosts cause miracles," and this is a synthetic judgment. But
unless intuitions are supplied, corresponding to the concepts of ghosts and miracles,
and unless they follow the schema of causation, such a synthetic judgment cannot
be justified. The only kind of dependence that can be known purely conceptually
is analytical dependence; intuitions for the schema are necessary if the asserted de-
pendence is to be synthetic. Yet the hypothetical judgment and the category of
causation are meant to apply, respectively, to synthetic (causal) judgments and to
the synthetic connection of phenomena whereby the occurrence of one is dependent
upon the occurrence of the other.
81 Critique of Practical Reason, 68 (r59). In Critique of Pure Reason, A 425 =
B 453, however, Kant states that moral philosophy, like mathematics, "can present
its principles, together with their practical consequences, one and all in concreto, in
what are at least possible experiences." But the contradiction between the two
passages is more apparent than real, as can be seen from the context of the latter
passage. Kant is there trying to show that the "skeptical method," i.e., the method
of the antinomies, is especially serviceable in speculative philosophy but useless in
natural science, mathematics, and morals, for none of these "lay claim to insight
into what is beyond the field of all possible experiences." The moral philosophy
which presents its cognitions in concreto is a philosophy already armed with a type
Practical C01lcepts and /udg;ment 157
morally good is not a natural property of an act standing in causal
or other categorical relations to others. Yet pure practical concepts
and principles must be applied, and they are applied even by the "com-
monest mind" with often more skill than by the moral philosopher.
The questions are: How is it done? What is it that will bridge the
conceptual gap between what ought to be and what is, so that the
concepts of the former may be applied in a definite way to, or withheld
from, at least parts of the latter so as to show that they are, or are
not, as demanded by the moral imperative? Without an answer to
these questions, the normative-descriptive distinction, upon which
Kant lays such enormous weight (26 [114]), marks an uncrossable
chasm which is incompatible with the very notion of the ought and
which is not, in fact, present in our ordinary moral concern as some-
thing uncrossable; we cross it every time we make a moral decision
or pass a moral judgment.
Kant's answer to these questions is his theory of the type or typic
of practical judgment. 62 Discussion of this has a place in the second
Critique analogous to that of the schematism in the first.
A practical principle does not, as such, give knowledge of any
empirical fact. While a practical judgment presupposes that what is
commanded or counseled is possible in the empirical world, this pos-
sibility is established by pure cognitive (theoretical) procedures. But
that an action possible for me in the world of sense is an action that
is morally possible 63 requires a different kind of judgment. We are not
concerned, in practical decision as such, with an action to which a
specific practical principle applies descriptively by virtue of its cogni-
tive component or correlate; we are concerned only with the decision
of judgment (though, of course, the typic was formulated after the passage in ques-
tion). The su11Tl11Um bonum, on the other hand, cannot be presented in concreto as a
cognition of practical reason, and there the method of the antinomies again comes
into its own.
62 Type: "That by which something is symbolized or figured; a symbol, emblem"
(Shorter Oxford English Dictionary). For full discussion of the meaning of this
term see Paton, The Categor;callmperative, pp. 160-61. Neither the New English
Dictionary nor Grimm's W orterbuch recognizes the use of "typic" (Typik) as a
noun; but it is clear that Kant wishes to distinguish between the type of moral
judgment (Typus) and the theory of the type (Typik). On symbol and schema cE.
Fortschritte der Metaphys;k, XX, 17~o, and Critique of Judgment, ~ 59.
Kant uses the terms "type of the concept" and ''type of the principle" as equivalent
(Critique of Practical Reason, 70 [161]).
63 Critique of Practical Reason, 57 (148): "To decide whether or not something is
an object of pure practical reason is only to discern the possibility of willing the
action."
158 Practical Crmcepts and Judg;ment
as to what action ought to be done among all those that are theoreti-
cally possible. We have, therefore, an analogy: a schema of a case
occurring according to a law is necessary for knowledge of the case,
while a schema of the law itselfu is necessary to connect, in practice,
possible events in sense experience with a cause under a law which is
not a law of natural connection. Because the schema of a case has an
intuitive component, it is always cognitive in function. But no intui-
tion is available to a practical law of what ought to be, for two reasons:
(a) it is a law of what ought to be, not of what is, and (b) only one
member of the moral motive-act nexus is within nature. A schema of
the law itself must be provided by the understanding or the faculty
of thought and not by intuition or imagination.
Just as a schema was the "third thing" that could mediate between
pure concept and pure intuition, the type must be a third thing that
can mediate between the concept of nature, all that is, and the concept
of what ought to be. The third thing in the practical judgment is the
concept of law itself as definitive of a realm or kingdom." Nature is
phenomena under law, and natural law provides a type or model 66
by which we can think the practical law in concreto.
We ask ourselves, in face of temptation, "What sort of world would
this be if everybody acted the way I wish to act? Would I be willing
to create such a world, or to live in it if it existed?" But the desire to
live in a world in which, for example, everyone developed his talents
is not the reason why I should develop mine; it is not the motive 61
of my efforts at improving myself, for the obvious and simple reason
6t Kant points out the danger of the expression "schema of the law" (ibid., 68
[160]). He uses the word "schema" in a broader sense in Reflexion 5611; in Meta-
physik der Sitten, § 45 (VI,468) it means hardly more than example. In the Founda-
tions neither "schema" nor "type" is used, but there is a brief reference to the prob-
lem under the name "canon of moral estimation" (414 [4d). Similarly, Zum ewigen
Frieden in der Philosopbie, VIII, 410: "canon of morally practical reason." Re-
flexion 7160: "analogon."
611 Foundations of the Metaphysics of Morals, 433 (51).
66 Perpetual Peace, VlII, 371 (Beck, 37): Natural law is a "model" for constitu-
tional construction.
61 Critique of Practical Reason, 69-70 (161). Failure to recognize the difference
between the type of moral judgment and the motive of moral action has led some
critics to say that in the categorical imperative Kant has committed himself to a
utilitarian or even egoistic doctrine (cl., for example, Mill, Utilitarianism, chap. i).
But Kant did not fall into this confusion, which he warns against. He was quite well
aware of the fact that if I lie, it does not mean that all other men will lie, and there-
fore the fear of the consequences of my lying does not include the fear of the con-
sequences which would follow from the existence of a world in which everyone lied.
ef. especially Foundations of the Metaphysics of Morals, 438 (57).
Practical Concepts and Judgment 159
that I am not the creator of a world, and in this world I know, all too
well, that not everyone is going to act the way I do, be it for good or
ill. But implicit in the notion of a moral order is that of an order of
interacting wills (the third category of relation), and the best model
we have for such a world is the order of nature under law. 68 That is,
what would not be possible in an order of nature under law is not
morally possible, though what is actual in nature (a matter which
empirical intuition must decide) has no judicative function in the
abstract determination of what is morally possible and necessary.6D
"The order of nature under law" means two things, one of which,
as Paton points out,1° is largely forgotten today. First, it means a
uniform sequence of phenomena under causal law, and its distinguish-
ing feature is the universal uniformity of nature. Nature, as a mechani-
cal system, was believed, in the eighteenth century, to be "governed"
by such laws having universal application. Before passing to the second
of the meanings of the "order of nature," we must inquire into the
practical significance of this concept of uniformity.
The first test of a maxim is the mere universalizability of the maxim,
i.e., the interpretation of it as a descriptive universal principle. Of some
maxims, Kant says that they destroy themselves if made universal. 71
68 From time to time Kant pushed the analogy very far. Thus he early compared
the good will in Rousseau to the force of gravity in Newton, seeing in each an organ-
izing principle in an orderly realm (Triiume eines Geistersehers, II, 330, 335; Re-
flexion 5419). He told Mendelssohn (April 8, 1766 [X, 72]) that this did not express
a serious opinion but was only an example of how far one could go in Erdichtung
where data are lacking for knowledge. But the thought evidently had a certain
fasc1nation for him, for we find it again in Metaphysik der Sitten, Tugendlehre,
S 14 (VI, 449) and Opus postumum, XXI, 35.
69 Though, of course, it has a contributory function, since every action is predi-
cated upon empirical knowledge, which provides the content for the abstract prin-
ciple of choosing means suitable to a given end.
70 The Categorical Imperative, pp. 150 ff., 161 ff. I hesitate to follow Paton in
only one point in his definitive and exemplary exposition: I do not think that he is
correct in asserting (p. 149) that only the second of these conceptions of the order
of nature plays a part in the typic. The concept of causal uniformity does have, I
think, at least a minor role, as I try to show in the text.
71 Kant says this repeatedly. In the Critique, however, he speaks also of actions as
being universalizable. (Note that in the first and second formulas in the Foundations
£42 r <38, 39)] he speaks of universalizable maxims, while in the typic it is the uni-
versalizability of the action itself which is mentioned.) This is an inaccuracy in the
Critique. If lying were universal, we would be able to get along far better than in
this world, where it is only frequent; we should simply interpret affirmative sen-
tences negatively and negative ones affirmatively. But if the ma:rim is to deceive
another person, the best way of doing it is by sometimes telling the truth and some-
times not.
160 Practical Concepts and Judgment
That is, the effectiveness of such a maxim is dependent upon the fact
that it does not correspond to a universal. or even general, description
of human behavior. It is not possible to will rationally that such a
maxim should be universal. even though the proposition "All men
should lie" is not logically self-contradictory. A maxim like "I should
lie" depends for its effectiveness upon the fact that it is not universal,
that its theoretical correlate "I lie" is not universalizable into a judg-
ment, "All men always lie"; for. if it were, there would be no such
thing as a lie at al1,12 One's lies show mendacity and cleverness only
because they are exceptions to a general rule. But general rules which
have exceptions are not laws of nature; the latter have no exceptions.
Attention should here be called to a matter already discussed above
(p. 12 I). The universalizability of a maxim is a negative test of its
validity as a law. But many maxims can in fact be universalized which
do not have the status of law. In this way, the type of the moral law
as a universal and uniform natural law is only a negative criterion for
moral judgment.
The second meaning of the order of nature is one according to
which all the laws and the phenomena under them are in such a rela-
tion that nature as a whole can be interpreted as an organic unity, 73
which suggested to the natural theologian that it had been designed
by a wise creator. For Kant, the inference to a wise creator is logically
invalid; but the thought of a teleological organization is required as a
regulative principle in our search for as yet unknown causal connec-
tions. Natural theology and natural teleology regard the world of
nature as such a realm. though for Kant only the Idea of it is a valid
methodological assumption. Moral teleology. on the other hand. sees
such a realm as an ideal to be achieved in action;74 it is a regulative
Idea for practice and not for knowledge only. This Idea is that of a
realm of ends, organized by the third category of practical relation.
that of community of persons under common law, the whole being
§ I. INTRODUCfION
It has been noted that the structure of Kant's exposition in §§ 1-6 and
in chapter ii is analytical and problematical. These pages inquire into
what would be the form of the law if the hypothesis is correct that
reason of itself alone can give an a priori practical law. While there
is a categorical finality in Kant's statements about the character of
practical reason as such, he is proceeding only analytically in §§ 1-6
with respect to pure practical reason. He is asking only: If pure reason
is practical, what is its law and what is the nature of the will that can
obey this law?
Nevetheless, a man who believes as finnly as Kant did that the moral
law is a law of pure reason cannot keep this belief entirely under
cover. There are several places where his belief in the reality of pure
practical reason is obvious, before he has reached the place in his
treatise where these assertions are systematically justified. 1 Even in the
Critique of Pure Reason, which does not pretend to show that freedom
is real but only that it is possible, he often breaks through this logical
reticence and comes out with assertions such as this: "I am justified
r
in making the assumption that moral laws a priori can determine the
will] in that I can appeal not only to the proofs employed by the
most enlightened moralists but to the moral judgment of every man,
in so far as he makes the effort to think such a law clearIy."2 The
"moral judgment of every man" is the true starting point of the Kantian
moral philosophy, and this fact should not be forgotten, however
synthetic, rather than analytical, the structure of the second Critique
is meant to be.
1 E.g., Critique of Practical Reason, 19.27,29 (1°5,115,117).
2 Critique of Pure Reason, A 807 = B 835.
"Transcendental Deduction" of the Principle I6S
Kant's readiness to appeal to what the ordinary man thinks in moral
matters does not strike the twentieth-century reader as being starding,
for to us this is the only place to begin. But the sharpness of the reply
to Titte13 shows that there was a spirit of novelty (at least in Con-
tinental philosophy) in going into moral problems from this direction,
instead of in the more usual way of erecting a world system that might
subsequently be used to explain, or often to explain away, ordinary
moral phenomena. In this procedure, Kant is indebted to Rousseau·
and to the conscientiousness of pietism, which manifested itself in the
profound but simple-minded righteousness of his parents, whom he
openly revered all his life. Yet this moral consciousness may be il-
lusory; duty may be a "vain delusion and a chimerical concept."5 We
do not know that true virtue can be found anywhere in the world,
even if the concept is valid; and no examples can prove that it does
exist. 6
The tone of Kant's argument suddenly changes in § 7. He no longer
attempts to restrict himself to hypothetical statements about what
would be true if pure reason were practical, but boldly asserts: "Pure
reason is practical of itself alone, and it gives (to man) a universal
law, which we call the moral law."7 To show the contrast between
the analysis and the assertion, compare the following two passages:
Sometimes we find, or at least believe we find, that the ideas of reason have
in actual fact proved their causality in respect to the actions of men, as appear-
ances. S
3 Critique of Practical Reason, 8 n. (93 n.); cf. Critique of Pure Reason, A 831 =
B 859: "Do you really require that a mode of knowledge that concerns all men
should transcend the common understanding and should only be revealed to you
by philosophers? Precisely what you find fault with is the best confirmation of the
correctness of these assertions."
4 "By inclination I am an inquirer. I feel a consuming thirst for knowledge, the
unrest which goes with desire to progress in it, and satisfaction in every advance in
it. There was a time when I believed this constituted the honor of humanity, and I
despised the people, who know nothing. Rousseau corrected me in this. This blind-
ing prejudice disappeared and I learned to honor man. I would find myself more
useless than the common laborer if I did not believe that this attitude of mine [as
an inquirer] can give worth to all others in establishing the rights of mankind" (XX,
44)·
5 Foundations of tbe Metaphysics of Mora/s,402 (18).
6/bid·,40 7 (23-24).
7 Critique of Practical Reason, 31 (120). Here appears the confusion between law
and imperative which we have already noticed occasionally. Kant should have said:
"It gives a universal law which we call the 'moral law,' valid for all rational beings,
and to man it issues a corresponding categorical imperative."
8 Critique of Pure Reason, A 550 = B 578.
166 "Transcendental Deduction" of the Principle
One need only analyze the sentence which men pass upon the lawfulness of
their actions to see in every case that their reason, incorruptible and self-con-
strained, in every action holds up the maxim of the will to the pure will, i.e.,
to itself regarded as a priori practical []l (Il 0 ) J.
What is it that authorizes Kant to make this change in the status of
the hypothesis that pure reason can be practical? A change in mood
does not of itself constitute a step in argument.
There are two reasons for it: the alleged "fact of pure reason" and
the somewhat equivocally titled "deduction" of the principle.
z. THE "FACT OF PURE REASON"
§
What was previously only a methodological standpoint, the assump-
tion of moral consciousness, now functions as an actual premise of
the argument, in spite of Kant's having acknowledged that it might be
illusory. He says:
The consciousness of this fundamental law may be called a fact of reason,
since one cannot ferret it out from any antecedent data of reason and since
I the law] forces itself upon us as a synthetic proposition a priori based on
no pure or empirical intuition [3 I (no)].
In order to regard this law without any misinterpretation as given, one must
note that it is not an empirical fact but the sole fact of pure reason [3 I (110)].
This Analytic proves that pure reason can be practical, i.e., that of itself
and independently of everything empirical it can determine the will. This it
does through a fact wherein pure reason shows itself actually to be practical.
This fact is autonomy in the principle of morality.1I
The moral law is given, as an apodictically certain fact. as it were, of pure
reason, a fact of which we are a priori conscious. 10
II
•
Critique of Practical Reason, 42 (131). In the next paragraph, this fact is said
to be identical with the consciousness of the freedom of the will.
10 Ibid., 47 (136). Note that this sentence says that it is "as it were a fact." On
pp.6, 31,42,4] (91,110,131, IJ2) Kant calls it a "fact"; and on pp. 47,55.91, and 104
(136, 145. 18 5, 199) it is called a "fact, as it were," or some similar qualification is
made. (Abbott omitted this qualification on his pp. 136 and J85.) This has suggested
to some that Kant was uncomfortable with this unusual use of the word "fact" and
became cautious after its earlier use. But though this is the most plausible interpre-
tation of the difference, his caution departed when he had finished the CTitiq~, for
we find the following: Critique of ludy;m.ent, § 91 (V, 468 [po-liD-freedom is the
fact; Metllphysik deT Sitten, Rechtslehre \ 6 (VI, 251)-practical law of freedom
is a fact; VorllTbeiten ZUT Tugendlehre, XXIII, J78-"It is res facti that this law is
in us and is indeed the highest"; Opus postumum, XXI, zI-the categorical impera-
tive is the fact of moral-practical reason.
Critique of PUTe Reason, B xxii and xviii, speaks of "practical data of reason" and
practical principles as "a priori data of reason." Reflexion 720J (i.e., Duisburg Frag-
ment 6) introduces the concept, perhaps for the first time, but without the word
"fact." But the moral law is here compared to a prior.i representations of space and
time as givens,
"Transcendental Deduction" of the Principle 167
With respect to these famous but obscure passages, we must ask
two questions: (a) What is the alleged fact of reason? and (b) Is it
a valid foundation for that which Kant attempts to build upon it?
a) \Vhat is the fact of pure reason? Kant himself does not seem to
have made up his mind on the best way of expressing it. The text
shows the following meanings: in the first quotation, it is consciousness
of the law; in the second and fourth quotations, it is the moral law
itself; in the third quotation, it is autonomy. Since Kant identified
freedom as autonomy with the moral law (33 [122]), perhaps the
second and third may be allowed to stand as one. But a prima facie
distinction exists between "consciousness of the moral law," which can
certainly be said to exist as a fact (whether we wish to call it a "fact
of reason" or not), and the law itself, of which we are conscious
(whose "factuality" is sub judice).
If this prima facie distinction is finally valid, Kant's argument re-
fuses to move, except in a circle. For everyone will grant that the
"fact" in the first sense exists, but it does not imply the "fact" in the
second of the senses. Yet it is the second of the facts that is essential
to Kant's argument.
b) Is it really a "fact" in the sense required? That the fact in the
second of the senses admittedly cannot be explained is, in itself, no
reason to reject it.ll Some facts must be unexplained; but why this
one, instead of some other that might be incompatible with it? Perhaps
we have a fundamental intuition or insight into its truth? I think that
Kant undoubtedly believed that we do, though he would not have
called it an "intuition"; but fundamental insights and intuitions do not
seem to be any different from the kind of evidence that Kant has
already admitted might be wrong. An appeal to insight Of intuition
is a confession of failure to find an argument or premise from which
some truth can be derived and an unwillingness to surrender it in
spite of that. In principle, some kind of intuition is indubitably nec-
essary, but that does not mean that some particular intuition is either
necessary or indubitable; it may indeed be wrong. Perhaps the moral
law is the kind of fact that must be assumed if we are to explain and
render intelligible our moral experience? 12 But if this is the meaning
12 The Deduction proper does not strictly follow the second of these paths but
adduces an independent warrant for the principle from the need of theoretical reason
to assume freedom. Hence it falls outside the scope of the censure of this sentence
(cf. below, pp. 173 f.). That the moral law is a "fact" in the sense in which, say,
mathematics and natural science are "facts" and therefore subject to the same kind
of justification is the interpretation of Hermann Cohen, Kll7lts Begrondung der
Ethik, p. 224. and his Marburg followers.
168 "Transcendental Deduction" of the Principle
here, it hardly deserves the name of fact but only of assumption, for
the experience it is supposed to organize is itself sub judice and could,
perhaps, be equally well organized in terms of some other assumptions,
e.g., those of psychoanalysis.
Thus if we permit the prima facie distinction between the two al-
leged facts to stand as a fundamental duality, Kant cannot make a
transition from the undisputed fact (that we are conscious of a moral
law13 ) to the disputed fact (that there is a law that can come only
from pure practical reason). Yet it is the latter whose factuality is to
be shown if the moral law is to be justified.
I believe, however, that this duality of meaning of "fact" does not
represent Kant's premises properly. There are two intimations of this.
First, there is his assertion that this fact (whatever it is) of pure reason
is the sole fact; and I do not think Kant means that it just happens to be
the only fact of pure reason. I think there is something about this fact
that makes him regard it as a priori unique. Second, there is the dif-
ficult and obscure hint that was never developed but is the clue to
the interpretation I shall present here. "Freedom and unconditional
practical law reciprocally imply each other," Kant says (29 [117])'
"I do not here ask whether they are actually different, instead of an
unconditional law being merely the self-consciousness of a pure prac-
tical reason, and thus identical with the positive concept of freedom/'
i.e., with the concept of autonomy.
To explore this possibility, we need to draw still another distinction
in the meaning of "fact of pure reason." "Fact of pure reason" may
mean a fact known by pure reason as its object, modo directo. Or it
may mean the fact that there is pure reason, known by reason re-
flexively. These may be distinguished as "fact for pure reason" and
"fact of pure reason."
When we think of the moral law as a fact for pure reason, i.e., as
an object of a particular and unique insight or intuition, the very
uniqueness of this fact is at least a ground of suspicion. The first
Critique showed that there are not any facts known by pure reason
without sensuous intuition. Facts are given to reason only by and
through intuition. Should there be any fact "for" pure reason, it
would be only a fact "as it were." If he is speaking of the fact "of"
pure reason, on the other hand, this is not subject to the suspicion we
13 If it should be argued that "to be conscious of" is like "to know" and therefore
must refer to a true proposition or actual object, then even the first of the facts is not
indisputable. It should then be stated: "We believe that there is a moral law," and
this states a fact but does not imply, even in ordinary speech, that there is a moral
law,
"Transcendental Deduction" of the Principle 169
feel about the fact "for" reason. Kant's point is that in any willing
there is a principle which is purely rational,14 and "if pure reason is
actually practical, it will show its reality and that of its concepts in
action."I:!
Only a law which is given by reason itself to reason itself could
be known a priori by pure reason and be a fact for pure reason. The
moral law expresses nothing else than the autonomy of reason (33
[ 12 2]); it is a fact for pure reason only inasmuch as it is the expression
of the fact of pure reason, i.e., of the fact that pure reason can be
practical. That is why the moral law is the sole fact of pure reason
and for pure reason.
If this seems to be a rather tricky argument, let us turn to the
moral phenomenon which it represents. A moral principle is not bind-
ing upon a person who is ignorant of the principle or law. On the
other hand, if a person believes that an imperative is valid for him,
then it is in so far forth valid for him, and he shows that reason is
practical even in the awareness of trus aspect of a valid claim. This
is true whether the imperative expresses a claim that is in fact valid
or not. Only a being with an a priori concept of normativity could
even make a mistake about this. To argue against it is to appeal to
normative grounds and is as ridiculous as to attempt to prove by reason
that there is no reason (12 [97]). "I say that every being that cannot
act otherwise than under the Idea of freedom is thereby really free in
a practical respect."16 But the Idea of freedom is expressed in the
moral law; hence to be conscious of moral constraint, i.e., of the law-
this is the fact of pure reason-ipso facto validates the practical claim
of a moral law, which is the fact for pure reason. It is indeed a very odd
kind of fact, and Kant is right in calling it a "fact, as it were."
To recapitulate: moral consciousness, consciousness of duty, is an
undisputed fact. Prima facie, it does not justify the assertion that duty
is not an illusory and chimerical concept. (If I believe, for instance,
that there is a God and that duty is determined by the will of God,
and if there is in fact no God, then it may remain a fact that I feel the
14 This is connnned by the analysis of even hypothetical imperatives, chap. vi, § 6.
But Kant goes even further and says that we become conscious of the moral law
whenever we construct maxims for the will. By this I think he means that in every
decision on a policy of life there is a putative rationality in the rules subsumed under
the principle, which, if fully elaborated, would require the rationality of the motive
as well as that of the means to its satisfaction (cf. Critique of Practical Reason, Z9
[1171).
15 Critique of Practical Reason, 3 (87). Cf. Critique of Pure Reason, B 430-31.
which must have been written only shortly before the second Critique.
16 Foundations of the Metaphysics of Morals, 448 n. (67 n.).
170 "Transcendental Deduction" of the Principle
call of duty, but it is not a fact that duty is an objectively valid con-
straint.) But because the moral law-the fact for pure reason-expresses
nothing but the lawgiving of reason itself, the fact of pure reason is
reflected in the fact for pure reason. For a being who thinks that there
is some obligation, there is some valid law. The metaphysical deduction
has shown what this law must be; it is the fact for pure reason.
S 3. THE DEDUCTION
Freedom
§ I. INTRODUCTION
Discussions of freedom are so frequent in Kant's works that the full
compass of the concept and its attendant problems cannot well be sur-
veyed in a running commentary on passages taken seriatim. In this
chapter, therefore, I shall attempt to examine Kant's ideas on freedom,
following an order which will be made clear in the second section.
Some of the difficulties in interpreting the Critique become more
manageable when we realize that its central doctrine of freedom of the
will involves two different concepts of freedom and two different con-
cepts of the will. Each of these pairs of concepts had a long and trou-
bled history before Kant, and each was used in his own earlier work.
His most important contribution in the second Critique was to show
that what is sound in each involves the other. But he did not help the
reader see that this is what he was doing, because he did not first estab-
lish, or even define, the two pairs of concepts and then bring them to-
gether; the unwary reader may not realize that Kant was using two sets
of concepts, because he shifts from one to the other without notice and
his language does not often indicate directly which he is using. Fur-
thermore, it is by no means certain that he was himself at all times
aware of the duality of his concepts and of the problem he was work-
mgon.
We must take the confusing fabric he wove, trace out the various
threads which form two patterns, and then, if at all, try to find a single
larger pattern composed of both. This program of work is summarized
in § 2 and occupies §§ 3-12, and it should be judged only in the light
of these later sections; §§ 13 and 14 discuss two points which are inde-
pendent of the major hypothesis of this chapter.
§ 2. TWO CO~CEPTIONS OF WILL AND FREEDOM
10 Votarbeiten zur Tugendlebre, XXIII, 383. Its autonomy is contrasted with the
heteronomy of Willkiir; the freedom of Willkiir, which is independence of material
of desire, is freedom in the negative sense, while the legislation, or autonomy, of
Wille is freedom in the positive sense (Critique of Practical Reason, 33 [l2.z]).
11 Critique of Practical Reason, 43 (132); d. Critique of Pure Reason, A 548 =
B 576.
Freedom 181
sense) and a free will are identical. The freedom of this will is demon-
strated by the fact that there are absolute obligations, which could not
be so if all the laws of practical reason were empirical, i.e., if practical
reason were not pure, i.e., autonomous. Though the two conceptions
of will and freedom are inextricably bound together, the metaphysical
difficulties are found mostly on the side of freedom in the negative
sense, where it comes into contact with the natural necessity of the
phenomenal world.
fringe on the laws of nature, and yet it is a basis for moral imputation.
To show this, he draws a distinction between the intelligible character,
the noumenal or transcendental subject, which is a causa noumenon,
and the empirical character, which is only its appearance in space and
time.4 2 The former is free, the latter is under the law of nature. The
former is thought (not known) as an unchanging substance in which
nothing happens (because it is not in time); the latter is its unrolling in
time under causal conditions of nature. We infer the nature of the for-
mer from that of the latter; we punish or reward the latter in the belief
that it is a manifestation of the freedom of the former. The freedom of
the empirical character is at first to be understood only negatively, as
not being necessitated by things in nature. The freedom of the former
is positive, for it originates a series of events in the world which would
not have happened, had the intelligible character been different. 43 Ev-
ery event in human conduct can be seen, therefore, in two ways: as a
necessary consequence of preceding events and as directly determined
by the intelligible character.
§ 7. CRITICISM OF THIS DOCTRINE
This is hard doctrine, and Kant admits that judgments under it "seem
at first glance to conflict with equity" (99 [193]). But is it only "at
first glance"? Why should a man repent the wrong he has done (98
[ 192])? How Can we hold a man responsible for his actions and yet
say, at the same time, that "before ever they have happened, they are
one and all predetermined in the empirical character?"44 That if we
knew all the empirical facts and the natural laws of their connection so
42 Critique of Practical Reason, 97 (Igo); Critique of Pure Reason, A 545 == B 573;
Reflexion 5608. Critique of Practical Reason, 1)8, 100 (191, (94), however, speaks of
our consciousness of having freely created our empirical character, and in this re-
spect it seems hardly to fall under natural law or to fit into the reality-appearance
rubric.
43 Critique of Pure Reason, A 554 ~ B 582, A 556:::: B 584; d. the very similar
passage in Critique of Practical Reason, 99 (193)' Hence it must not be simply as-
sumed that the intelligible-empirical distinction coincides with that between Wille
and WilIkiir. The intelligible character acts timelessly, manifesting its acts in the
conduct of the empirical character. The Wille does not act but legislates for (is a
law for) the Willkiir. Kant also speaks of personality (intelligible personality) as
an archetype the respect for which constitutes the motive for the Willkur (ibid.,
87 [180]). The freedom of Wille is distinguished from that of W illkur as libertas
noumenon from libertaJ phaenomel1on; yet the latter is not to be defined empirically
(d. Metapbysik der Sitten, VI, 216; and Opus postumum, XXI, 470), for this gives
only a comparative or psychological sense of freedom. Kant's statements are so
cryptic that it is hard to know whether he is entirely consistent or not.
44 Critique of Pure Reason, A 553 :::: B 581.
192 Freedom
that "his future conduct could be predicted with as great a certainty as
the occurrence of an eclipse," we could nevertheless still assert that the
man is free" (99 (193))? This requires us to accept a secularized ver-
sion of the classical theological quandary of man's freedom and God's
foresight, and it seems no more intelligible in structure and equitable
in outcome than that hoary mystery.
If by "freedom" we mean noumenal causation and assert that we
know no noumena, then there is no justifiable way, in the study of phe-
nomena, to decide that it is permissible in application to some but not
others of them to use the concept of freedom. The uniformity of hu-
man actions is, in principle, as great as that of the solar system; there
is no reason to regard statements about the freedom of the former as
having any empirical consequences. If the possession of noumenal free-
dom makes a difference to the uniformity of nature, then there is no
uniformity; if it does not, to call it "freedom" is a vain pretension.
There seems to me to be only one way out of the dilemma. There are
faint suggestions of it in two widely separated works of Kant, but he
never fully developed them. He does not seem to have felt the paradox
in his own views that all his critics and most of his disciples felt.
The first suggestion is this: Instead of regarding the world as consist-
ing of two realms-a phenomenal under one set of laws and a noumenal
under another-can we not think of one world under two aspects, the
aspects to be defined methodologically with reference to the purposes
we have in holding these two perspectives on a common world, and
not ontologically? Kant intimates such a two-aspect theory, instead of
the more commonly known "two-world theory," in his contrast be-
tween the observing, theoretical attitude and the acting, practical atti-
tude 45 and in his statement that "supersensuous nature is nothing else
than nature under the autonomy of pure practical reason" (43 [132))'
The other suggestion, necessary to the full development of the pre-
ceding one, is his conclusion in the third Critique 46 that the distinction
between natural and moral law is dependent on the peculiar nature of
our understanding. We can read these passages as a suggestion to re-
gard the two kinds of laws as co-ordinate, not one as subordinate to the
other in constitutive authority in experience. The only evidence we
have that Kant entertained such a view is in § 70 of the Critique of
Judgment, where he speaks of the thesis of the complete mechanical
determination of nature, in its antinomic relation to that of teleological
causation, as a regulative Idea even in respect to nature. It would be
easy to extend this to the antinomic relation between freedom and nat-
45 Critique of Pure Reason, A 550 ::::: B 578; cf. chap. iii, above.
46 Critique of Judy;ment, § § 70 and 76.
Freedom 193
ural causation. The Idea is expressed as a maxim: "All production of
material things and their forms must be judged to be possible according
to merely mechanical laws."
Had Kant said, in the solution to the third antinomy in the first Cri-
tique, that the tfue meaning of the antithesis is that it is a maxim for pro-
cedure and not a constitutive principle of nature, then we would have
had two maxims: "Always (in science) search for mechanical causes
and allow no non-natural causes to enter into the explanation of natural
phenomena," and "Always (in ethics) act as if the maxim of the will
were a sufficient determining ground of the conduct to be executed or
judged." Neither of these is a declarative a priori statement; they tell
us, rather, what we must do in order to be a spectator or an actof, but
one cannot be both at the same time and with respect to the same item
of conductY
It may well be that we would, under these rules, sometimes hold a
man responsible for actions that he could not have avoided doing, for
human freedom is far more limited, I think, than Kant held it to be. In
this event, we are simply unjust judges. It may well be that we give an
abstract, schematic causal explanation to some event which did not, in
fact (though we shall never possibly know it), have a sufficient natural
cause; in this event, we are dogmatic (unavoidably so) in our scientific
work. But the alternative to sometimes being unjust is that of always
being unjust when we hold a man responsible for any of his actions
when, if Kant is correct, none of them could have been left undone in
the course of nature and history as constituted by natural law.
The solution proposed here involves reading back into the Transcen-
dental Analytic of the first Critique some of the conclusions of the
Transcendental Dialectic. Specifically, it requires that the sharp distinc-
tion between constitutive category and regulative Idea be given up,
that even the categories be regarded as devices for the regulation of
experience and not as structures necessarily given in a fixed constitu-
tion of our experience of nature, and that the Analogies of Experience,
which Kant called "regulative" in a very modest sense,48 be reinter-
preted as regulative in the full sense of the Dialectic. 49
47 Cf. above, chap. iii. The second of the maxims need not be interpreted in the
manner of the fictionalists, i.e., on the assumption that the "as if' intrOduces a con-
trary-to-fact condition. There is no more need to assume that the latter is any more
fictional than the fonner or that the former is constitutive while the latter is merely
regulative.
48 Critique of Pure Reason, A J 79 = B 222.
40 Such an interpretation of the Transcendental Analytic, in fact, can be recom-
mended on purely epistemological grounds, though it no doubt distorts Kant's own
estimate of the relation of the functions of understanding and of reason.
194 Freedom
If we undertake to make these revisions, we are allowed to regard
the moral or practical realm as a perspective of the realm of experience,
which, through other regulative ideas (categories), is seen as the realm
of nature. We will then no longer have to think of science as dealing
with appearances (in some ontoiogically pejorative sense) and morals
as dealing with noumena (in some epistemologically pejorative sense).
Both can keep their a priori structures intact; both will claim to cover
all the relevant experience;5o but each will be carried out for different
purposes and only occasionally will come into conflict with each other
-a conflict to be settled by moral scrutiny sensitive to all the myriad
facts of life-instead of invariably doing so, as, according to the ortho-
dox Kantian theory, they must. That one of these realms is a limiting
case of the other, that the categories of one of them can be derived
from those of the other, is a view, classically developed by Fichte and
strongly represented in both idealism and pragmatism, which goes far
beyond anything Kant said or probably would accept, though the germ
of it is present in the third Critique.
§ 8. FREEDOM AS AN ACTOR-CONCEPT; SPONTANEITY
That we have a right and even find it necessary to make use of the
concept of freedom in our actions, regardless of how the antinomy is
resolved and whether freedom is compatible with natural causation or
not-in short, that the actor in a moral situation must act as if he were
free and thereby shoulder all the responsibilities that he would have if
it were theoretically proved that he is free-is shown by Kant in several
ways. It is not shown, however, by appealing to what Kant calls com-
parative freedom, i.e., empirically observed relative independence from
outward stimuli and inward impulse through the exercise of intelligent
foresight. It is shown only through an elaboration of the inward phe-
nomenon of choice and a regression upon its conditions.
In the second edition of the Critique of Pure Reason there emerges
a conception of one's own existence which is not present, at least so
explicitly, in the first edition. It is the conception that we have a direct
experience of our own spontaneous activity as a substance.1I1 This ex-
50 Cf. Critique of Judgment, Introduction, II: The realm of theoretical reason and
that of practical reason do not limit each other in their legislation, but they per-
petually do so in the world of sense.
III Critique of Pure Reason. B 157-58 n.: "I cannot detennine [i.e., know categor-
ially] my existence as that of a self-acting being; all that I can do is represent to
myself the spontaneity of my thought. that is, of the determination; and my existence
is still determinable only sensibly, that is, as the existence of an appearance. Sut it
is owing to this spontaneity that I entitle myself an intelligence." Similarly, Founda-
tions of the Metaphysics of Morals, 451 (71); Critique of Practical Reason, 56 (J47).
Freedom 195
perience is neither a sensuous intuition:;:? nor an abstract thought. Kant
never tells us what its epistemic character is, but that it occurs is a fact
to which the epistemology of the Critique of Pure Reason does scant
justice.
Such self-awareness of one's spontaneity and the attendant presup-
position of freedom is found even in the act of theoretical thinking,
though we can use this conception of the self as a thinking substance in
no theoretical explanation of the inner life. 53
But the clearest evidence of one's spontaneous freedom-sometimes
said to be the only evidence 54-is one's awareness of obligation, which
is a necessitation of a wholly different kind from all natural necessita-
tion and which produces a unique kind of feeling (92 [185]). It has
this revelatory function when, in fact, the moral law is not obeyed
but only acknowledged. 55 As wholly different from natural determina-
tion, it cannot be understood in theoretical terms. The freedom that
can be understood, he says in his review of Ulrich, is of no use in eth-
52 When Kant, in the precritical period, believed that there was an intellectual in-
tuition, it was to this faculty that he ascribed consciousness of the self (see Reflex-
ionen 4228, 4336, 600 I). With the denial of intellectual intuition to man, no so-to-
speak official position is taken on knowledge of one's own spontaneity, but the ex-
perience is not denied. On the whole question, with a collection of sources, d.
Heimsoeth, Studien zur Philosophie Immanuel Kants, pp. Z45 ff., and Ingeborg
Heidemann, Spontaneitiit und Zeitlichkeit, pp. 173 ff.
113 The theoretical uselessness of this conception is the main point in the Paralo-
gisms of the first Critique; even there the practical importance of the concept is
recognized and preserved (B 431-32). Cf. Foundations of the Metaphysics of
Morals, 448,451-52 (67,70-71), on the awareness of spontaneity. In Besprechung
von Schulz's Sittenlehre (1785), VIII, 13, Kant remarks that the determinist in meta-
physics claims freedom for himself in the conduct of his own thinking and a for-
tiori in his action. Cf. also Beantwortung der Frage: 1st es eine Erfahrung dass
wir den ken? (Cassirer ed., IV, 519-10), where it is denied that the awareness of
thinking is an Erfahrung, and it is called merely "transcendental consciousness";
similar! y Critique of Pure Reason, A 117 n.
114 Critique of Pure Reason, B xxxiii, B 430-F; Critique of Practical Reason, 42
(13 1 ) •
55 The pair of examples in the Critique of Practical Reason, 30 (1I8), brings out
this contrast between natural necessitation, which is effective, and a moral necessita-
tion, which mayor may not be effective. The first example, concerning the man
who says his lust is uncontrollable, resembles one in Rousseau: "Let us suppose the
maddest of men, the man who has his senses least under control; let him see the
preparations for his death; let him realize that he will cenainly die in torment a
quarter of an hour later; not only would he from that time forward be able to resist
temptation, he would even find it easy to do so" (Emile, Book IV ["Everyman" ed.,
pp. 289-9°)). The second example anticipates one to be used in the Methodology
(155-5 6 [154])'
196 Freedom
ics, and the freedom needed in ethics cannot be understood. 56 All that
we can do is to comprehend its incomprehensibility::'· and accept it as
guaranteed by the fact of pure reason, the moral law which reveals it
to US. 58
With this argument, that the consciousness of the moral law itself
proves the reality of freedom,59 we are brought to a somewhat differ-
ent conception of freedom, which we must now examine.
§ 9. FREEDOM AS SUPREME LEGISLATION; AUTONOMY
Let us suppose that the Critique of Pure Reason has proved that it is
not self-contradictory to say that there is in man a causa noumenon or
a faculty of initiating a new causal series in the world. This faculty
would be called "freedom in the negative sense"60 or freedom from
nature. But freedom is not lawless caprice, any more than it is lawless
in its effects found in nature and historv.
-'
But what limits freedom and
renders it lawful? The Critique of Pure Reason says that reason places
a limit on freedom,61 but how it does so is left unexplored and unex-
plained.
To answer this question is one purpose of the Foundations of the
Metaphysics of Morals. The problem is to determine a law that the will
can obey without losing its freedom through that very act of obeisance
to law. The will can "obey" laws of nature in the sense of intentionally
using them, in the form of means-end statements, in seeking men's goals
in nature. But to be free even from the importunities of the desire for
happiness-a goal given by man's natural existence-requires that the
law of reason be not such a law borrowed from nature and hence, in-
directly, leaving action under the domination of nature. The law must
be given by reason. Just as the will (Willkur) considered as a faculty
in man may be free in the negative sense of spontaneous activity, prac-
tical reason is spontaneou:; in the sense of giving law instead of subject-
56 Ak., VIII, 458.
57 Foundations of the Metaphysics of Morals, 463 (84).
58 Cf. above, chap. x, and Critique of Practical Reason, 31 (no).
59 It is likewise taken as a route to the consciousness of self, independent of em-
pirical conditions, and thus serves to supplement the failures of rational psychology
as a doctrine of the soul, though without making any contribution to theoretical
knowledge (cf. Critique of Pure Reason, B 43<>-31; Critique of Practical Reason,
105-6 [zoo-zol]).
60 Ibid., A 553 = B 581; Critique of Practical Reason, n 5. 6, 8 (z9. 33 [116-17,
Ill]).
61 Critique of Pure Reason, A 569 = B 597. It exercises this control even in issuing
hypothetical imperatives, i.e., in its logical as well as in its real use (ibid., A 54B =
B 576).
Freedom 197
ing itself to an alien law. It gives a law to Willkur which it has freely
legislated-legislated by the necessity of its own nature. The faculty of
lawgiving is will in the sense in which pure practical reason is will, and
its legislation is "freedom in the positive sense," or autonomy. In the
choice of this language, Kant, as so often, thinks in political metaphors
and analogies, and especially in terms of Rousseau's political theory.
This meaning of freedom is quite different from the one we have
been examining heretofore. Here it is not a question of whether the
acting subject can initiate a new causal series, but it concerns the
source of the law which the subject follows in such an initiation. The
consciousness of the moral law, and not obedience to it, is the evidence
of freedom in the positive sense (47 [136]). Freedom and uncondi-
tional law reciprocally imply each other; the latter is the ratio cogno-
scendi of the former, the former the ratio essendi of the latter (4 n.
[88 n.]). Freedom analytically pertains to reason. 82 The moral law
reveals to me a "marvelous faculty," a purely intellectual principle
for the determination of my existence. 83
Here we may appropriately notice one of Kant's best-known argu-
ments for freedom, in which he says that a being that cannot act except
under the Idea of freedom is really free in a practical respect and is
obligated by the laws which follow from that Idea, regardless of
whether we can prove theoretically that he is free or not. 64 Which
freedom is warranted by this argument? Not the freedom of spon-
taneous action, for consciousness of this could be illusory. Everyone,
including the fatalist or determinist, acknowledges, with Mephistoph-
eles, that "Ou glaubst zu schieben"; yet the fatalist or determinist adds,
"aber du wirst geschoben." The argument applies to freedom in the
sense of obligation-creation, not obligation-execution. For we are not
directly obligated to do anything, to initiate a specific causal series
in time. We are obligated directly and unexceptionably only to take
up and act upon a certain maxim and to reject the maxims and actions
which are incompatible with it, and even this thought shows a legisla-
tion of reason which is not derived from nature, since nature presents
no "ought." To think one's self obligated to do something is to be
really obligated to do something (though we may not be, in fact,
obligated to do precisely that which we thought we were obligated
62 Critique of Practical Reason, 31 (120): the moral law "would be analytic if the
freedom of the will were presupposed." Cf. Foundations of the Metaphysics of
Morals, 420 n., 426 (37 n., 44).
83 Critique of Pure Reason, A 431-32'
84 Foundations of the Metaphysics of Morals, 448 n. (67 n.).
198 Freedor.n
to do) .65 The thought of obligation is, therefore, self-guaranteeing
in a way in which the feeling of spontaneity is not self-guaranteeing.
Hence one can say that freedom in the sense of legislation is real, even
though he admits that in the entire history of mankind no free action
may ever have been performed.
§ 10. THE SYNTHESIS OF THE TWO CONCEPTIONS
Kant regrettably did not single out these two conceptions and then
formally show their relationship to each other. He deals with them
together, without clearly showing, at any moment, which it is that
he is talking about. Had he followed a truly synthetic method, he
would have set up the two concepts and then related them. As it is,
however, we have to find the relationship between them by an analysis
of the complex mixture. Fortunately, it is not difficult to do so, and
Kant himself does it, all but explicitly, in § 6 (29-30 [116-17]). I
shall now paraphrase this passage, using the two conceptions of will
and of freedom, in order to show the synthetic or bridging function
of this important section.
Granted that human beings have a free will (Willkur)66 in the
practical and phenomenological sense, as a faculty of spontaneously
initiating a new causal series in nature, what is the law of its action?
A law must have a content, which may be taken from nature, and a
form, the form of universality, which can be prescribed only by rea-
son. 61 If the Willkur obeys a law because of its content, it can be free
in the practical sense, for will shows its freedom even in obeisance to
hypothetical imperatives and, indeed, in actions which are evil. But if it
is to be determined necessarily, i.e., irrespective of the desires which
are the material cause of its willing, as it is if there is duty, it must be
determined not by the content but by the form of law. The form of
law is universality, fitness for universal legislation. In giving such a
law, reason is not responding to the promptings of nature. It is there-
fore a spontaneous legislator and is free.
Hence a free, i.e., spontaneous, WillkuT, when it is good, is deter~
mined by a free, i.e., autonomous, Wille, or pure practical reason,
which gives it a law. 68 It can obey only this law without jeopardy to
85 A prima facie obligation can be abrogated only by another prima facie obliga-
tion or real obligation. It cannot be simply denied, leaving a kind of moral vacuum.
66 Kant should have used the word Willkur in this section, but he did not.
61 This is true of both natural and moral laws (cf. Critique of Practical Reason,
16 [1141>.
68 Foundations of the Metaphysics of Morals, 455 (74): The good will is a law
for the bad will as a member of the world of sense. Unfortunately, Kant used Wille
for each in this passage.
Freedom 199
its freedom. Indeed, it gains in freedom, by now being an autonomous
as well as a spontaneous will. Thus is added to the negative concept
of freedom (spontaneous independence of foreign laws) the positive
concept of freedom (autonomous self-legislation). There is a loss only
of lawless freedom: "Where the moral law speaks there is no longer,
objectively, a free choice [Wahl] as regards what is to be done."68
The Willkur can obey the law of Wille without losing its own
freedom only because they are not two faculties externally related to
each other. They are two aspects of practical reason, differing as the
legislative and executive functions. The former function binds the
latter; the former is the pure form of the latter. It gives a law that the
latter 'Would obey, were it a holy will, i.e., if the Willkur fully realized
its potentialities of purity. The law is found not by seeking something
outside the Willkur but by a regression upon the conditions of its
full freedom, conditions that are not actually realized in the natural
man. Thus, Kant explicitly says, we find Wille by a regression upon
the conditions of Willkur.70
That the will of man-because of the fact that it is not a simple
notion and has two distinguishable aspects-can be both obligation-
creating and obligation-executing is one of the most dramatic theses
in Kant's philosophy, as dramatic as, and analogous to, the Copernican
Revolution in his theoretical philosophy. So long as the origin of law
was ascribed to nature or experience or God, no matter how rigoristic
and certain it might be, obedience to it had to be heteronomous and
restrictive of freedom. That the moral law was rational and was to be
discovered a priori and was in some way binding on man quite apart
from the rewards that were believed to accompany obedience to it-
all this was well-known doctrine in the German academic philosophy
of Kant's time. But, because of the abstract ontology of perfection,
no specific law seemed derivable from the putative sources of the moral
law, and obedience to it was always in danger of being explained only
eudaemonistically. Because Kant discovered the law in the concept of
rational will (not abstract perfection or "will in general") 71 and was
able to derive its formula from the concept of its source and because
this source was an idealization of the will as a faculty of spontaneity
in man, Kant did not have to seek any outward motivation for obedi-
ence to it. Rational personality as lawgiving expressed an "is" which
is ipso facto an "ought" for partially rational beings; the law and its
It has often been objected that there are two paradoxes in Kant's
ethics: (I) Kant is, or is reputed to be, "individualistic" in his ethics,
while the moral person is, for him, only an abstraction that is meant
in some way to dominate and restrict the inditidual person. (2) Kant's
ethics is so autonomous that the social or universalistic aspect of
morality is left unfounded because it is an outward restriction on the
freedom he insists upon. 76
Quite apart from the fact that these two alleged paradoxes cancel
each other out, it is possible to show that neither is valid and that both
arise from a misinterpretation of the point we are now discussing.
Each, if fully expanded, would entail the supposition that in Kant's
doctrine there are two wills, and which paradox is drawn depends
upon the critic's belief concerning which is the more important to
Kant (which happens always to be the one that is less important to
the critic-for such are the ways of philosophical polemics). (I) If the
Wille or pure practical reason as an abstract epistemic or moral concept
is emphasized, then the individual human W illkuT is restricted and is
not free. (2) If the legislation is thought of as issuing from an individual
WillkuT, it is not possible to see how the laws issued will meet the
requirements of social universality and harmony.77
But there are not two wills. There is one will with its formal condi-
tion, which is universally valid reason, and its material condition arising
from the specificity of its involvement in the world at particular times
and places. And the two paradoxes are not paradoxes of Kant's ethics
so much as manifestations of the human predicament in which we find
in ourselves individualized manifestations of universal mandates and
injunctions. Man is the only being in the world that not only is a
manifestation of some universals but ought to be an instance of others;
he is an individual that gives no valid laws to others that he does not
lay upon himself, that gives no privileges to himself that he does not
allow to others. Had the Kantian teaching missed what is true in these
paradoxes, it would have been less true and less responsive to the
paradoxical aspect of human life itself, in which man finds himself
neither a brute nor a god, neither a mere particular nor a mere
universal.
But, of all the misinterpretations of Kant, perhaps none is more ob-
76 It is noteworthy that the same "paradoxes" have been found in Rousseau. Was
Rousseau the father of an impersonal fascism or of democratic individualism run
wild? An affinnative answer to either of these questions shows a misunderstanding
of Rousseau analogous to the misunderstanding of Kant explored here.
77 Cf. George Santayana, Egotism in German Philosophy (1916; zd ed., New York,
1940 ),Pp· 5~51.
101 Freedom
viously wrong or more widely accepted than the accusation that Kant
represents something vaguely called "Prussian philosophy," in which
blind obedience to law is so esteemed as an absolute virtue that neither
political nor moral freedom is allowed to be more than a name. The
accusation is too ridiculous to deserve serious consideration on its own
merits and should be refuted only because of its widespread accept-
ance. 78 But the refutation is easy: it not only separates two wills within
the individual but puts each in a different person, attributing rights to
one and duties to the other. 79 It forgets that all moral discipline is, for
Kant, self-discipline and that self-discipline, while not the whole of
morality, is a necessary condition of it. It forgets that Kant taught that
all just government is self-government. The same man, by virtue of
the same faculty in its positive and negative use, is both subject and
legislator in the realm of ends and in the just state.
§ I I. SUMMARY
Let us pause and summarize the results of our argument, results
anticipated in § 2 above, and relate it to some other concepts in Kant's
ethics.
We have distinguished two meanings of will: Wille as practical
reason, the legislative function, and WillkUr, as the executive faculty
of man.
The freedom of the former is autonomy; it gives a law to WillkuT.
This law is determined by the nature of Wille and not by anything
else in the world, including human nature or the will of God. The
moral law is a synthetic a priori statement of what a WillkuT would
necessarily do if it were exclusively rational; it is a law or imperative
of duty for a Willkur which does not do by nature what the law
78 John Dewey is not responsible for the extremes of this view, for he acknowl·
edged Kant's "individualism" and held only that such "Prussianism" arose because
"the two worlds of Kant were too far away from each other" and could be con-
nected only through the remnants of the idealistic theory of history and the state
(cf. German Philosophy and Politics [New York, 19[51, p. 1%2, and the defense of
Kant by Julius Ebbinghaus, "Interpretation and Misinterpretation of the Categorical
Imperative," Philosophical QUaTterly. IV [[954], 97-108. On the latter see also K.
Kolenda, "Professor Ebbinghaus' Interpretation of the Categorical Imperative,"
"Philosophical Quarterly, V [[955],74-77).
79 Perpetual Peace, VIII, 348 n. (Beck l [-I 2 n.): "With regard to the most sub·
lime reason in the world that I can think of with the exception of God (say, the
great Aeon). when I do my duty in my post as he does in his, there is no reason,
under the law of equality, why obedience to duty should fall only to me and the
right to command only to him." God is excepted only because of the inapplicability
of the concept of duty to a holy Willi but the same law applies both to man and
to God.
Freedom l03
contains "causation" is formally invalid but that if "X creates Y" and
"The appearances of Y cause the appearances of Z," we can say
nothing of the relation of X to the appearances of Z.
But I do not see how this meets the issue. For X is also the creator
of Z. Since the appearances of Y and Z could not be different from
what they are without a different noumenon underlying each90 and
since each is created by X, it is not easy to see how X can be let off
so easily and exculpated for the appearances of Z.
Kant himself seems to have felt that his argument here is not entirely
convincing and to have put it forward only faute de mieux. The last
paragraph of this section (103 [197]) can be read only as evidence of
such dissatisfaction.91 As a puzzle from theology and classical ration-
alism, I suspect that the problem did not interest Kant at this time as
much as it had earlier. The relation of God to the world is not an
object of theoretical knowledge but only one of practical belief. This
particular aspect of the relation does not have positive moral con-
sequences which make it necessary for Kant to take a stand on it, and
such puzzles are dismissed in the Strife of the Faculties.
§ 14. FREEDOM AS A POSTULATE
Freedom is the only one of the Ideas of pure reason that we can
know. It is proved by the apodi~tic law of practical reason. All other
Ideas gain reality (i.e., are known to have objects) only through their
connection with it. 9 :! These Ideas are those of God and the immortality
of the soul. They are called "postulates" because they are dependent
upon the need of human reason to establish the possibility of the
highest good; they are not directly necessary to morality or revealed in
the fact of pure reason, and they cannot be theoretically justified.
It is a little surprising, therefore, to find Kant calling freedom in
the positive sense also a postulate of pure practical reason. 93 This is to
be explained in three ways. First, there is the wide latitude that Kant
90 =
Critique of Pure Reason, A 556 B 584.
91 One of the first commentators on the Critique (Brastberger, Untersuchungen
tiber Kants Kritik der praktischen Vernrmft, p. 156), who was almost sentimentally
sympathetic to and uncritical in his acclaim of Kant's book on most points, rejects
Kant's solution to this problem or, rather, confesses that he cannot find what Kant's
solution is, Crusius, who in so many matters concerning freedom anticipated Kant,
declared the problem to be insoluble.
92 Critique of Practical Reason, 4 (88). The postulates "give objective reality to
the Ideas of speculative reason in general" (ibid., 131 [229]). In this sense freedom
may be, and is, considered a postulate.
93 Ibid" 131 (ZZ9). Elsewhere he says that it is freedom in the negative sense which
is a postulate (cf. Zum ewigen Frieden in der PhiJosophie, VIII, 418).
208 Freedom
permitted himself in using the word "postulate." The variety of its
meanings will be discussed below in chapter xiii.
Second, there is the specific relation of freedom to the summum
bonum rather than to the moral law itself. With reference to the
highest good as an object of a necessary human need, whatever it is
necessary to assume in order to achieve it may be considered a pos-
tulate. Though the Analytic has given better reasons to assert freedom
than that it is necessary to the highest good, when the highest good
is the topic of discussion, whatever we have to assume in order to
show its real possibility is to that extent a postulate. We cannot con-
clude that the doctrine of the Analytic is in any way abandoned or
even modified because what was there "deduced" is here "postulated."
The doctrine of freedom as a postulate is, in fact, earlier than the
deduction of freedom. In the Dialectic of the second Critique we hear
the last echo of a doctrine developed in the Critique of Pure Reason
which put God, freedom, and immortality on exactly the same level.94
Third, there is a difference in the meaning of freedom as a condition
of the moral law and as a condition of the summum bonum. Freedom
in the latter sense is an object of faith, not a scibile;95 it is the faith
( Vertrauen) in the achievability of the summum bonum, i.e., the belief
in virtue (Glaube an die Tugend) as adequate to achieve the highest
good. In this sense, freedom is not mere autonomy but "autarchy of
will."96
94 Critique of Pure Reason, B 395 n. (not in A). Albert Schweitzer (Die Religions-
philosopbie Kants, chap. ii and p. 134) argued that the original plan of the second
Critique was that it would continue the doctrine of the cosmological Ideas of the
first, and in the order of their derivation in the Dialectic but that Kant then dis-
covered the special prerogative of the Idea of freedom. This caused him to take that
Idea from the Dialectic as planned and put it into the Analytic.
95 Critique of Judgment, § 91.
96 Fortschritte der Metaphysik, XX, 295.
XII
The ("~esthetic"
of Pure Practical Reason; Com-
mentary on Analytic, Chapter III; Part
of Dialectic; and Methodology
§ I. THE PSYCHOLOGICAL AND THE ETHICAL PROBLEM
Kant has repeatedly assened the necessity of distinguishing between
the subjective, psychological factors in conduct, which can be learned
through observation, and the objective a priori law or norm of morali-
ty, which can be discovered only by pure reason. The Foundations of
the Metaphysics of Morals tried to show that this distinction is im-
plicit in ordinary moral consciousness, however unsophisticated it may
be and however little self-conscious discipline of reason may be found
in it. In that work Kant attempted to give this distinction a definitive
form so that the a priori principle could be discerned in its simple
purity. The Critique of Practical Reason, in following a synthetic or
progressive method, does not begin with the consciousness of duty in
order to show that it has features which cannot be exhaustively ex-
plained in empirical psychological terms, but with a delineation of the
formal and material elements in all willing. Only at the end of chap-
ter i of the Analytic is the statement of the supreme principle of pure
practical reason attained. It is not based upon any empirical data given
by or about human nature; it is a law for rational beings generally.
It is, moreover, a norm or a regulative principle for men and for any
other rational beings in the world, if there are any, who are not pure
rational beings. That is, it is a priori descriptive of a fully rational
being and a regulative concept for a partially rational being; hence it
is meant to be applied to human beings.
"Apply to" may mean two different things. It may mean that this
law is applied as a norm in judging the conduct of men, being a
standard of comparison between what is and what ought to be. And it
may mean that one renders obedience to it, so that it is the supreme
factor in the determination of choice and the conduct issuing from it.
The former of these meanings is explored chiefly in the Metaphysics
209
210 The "Aesthetic" of Pure Practical Reason
of Morals, where it lends itself to a full casuistical treatment. Except
incidentally, it is only the second of the two which is studied in the
Critique of Practical Reason. There are two separate but closely con-
nected discussions of it, one in the chapter "The Incentives of Pure
Practical Reason" and the other in the Methodology. Some corollary
problems are discussed in the chapter on the resolution of the antinomy
arising from the Idea of the summum bonum.
The problem of these chapters and sections is a singularly difficult
one for Kant, because in the rest of his treatment of the moral prin-
ciple he is writing about rational beings in general, and the path from
a priori knowledge to a posteriori application is always (not for Kant
alone) a hard one; it is one that must be trodden with exceeding care,
for Kant realized here and in his other works that a mistake in the
analysis of the a posteriori factors could cast doubt on the validity
of the a priori features of experience that were his chief concern. In
none of his works is this transition made in such a manner as to give
complete satisfaction to Kant's critics, and their reluctance to accept
his psychology has favored a skepticism toward his theses in pure
philosophy. It is, nevertheless, a passage that must be made if pure
reason is to be practical.
Man is a rational being, but he is also a being of need, impulse, and
sense. He is, or may be, a free agent, but he is also a part of the
mechanism of nature. His self is noumenal, but he also has an empirical
character that sets him off from all other rational beings and all other
empirical objects. How, then, can we describe his relation, actual and
ideal, to the pure moral law? How can the moral law or his con-
sciousness of it be a determining factor in his conduct, as it is if he can
properly claim to be a moral agent?
Before venturing upon Kant's actual answer to this question, let us
try to become clear as to what would constitute, in his terms, a satis-
factory answer to it. Kant says that this question is equivalent to that
of how the will can be free, and to that question, like all others con-
cerning fundamental powers or faculties, no answer can be given. We
cannot find any higher principle that will throw light on such facts
as the sensuous character of our intuition, the practicality of reason,
the fact that our reason is discursive rather than intuitive, and the
freedom of the will. But we can go this far: we can explain their
inexplicability and show that attempts to explain them away are not
successful. Though it displaces the inexplicable mystery of man by
only one step, the Critique Of Practical Reason does attempt an ex-
planation, in psychological terms, of how the knowledge of the moral
The "Aesthetic" of Pure Practical Reason 2 It
9 Triiume eines Geistersehers, II, 330, 335; cE. above. p. 159. n. 68.
214 The "Aesthetic" of Pure Practical Reason
This is a speculative metaphysics of moral feeling, in which meta-
physics is only a hypothetical extension of the type of explanation
valid within empirical experience. But if we remember what "meta-
physics" meant to Kant before and after 1770 and the change which
came in his conception of its scope and method about this time,10 we
shall not be surprised at the developing intellectualism present in a
very significant statement that Kant made in commenting upon an
ethical essay by his former pupil, Marcus Herz. In 177 3 he wrote:
"The concept of morality must please in the highest degree, must have
moving power; and though it is indeed intellectual, it must have direct
relation to the basic incentives of the wi11."l1
This view, not further developed at that time and, indeed, so briefly
indicated as to make its full interpretation hazardous, is probably the
first intimation of the doctrine of the critical writings of the 1780's.
In the intervening period, Kant seems to have been so torn between
two incompatible doctrines of the moral motive (if, indeed, he had
time to think of them at all) that when we come to the Critique of
Pure Reason we find him espousing each of them on a single page.
Thus the necessity of rewards and punishments as a spur to morality is
asserted in a passage which sounds like a religious eudaemonism:
It is necessary that the whole course of our life be subjected to moral
maxims, but it is impossible that this should happen unless reason connects
with the moral law, which is a mere Idea [and not yet a maxim], an opera-
tive cause which determines for such conduct as is in accord with the moral
law an outcome that is in exact accordance with our supreme ends [i.e.,
happiness]. Thus without a God and a future world invisible to us now but
hoped for, the glorious Ideas of morality are indeed objects of approval and
admiration but not springs [Triebfeder] of purpose and of action. 12
Yet almost immediately and with no obvious consistency, he denies
that the prospect of future happiness makes the moral disposition
possible. The moral disposition is one of worthiness to be happy, to
which the hope of happiness is added by these "postulates"; but the
desire for happiness does not generate the moral disposition. 13 The
10 Cf.above, p' 7 fr.
11 To Herz, end of 1773 (X, I'H)' The Vorlesungen iiber Metaphysik ([Politz
ed., 1821], p. 242), of uncertain date, speaks of the moral feeling as that through
which the motives of the understanding gain moving power. The student's lecture
notes, however, even if accurately datable, cannot be depended upon to be accurate
where any subtle issue is at stake, as every professor will ruefully acknowledge.
12 =
Critique of Pure Reason, A 812-13 B 840-41.
13 In a similar passage in the Refiexions on his essay on the sublime and beautiful,
Kant has recourse to the mysterious working of future rewards and punishments on
The "Aesthetic H of Pure Practical Reaso1l 2 15
second of these views is the one that is to appear again and again, and
the religious eudaemonism of the quoted passage is transcended, never
to be asserted again. Still we look in vain in the first Critique for an
analysis of the desire to be worthy of happiness, which is a truly moral
and not eudaemonistic desire. Without such an analysis, the view of
the Critique of Pure Reason seems to be an incompatible mixture.
For an analysis of this desire, we must turn to the celebrated Duis-
burg Fragment 6, written, I believe, just after the first Critique. H
Here the desire for happiness is still fundamental, though here again
Kant recognizes that such a desire does not eo ipso generate virtue.
But happiness is here given a moral definition, under which it can be
desired morally. Happiness is well-being only insofar as it is de-
pendent upon our choice and not upon accidental circumstances; it
is, accordingly, nothing but "well-ordered freedom." Well-ordered
freedom, the exercise of which occasions self-satisfaction, is the unity
of all actions under general laws and is equivalent to morality. Hence
Kant concludes that morality is a necessary condition of happiness
and makes it possible, but does not have happiness as its purpose. It is
the a priori form of happiness, the content being contingent upon
circumstance. Virtue and only virtue can bring about happiness, but
it does not necessarily do so, and, even when it does, that is not the
source of its worth. But to feel that we are the authors of a state of
being worthy of happiness (i.e., to have its a priori condition) is itself
a positive feeling of self-contentment, and this constitutes the human
worth of morality and is a necessary factor in happiness. The inner
applause is a sufficient motive and is an "intellectual pleasure" in the
enjoyment of freedom. l :!
This fragment constitutes a transition from a dualistic to a monistic
theory of the moral motive. Previously, except perhaps in the one
sentence of the letter to Herz, the cognitive and conative factors had
been kept separate, with the consequence that pure reason could not
consistently be asserted to be intrinsically practical. In Fragment 6
the moral disposition which can remain pure even under their influence. He says:
"If, however, in some supernatural way there is some influence on the purity of our
nature, future rewards no longer have the character of motives" (XX, 18). This is
no longer an expedient that Kant could take in the Critique, though the thought
remains in the estimate of the purity of Christian morals and his rejection of a
hedonistic interpretation of Christian motives (113 n. [110 n.1).
14 Reflexion 7102 (Ak., XIX, 17Crth; Reicke, 1,9-(6).
15 The term "intellectual pleasure" is later rejected (Critique of Practical Reason,
117 [ 21 3]).
216 The "Aesthetic H of Pure Practical Reason
the conative element is itself intel1ectualized 16 and formalized, and
moral happiness, i.e., a happiness in morality and not a happiness re-
sulting from it, becomes the goal. But when, in the next few years,
Kant formulated the doctrine of autonomy, even moral happiness be-
came a corollary effect and not a motive as a separate and distinct fac-
tor in moral purpose. Weare now ready to study this last and most
momentous change in Kant's views.
§ 4. INCENTIVES
2olbid.,¥>z (83).
The "Aesthetic" of Pure Practical Reason 117
21 Critique of Practical Reason, 76 (168). The statement "The law itself must be
the incentive" is not quite accurate and must be qualified (cf. below, § 7).
22 Foundations of the Metaphysics of Morals, 449, 455 (68,75).
23 Vorarbeiten zur Tugendlehre, XXIII, 378.
24 Erste Einleitung in die Kr;tik der Urteilskraft, XX. 230 n.; Critique of Practical
Reason,9 n. (94 n.).
218 The "Aesthetic" of Pure Practical Reason
he says, are all that he needs to "borrow from psychology" for this
Critique.
That man is a living being-indeed, that there are men in the world-
affected by desires and susceptible to pleasure is an empirical fact. The
relation between pleasure and desire is to be determined empirically
when desire is given an empirical content (as hunger, thirst, and the
like). It is essential that the psychologist not give definitions which
are implicit hypotheses about matters which can be settled only em-
piricalIy.25 They must be defined in such a way as to leave open for
empirical determination the relations which subsist between the defined
feelings. They must be defined, he says not quite accurately,26 only
"in terms belonging to the pure understanding, i.e., categories, which
contain nothing empirical." There are three such definitions:
Life is the faculty of a being by which it acts according to the laws of the
faculty of desire. The faculty of desire is the faculty such a being has of
causing, through its representations, the reality of the objects of these ideas.
Pleasure is the representation of the agreement of an object or an action with
the subjective conditions of life [9n. (94n.)].
We have already mentioned the distinction between the lower and
the higher faculty of desire. 27 The latter is not one of experiencing a
subjective need which will produce pleasure only indirectly through
success in bringing an object of desire into existence, but it is the
faculty of desire whose object is one that can be represented only by
pure reason and caused by virtue of this representation. That is, an
object of the higher faculty of desire is not a thing considered
materialiter but a thing considered formally as under a law and pleasing
only because of its conformity to and generation through the law.
Since only reason can represent a law, reason is the higher faculty of
desire, and its a priori principle is a purposiveness which is an impera-
tive for all actual purposing. 28
The question of the present chapter may therefore be phrased thus:
How can reason be the higher faculty of desire and not merely provide
norms for judgment, post facto?
2:; This is in conformity to Kant's theory of the philosophical uses and dangers of
definition (cf. my "Kant's Theory of Definition," Philosophical Review, LXV
[1956],17HI).
26 He means not categories, but predicables (cf. Critique of Pure Reason, A 8z =
B (08). But even this is not correct. Pleasure is an empirical, not a transcendental,
predicate (ibid., A 343 = B 401).
27 Cf. above, chap. vii, § 4.
28 Critique of Practical Reason, 24-Z5 (112); Erste Einleitung in die Kritik der
Urteilskraft, XX, 146.
The "Aesthetic" of Pure Practical Reason Z 19
§ 7. AN ALLEGED INCONSISTENCY
Serious inconsistencies have been alleged to be present in Kant's dis-
cussion of the incentive to morality. He states:
The moral law should directly determine the will. If the determination of
the will occurs in accordance with the moral law but only by means of a
feeling of any kind whatsoever which must be presupposed in order that the
law may become a determining ground of the will bl (164)]
we have to do with legality, not morality. These are the opening
sentences of chapter iii and must, because of their position, be given
pre-eminent importance and weight.
Then he tells us that respect for the law is the incentive for moral-
ity,33 that respect is not the incentive to morality but is morality itself
(76 [168]), and that the law is the incentive (72 [164])'
It is regrettable that Kant was not more careful; though, had he been
so, the race of Kant-commentators would have been unemployed. But
it is unjust to do more than complain of his carelessness here and to
allege serious and insurmountable inconsistencies where the develop-
ment of the argument itself makes sometimes one and sometimes an-
other choice of words appropriate, though the statements taken out of
their context certainly seem incompatible.
In spite of what Kant says, the law itself is not the incentive. A law
is just not the sort of thing that can be an incentive. At most, con-
sciousness of a law can be an incentive. If the law itself were a deter-
minant of conduct, without the intervention of consciousness (which
32 Cf. Critique of Practical Reason, 86-87 (180); Metaphysik der Sitten, VI, 402-
3; Critique of Judgment, § 27. He had previously (Beobachtungen uber das Gefuhl
des Schonen und Erhabenen, 2. Abschnitt) distinguished among virtues those which
are sublime from those which are beautiful, obviously following Burke, Inquiry into
the Origin of Our Ideas of the Sublime and Beautiful, Part III, Sees. X and XI. On
the long history of these aesthetic concepts in morals before Kant and in Kant's
early works, see Giorgio Tonelli, "Kant: dall'estetica metolfisica all'estetica psico-
empiric a," Memorie della Academia delle Scienze di Torino (1955), pp. 77-4ZI,
passim (use the exhaustive index).
33 Critique of Practical Reason, 78 (171); at 85 (178): "respect for duty,"
222 The "Aesthetic" of Pure Practical Reason
means, for us men, also feeling), it would not be a practical law, and
men would not be free agents. Kant frequently speaks of laws' doing
things, when he means that the world is such that the laws describe it
(natural laws) or when he means that consciousness of the law will
induce a rational being to do what the law demands. u
Hence we must ask: What is the nature of the consciousness of the
law, such that it can be the incentive? If the Critique leaves any room
for doubt on this, it is removed by the Metaphysics of Morals. There
Kant says: "The respect for the law, which is, subjectively, called the
moral feeling, is identical with the consciousness of one's duo/."35 We
do not, therefore, have a theoretical knowledge of our duty and of our
will at potential variance with it, with the subsequent addition of the
feeling of respect. To know one's duty may not be, as Socrates be-
lieved, to do it. But to know what the law requires and to feel respect
for it (if not actually to obey it)36 are, for Kant, identical; even knowl-
edge of a moral law is not aseptically theoretical. Hence a false con-
trast between "law" and "respect" falls away when we read for the
former "consciousness of law," which is the only mode in which a
practical law can be effective and yet be distinct from a natural law.
But, then, what of the opening sentences? The second sentence is
often read as if Kant, when he wrote it, thought that the law must di-
rectly determine the will without any feeling intervening between the
law and the decision-a view at variance with the remainder of the
chapter. Yet he did not believe this even when he wrote this sentence.
For though the word "presupposed" is there precisely to guard against
another error (that of the moral-sense or moral-feeling school), it
should also guard against the opposite error of supposing that the de-
termination must or can be without any subsequent feeling. The de-
termination of the will by law does not require any feeling which must
be presupposed; that means that it does not require any feeling inde-
pendent of, or prior to, the consciousness of the law. The feeling of
respect, like any other feeling, need not be presupposed as something
in the natural man awaiting an object which can be found only in the
law; the feeling of respect presupposes the law and our consciousness
3. Cf. ibid., 29 ("7), the suggestion that the law is only the self-consciousness of
practical reason.
85 Metaphysik der Sitten, VI, 464.
36 Even the sinner, who transgresses the law, shows his reverence for it by trem-
bling before its majesty (Critique of Practical Reason, 80 (171)). Cf. Religion within
the Limits of Reason Alone, VI, 46-47 (Greene and Hudson, 41-42) on the re-
establishment of the effectiveness of respect for the moral law, "which we have
never been able to lose."
The "Aesthetic" of Pure Practical Reason 113
of it (or is, indeed, equivalent to the latter). The first sentence, there-
fore is not incompatible with, but anticipates, the later statement: "In
the subject there is no antecedent feeling tending to morality."31
statement does not conflict, except verbally, with the statement that
there is "no antecedent feeling tending to morality"; it is as if one were
to say in English, "A man must have feeling before he can have a feel-
ing"-perhaps a not very lucid way of saying a simple thing, but cer-
tainly no evidence of fundamental confusion in the speaker's mind.
Yet, in spite of this, there is a difference between the analyses of the
actual moral feeling (i.e., the feeling that is felt, and not the potential-
ity for it) as given in the Critique and as suggested in the Metaphysics.
The Metaphysics of Morals says that the moral feeling is either pleas-
ure or pain, pleasure if there is felt accordance with the law and pain if
there is not. The Critique, on the other hand, says that there is some-
thing analogous to both pleasure and pain even if the action does con-
form to the law, because even in conformity there is the pain of humil-
iation arising from the thwarting or striking-down of inclination.
But there is no real inconsistency here. The Metaphysics of Morals
makes a sharp contrast between pathological and moral pleasure and
pain and discusses only the latter. Now there is no moral pain in obedi-
ence to the law; there is at most a pain arising from the thwarting of
pathological feeling. But, because of the difference drawn between the
QVo kinds of feelings, Kant can discuss one without the other in the
later work. In the former, he discusses them together and thinks of
pathological feeling as the material cause of moral feeling. "Sensuous
feeling," he says (75 [168]), "which is the basis of all our inclinations,
is the condition of the particular feeling we call respect." In the Meta-
physics of Morals this relationship is neglected, and therefore there is
missing this peculiar mixture of pleasure and plain. That he is discussing
moral feeling, as actual, in a slightly different sense in the two works is
shown by the fact that in the Metaphysics moral feeling is not equated
with respect, while in the Critique it is.
t2The form of the celebrated apostrophe to duty (86 [18o}) may have been sug-
gested by Rousseau's apostrophe to conscience in Emile, which Kant admired,
On Kant's text see H. Romundt, "Vorschlag zu einer Anderung des Textes von
der Kritik der praktischen Vernunft,' Kant-Studien, xm (1908), 313-140 and re-
plies by B. Bauch and P. Natorp, ibid., pp. 315-16. Kant made little use of this pop-
ular trope. since it was hardly in keeping with the Sachlichkeit of his style. There is
an apostrophe to sincerity in Religion within the Limits of Reason Alone, VI, 190 n.
(Greene and Hudson, 178n.).
tS Foundations of the Metaphysics of Morals, 400 (16).
226 The "Aesthetic" of Pure Practical Reason
position to understand the full meaning of this in all its ramifications,
when Kant says: "Respect for law is not the incentive to morality, but
is morality itself" (76 [168]). Morality is a disposition or settled habit
to act in accord with law because it is law; and one who respects law,
in the sense described here, is one who acts from duty. Duty is the ne-
cessitation or constraint to an action by a law in the consciousness and
conduct of a person who does not, by nature, necessarily act as the law
requires. A non-sensuous rational being would not have to have any in-
centive (72 [164]) to such obedience, and therefore the concept of
duty is not applicable to such a being; such a being would have a holy
will. The "ought" of the moral imperative is an "is" for such a being. 44
But man, as a sensuous being, is "under the discipline of reason" (82
[175]) and is not endowed with a holy will. His will is at best virtuous.
Even so, man is a person and, as such, is holy. He is an end-setting
being and therefore an end in himself. He has dignity and is not like a
thing having only a price. Personality-that attribute which distin-
guishes rational beings from things-is freedom from the mechanism of
nature through a capacity to be subject to laws given by itself (auton-
omy). An empirical self, having only negative freedom, is "subject to
his own personality so far as he belongs to the intelligible world" (87
[ 180]). Personality is sublime, and whenever we respect any man, we
respect the law of the intelligible world which he more or less ade-
quately represents in the phenomenal world (77-78 [170]). Categorical
imperatives for the empirical self are laws of the intelligible self; the
"should" of the former is the "would" of the latter. While man, em-
pirically regarded, is "unholy enough," personality and humanity in
him are holy.
With this explanation of the provenance of duty, there is a signif-
icant introduction of the concept of personality. Whereas Kant had
emphasized, in the first Critique, the impossibility of a rational psychol-
ogy of the soul as substance and had made the self only an appearance
-an emphasis continued in the second Critique-there is here and else-
where a transition to a more metaphysical interpretation of the self not
only as thought but as directly experienced as a being noumenal in
character. 411 The phenomenalistic interpretation of the self. though it is,
as it were, the official Kantian doctrine, shows signs here and there of
being transcended in a metaphysical personalism and activism, since the
spontaneity felt in self-consciousness separates man from nature. While
a "metaphysics" in the classical and not in the Kantian sense seems to
u Ibid., 449, 455 (68,75); d. Critique of Judgment, § 76.
45A full study of this, with a collection of relevant passages, is given in Heim-
5Oeth's Studien zur Philosophie Immanuel Kll1lts, pp. 227 fl.
The "Aesthetic" of Pure Practical Reason 117
§ 10. VIRTUE
Virtue is a "naturally acquired faculty of a non-holy will" (33
[121]). It is th" moral disposition in conflict or, as Abbott eloquently
translates it, "the moral disposition militant."48 Without the two poles
of sensuous inclination and pure rational principle, it could not arise.
46 Critique of Pure Reason, A 365, absent from B; cf. B 431-32, absent from A.
41 Religion within the Limits of Reason Alone, VI, 27~28 (Greene and Hudson,
22-13). This is identified with moral feeling in the sense of the Metaphysik der
Sitten, VI, 399. instead of in the sense of the second Critique (cf. above, p. 224).
48 Critique of Practical Reason, 84 (t 78) • Abbott's translation here aptly preserves
the military symbolism of the page.
228 The "Aesthetic" of Pure Practical Reason
We stand under the discipline of reason; we are conscripts in the moral
host, taking orders, and not volunteers, acting from our individual
merit. We are members of the realm of ends, but we are subjects in it
and not, at first, sovereigns of it (82 [175])' In man, virtue is always
defective, and virtue itself, like personality, is a mere Idea to which no
perfectly adequate empirical representation can be found. 49
In many of Kant's examples of virtue, the necessity of acting out of
respect for law and not from inclination is presented as if the action
had to take place without, or even always against, inclination. But this
is not what the examples were chosen to illustrate, and this is no part
of Kant's theory; even acting from inclination is not in itself evil, but
only the taking into one's maxim of action and insubordination of sensu-
ous incentives to rational incentives-this insubordination itself being
free and spontaneous and a matter of principle-is the root of moral
eviPo
The purpose of the examples of virtuous action can be understood
only in the light of their polemical or pedagogical use. An act in which
a man did what duty required of him and to which his inclination also
pointed might easily appear-and certainly would appear to a man with
Kant's skeptical hatred of cant-to be an action done in accordance
with law because of his inclination. Even where inclination seems to be
against the law, the mystery of the human heart is so deep that we can
never be sure that we will not, upon more penetrating self-analysis,
come upon "the dear self" as the real determinant. 11l But where we find
the dear self in this role, there is no morality; hence the best illustra-
tions of genuine morality will be found where the dear self is obviously
opposed, though that does not mean, as we shall soon see, that the best
morality is one in which there must be active and painful self-abnega-
tion. Virtue does not entail a renunciation of happiness, but only a
willingness "to take no account of happiness when duty is in question"
(93 [I 86]).
But those who read Kant's examples and fail to understand the text
that they were meant to illustrate have always regarded his ethics as
singularly repellent and have thought that if they represent typical and
most estimable virtue, then it is an ethics which they must reject. 52 In-
clinations and feelings, however, have a legitimate place in Kant's ethics
which is not brought out by these examples. They may even be seen as
491bid., 127 n. (214 n.); Critique of Pure Reason, A 315 = B 372; Critique of Judg-
ment, § 57, Remark 1.
ISO Religion within the Limits of Reason Alone, VI, 36 (Greene and Hudson, 3x)'
lSI Foundations of the Metaphysics of Morals, 407 (24).
1S2 Cf. above, p. 120, n. 20.
The "Aesthetic" of Pure Practical Reason 129
§ I I. MORAL SATISFACfION
We said above, in § 7, that in moral action we find a satisfaction in
the experience of our own autonomous spontaneity and that this is the
positive element in respect as the moral feeling. 56 Man is humiliated by
his vision of the law, before whose majesty even the boldest sinner
trembles (80 [I 72]); but he is elevated in his consciousness that this
humiliation is the mark of his higher vocation, for it is humiliation of
himself by himself, of his heteronomous natural being by his autono-
mous intelligible being (cf. 76 [1691 on Fontanelle).
In morality, therefore, in the consciousness of determination by law,
there is a feeling of joy, since reason's interest is being furthered (116
[z I 2]). Freedom itself, as the exercise of moral will, thus becomes sub-
ject to enjoyment, which is not happiness but essential to it. lS1 This joy,
however, cannot define virtue, since it arises only from a prior con-
sciousness of virtue, which, formally and abstractly considered, has
113 A holy will has no incentives, no duty, and no virtue; it is only a standard or
Idea by which we judge human incentives, duty, and virtue. Kant observes that the
Stoic sage was believed to be so perfect that he was "elevated above duties though
he propounded duties to others," but that this error arose only because the Stoics
failed to conceive the law "in the same purity and rigor as does the precept of the
Gospel" (Critique of Practical Reason, 117 n. [114 n.J).
5f On "sweet and sour duty," cf. Metaphysik der Sitten, VI, 377, 391, and Lectures
on Ethics, p. 199.
illS Even the inclinations refined to civility have nothing stable or of absolute
wonh in them: "Everything good that does not grow from its connection with a
morally good disposition is nothing but empty show and glittering misery" (Idea
for a Universal History, Seventh Thesis, VIII, 16). In such judgments, Kant shows
some of the passion and often uses some of the words of Rousseau.
116 The Critique of Judgment, § 29, General Remark (V. 171 [III-U», emphasizes
the positive side of the moral feeling more than the second Critique does. In fact,
the discussion of se1f-contentment in the second does not directly identify it as a
component in the moral feeling, and the whole treatment of it is not found in the
chapter on the incentives but is introduced later into the chapter on the summum
bonum. Nevertheless, it is cleat that the joy in morality is a part of it, the humilia-
tion of natural man before the law being a necessary precondition.
51 Critique of Practical Reason, 118 (liS) and Fragment 6 (Reftexion 7201).
230 The" A esthetic" of Pure Practical Reason
nothing to do with happiness itself. It was the fundamental logical error
of the eudaemonists to recommend virtue as a source of happiness with-
in one's self, arising from self-approbation predicated upon conscious-
ness of virtue, without giving a criterion for approbation so that con-
formity to it could indeed be a source of joy. But the error of Epicurus
and the moral-sense theory (38, 116 [128, 21 2]), already pointed out
by Hutcheson and Price,II8 should not cause us to commit the opposite
error of denying a positive joy to be possible in moral conduct.
Hume59 had complained of the lack of a name for this feeling of
moral well-being, and Kant, as if directly answering Hume, says:
Do we not have a word to denote a satisfaction with existence, an analogue
of happiness which necessarily accompanies the consciousness of virtue, and
which does not indicate a gratification, as "happiness!) does? We do, and
this word is "self-contentment," which in its real meaning refers only to
negative satisfaction with existence in which one is conscious of needing
nothing. Freedom and the consciousness of freedom, as a capacity for fol-
lowing the moral law with an unyielding disposition, is independence from
inclinations, at least as motives determining (though not as affecting) our
desiring; and, so far as I am conscious of freedom in obeying my moral
maxims, it is the exclusive source of an unchanging contentment necessarily
connected with it and resting on no particular feeling. 60
That this feeling is the stronger the more purely the law is presented.
is a fundamental thesis in Kant's educational theory.61 Other impure
motives may well work in the direction of legally correct action (81
[174]), but they may cloud the vision of the man for the only thing
that can bring him to morality and lasting equanimity. Thus, Kant
continued:
Even the feeling of sympathy and warm-hearted fellow feeling, when pre-
ceding the consideration of what is duty and serving as a determining
ground, is burdensome even to right-thinking persons, confusing their con-
sidered maxims and creating the wish to be free of them and subject only
to law-giving reason. 62
:>8 Cf. above, p. (06.
119 Enquiry concerning the Principles of Morals, Appendix IV (ed. Selby-Bigge
[zd ed.J, p. 3(4).
60 Critique of Practical Reason, ((7 (zr4); on contentment d. Wolff, Verniinf-
tige Gedancken 'Clon Gott, der Welt und der Seele des Menschen, § 463.
61 Critique of Practical Reason, I~Z, 1~7 (25t, 255); d. Foundations of the Meta-
physics of Mora/s, 41()-u (27); Uber den Gemeinspruch •.. , VIII, 288.
62 Critique of Practical Reason, 1I8 (214). italics added. Similarly, Foundations
of the Metaphysics of Morais, 428 ('J6). But contrast the remarks in Religion within
the Limits of Reason Alone, VI, 58 (Greene and Hudson, 51) on the vanity and
sin of wishing to be free from all inclinations.
The "Aesthetic" of Pure Practical Reason 13 I
It was no doubt in reaction to this that Friedrich Schiller wrote:
In the Kantian moral philosophy the idea of duty is presented with a hard-
ness which frightens away all the graces and which could mislead an obtuse
mind to seek moral perfection on the path of a dour and monkish asceticism.
However much the great philosopher tried to avoid this misunderstanding,
which must have been disturbing to him above all, with his serene and free
mind, I still think that he himself gave obvious occasion to it through the
sharp and shrill contrast between the two principles which work on the will
of man-an occasion which perhaps can hardly be avoided in his theory.63
To this Kant replied in a passage in the Religion which is not in any
way inconsistent with the teaching of the Critique, though, if it had
been said in the Critique, the "occasion" would not have been so ob-
ViOUS:
Now if one asks, What is the aesthetic character, the temperament, so to
speak, of virtue, whether courageous and hence joyous or fear-ridden and
dejected, the answer is hardly necessary. This latter slavish frame of mind
can never occur without a hidden hatred of the law. And a heart which is
happy in the performance of its duty (not merely complacent in the recogni-
tion thereof) is a mark of genuineness in the virtuous spirit-a genuineness
even in piety, which does not consist in the self-inflicted torment of a re-
pentant sinner (a very ambiguous state of mind, which ordinarii y is nothing
but inward regret at having infringed upon the rules of prudence 64 ), but
rather in the fum resolve to do better in the future. This resolve, then, en-
couraged by good progress, must needs beget a joyous frame of mind, with-
out which man is never certain of having really attained a love for the good,
i.e., of having incorporated it into his maxim.6Ii
Let this lay the ghost of an old error. But it has often been said that
the Kantian doctrine is "false to human nature," in that it neglects or
denies the role of spontaneous feeling precisely because it puts them in
63 0 ber Anmut und Wurde. Note that Schiller says that this is a misunderstanding
of Kant's views and that Kant regarded Schiller as a philosophical ally, not an op-
ponent (Cf. Religion . .. , VI, 13 n., first sentence [Greene and Hudson, 19 n., sec-
ond sentence]). Schiller's famous verse can be regarded only as a joke and not as
presenting Schiller's real views of the Kantian philosophy. There Was a point that
actually divided them, however, though it is not clearly brought out in their Ausein-
andersetzung, to wit, whether the collaboration of rational and sensuous elements is
essential to vinue (Schiller) or, at most, one of the duties (Kant, Metaphysik der
Sitten, Tugendlehre. S 48 [VI, 473]-to unite the virtues and the graces is a duty).
For perhaps the best full study of the Kant-Schiller controversy and the extensive
literature that has grown from it see Hans Reiner, Pflicbt 'lind Neigung (Meisen-
heim, 1951), pp. z8-49.
64 On repentance see Critique of Practical Reason, 98 (19Z).
6Ii Religion within the Limits of Reason Alone, VI, 13 n. (Greene and Hudson,
(9 n .).
232 The "Aesthetic" of Pure Practical Reason
second place, after the rational determination. 66 It therefore seems to
make impossible, as well as undesirable, demands upon the natural man.
Not, perhaps, transcendentally impossible demands, because of Kant's
doctrine that "thou canst, because thou shouldst," but demands that
living men cannot be reasonably expected to acknowledge.
There are three answers to this. One is that it is an irrelevant criti-
cism, since for Kant an ethical theory is not to be judged by the facts
of human nature (about which, indeed, we know far too little and into
which we surreptitiously introduce unexamined pessimistic or optimis-
tic moral judgments according to our taste).67 Second, as a corollary to
this, we must remember that such judgments of human frailty are not
alien to Kant himself. His derogatory judgments of the moral pros-
pects of human nature and society are, in fact, sharper than those of
either Aristotle or Dewey, two other great proponents of reasonable
morality.68 And, third, Kant saw the need, for us men, of supplement-
ing the pure moral motive with others more natural to man, in spite of
his warnings against this in the Critique, where he was examining pure
practical reason. Love is the most important of these motives:
When it is a question not merely of the representation of duty [which is
the task of the Critique J but of its execution, if one asks about the subjective
ground of action on which, if presupposed, one can at least expect what man
will do and not what on the objective ground he ought to do, it is love which
66 Dr. Johnson was no great philosopher, not even a competent one. But he was a
man of solid understanding whose testimony on a point of moral judgment carries
a weight it would not have had if it had issued from a degree of philosophical so-
phistication he neither had nor admired; thus does the wise layman sometimes serve
as the impartial observer in philosophical disputes. Boswell writes: .. 'Sir,' said Mr.
Johnson, 'I can lay but little stress upon that instinctive, that constitutiona~ goodness
that is not founded upon principle. I grant you that such a man may be a very good
member of society. I can conceive him placed in such a situation that he is not much
tempted to deviate from what is right; and so, as goodness is most eligible when
there is not some strong enticement to transgress its precepts, I can conceive him
doing no harm. But if such a man stood in need of money, I should not like to trust
him. And even now, I should not trust Mr. Dempster with young ladies, for there
is always a temptation" (London Journal, July 11, (763). The same is given in
slightly altered fonn in the Life ("Modern Library" ed.), p. 268; cf. also Price,
Rewew of the Principal Questions of Morals, :ed. D. D. Raphael (Oxford. (948),
p.I9I.
67 Uber den Gemeimpruch ... , passim, but especially VIII, 276-77.
68 Foundations of the Metaphysics of Morals, 407 (23); Idea fOT a Universal His-
tory, Sixth Thesis (VIII, 23): "Out of such crooked lumber as man is made of,
nothing truly straight can be builded." Similarly, Religion within the Limits of
ReMon Alone, VI, 100 (Greene and Hudson, 92),
The "Aesthetic" of Pure Practical Reason 133
is an indispensable supplement to the imperfection of human nature, as a free
assumption of the will of another under one's own maxim. 69
This love cannot be commanded; it is not a duty, because it is self-
contradictory to require of a man that he do something out of inclina-
tion. 70 It must, therefore, be distinguished from the love which can be,
and in the gospel is, commanded, which is practical love or the ready
willingness to fulfil our total obligation to others. Kant, with his con-
viction that the true kernel of all religious truth is moral, regards this as
equivalent to the love commanded by Christianity.71
§ n. MORAL EDUCATION
By what course is the human race brought to a state in which the
moral disposition may rise and flourish? How can the child, in a civi-
lized state, be brought from innocence to moral maturity? These two
questions are closely related~ and though in the Critique Kant deals
only with the second, the first was very much on his mind during the
years immediately preceding the writing of it.
In the Age of Enlightenment, religion as the moral teacher of man-
kind was gradually pushed from its pre-eminent position. Though Kant
esteemed religion as the moral teacher of mankind, religion was, for
him, only its teacher, not its source. The great Enlightenment philoso-
phers sought a generally naturalistic origin of morality, of which reli-
gion might be seen as only a vehicle. Lessing, Herder, Kant, and Schil-
ler were at one in attempting a natural history of freedom, and, with it,
of morality, though for Kant, at least, neither natural history nor di-
vine revelation was adequate to explain its peculiar form and unique
authority. These attempts led to their philosophies of history, in which
the education of the human race is the chief theme, appearing even in
the titles of some of their books.
69 Das Ende alJer Dinge, VIII, 337. A similar office. is assigned to sympathy, which
is "an impulse, planted in us by nature, to do that which the representation of duty
alone would not be able to accomplish" (Metaphysik der Sitten, Tugendlehre, § 35).
But in moral education it is bener to try to weaken the inclinations than to ally them
with consciousness of duty, an alliance which may indeed make for better actions
but will not produce better men (Reflexion 6722; cf. Critique of Practical Reason,
88 [182]). Even to let other motives co-operate with the moral law is risky (ibid.,
72 [.64]). The difference in these passages cannot be denied; but the emphasis upon
a kind of moral synergism is more obvious in the practical than in the purely theo-
retical investigations of morals.
70 We cannot he commanded to do anything from inclination Critique of Prac-
tical Reason, 83 [l76J and Das Ende aller Dinge, loco cit.
11 On practical love cf. Metaphysik der Sitten, VI, 449, and below, p. 243, n. [3,
234 The "Aesthetic" of Pure Practical Reason
That the human race must advance itself is clearly and sharply stated
by Kant:
Nature has willed that man should, by himself, produce everything that
goes beyond the mechanical ordering of his animal existence, and that he
should partake of no other happiness or perfection than that which he him-
self, independently of instinct, has created by his own reason.
It is
just as if she had willed that, if man ever did advance from the lowest
barbarity to the highest skill and mental perfection, and thereby worked him-
self up to happiness (so far as it is possible on earth), he alone should have
the credit and should have only himself to thank-exactly as if she aimed
more at his rational self-esteem than at his well-being. 72
There is a natural predisposition to morality in man, but not natural
morality. From a state of natural innocence, men fell into evil; history
is his gradual working himself out of it. The Conjectural Beginning of
Human History treats the Book of Genesis as an allegory of this fall
and its consequences: "The history of nature begins from the good, for
it is the work of God; the history of freedom begins from evil, for it
is the work of man."73 The Idea for a Universal History traces the his-
tory of man's social arrangements, which are his measures for control-
ling the evils naturally arising from his "unsocial sociability" or need
for, but antagonism to, his fellows. As Heraclitus said, conflict is the
origin of all things; and the product is civilization and ordered state, in
which man assumes the role of unselfishness, respect for others, and de-
cency-a role which fools nobody and which is called "permitted moral
pretence."74 From it, however, true morality can arise when reason,
gradually perfected as an instrument, becomes aware of its higher voca-
tion. Thence a.rises the genuine interest in morality, and thereafter
"moral pretence" must be combated. 75
Assuming that genuine moral ideas have a certain currency in soci-
ety, how is the child brought to recognize them and respect them?
This is the problem of moral education proper. It is, for Kant, a unique
and irreducible division of pedagogy, not merely because of its method,
72 Idea for a Universal History, Third Thesis, VIII, 1,)-10.
§ I. WHAT DIALECTIC IS
Both theoretical and practical reason have a dialectic and on the same
grounds, viz., as reason they seek the unconditioned for all that is
conditioned, but they cannot find it as an object of knowledge, though,
in default of a critique, it appears that they have found it (107 [101]).
Dialectic is the exposure of the illusion that the unconditioned, as re-
quired by reason, is an object of some definite and specific cognition.
The Critique of Pure Reason gives a full account of what dialectic
is. General logic, Kant tells us there, is sometimes employed as if it
were an organon for the production of knowledge instead of as a
canon for the institution and evaluation of inferences among cogni-
tions which have their source in experience. 1 As such an organon, it is
an art of giving to ignorance and sophistry the appearance of truth.
But instruction in such an art is beneath the dignity of philosophy;
and dialectic, he says, has come to mean the criticism of the 'illusions
arising from the sophistical art.
Transcendental logic likewise produces illusion when the categories
of the understanding, which give knowledge only when related to
experience, are used as an organon for knowing things in general,
whether given in experience or not. 2 The transcendental illusions which
arise in this way are not arbitrary and intentional, like those produced
by the sophistical dialectician for forensic purposes, but arise necessari-
ly from the inescapable, yet insatiable, requirements of thought. The
conditions of all judgment are the categories, but out where experience
cannot reach they are mere Ideas, rules, or maxims for the employment
of our reason. Unchecked by experience, they inevitably appear to be
1 Critique of Pure Reason, A 61 = B 85.
21bid., A 63 =B 88.
139
240 The Dialectic of Pure Practical Reason
objective truths of the highest kind, since no experience can ever
refute them. They seem to be true, that is, until their illusoriness is
exposed by criticism, which prevents them from deceiving us, though
it cannot prevent them from arising, any more than a knowledge of the
laws of light can prevent our seeing optical illusions.3
The master method of exposing dialectical illusion is what Kant
calls the "skeptical method," which consists in "watching, or rather
provoking, a conflict of assertions, not for the purpose of deciding
between them, but of investigating whether the object of controversy
is not perhaps a deceptive appearance."4 The aim of the skeptical
method is not skepticism but certainty, since, by exposing illusion
where it exists, it will prevent it from infecting the entire organism
of reason with dubiety. By restricting the pretensions of theoretical
knowledge, it opens the way for the practical use of reason, and the
antinomy, which destroys speculative or theoretical metaphysics about
the supersensuous, is thereby the "most fortunate perplexity" into
which pure reason could ever fall (107 [203]).
§ 2. THE DIALECTICAL ILLUSIONS OF PURE PRACTICAL REASON
One might well be surprised that there is a dialectic of pure prac-
tical reason. In the first Critique Kant says that moral philosophy does
not lend itself to the skeptical method, since it can "present its prin-
ciples, together with their practical consequences, one and all in con-
creto, in what are at least possible experiences, and the misunderstand-
ing due to abstraction is avoided."1i While the going beyond sense
experience is the occasion for the theoretical dialectic, leading theory
into a "chaos of uncertainty, obscurity, and instability," the power of
practical judgment shows itself at its best where everything sensuous
is excluded. 6 There is no need to see whether pure practical reason,
like theoretical reason, "presumptuously overreaches itself"; it does
not "lose itself among unattainable objects or contradictory concepts,"
and pure practical reason stands in need of no critique (3, 16 [87, 102]).
Yet we are promised in the Introduction a Dialectic as the exhibition
and resolution of illusion in the judgments of practical reason (16
[ 102]). And this does not refer to the illusion that practical reason
in its empirical function overreaches itself and usurps the place of pure
practical reason; 7 for the section which fulfils this promise is clearly
3 Ibid., A 297 = B 353-54. 4 Ibid., A 423 = B 45 1• II Ibid., A 425 = B 453.
6 Foundations of the Metaphysics of Morals, 404 (11).
7 This is the sense of "dialectic" in the Foundatitms (405 [21]): from the conflict
of reason with the senses a "natural dialectic arises, a propensity to argue against the
stem laws of duty."
The Dialectic of Pure Practical Reason Z41
22]bid., 119 (11 5). The argument has been interrupted by a discourse on the con-
nection between happiness and virtue not mediated by the intelligible world (pp.
115-19 [ZII-I5]). These paragraphs might well be set apart from the rest and called
a "Remark." We have already commented upon them in chap. xii. The connection
between virtue and happiness in the highest good is made possible in an intelligible
world of perfect beings without the mediation of God, according to the Critique of
=
Pure Reason, A 810 B 838, where it is called a "system of self-rewarding morality."
But for us men this is not sufficient, or, rather, we are not sufficient to it.
23 Each proposition in an antinomy is "not merely in itself free from contradiction
but finds conditions of its necessity in the very nature of reason-only that, un-
fortunately, the assertion of the opposite has, on its side, grounds that are just as valid
and necessary" (ibid., A 4z I = B 449).
248 The Dialectic of Pure Practical Reason
This is a real antinomy, since the propositions are contradictory,
not contrary; each expresses an inescapable interest of reason (moral
and theoretical); and each is a true formula for one of these interests.
Furthermore, the resolution of this antinomy conforms to that of the
third antinomy in the first Critique. The antithesis is and remains
(unfortunately for the virtuous) true of the sensuous world; the wind
is not tempered to the shorn lamb, and the rain falls on the just and
the unjust alike. The thesis is false in "a system of nature which is
merely the obj ect of the senses" (II 5 [21 I]). But it may be true if
there is an intelligible world in which there is a moral government
distributing happiness (in the sensuous world, under natural condi-
tions)24 according to laws holding in the intelligible world (laws of
the moral intentions of persons). Hence the pursuit of virtue does not
rule out the hope, based upon the possibility, of a proportionate
happiness.
Another statement of the antinomy 25 is possible, and it has the merit
of keeping the summum bonum in the center of attention throughout.
II. THESIS: The summum bonum is possible. Proof: The rnorallaw requires it.
ANTITHESIS: The summum bonum is not possible. Proof: The connection
between virtue and happiness is neither analytical nor synthetic a priori
nor empirically given.
RESOLUTION: The antithesis is true of the sensible world if the laws of
nature have excusive sovereignty; the thesis may be true of the intelli-
gible world because the synthetic connection of virtue (as ground) to
happiness (as consequent) is not absolutely impossible.
In either formulation we are enabled to see that the concept of the
highest good is not impossible. Kant, believing that it is morally nec-
essary, then turns to those conditions which would make it actual
if superadded to morality, which is the only one of the conditions
that lies within our control and competence.26
24 Kant says that the effect must be in the sensuous world. He therefore does not
base the postulate of inunortality upon the need for happiness in another world
as the locus of this reward. But in this life the adjustment is not made. The four
paragraphs which I have suggested should be set apart as a "Remark on a Self-re-
warding System of Morality" do not speak of the moral contentment in this life as
being happiness. This would suggest that the happiness in the summum bonum must
be that of an afterlife, not under a system of nature. So, even before we are ready
to take up the postulate of immortality, we run into an obscurity, the first of many
we shall find in this notion (cf. below, chap. xiv, ~ 4).
25 Slightly modified from a proposal by August Messer, Kants Ethik, p. 88.
26The twofold division of his task, mentioned on p. 1I0 (205), seems to refer to
this; but this division did not actually serve to articulate the succeeding sections
of the Dialectic. Perhaps Kant had intended a more extensive discussion of the postu-
late of freedom prior to the introduction of the other two postulates; then this
division would apply.
The Dialectic of Pure Practical Reason 249
§ 5. THE PRIMACY OF PURE PRACTICAL OVER THEORETICAL REASON
The question is: Have we a right to use concepts of reason, and
to assert objects for them, which are beyond the sphere of knowledge
marked off for and by theoretical reason? It is not sufficient to point
out merely that the interests of practice extend beyond those of theory.
Theory alone does not decide between the thesis and the antithesis of
the third antinomy, and theory would leave us in a state of vacillation
which could be ended, if theory were the only use of reason, by a deci-
sion forced upon us by the needs of practice. 27 This decision would in-
deed be a decision in favor of the thesis. The failure of theoretical reason
to decide between thesis and antithesis would pennit us to employ the
requisite "faith" in the claims of dogmatism but could not give faith
the "form of science," i.e., show its systematic rational strucrure. 28
There is not now, after the Analytic's demonstration of the function
of pure practical reason, any remnant of arbitrariness or irrationality
left in the decision that must be made. We do not have the contrast
between "reason" and "faith" but that between "theoretical reason"
and "rational faith," i.e., an a priori faculty that goes beyond theory.
A contrast exists, but not necessarily an opposition, because (a) there
is only one reason with two interests, not two reasons with opposing
interests (12 I [2 I 7] ), and (b) the practical does not require us to go
against anything that the thtoretical reason has established or can
establish. Put in classical theological terms: the things of faith are be-
yond the things of theoretical reason but not beyond the things of
reason simpliciter; and there are no valid things of faith that are against
the things of theoretical reason.
Theoretical reason pushes inquiry to the farthest limits in striving
to reach the unconditioned. It cannot do this by attaining knowledge
of the unconditioned, but its interest is whatever furthers the exercise
of this function. 29 Its true interest, therefore, is not in some special
cognition, or definition, of the unconditioned but solely in the restric-
tion of speculative folly. The interest of practical reason is the "de-
termination of the will with respect to the final and perfect end" (120
27 Critique of Pure Rellson, A 475 = B S03.
281bid., A 470::;:: B 498.
29 There is an inconsistency in Kant's use of the word "interest," He identifies
interest with (a) the end of the use of a faculty and (b) a principle under which
its exercise is advanced. Thus, of theoretical reason, a is "the knowledge of objects
up to the highest a priori principles," and b is the establishment of limits which
cannot be breached by "monstrosities of reason" which interfere with its orderly
progress. Its interest in sense b is the restriction of speculative folly (Critique of
Practical Reason. no, IZ J [117]). The interest of theoretical reason appears to the
dogmatic metaphysician to be iI; to the critical metaphysician like Kant, it is b.
2. 50 The Dialectic of Pure Practical Reason
[216]), and whatever shows this end to be possible furthers the prac-
tical interest.
The two interests may superficially appear to be in conflict, but
they are not opposed to each other if the interest of theoretical reason
is not some end which it cannot achieve (knowledge of things in
themselves) but only the guaranty of the orderly progress of science,
and if the interest of practical reason is not presented as cognition
but as determination of actions. Then judgments about what furthers
the latter interest can be compatible with the former interest. These
judgments will not be cognitive judgments known theoretically to be
true; they will be only postulates.
The Critique of Pure Reason itself prepares the way for the assertion
of the primacy of the practical over the theoretical interest of reason.
The interest of speculative reason in the Ideas needed by pure practical
reason is very small;30 yet the Ideas needed by speculative reason for
its own purpose are the same as those needed by pure practical rea-
son;31 and theoretical reason can present them only as problematic
concepts.S2 Practical reason, if there were a conflict between it and
theoretical reason, would enter the contest with the "right of posses-
sion" on its own side.ss In the absence of opposition, the question
concerns solely which maxim should be followed: to refuse stubborn-
ly to admit any non-empirical objects (the canon of Epicurus) or to
"admit certain theoretical positions" inseparably bound with a priori
practical principles but going "beyond any possible insight of the
speculative reason (although not contradictory to it)" (120 [:u6]).
The doctrine of primacy prevents what is a difference from becoming
an incompatibility, by establishing an order of subordination instead
of co-ordination. And in this subordination the practical has primacy
because "every interest is ultimately practical."3~
If theoretical reason can be assured that these "theoretical positions"
are required by the practical interest and exceed, but do not conflict
with, its own interest, this primacy is assured. Kant claims that the
requirement is shown by the resolution of the antinomy of practical
reason and the necessity of the conception of the summum bonum.
Sections 6-9 of this chapter of the Critique are devoted to the establish-
ment of the absence of conflict.
30 Critique of Pure Reason, A 74)8 = B 816. =
31/bid., A 7¢ B 814.
32/bid., A 796 = B 814: "Reason has a presentiment of objects which possess a
great interest for it. But when it follows the path of pure speculation. in order to
approach them. they fly before it."
s3/bid., A 777 = B 805.
=
IU Ibid., A 816 B 844; Critique of Practical Reason, 111 (u8).
The Dialectic of Pure Practical Reason 15 I
§ 6. THE POSTULATES OF PURE PRACflCAL REASON 35
§ I. INTRODUCTION
In chapters xi and Xlll it was mentioned that Kant uses the word
"postulate" rather loosely and that his list of practical postulates varies
from place to place. There is some diversity even within the Critique
of Practical Reason. The possibilityl and the actualit y 2 of God, free-
dom, and immortality; the possibility of the summum bonum;3 and the
moral law itself4 are named as postulates. But, in spite of this diversity,
Kant's meaning and intention in the Critique are clear: the three
postulates, officially and deliberately so called, are the actual freedom
of the will, the immortality of the soul, and the existence of God.
We have already discussed the concept of freedom and the sense
in which freedom is a postulate. 5 There remains, therefore, in this
chapter to discuss only the two remaining postulates. We shall begin,
however, by examining once again the concept of postulate in the
light of how a postulate is "demonstrated," then examine the two
postulates themselves, and conclude with a discussion of the relation
between ethics and religion. The very last section comments upon
Kant's conclusion to the whole work, which, though it follows the
Methodology, follows smoothly upon the Dialectic.
1 Critique of Practical Reason, 134 (132): " ..• Through the practical law, which
requires the existence of the highest good possible in the world, there is postulated
the possibility of those objects of pure speculative reason,"
2 Ibid., next sentence: "Practical reason inexorably requires the existence of those
objects for the possibility of its practically and absolutely necessary object, the
highest good." Ibid., 125 (221): " .•. The postulate of the possibility of a highest
derived good (the best world) is at the same time the postulate of the reality of a
highest original good, namely, the existence of God" (all italics supplied).
3 Cf. the just preceding quotation.
4 Ibid., 46 (135); denied at 131 (229). 5 In chap. xi, ~ 14.
260 The Postulates of Pure Practical Reason
§ .:. IHI'. SfRL'CrCRE OF A 1\lORAL ARGUME;\" r
Kant does not use the words "moral argument" in the second Cri-
tique, but it is the name by which he refers to this argument in the
third. Although it is not an entirely suitable name, it has become so
widely accepted that it is useless to try to reject it; but we must try
to be very clear in deciding what a "moral argument" is; else we shall
find the term so vague that it covers some arguments that are logically
quite different from the argument first called a "moral" one.
Ic is universally recognized that a value-judgment or a practical
proposition containing an "ought" cannot be derived by an argument
from premises which are exclusively factual. Enthymemes in which
this is done always require a hidden value-premise, which is omitted
simply because it is thought to be obvious. Only a philosopher would
object to an argument that said, "You ought to take an aspirin because
you have a headache." But even he would make himself tedious if he
insisted upon completing such syllogisms. Still, such an argument is
not valid unless the value-premise is supplied.
A moral argument is one in which a factual conclusion is reached
from a value-premise, either immediately or taken in conjunction with
a factual premise. 6 Such an argument likewise is formally invalid. It
cannot be expected that a philosopher of Kant's skill and stature should
be guilty of such a sophomoric blunder; yet the ordinary interpretation
of the moral argument attributes this nonsense to him.
Italicizing two words in Kant's definition of a postulate shows him
to be aware of the danger of this kind of error, though whether he
finally avoided it in its most subtle and insidious form is not a matter
that can be decided by examination of a single sentence of his text. He
says a postulate is "a theoretical proposition which is not as such
demonstrable, but which is an inseparable corollary of an uncondi-
tionally valid practical law" (122 [219])' We must now examine this
"as such."
Whether a proposition is theoretical or practical depends ultimately
upon its function, not upon its formal structure or content, though we
have the means, syntactic or semantic, to make these functions clear.
"A causes B" is a theoretical proposition; "Since you wish to have B,
do A" is a practical proposition; "A is a means to B" has some of the
characteristics of both. Just as we can move from theoretical to prac-
tical propositions by supplying a practical premise (e.g., "If you want
X, you ought to do what will lead to X"), it would appear that we
ferent, for the existence of God is not connected with intuitions but only with con-
cepts. To be really possible in the senSe meant in the second Critique is to be (0)
logically possible and (b) related necessarily to some other fact (viz., the moral)
whose reality is given. Thus the fact of pure reason is the practical corollary of in-
tuition (cf. above, pp. 152 f., 172 f.) in converting mere concepts of the logically
possible into cognitions that the logically possible is really possible (cf. Critique of
Practical Reason, 66 [157]).
33 Critique of Pure Reason, A 813 ::::: B 841.
34 Ibid., A 818 = B 846, A 589 :::::: B 617, A 634 :::::: B 662; Reflexionen 61 10,6236.
35 Only in Religion within the Limits of Reason Alone, VI, 101 (Greene and
Hudson, 92) does Kant seem to have been able to argue for the existence of God
exclusively from the first component of the summum bonum. The highest moral
good, he there tells us, cannot be achieved by an isolated individual, but only in a
moral commonwealth. The establishment of this moral commonwealth cannot be
the work of man, who can only make himself worthy of membership in it; its actual
establishment can be only by the grace of God. Virtue is worthiness to this grace
(which now replaces worthiness to be happy), and the summum bonum is the
Kingdom of God to be established by God. Hence the moral command to seek the
Kingdom of God requires that such a moral legislator exist or justifies the belief
and hope independently of the rewards it may contain. The last time the moral
argument is given in its classical fonn, involving both the components of the
summum bonum is in Critique of judg;mem, § 87, where it is presented with only
one modification. Here the natural desire of man for happiness is not taken as a
prior condition for the definition of the summum bonum, but the existence of God
is taken as the condition under which a man may morally .set before himself any
274 The Postulates of Pure Practical Reason
the development of the theory of autonomy, there is a decisive modi-
fication of the peculiar feature of the argument of the first Critique.
The belief in the existence of God is based upon the putative necessity
of the second element in the summum bonum (124 [220]); but he now
succeeds in explaining more fully how it can be present and not func-
tion as a motive-a matter that made the conclusion in the first Critique
seem heteronomous, as pointed out above. Now the reason for assum-
ing the Slmrmum bonum and its conditions is not that man naturally
desires happiness and happiness is a component of the summum bonum;
it is that the moral command would not be just ineffective, it 'Would be
null and 'Void, if it commanded the impossible, and the summum bonum
would be impossible (so far as human mind can comprehend it) (145
[244]) if God did not exist.
We are now ready to summarize Kant's argument for this postulate.
I. Happiness is the condition of a rational being in the world in whose
whole existence everything goes according to wish and will.
2. Man's will is not the cause of nature and does not bring nature
into complete harmony with the principles of his will.
3. There is, therefore, no ground in the moral law (or in nature) for
expecting a necessary connection between the morality and happiness
of men.
4. But such a connection, in the concept of the summum bonum, is
postulated in the command that we ought to seek the summum bonum.
5. The highest good must, therefore, be possible.
6. Therefore, a cause adequate to it must be postulated.
7. Such a cause must be the Author of nature, acting through under-
standing and will. Such a being is God (IZ 5 [n 1]).
Let us begin our examination of this argument with statement 3. The
disproportion of worthiness to be happy and actual happiness can be
and often has been taken as evidence against the existence of God, at
least in the sense of statement 7. How then does Kant manage to build
this premise of his opponents into an argument for the existence of
purpose as necessary. even if the realization of it is not within his power. This pur-
pose is not happiness desired and subsequently restricted to the moral condition of
wo~~hiness. but happiness only insofar as it is morally earned. A similar thought is
in U beT den Gemeinsprucb ... , VIII. 180 n.
Though Kant supplemented the moral argument in various ways. as we have seen.
he did not renounce it (d. George A. Schrader. "Kant's Presumed Repudiation of
the Moral Argument in the Opus postumum," Philosophy. XXVI [l95d. 218-41).
That the "new" doctrines of the Opus postumum can be found side by side with the
moral argument is shown also by Walter Reinhard. UbeT das Verha/mit von Sitt-
lichkeit una Religion bei Kant (Bern: Haupt, 19%7).
The Postulates of Pure Practical Reason 275
God? The answer is found not in any alleged theoretical fact or hy-
pothesis of the kind ordinarily used to "solve" the problem of natural
evil, but by the practical premise (4), viz., "Seek to realize the highest
good."
We have already commented extensively on this alleged impera-
tive. 36 We have argued that, as an imperative, it is a command only that
we seek virtue, let the eschatological chips fall where they may. But
Kant regards the second component of the summum bonum as essential
because he holds an ideal of the rationality of morals. This is described
at the beginning of chapter ii of the Dialectic. Happiness, he says there,
is required in the summum bonum "not merely in the partial eyes of a
person who makes himself his end but even in the judgment of an im-
partial reason, which impartially regards persons in the world as ends-
in-themselves." (Here we meet Kant's version of the English philoso-
phers' "disinterested observer.") "For to be in need of happiness and
also worthy of it and yet not partake of it could not be in accordance
with the complete volition of an omnipotent rational being, if we as-
sume such only for the sake of argument."31
This seems innocent enough; but notice that it completely displaces
premise 4, the practical clause of which made the argument of the sec-
ond Critique a moral argument. 3S The alleged command to seek to
establish the summum bonum now contributes nothing to the concep-
tion of the distribution of happiness in accord with worthiness. The
argument based on this conception of the summum bonum as rational
is a revision of the teleological argument, which is purely theoretical.
It is not, of course, a physicoteleological argument, but only an anal-
ogy to it. It is a teleological argument, based not on the moral com-
mand in question but on the moral phenomenon as requiring a designer
for the adjustment of twO disparate things to each other. This is not
made clear in the Critique of Practical Reason, where the moral argu-
ment is compared with the physicoteleological to the detriment and cen-
sure of the latter (138 [236]). But it dominates the final theological
sections of the Critique of Judgment, and the moral argument of the
38 Kant has not begged the question by this statement, as it might appear. The
conception of God here is not used to establish the concept of the surmnum bonum.
The conception of an impartial observer is the essential one; it follows that if there
were an impartial observer with the requisite power, the SU1mJ'lUm bonum would
be made real.
176 The Postulates of Pure Practical Reason
second Critique had already been assigned to the obscurity of a diffi-
cult footnote by the time Kant wrote the treatise on religion. The ex-
plicit formulation of the new teleological argument is given in the
Metapbysik der Sittel1, in the sample of moral catechism Kant works
out. The passage is as follows:
TFACHER: \Vhen we are conscious of a good and acti\'e will through
which we hold ourselves to be worthy (or at least not unworthy) of hap-
piness, can we base on it the certain hope of partaking of this happiness?
PCPlL: Ko, not on that alone . . . . Our happiness remains only a wish
that cannot be a hope unless some other power is added.
TEACHER: Has reason perhaps grounds in itself to believe in God, i.e., to
assume as real a power that apportions happiness according to desert, or-
dering nature and ruling the world with supreme wisdom~
PUPIL: Yes; for we sec in the works of nature which we can judge such
an extensive and profound wisdom that we cannot explain It except through
the inexpressibly great art of a Creator; and frolll this we have reason to
promise ourselves a not less wise government as respects the moral order,
which is the highest ornament of the world; to promise ourseh'es that if we
do not nuke ourselves umvorthy of happiness through trespassing against
our duty, we may also hope to participate in it.:m
This passage is especially noteworthy. There is no discussion of the
SU11munn bonum in the Metaphysik der Sitte7l. The proof of God's ex-
istence is an argun1ent from design, pure and ~imple. As a theoretical
argument, it "always deserves to be mentioned with respect,"40 but it
is neither theoretically coercive nor independent of the ontological ar-
gument. Yet it is, if my analysis is correct, the hidden sense of the
moral argument, since the practical premise of the moral argument-
the command to seek the whole Sll1ll1ml111 bonum-merely calls atten-
tion to the internal heterogeneity of this concept without placing us
under any obligation to seek the second of the components.
The shift from the practical to the theoretical argument, however,
contributes nothing to the theoretical fruitfulne~s of the concept of
God; it may indeed lessen it. For theoretical arguments, as shown in
the first Critique, could lead, if they were valid, only to the cosmologi-
cal concept of a cause, which is less than what is meant by God, or
they enrich the concept with anthropomorphic predicates, reasoning
by analogy.41 These analogies, however, never lead to the superlative~
39 Metapbysik der Sit ten, VI, 482; d. also Critique of Judgment, § 91 (V, 479
IHz]), where we are told that the teleological argument is a desirable confirmation
of the moral.
40 Critique of Pure Reason, A 623-24 ::::: B 65 I-51.
41/bid., A 817::::: B 845; Critique of Practical Reason, 140 (23 8).
The Postulates of Pure Practical Reason 277
demanded by the concept of God. 4 :.! If the anthropomorphic elements
in the physicoteleological conception are removed, nothing is left of
the concept of God but the mere name.
The moral argument, paradoxical though it may seem, leads to a less
anthropomorphic conception of God than that of natural theology.
For all the predicates essentially attributed to God are predicates which
define merely a rational being endowed with a will (131 n. [22 8 n.1),
and these concepts are not empirical, psychological concepts. We do
not need to consider the peculiar nature of human understanding or
will-that the former is discursive and that the latter is sensuously af-
fected-in the definition of moral personality, but only the canonical
relation of the former to the latter. 43 This relation is the starting point
for the (moral) argument for the existence of God, and this is the only
essential content that the concept of God has. The only conclusions
that can be drawn from the concept concern these two in their inter-
rela!:ions; they are all moral in their import. If we try to "sensualize"
the conception of God, we weaken its moral force by mixing empirical
concepts drawn from human nature with the purely rational concepts
of a moral being in general.
The only valid theology, therefore, is moral theology; God is a con-
cept belonging not to physics (or its extension, metaphysics) but to
morals (138 [236]). Until the moral motive was explored, metaphysics,
based on the study of nature, did not need a rational theology, and no
trace of one is found before moral consideration generated it.H Simi-
larly, the ultimate teleology of the world is moral, not natural. The
final purpose of creation is moral; it is the summum bonum (130 [22 8]).
Men serve and glorify God-in the figurative sense, the only one Kant
will permit-by respect and obedience to his command.
The theological discussion in the second Critique ends here without
developing the most interesting philosophical and ethical conclusions
which are implicit in it. The theoretical interest in the concept of God
is to find an absolute and sufficient ground for the unity of the world,
a cause of all causes, and a purpose of all purposes. If we examine this
45 Critique of Practical Reason, 131 (129). It has quite other premises, from which
it follows much more clearly and convincingly, in Foundations of the Metaphysics
of Morals, 4 2 9 (47).
48 Critique of Judgment, § 86 (V,44 2-43 [293-94]).
47 Ibid., and Uber den Gebrauch teleogischer Prinzipien, VIII, 182 f.
48 Cf. Critique of Pure Reason, A 816 = B 844.
Tbe Postulates of Pw'e Practical Reason 179
philosophy4!J and to show that practical and theoretical reason are fi-
nally one and the same, The system that brings them into unison, how-
ever, exists only for reflective judgment, i.e., as a guiding maxim for
the systematization of experience, not as a determinative principle from
which specific natural a.nd moral consequences can be drawn. If it were
not a merely regulative principle for judgment, we should have the
double absurdity of a theological physics and a theological morality.:iO
that Kant's theology was moral and that of the deists was natural, this connection of
morals and religion, whereby the latter contributes nothing to the content of the
former, is common to both. Thus Kant says (Streit der Fakultiiten, VII, 36) that
there is no material difference (difference in object) between morality and religion,
but only a formal difference; and Tindal characteristically wrote: "Acting accord-
ing to the reason of things considered in themselves [is morality; religion is] acting
according to the same reason considered as a rule of God" (Cbristianity as Old as
Creation; quoted from Leslie Stephen, History of English Thought in the Eight-
eenth Century, I, '44).
58 Perpetual Peace, VIII, 367 n. (Beck, 31 n.): There can be only one true religion
because there is only one true morality. Similarly, Religion within tbe Limits of
Reason Alone, VI, 107 (Greene and Hudson, 98), and Streit der Fakultaten, VII, 36:
there is only one religion, of which Christianity is the plainest (schliclneste) form.
59 Allgemeine Naturgescbicbte und Tbeorie des Hi1'1mle/s, concluding paragraph;
Beobachtungen tiber das Gefiibl des Schon en und Erhabenen, II, 208-<); Traume
eines Geistersehers, II, 332; Einzig 1110glicher Beweisgrund ... , II, 141. A similar
conjunction in Seneca, one of Kant's favorite authors, has been pointed out by
Vaihinger in Ad Heft,·iam 111atrem de comolatione cap. viii (cf. "Ein beriihmtes
Kantwort bei Seneca?" Kant-Studien, II [1898], 491 ....()3); but, as Vaihinger in-
dicated, the Seneca passage is more like those in Kant's precritical writings than
that in the Critique.
282 The Postulates of Pure Practical Reason
with the grand simplicity of this passage. That the two were deeply
connected in Kant's own life of feeling, no doubt having first been
joined by Kant's mother,60 is sufficiently attested by the early biog-
raphers who knew him personally.
But there is a marked progress in this passage, not only stylistically
but also philosophically; it arises from the fact that, when writing the
earlier ones, Kant had not yet succeeded in sharply distinguishing be-
tween the legislations of the two realms. In his earlier works he had
thought of the starry heavens as a possible abode of moral beings of a
higher order than man or as the place of our souls after death. All that
is now transcended in the complementary conceptions of an unmoral
astronomy and a non-naturalistic ethics; all that remains of the older
conception of the relation of the moral law to the natural law is the
Typic. The stark contrast between the two, not some simple harmony
hazarded between them, gives force to their bold contexture. What
had previously been thought of as a synthesis from the standpoint of
the physical object-the evidence of natural teleology, the natural
origin of the moral disposition-he now sees as a synthesis of contrasts
within man's own rational nature. The starry heavens seem sublime
because man first feels reduced by them to impotency, only to rise
above them again when he knows that his rational nature, which
comprehends them, is not subdued but heightened by the magnitude
and power revealed in them. He erroneously attributes a sublimity to
nature which actually belongs only to his own rational being;61 the
sublimity ascribed to nature is a clue to his own superiority to nature,
for we convert "respect for the idea of humanity in our own subject
into respect for the object."62 What the starry heavens awake in us
only indirectly is produced directly by the contemplation of the moral
law, sublime in itself, and of the moral agent who embodies the law;
the humility thus induced in man is itself sublime. 63
There is no tendency remaining now to use natural concepts in the
articulation and elaboration of moral ideas or to confuse two things,
to the detriment of each. But the two things are not set before us
merely as a literary peroration. Even on the last pages Kant extracts
60 In the famous statement on his reverence for his mother (jachmann, Immanuel
Kant geschiJdert in Briefen an einen Freund [1804], neunter Brief), he speaks of her
as having "planted and nourished the seed of the good" and "opened [his] heart to
the impressions of nature," the same combination as in the present apothegm.
61 Critique of /ud?;mem, § 16 (especially V, 25 6 [95]).
62 Ibid., § 27 (V,lS7 [96]).
s31hid., § 18 (V, 264 [103]); Critique of Practical Reason, 117 (213); Beobach-
tung en uber das Ge[iibl des Scbonen und Erhabenen, II, 215.
The Postulates of Pure Practical Reason 283
a lesson of use to moral philosophy by drawing an analogy from the
history of astronomy.
"Admiration and respect," he says, "can indeed excite to inquiry, but
they cannot supply the want of if' (162 [261]). The consequence of
passive admiration and an admixture of human wants in subsequent
astronomical investigation was astrology.64 A like undisciplined elab-
oration of the "noblest attribute of human nature" led, in morals and
religion, to fanaticism and superstition. With the perfection of the
method of scientific investigation in astronomy and physics, there was
brought forth "a clear and henceforth unchangeable insight into the
structure of the world" (163 [26 I]). Perhaps even more important, it
gave a model to philosophy, to set it on the secure path of science. 6s
This method he now recommends in the investigation of morals. "Sci-
ence," in the sense of critically and methodically directed inquiry, is
"the narrow gate that leads to wisdom."
The Critique of Practical Reason is meant to be the key to that
narrow gate.
64 He has drawn a similar analogy in Prolegomena, IV, 366 (Beck, 115): "Critique
stands in the same relation to the common metaphysics of the school as ... astron-
omy to the astrology of the fortune-teller."
65 Cf. Critique of Practical Reason, 108 (2°3); Critique of Pure Reason, B xxiv.
Bibliography
Bibliog;raphy
295
296 Indexes
AKAD£HI£ ABBOTT AKADEHIE ABBOTT
EDITION· TRANSLATION COHHENTAIlY EJ)lnON* TRANSLATION COHHENTAlIY
18z 163. 1 35
38, 166, 178, 18z 221,225
181 191, 2H, 226
195 2Z0
254 46 , 233
119. 130• 13 1 , 137. 52
IS? H,212
90 55,67, 211
1}1 55, 66, 7 8, 81, 12 7.
13 0 128
36 , 13 0 , 13 2 , 1<tH, 17 8 47. 55.69
13S 66, 166
IH 195,218,220
60, 1Z7 66, 120, 141
60, 1%7 100, 107, 115, 120,
60 IH. 228
60 93 18 7 21 9
243 94 18S 162
13 6 , 137 95 189 186
39,137, 13 8 , 139 96 189 190
14 1, 14 2, 149, 273 96 190 179, 190
1%6, 141, 14z, 144, 97 190 179, 190, 19 1
15 2 97 19 1 179
146 98 19 1 190, 19 1, z7 0
15 6 9B 192 116, 191, 231, z70
113, 1%9, 158 99 192 188,270
158, 16z 99 193 30, 188, 191, 192
15 8 , 235 100 193 20 4
157, 161, 162, 235 100 194 188, 191, 205,206
235 101 195 190. z06
16 3, z3S 101 196 206
219, ZZI 102 196 206
217, %21,126, 233 10 3 197 207
211 104 199 25. 166
100 105 200 19<>
100 106 200 19<>
ZZ3 106 101 16
ZZ3, :n5 10 7 202 18 3
117. 221, 126 107 20 3 25, 186. 240
129 108 20 3 211, 283
:n6 109 204 z42. 246 , 247
126 109 20S 24 1 , 24 2 , 243
22[ 110 20 5 242, 248
35.9 1 110 206 268,275
203. :u6, 221, 229 III 20 7 246
223 112 20S 246
230 113 209 246, 26z
50, 1%5, 226, 228 114 210 244, 245
280 115 211 223, 224, 248, 269
2B,28o 115 212 2240 247
138,228 I16 2IZ 101, 2%4, 229, 23 0 ,
I2S. 135, 211 247, 27 1
163 u6 21) 220, :ni, 2-47
Indexes 2.97
AItADEIllE AaJlOTT A~DE"IE A.IIOTT
EDITION- TaANSLATION CO.... ENTARY EDlnON- TaANStATION COIlJOIIYARY
Midway Reprints
Midway Reprints are limited-edition, paper-covered books. The purpose
of this series is to keep in print titles for which there is a steady but
small demand-out-of-print titles, as well as books which might oth-
erwise have been declared out of print. In order to keep the prices as
low as possible, Midway Reprints will not be issued in clothbound
editions.