303 8 5MB
English Pages 233 Year 1965
Studies in The Philosophy o f Kant
Lewis White Beck BURBANK
PROFESSOR O F
I N T E L L E C T U A L AN D
MORAL
P H I L O S O P H Y IN T H E U N IV ER SIT Y O F ROCHESTER
TH E
BO B BS -M ER R ILL
COMPANY,
INC.
A Subsidiary of Howard W. Sams & Co., Inc. p u b l i s h e r s Indianapolis New York Kansas City •
•
Copyright © 1965 by T h e Bobbs-Merrill Com pany, Inc. Library of Congress Catalog Card N u m b er 64-66063 Printed in the United States of America FIRST EDITION
FOREWORD
During the past twenty years much of my writing has dealt with the philosophy of Immanuel Kant. Not all of it has been in English; the papers have been scattered among several journals, and they have not appeared in any logical sequence. It is therefore gratifying to me to bring many of them together into a single volume, one language, and a reasonable order. Even so, several articles have been left in the decent obscurity of the learned journals. T h e papers included here do not, of course, survey all parts of Kant’s philosophy, but by disregarding chronology I have brought related papers together. Thus the first is introductory; essays III through VII deal with the problems of the synthetic a priori; VI through X are Auseinandersetzungen with leading contemporary critics of Kant; and X through X III concern Kant’s moral phi losophy. T h e only purely historical paper is the second. Except for translating two of the papers, correcting a few obvious errors, and revising many of the bibliographical references, I have made no changes in the papers since they were published. T h e reader may discern, as I did in preparing this volume, some shifts in empha sis and even changes in point of view from the earlier to the later essays. T o have tried to gloss over them or to resolve the differences would have meant an undertaking that might have become still another book. I have therefore let them stand, as evidence of how hard it has been to reach settled conclusions on some important ques tions in philosophy and in the interpretation of Kant. In four in stances, however, I have added brief notes to warn others against errors I now know I made. T h e editors of the following journals have permitted me to reprint papers which first appeared in their journals: Annales de philosophie
vi
Foreword
politique, Kant-Studien, M i n d , T h e Philosophical F orum , T h e Philosophical Review, Philosophy and Phenom enological Research. Mr. Arne Unhjem and the editors of T h e Philosophical Forum have
generously allowed me to reprint Mr. Unhjem’s translation of Kant’s letter to Herz. I wish to express my thanks to each editor and to Mr. Unhjem. I owe special thanks to the editor of T h e Open Court Pub lishing Company, Dr. Eugene Freeman, and to Professor Paul Arthur Schilpp, the editor of the series, T h e Library of Living Philosophy, for permission to print my essay, “ Lewis’ Kantianism,” prior to its publication in their volume, T h e Philosophy of C. I. Lewis. T h e Open Court Publishing Company retains all rights to this essay, on which it holds the copyright. Over many years I have used divers editions of Kant’s works, and in my publications I have had to cite sometimes one, sometimes an other. T h e edition I prefer, however, is that of the Prussian Academy of Sciences (Berlin, 1900-1955) in twenty-three volumes. I have, therefore, in this book converted all references to accord with the Academy edition. It would not have been possible to do this without the assistance of the great new index to this edition now being pre pared by my good friends in the University of Bonn— a mammoth undertaking which, when complete, will render all other editions of Kant’s works obsolete. In many citations, I have also given references to standard English translations, identified by the name of the trans lator. Citations to the Critique of Pure Reason are given in the normal manner, viz., “A ” refers to the first edition and “ B ” refers to the second. T h e translation of these passages is usually taken from that of Norman Kemp Smith. References to my own translations of Kant are always to the editions in the Library of Liberal Arts, except where it is noted that I refer to works in the Chicago edition which have not been included in the Liberal Arts series. L E W I S
Day of N ational M o u r n in g N o v em b er 25, i ç 63
W H I T E
B E C K
CONTENTS
i K a n t ’s T h e o r e t i c a l a n d Practical P h i l o s o p h y 3
•♦
n
K a n t ’s L e t t e r to M a r c u s H e r z , F e b r u a r y 2 1 , i y y i 54
«• • m
K a n t ’s T h eo r y of D efin itio n 61
iv Can K a n t ’s Synthetic Jud gm en ts B e M a d e A n a lytic? 74
v
O n t h e M e t a - S e m a n t ic s o f th e P r o b l e m o f th e S y n t h e t i c A P r i o r i 92
vi R e m a r k s on t h e D i s t i n c t i o n B e t w e e n A n a l y t i c a n d S y n t h e t i c 99
v ii L e w is’ Kantianism 108
N i c o l a i H a r t m a n n ’s C r it ic is m o f K a n t ’s T h e o r y o f K n o w l e d g e
125 ix O n H e n r y M a r g e n a u ’s K a n t i a n i s m 158
X Sir D a v i d R o s s o n D u t y a n d P u r p o s e in K a n t
165 xi A p o d i c t i c I m p e r a t iv e s
177 xii T h e F a ct o f R e a s o n : A n Essay o n J u stific a tio n in E t h i c s 200
xm K a n t ’s T w o C o n c e p t i o n s o f t h e W i l l in T h e i r P olitical C o n te x t
215 A PPEN D IX
K a n t ’s L e t t e r to M a r c u s H e r z , F e b r u a r y 2 1 } 1 J J 2 T r a n s la t io n by A r n e U n h j e m 230
IN D EX
I
KANT’S THEORETICAL AND PRACTICAL PHILOSOPHY K A N T A N D H IS P R E D E C E S S O R S
There is a saying among philosophers, “ You can philosophize with Kant or against Kant, but you cannot philosophize without him.” Much truth lies in this adage, even though few thinkers today would call themselves Kantians without much redefinition and qualifica tion. Nor is its truth greatly diminished by the fact that idealism, as it developed from Kant’s philosophy, has been on the defensive for nearly a century. But, inasmuch as Kant saw more clearly than any one else the ultimate problem of modern thought, and devoted to its solution the greatest philosophical genius of modern times, any thinker who ignores Kant proceeds at unnecessary risk. T h e ultimate issue which Kant faced consists in the logical in compatibility between the objective and subjective conditions of scientific knowledge. It is the disharmony between the object of science and the human ends science is made to serve. In the Renais sance, after Galilei, Descartes, and Newton had banished purpose from nature, nature came to be seen as a vast mechanism. With the replacement of Aristotelian ideas by mechanistic conceptions, science began to achieve unprecedented control over nature. A similar change of viewpoint in Hobbes, Spinoza, and Harvey with regard to man’s own body and mind opened the way for analogous advances in the control of man. But control for what? It is man who develops science and who through it controls nature for his own purposes. There lies the paradox: man is understood as a machine, but the use of his knowl edge of himself and of the external world is thoroughly purposive. t h i s e s s a y first a p p e a r e d as the I n t r o d u c t io n to m y v o l u m e of translations,
K a n t ’s C ritique of Practical Reason and other Writings in M oral Philosophy ( C h icag o : U n iv e r s ity o f C h i c a g o Press, 1949).
3
4
Studies in the Philosophy of Kant
T he problem is more urgent than any other in modern philosophy because the two incompatible convictions— the idea of the world as a vast impersonal order and the idea ^of knowledge as power— are fundamental to our world view and equally deep-rooted. T o the ex tent that Western civilization is based on science, it rests on a paradox. Philosophers before Kant who were aware of this problem at tempted to solve it in a variety of ways, and many who were not explicitly conscious of its full implications nevertheless developed philosophies which even now sometimes serve as frameworks for at tempted solutions. These ventures involved one of four strategies: 1) The problem was denied by exempting man from the laws of nature through ad hoc hypotheses (Descartes, many orthodox Christian philosophers). 2) T he problem was declared irresolvable and transferred to a higher court of faith (Malebranche, sometimes Descartes, and many orthodox philosophers). 3) T he problem was declared illusory because purpose is not ultimate even in man (Spinoza and H obbes). 4) T h e problem was declared illusory because mechanism is not ultimate even in nature (Leibniz and Berkeley). All these strategies have one failing in common. Each allays the conflict only by weakening one or both of the contending forces. None of them accepts with “ natural piety” the competing claims of man as knower and agent and of nature (including man) as known and mechanical. Kant’s consummate greatness as a thinker is nowhere shown more indubitably than in his acceptance of both knower and known as facts not to be compromised. Though in his philosophy there are elements of each of these four inadequate attempts, they are inte grated into a new philosophy which does justice to their antagonistic demands. In the eighteenth century, the Age of Enlightenment, the main stream of thought was guided by a faith in human reason. It was intellectualistic in its attitude toward revelation and tradition and skeptical of things irrational; it was little interested in the past and exceedingly optimistic about the future. Kant was in many respects an exemplary philosopher of the En lightenment. His What Is Enlightenment? is a document of much
Studies in the P h ilo s o p h y of K a n t
5
diagnostic value to the historian of ideas. Yet he was also the nemesis of the Enlightenment as a historical epoch. More than any other philosopher, he placed limits on knowledge without falling into the irrationalism of the forerunners of Romanticism. He exposed the superficiality of the humanistic and intellectualistic optimism of the time without becoming an apologist for the past. Together with Hume and Rousseau, he subjected current ideas to a searching examination and found them wanting. Hume’s demon stration of the nonintellectual foundations of science and Rousseau’s nullification of contemporary institutions prepared the way for fundamental changes. These two critics, however, did not construc tively replace what they had rejected. Hume was finally left with only a contemplative skepticism, which Kant turned into a justification of science; Rousseau prepared men’s minds for the Revolution, but it was Kant’s deepening of Rousseau’s criticism of law imposed from above that gave philosophical dignity to liberté , égalité, fraternité.1 Indeed, George Herbert Mead, in his Movements of T h o u g h t in the N in e te e n th Century, gave to Kant the title usually reserved for Rousseau, “ the philosopher of the Revolution.” 2 This dual relationship of Kant’s thinking to the Enlightenment can best be accounted for by his two dominant interests— his interest in natural science and his religious allegiance. More than any other philosopher of the period except perhaps Leibniz, Kant was attentive to the results of the scientific explora tion of nature. Even when placing restrictions on science, he seems always to have thought as a scientist. His earliest works were purely 1 Kant formulates these ideals as liberty, equality, and independence. See On the Saying, “ T h a t May B e T r u e in Theory B u t It Does N o t H o l d in Practice,” II (Academy edn., VIII, 290). See also W illiam Hastie, K a n t’s Principles of Politics (Edinburgh: Clark, 1891), p. 35. 2 K a n t’s attitude toward the French and American revolutions is complex, be cause his rejection of violent revolution in general (see Hastie, pp. 50-56) went with an enthusiastic approval of the ends sought in these revolutions. See Strife of Faculties, Part II, §6 (trans. Robert E. Anchor, in K ant on History, “ Library of Liberal Arts,” No. 162 [New York: Liberal Arts Press, 1963], p. 144). His position with regard to the French Revolution has been pictured in detail by Karl V o r länder, “ Kants Stellung zur französischen R evo lu tio n ,” Philosophische A b h a n d lungen H erm a n n C o h en . . . dargebracht (Berlin: B. Cassirer, 1912), pp. 246-81; and by Paul Schrecker, “ Kant et la Révolution Française,” R e v u e p h ilo so p h iq u e, X X V I I I (1939), 394-425. Schrecker says: “ T h e Revolution and even more the echo it set up in the world are to the Kantian ethics what the discovery of the circulation of the blood is to the mechanism of Descartes: a confirmation, as it were, an experimental confirmation, of the fundamental principles of the theory.”
Studies in the Philosophy of Kant
6
scientific in character, and the number of times he compares his procedure to that of the scientist is remarkable. T h e Critique of Pure Reason is a criticism of the metaphysical claims of science, but it is also a defense of science against its internal enemies. T he counterweight to science in Kant’s thought is religion. Pietism, an anticlerical movement founded by Spener in the seventeenth century, emphasized the religious and moral responsibility of the individual, the directness of his contact with God, and the impor tance of good works and quiet faith over ritual and dogma. T h is religious attitude was instilled into young Kant by his Pietistic parents. But at the same time he learned to dislike the externalized type of Pietism with which he became painfully acquainted in school. Much of his later religious and ethical thought can be under stood as a revolt against the externalized and formal Pietism of the Collegium Fridericianum as well as a defense of the sturdy and un pretentious devotion of his family. For a proper appreciation of Kant’s historical position it is neces sary to understand his relation not only to these dominant cultural factors but also to four previous philosophies— the rationalism of the Leibniz-Wolff school, German “ popular philosophy,” British psychological ethics, and the unclassifiable philosophy of Rousseau.3 Kant’s philosophical training was that of a Wolffian. Christian Wolff, Professor of Philosophy in the University of Halle, had de veloped the rationalism of Leibniz into an elaborate and compre hensive system of ideas. Although Kant was almost never an encyclo pedic systematizer in Wolff’s manner, his formalistic emphasis is undoubtedly in part attributable to his training. Much of his termi nology is derived from textbooks of the Wolffian school, regularly used by him for his own lectures. His marginalia to the textbooks of Alexander Baumgarten, a leading Wolffian, are important sources of information on Kant’s development. However, the substance of Kant’s philosophy differs widely from that of Wolff. T he Inquiry of 1764 renounces the Wolffian method of synthesis, and the Dissertation of 1770 restricts the scope of meta physics as conceived by Wolff. It is this kind of metaphysics which T h e extremely important relationship between Hume and Kant will be dis cussed in the following section. T h e relation to Rousseau has been brilliantly treated by Ernst Cassirer, Rousseau, Kant, Goethe, trans. J . Gutmann, P. O. Kristeller, and J . H. Randall, Jr . (Princeton: Princeton University Press, 1945). 3
Studies in the P h ilo s o p h y of K a n t
7
is repudiated in the first Critique. In ethics Kant rejects the empty formalism, the perfectionistic ideal, and the utilitarian basis of Wolff’s system. T h e followers of Wolff, the so-called “ academic philosophers,” were opposed by the “ popular philosophers” who rejected the logical rigor of the academic philosophy. Though some of them, influenced by the “ encyclopedic” tradition, retained elements of rationalistic metaphysics, they generally preferred appeals to feeling, common sense, and “ sound human reason” as criteria of truth. Through their eclecticism and their uncritical enthusiasm for progress and enlight enment, they gained acclaim and popular following. Windelband correctly remarks that the emptier and more superficial their meta physics became, the greater the role they ascribed to utility; and, as a result, they fell into “ the most jejune philistinism and sensible, prosaic commonplace.” 4 Kant, especially in the second part of the Foundations and in What Is Orientation in Thinking,? opposed this group even more vigorously than the Wolffians did. British psychological moralists appealed so greatly to Kant that he has often been regarded as their disciple.5 But his early acceptance of their theories was not that of a disciple, and the changes he intro duced were of central significance. His later rejection of their theory of moral sense and his radically new orientation to moral feeling are apparent throughout the ethical works of the critical period. Kant’s relation to Shaftesbury and Hutcheson is somewhat like that of Socrates to the Sophists. Both the British moralists and Kant based their ethics on human nature. But Kant, like Socrates, searched for the essential character of man’s nature, a universal law determining its particular manifestations.6 In the announcement of his lectures for the winter semester of 1765-66, he says this explicitly: I shall m ak e distinct the m e t h o d b y w h ic h m a n m ust be studied, not just m an as he is distorted by the variable form w h ich his chance co n d itio n
4 W ilh elm W indelband, A History of Philosophy, trans. James Hayden T u fts (2nd edn.; New York: T h e Macmillan Co., 1919), p. 507. 5 As late as 1771, his friend and former pupil, Marcus Herz, wrote that Adam Smith was said to be K a n t’s “ favorite” (Academy edn., X, 121). T h e degree to which Kant was a disciple has, however, been exaggerated. For a critical evaluation of the relationship see Paul A rth u r Schilpp, K an t’s Precritical Ethics (Chicago: N o rth western University, 1938), and below, pp. 45-47. 6 Later, in the Critique of Pure Reason (A 314-15=1* 371-72), he expresses his admiration for this feature of the Socratic-Platonic philosophy.
8
Studies in the Philosophy of Kant impresses upon him and as he has almost always been misjudged even by philosophers,7 but rather the abiding nature of man and its unique posi* tion in creation. . . .8 >.
This “ abiding nature of man,” differently conceived, becomes the central topic in Kant’s later ethical works. T h e philosopher who had the greatest influence on K ant’s ethics was undoubtedly Rousseau. Kant’s admiration for Rousseau is most clearly expressed in unpublished fragments in which he speaks of Rousseau’s “ noble sweep of genius” and of the beauty of the style as so disturbing that he has to read him a long time before he can be rea sonable in his approach.9 According to a well-known anecdote, Kant missed his customary walk on the day ILmile arrived. Although the published writings on moral philosophy mention Rousseau only a few times, and in the works in the broader field of social philosophy explicit references to him are generally somewhat critical, Rous seau’s influence on Kant is obvious. Hendel finds the dominant motifs of Rousseau’s philosophy and life in the “ ideas of obligation, contract, equality, freedom.” 10 These are likewise central in Kant. For both, the social contract is not a his torical fact but a principle of justification, a political postulate.11 Freedom for both is not just political freedom but a symptom of reason’s dominance and, as such, inseparable from moral obligation. Rousseau describes immoral action as a violation of the contract by which the individual is bound to the whole. Immoral action re stricts equality by partiality, and it is possible because reason does not free man from the importunities of the senses. T h e will which en genders moral actions must be independent of personal contin gencies. This, for Kant, is the good will, and it is clearly anticipated in Rousseau’s general will. 7 T h e allusion is clearly to Shaftesbury and Hutcheson, whose views he has just previously said are incomplete. 8 Academy edn., II, 3 1 1 . 9 Academy edn., X X , 30. 10 Charles William Hendel, Jean-Jacques Rousseau: Moralist (2 vols.; London: Oxford University Press, 1934); reprinted in one vol. in “ Library of Liberal Arts,” No. 96 (New York: Liberal Arts Press, 1963), II, 165. 11 See On the Saying, “ That May Be True in Theory But It Does Not H old in Practice,” II (Academy edn., V III, 297). See also Hastie, Kant’s Principles of Politics, P- 49-
Studies in the P h i l o s o p h y of K a n t
9
These two analogous concepts reveal at the same time a marked difference in point of view. For Rousseau, the general will is a politi cal ideal which is to serve as a check upon the will of the majority. This ideal underlies the original contract itself, and from it all posi tive laws derive their justification. In its capacity as the ultimate legal norm, the general will corresponds to “ the right” in Kant’s philos ophy, to which all positive laws must be accommodated. But the good will has a far wider competence than the general will, and this wider competence befits its higher metaphysical status as a fact of pure reason. It is the direct evidence for and the sole condition of morality. Its expression is the categorical imperative and (in politics) the criterion of publicity.12 Each of these points of agreement between Rousseau and Kant involves both the universality neglected by British moralists and the concreteness lacking in Wolff’s universals. Rousseau sought to discover man’s nature by historical analysis. As Kant says, he did not so much wish that man should return to the state of nature as that, from his present position, he should look back upon his natural condition as a means of discovering in himself the universal— the essence of what he is, apart from the various distortions introduced by society.13 In this way Rousseau succeeded at the precise point where Kant had found Hutcheson and Shaftesbury wanting: A m o n g the m u l t it u d e o f forms assumed b y man, R ousseau first dis covered
m a n ’s d e e p ly
hidden
nature
and
the
con cealed
law
by
the
o b serva tion o f w h ic h p r o v id e n c e is j u s t i f i e d . . . . G o d is justified b y N e w ton a n d Rousseau, a n d more than ever is P o p e ’s thesis true.14
Kant’s indebtedness to Rousseau is best stated in a fragment probably dating from the ’sixties: B y in c lin a tio n I am an inquirer. I feel a c o n su m in g thirst for k n o w l edge, the unrest w h ic h goes w ith the desire to progress in it, a n d satisfac tion at every a d v a n c e in it. T h e r e was a time w h en I b e lie v e d this consti tu ted
the h o n o r o f h u m a n ity , a n d I despised the peop le, w h o k n o w
n o th in g . R o u s se a u corrected m e in this. T h i s b i n d i n g p r eju d ice disap peared. I learned to h o n o r m an, a n d I w o u l d find m yself more useless than
12 See Perpetual Peace, A p p e n d ix II. 13 See Anthropology, II, E (Academy edn., VII, 326). 14 Academ y edn., X X , 58-59.
Studies in the P h ilo s o p h y o f K a n t
IO
the c o m m o n laborer if I d i d n ot b elieve that this a ttitu d e of m in e [as an investigator] can give a w o rth to all others in establishing the rights o f m a n k i n d . 15
In this fragment Kant reflects Rousseau’s conviction of the supe riority of uncorrupted natural feeling over vain pride of intellect, his pessimism concerning progress through enlightenment, and his faith in democracy founded upon moral egalitarianism. It also fore shadows Kant’s doctrines of the limits of human reason. T h e idea of limits of human reason is to be interpreted morally as the primacy of practical reason, a possession of all men and not just of the enlight ened few. From the concept of the limits of human reason— fully developed only twenty years later— flows the philosophical justification for the other views Kant shared with Rousseau. First among these is the moral argument for the existence of God anticipated in £m ile. In order to voice his protest against contemporary naturalism, Rous seau, lacking speculative power, had to fall back on personal faith. Kant, by formulating and defending a metaphysics that was both a priori and practical, developed Rousseau’s insight into an in dispensable part of his own more critical philosophy. In the 1760’s, when Kant was studying Hutcheson and Rousseau, he was working on what he intended to be his definitive treatise on ethics, which, with characteristic optimism in matters of authorship, he expected soon to finish. From the fragments that have come down to us we might reasonably suppose that it would apply the method of the Inquiry, searching out by analysis the hidden nature of man of which he speaks in the announcement of his lectures for 1765-66. We might expect that the projected work would emphasize the indemonstrability of the ultimate principles of ethics and would sup port them by arguments not unlike those found in Hutcheson and Rousseau. Or Kant might have developed a speculative metaphysics, also in the manner of the Inquiry, which would provide a context for the more empirical ethics. But something unexpected happened. Kant read Hume, and, before he could go forward with his ethical works, ultimate ques15 Academy edn., X X , 44. T h i s revealing passage has been subjected to an e x haustive analysis from the existential standpoint by G erhart Krüger, Philosophie und Moral in der Kantischen Kritik (Tübingen: Mohr, 1931), pp. 60 ff.
Studies in the P h i lo s o p h y of K a n t
tions had to be answered. T h e competence of reason had been radically questioned, and, before the mind could enjoy the luxury of metaphysics, or the security of rational certainty in science or of moral certainty in religion, reason’s authority in experience and science’ relation to the spirit had to be determined. Only upon a basis so secured would it be possible to found an ethics more than merely edifying. T h e result of this fundamental investigation was the Critique of Pure Reason. THE
E T H IC A L IM P O R T OF T H E
C R IT IQ U E OF PU RE REASON
T h e explicit task of Kant’s first Critique was to answer the ques tion— restated in more exact terms— which had occupied him in the Inquiry of 1764: “ How are synthetic judgments a priori possible?” Since all rational knowledge consists of a priori synthetic judgments, failure to answer this question could not but affect every depart ment of thought. A synthetic judgment is one whose predicate is not contained in the concept of its subject. It synthesizes diverse elements into one proposition. For instance, “ This table is of oak” is a synthetic judg ment, while “ This table is a piece of furniture” is analytic, because we find the predicate of the latter by merely analyzing the concept of table. An a priori16 judgment is one which applies to all possible ex perience of a relevant kind without being derived from any par ticular experience. If a judgment is derived from particular experi ences, i.e., if it is a posteriori, we cannot know that it is universally and necessarily true. Now, since mathematics and natural science make statements which we accept as being universally true, their validity cannot be derived from experiences, however often repeated. All this was well known to Kant’s predecessors in the rationalistic school. But Kant discovered that the basic propositions in these fields of knowledge were both synthetic— going beyond the subject — and a priori— requiring no experience to amplify the concept of the subject. Before Kant, “ a priori synthetic judgment” would 16 “ A p riori” was originally used in formal logic to denote the evidence for a deduction, wherein the conclusion is known from prior grounds. A more critical account of K a n t’s concept is given below, pp. 102-4.
12
Studies in the Philosophy of Kant
have been considered a contradictio in adjecto. If a judgment was synthetic, experience seemed necessary to lead us beyond the con cept of its subject. In order to find oi^t what predicate not included actually qualifies it, the object denoted by the subject of the proposi tion would have to be given in perception. T hat is to say, a synthetic proposition would have to be an empirical, an a posteriori, proposi tion. On the other hand, if a judgment was a priori, it was thought that it had to be analytic, for only by analysis of a given concept could a judgment be made which did not require experience. T hus the rationalistic philosophers who emphasized reason’s faculty of reaching true conclusions without sense experience were forced to the conclusion that all final knowledge is analytic. Kant, originally trained in the rationalistic philosophy of Wolff and himself an able practicing mathematician, only gradually per ceived the inadequacies of this view. In the Inquiry of 1764 he saw the divergence between mathematical and metaphysical knowledge and made the first hesitant step toward coming to terms with empiricism. In the Dissertation of 1770 he discovered that mathe matics— the paradigm of rationalistic certainty— was concerned only with appearances. This was a radical change of view which fore shadowed the further retrenchment of speculation in the Critique of Pure Reason eleven years later. Probably it was only after Kant had begun what might be called his own “ palace revolution” within rationalism that he came to know of Hume’s work. Hume awoke him from his “ dogmatic slumber” — Kant’s term for his early faith in the power of reason to give meta physical knowledge. Yet even if he had not accepted Hume’s argu ment against the possibility of metaphysics, it is probable that Hume’s strictures on natural science would have aroused him. For both rationalism and empiricism, if carried to the ultimate, deny the necessity and universality of natural science, and Kant’s convic tion of the certainty of Newtonian mechanics was too deep to be shaken by any negative conclusions drawn from speculations con cerning the human mind. Kant asked how synthetic judgments a priori are possible, but that they are possible he never seems to have questioned. Instead, he revised the principles of both rationalism and empiricism in the light of a hard fact to which these two schools
Studies in the P h i l o s o p h y of K a n t
13
had not given due regard— the validity and certainty of the synthetic judgments of geometry and physics. This revision Kant made in his well-known “ Copernican Revolu tion.” In a famous passage in the Preface to the second edition of the Critique of Pure Reason, Kant compares his new epistemology to the Copernican hypothesis. T h e predecessors of Copernicus had had difficulty in explaining the apparent motions of the planets on the supposition that they all revolved around the earth. Before Kant, it was similarly impossible in philosophy to explain how there could be a priori knowledge of things on the assumption that knowl edge is passive conformity to the object. “ Failing of satisfactory progress in explaining the movements of the heavenly bodies on the supposition that they all revolved round the spectator,” Kant says, “ he [Copernicus] tried whether he might not have better success if he made the spectator to revolve and the stars to remain at rest.” 17 Similarly, if the phenomenal characteristics of objects are explained in terms of the behavior of the knowing mind, it is possible to see how knowledge of them can be a priori, for, as objects of knowledge, they must conform to the structure and activity of the knowing mind which make knowledge possible. One advantage claimed for the rationalistic method lay in the fact that, whereas empiricism had to stop at the limits of sense experience, rationalism was perhaps even more fruitful beyond these boundaries. Under the influence of Hume’s criticism of causation, Kant discovered that the rationalistic method accomplished too much: it not only proved theses which transcended possible experi ence; with equal cogency it proved their antitheses too. T h e dis covery of these antinomies (i.e., conflicts of principles) was the true beginning of Kant’s new philosophical development. T h e antinomies arise from the inadequacy of reason to meet its own demands on 17 Critique of Pure Reason, trans. Norman Kem p Smith (2nd edn.; London: Macmillan & Co., 1933), B xvi. It has appeared to some that K a n t’s procedure could be more aptly compared to that of Ptolemy, since Kant placed the spectator in the center of the epistemological world. Kant, however, was interested in explaining what we may by analogy call the “ apparent motions” of the object, and in doing this it is important, as Copernicus showed, to consider the real nature (“ m otion”) of the spectator. Hence the Copernican Revolution itself implies the distinction between appearance and reality, just as Copernican astronomy distinguished real from apparent motion. See H. J. Paton, K an t’s Metaphysic of Experience (2 vols.; New York: T h e Macmillan Co., 1936), I, 75-76.
14
Studies in the P h i l o s o p h y of K a n t
thought. Let us see how these antinomies result from the diversity of our cognitive faculties and what they imply concerning the limits of knowledge. T h e mind has three cognitive faculties: sensibility, understanding, and reason. Sensibility supplies empirical content or sensation. U n derstanding links sensations into perceptions and these into objects of knowledge and series of objective events. Sensibility and under standing supply us with a constellation of phenomena under laws; the experience formed by them is empirical nature. Nature com prises series of phenomena in space and time, and these series can and indeed must be infinitely extendable, for every phenomenon has other phenomena as its conditions. There is, in sensibility and under standing and in the world of experience they underlie, no way to arrest this infinite regress of conditions. Reason, however, demands a totality of these conditions, for otherwise all is contingent. This demand cannot be met by a faculty which, like understanding, merely seeks out proximate causes, and proximate causes of proxi mate causes, and so on. Extending the phenomenal series infinitely (in the antitheses of the four antinomies) is met by the demonstra tion (in the theses) that the extension under the rules of under standing is inadequate to reason’s needs; and, if reason is to reach satisfaction, it must speculate beyond any possible experience to find the unconditioned. In this speculation, reason negates the restriction of the categories to the world of possible experience. A category and its schema are constitutive of nature, for nature is simply phenomena under the laws of understanding; but the object of speculation cannot be in time, and therefore the schemata of the categories do not apply to it. It cannot be in time, for then it would be conditioned in an infinite regress; but, as it is not in time, the categories cannot constitute it. In spite of this, the categories control our thought about it and are consequently in this function called regulative ideas. If the difference between the constitutive categories and the regulative ideas is over looked, the antinomies cannot be resolved. T h e rationalistic philos ophers had not drawn this distinction, and therefore their speculative metaphysics was ripe for the Kantian critique, which really com mences with the demonstration of the antinomies. While the Copernican Revolution alone might have strengthened
Studies in the P h ilo s o p h y of K a n t
15
Kant’s predisposition toward rationalism, discovery of the antinomies showed this philosophy to be inadequate at its crucial point. T h e Copernican Revolution is a confession that the object is not the determining factor in knowledge; the natural dialectic of reason is proof that speculative theory is not the determining feature in meta physics. If the object determined knowledge, a priori knowledge would be impossible, and scientific knowledge would be out of the question. If theoretical reason were the only faculty, a true and ade quate metaphysics would be impossible, and the supersensuous world would be a mere extension of the world of appearances. And, if this were the true state of affairs, morality would be impossible. For morality, Kant argues, makes demands on men as free agents— demands at variance with the mechanistic world picture of science. It makes demands conflicting a fortiori with a supersensuous world conceived of as a mere extension of nature. T h e antinomies show that this extension is impossible, and for that reason Kant regarded them as the “ most fortunate perplexity” 18 into which pure reason could ever fall. T h e antinomies strictly limit theoretical reason to the world of space and time, nullifying all speculative flights from the results of science and all attempts to use scientific method in speculation beyond the limits of sense. But their resolution permits an altogether different use of reason. T h e occurrence of the antinomies is indica tive of reason’s broader competence as a faculty not exclusively devoted to cognition. This is very clear in the third antinomy, the one most directly con cerned with ethics. This antinomy arises from the conflict in the idea of causality and is resolved by the distinction between the world of appearances and that of things as they are in themselves. Kant made the discovery that the thesis, which asserts the reality of non mechanical causes, and its antithesis, which asserts the sufficiency of natural causality, may each be true if their respective scopes are sharply distinguished. T h e field of application of each is defined by the nature of the argument supporting it, and neither can be validly employed beyond 18 Critique of Practical Reason, Bk. II, ch. 1 (Academy edn., V, 107; trans. L. W . Beck, “ Library of Liberal Arts,” No. 52 [New York: Liberal Arts Press, 1956], p.
111).
Studies in the Philosophy of Kant
i6
the area to which the respective proofs extend. T h e proof of the thesis presents the claim of reason, which requires a sufficient cause of every phenomenon. This sufficient cause cannot be found within phenomena, because a phenomenal cause is the effect of prior events and hence not by itself a sufficient explanation of subsequent phe nomena. T he proof of the antithesis, on the other hand, represents the interest of the understanding in applying the law of causality to a series of events in time. T h e argument shows that the assumption of a free cause among phenomena would interrupt the continuity required by natural law. There is no contradiction, however, when the thesis is applied to the relationship between noumena (things-inthemselves) and phenomena (appearances) and the antithesis to relations among phenomena. These separate and distinct but com patible applications are all that is legitimized by these two proofs. T he solution to the third antinomy, therefore, is achieved through a distinction between the world of appearance and the world of supersensuous reality. This dualism is the necessary presupposition of Kant’s ethical theory. Without it, science would be the only occu pation of reason. With it, science is limited in two respects: a bound ary is fixed beyond which scientific knowledge cannot aspire and the possibility is established that natural law may not be the only form of causality. First, by rejecting the presumptions of theoretical reason to a scientific metaphysics, a practical extension of pure reason is made possible. “ I have therefore found it necessary,” Kant says, “ to deny knowledge, in order to make room for faith. T h e dogmatism of metaphysics . . . is the source of all that unbelief, always very dog matic, which wars against morality.” 19 T he limitation he places upon theory may be taken negatively or affirmatively, with respect to what it forbids or what it permits. Nega tively, it means that reason is incapable of knowledge of God, free dom, and immortality. Regarded in isolation, this stricture has been the occasion of positivistic and fictionalistic interpretations of Kant — the belief that reason’s true vocation is found in knowledge and that whatever is assumed in the light of practical demands is fictional and subjective. Yet such interpretations are in complete discord both 19
Critique of Pure Reason, B xxx, trans. Kemp Smith.
Studies in the P h i l o s o p h y of K a n t
17
with the general character of Kant’s philosophy and with his person ality as a man of deep faith. In his ethical and religious works, however, the affirmative inter pretation— the emphasis on “ making room for faith” instead of on “ denying knowledge”— is of more central importance. Morality, even in the Critique of Pure Reason, is a given fact whose necessity ^and universality require legitimation just as the apriority of science does. T h e denial of the possibility of knowledge is not a reluctant admission of reason’s impotence but an expression of the moral in junction that we ought not to know.20 Knowledge of the intelligible world would destroy the possibility of free actions and of the faith that moral demands can be met. Thus instead of providing us in a “ stepmotherly fashion with a faculty needed for our end,” “ inscru table wisdom is not less worthy of veneration in respect to what it denies us than in what it has granted to us,” for only through our theoretical ignorance of our destiny is there proper scope for our practical faith.21 T h e second way in which science is limited is by defending a causality not under natural law. T h e resolution of the antinomy of causality is necessary for Kant’s ethics, for the antithesis, which implies the impossibility of morality, is the inevitable consequence to which theoretical understanding, unrestrained by criticism, leads us. Indeed, Kant says that, if the antinomy could not be resolved, it would be morality and not nature that we should have to surrender.22 'tWith the demonstration that natural causality is not the only think a b le kind of causality, Kant voided the chief argument against free dom; but this refutation is not by itself enough to establish freedom. T h e reality of freedom must be shown by indicating it as a necessary a priori condition of a type of experience. In this case the experience to be considered is morality with its universal and necessary injunc tions. Morality must be defended against subversive empiricism in the same way mathematics and physics were defended. This mode of arguing for freedom as the ratio essendi of morality is in its turn a Copernican Revolution in ethics. There is a perfect 20 See R ichard Kroner, Kants W eltanschauung (Tübingen: Mohr, 1914), p. 20; trans. John E. Smith (Chicago: University of Chicago Press, 1956), p. 24. 21 See C ritiq u e of Practical Reason, Academ y edn., V, 146-48. 22 C ritiqu e of P u re Reason, B x x ix , A 537=B 565.
Studies in the Philosophy of Kant
i8
parallelism between the mode of argument and the conclusions in the theoretical and practical phases of Kant’s philosophy. In both, reason appears as the lawgiver and as bound by the laws which it gives. Kant clearly compares these two legislative functions: T h e legislation of human reason (philosophy) has two objects, nature and freedom, and therefore contains not only the law of nature, but also the moral law, presenting them at first in two distinct systems, but ulti mately in one single philosophical system. T h e philosophy of nature deals with all that is, the philosophy of morals with that which ou ght to b e .23
Apriority of knowledge can be maintained only by rooting it in understanding; apriority of duty can be preserved only by basing it on an equally pure, but acting, reason. Just as empiricism in epistemology destroys certainty, so empiricism in morality destroys its obligatory character. Any ethics deriving from the idea of the good as happiness dislodges the person from his autonomous position as legislator and destroys both the dignity of the agent and the necessity inherent in moral law. Hence, in order to justify the phenomenon of moral necessitation, moral will must be identified with pure but practical reason. T h is pure reason is the same reason that was discovered in the Critique of Pure Reason, but it is here acting in a different capacity. It is no longer theoretical, no longer loses itself in transcendent speculation. Only in action can it be adequately manifested. T he ideas of reason remain transcendent and problematical to thought, while in action they are concretely effective. Kant says that in the “ ought” reason “ frames to itself with perfect spontaneity an order of its own accord ing to ideas.” 24 Caircl has succinctly stated the continuity and differ ence between the two functions of reason: “ J ust because reason cannot find its ideal realized in the world, it seeks to realize that ideal for itself.” 25 Not only does the first Critique thus erect the framework in which all Kant’s subsequent thought naturally fits, but there are several Ibid., A 840=1* 868, trans. Kemp Smith. 24 Ibid., A 548 = B 576, trans. Kemp Smith. 25 Edward Caird, The Critical Philosophy of Immanuel Kant (New York: T h e Macmillan Co., 1889), II, 164. Kant asserts this continuity throughout the second Critique. Note especially his statement that all interest, even that of theoretical reason, is ultimately practical (Critique of Practical Reason, Academy edn., V, 121). 23
Studies in the Philosophy of Kant
19
clear indications in it of the specific ethical theses developed later. In 1781 Kant had already passed beyond the moral doctrines of the precritical period, though this transition was a gradual evolution rather than a radical change. T h e following specific ethical doctrines are anticipated in the first Critique: 1) Moral laws are principles of the possibility of experience, the imperatives being objective laws of freedom.26 2) T h e laws of morality are not empirical and prudential, i.e., they do not show how happiness is obtained, but they contain the a priori conditions of worthiness to be happy. This worthiness, unlike happiness itself, necessarily constitutes a system, a “ corpus mysticum of the rational beings in it [the world].” 27 3) T he highest good, defined as the proper proportion between happiness and virtue, gives practical confirmation to other ideas that were only problematical to speculative reason, viz., the existence of God and the immortality of the soul. T h e way in which immortality is postulated in the first Critique should be particularly noted: happiness and virtue do not correspond in this life, though reason demands that they should; hence there must be another life. In the second Critique, however, the argument is more strictly moral: we are required by moral law to be perfect, and, as this is impossible for a finite sensuous being, a continuation of moral progress in infinitum is postulated— a consequence of Kant’s view that an obligation is invalid unless it can be fulfilled.28 APPRO ACH TO T H E C R IT IQ U E OF P R A C T IC A L R E A SO N "
Kant’s chief works on ethics— the Foundations of the Metaphysics of Morals and the Critique of Practical Reason — may almost be said to be afterthoughts. T he Critique of Pure Reason, as first planned in 1772 and published nine years later, was designed to lay the Critique of Pure Reason, A 802 = B 830, A 807 = B 835. 27 Ibid., A 8o8=B 836, trans. Kemp Smith. 28 In this change, Kant has anticipated the objection often made that happiness, being a state in which all desires are satisfied, is conceivable only for a sensuous being and hence is incompatible with the idea of immortality. T h e moral argu ment for the existence of God undergoes an analogous refinement in the Opus postumum. See below, pp. 51-52. 26
20
Studies in the Philosophy of K an t
foundation of philosophy in its entirety— for a metaphysics of nature as well as for a metaphysics of morals. But, instead of following the order of development as foreshadowed in the first Critique, Kant first wrote a Foundations of the Metaphysics of Morals (1785), and then, as though the work of laying foundations were still unfinished, a Critique of Practical Reason (1788) followed. T his duplication or triplication is all the more surprising as Kant, while developing the concept of practical reason in the Foundations, denied the necessity of a critique of practical reason. In 1797 the Metaphysics of Morals — the book that might have been expected to follow on the heels of the Critique of Pure Reason — at last appeared. But this final work turned out to be less significant than its predecessors. These changes in Kant’s plans require some explanation. T h e most obvious change in thought occurs between the first Critique and the Foundations. In the former, ethics is not a part of transcen dental philosophy; in spite of its a priori form, it must have reference to particular empirical conditions in man which we cannot suppose to be universal and necessary characteristics of rational beings gen erally. These conditions are feelings of pleasure and pain and desire. It was to be the task of the Metaphysics of Morals to take the a priori forms enunciated in the Critique and apply them to the empirical nature of man, just as the Metaphysical Foundations of Natural Sciences did this for nature in general. T h e Foundations, however, begins from quite a different point of view: a metaphysics of ethics is developed wholly a priori. Later, the applicability of this meta physics to man is proved in the Critique of Practical Reason. Here Kant asks how pure reason can be practical. T he Foundations follows the analytic method outlined by Kant in the Inquiry. It reproduces the natural order of thought by begin ning with a phenomenon that requires explanation, whereas the Critiques begin with principles and “ elements” and only subse quently arrive at the experiences to be explained. Why was the second Critique written? T h e most obvious reason is that Kant wished to answer various critics of his ethical theory as previously expounded. Since Kant was not given to petty polemics, a reply to each critic on each point did not appeal to him. T h e objections indicated that his work had not been understood; the Critique of Practical Reason may be considered as K ant’s new at tempt to explain his ethics from the ground up. T his is borne out by
Studies in the P h i l o s o p h y of K a n t
21
evidence that the work we now know as the Critique of Practical Reason grew out of the revisions Kant projected for, but did not include in, the second edition of the Critique of Pure Reason. But of more importance than any answers to criticism are the problems which, in Kant’s own opinion, must have needed more explanation and development than they had previously received. These may be listed as follows: 1) T h e phenomenon of morality and the necessity expressed by the imperative are simply accepted in the first Critique as facts. T h e full seriousness of this acceptance is not evident until we reach the somewhat exhortatory tone of the Foundations. 2) How man as a member of the world of sense can take an in terest in morality, i.e., how the moral law can be effective, is hardly discussed at all in the first Critique, though a framework for an answer to such a question was provided. Therefore, the Critique of Practical Reason asks explicitly: How can pure reason be practical? This was regarded as an empirical question in the first Critique, but the Foundations showed how it could be treated as a transcendental question. It required an answer before the Metaphysics of Morals with its encyclopedic delineation of human duties could be written. 3) While the conflict of theoretical and practical claims is solved — in principle, at least— in the Critique of Pure Reason and pre supposed in the Foundations, the essay What Is Orientation in T h in k in g ? and the second Critique provide the definitive exposition of the doctrine of the primacy of practical reason over theoretical reason. T h e problem of the identity of reason in its dual employ ment, explicitly excluded from discussion in the Foundations, is of central importance in the Critique of Practical Reason. 4) T h e doctrine of immortality, inadequately developed and de fended in the first Critique, undergoes a thorough revision in the second. These various points must be examined in the light of the whole system of which they are integral parts.
THE
PHENOM ENON
OF M O R A LITY
T h e use of the transcendental method characterizes all phases of Kant’s work. He begins like a scientist examining an object or an experience. He notes what aspects are explicable in terms of what
Studies in the Philosophy of Kant
22
is given and sets out to find a reason for those aspects not adequately explained in these terms. T hat is, he makes a regress upon the condi tions of the datum, requiring that the conditions explain adequately the peculiarities of the conditioned. In the first Critique he investi gates the contents of the sciences and finds that their a posteriori material and their a priori form require quite different modes of derivation and explanation. And in turn the a priori conditions, which cannot be empirically given, require a justification (a “ deduc tion” ) to show that they actually serve as conditions of this experi ence. Through this dialectical development, the original experience becomes clarified, the a priori form becomes more explicit and highly articulated, and the whole experience grows richer in implications. T he Critique of Practical Reason does not reveal this method as clearly as the Foundations of the Metaphysics of Morals. T h e former begins with definitions and theorems, and seems, at first, to “ deduce” morality in the ordinary sense of deduction— i.e., to derive it from something more universal. But, farther on, the reader comes upon Kant’s repeated assertions that we cannot answer the question as to why man should be moral. This shows that the deduction actually goes in the other direction, from the phenomenon to the transcen dental conditions which underlie it. Though the moral phenom enon, “ chill duty,” may be, as Hegel says, “ the revelation given to reason,” it is not “ the final undigested lump left in the stomach” 29 but rather the starting point of the investigation. For this reason Kant places a high value on the ordinary moral convictions of mankind. Though innocence may be misled, in moral matters it is a useful guide. Duty, being obligatory upon men, cannot lie beyond their comprehension. Thus Kant takes his departure from the ordinary practical knowledge of morality. Philosophy is to eluci date moral common sense and to strengthen it by distinguishing its essential from its accidental features. Lectures on the History of Philosophy, trans. E. S. Haldane and F. H. Simp son (3 vols.; London: Routledge & Kegan Paul Ltd., 1892-96), III, 461. T here is Kantian authority for this Hegelian judgment. When Kant says that deduction of the categorical imperative is impossible, he is thinking of an ideal that he did not reach even in the first Critique; in both Critiques deduction is the exhibition of a formal supreme principle as underlying an actual experience. Messer, in pointing out this fact, remarks that while Kant is usually criticized for forcing a parallelism on the arguments of the two Critiques , here is an actual parallel that he ignores or denies (August Messer, Kants Ethik: Eine Einführung in ihre Hauptprobleme und Beiträge zu deren Lösung [Leipzig: Veit, 1904], p. 109). 29
Studies in the Philosophy of Kant
23
“ Concerning the ruling ideas in the practical part of Kant’s system,” said Schiller, “ only philosophers disagree, but men have always been unanimous.” 30 These ideas, summarized in the first section of the Foundations, are basic to Christian ethics, especially to Pietism. In this section Kant gives an unusually concise anatomy of moral ity. A moral action, he asserts, is one done solely from duty, not for any specific purpose. T his is the most remarkable and also the most debatable of his basic ideas. Benevolent actions are enjoined by Christianity, and they were made the basis of those British ethical systems which remained closest to common sense. Why, then, does Kant deny moral value to actions motivated by a feeling of altruism, animated as they are by a worthy motive? Although benevolent actions are obligatory, their moral value lies in submitting to the obligation, not in actually achieving the end.31 If the end of the benevolent action were the root of the obliga tion, failure to achieve it would nullify the morality of the action regardless of the intention. T h e English philosophers remained close to common sense in not actually drawing this conclusion. Kan tianism drew the conclusions latent at this point in their philosophy. Kant’s conception of the nature of man would not permit him to derive man’s obligation from anything in the world of sense. By doing so, he would surrender the apriority of duty. T h e a priori character of morality, however, is not just an implication of Kant’s system, and, in insisting upon it, he is not a mere stickler for prin ciples. Rather, the moral a priori indicates a fact largely overlooked by the ancient writers and explained away by Kant’s forerunners, Quoted in Kroner, Kants Weltanschauung, trans. John E. Smith, p. 20. An alternative to the dichotomy between actions for ends and actions from specifically moral motives has often been suggested in recent British ethics. This concept of prima facie duties has been proposed as a way out of a peculiar prob lem in K ant’s ethics: since a certain motive involving a specific attitude toward obligation constitutes the necessary condition of morality, we cannot have an obligation to be moral, for that would be an obligation to have an obligation, and so on in infinitum. Moreover, the feeling that we have an obligation is not a feeling that can be commanded. Kant’s answer to this problem is more abstract than that of his British critics, since he makes the phenomenon of constraint in general the basis of specific duties, while they begin with intuitions of particular duties. But, in finding distinct sources for the goodness and rightness of an action, their as sumption of the quality of goodness is not unlike K ant’s own, and their ascription of rightness is not immune to K ant’s criticism of teleological theories. There is a brief, but excellent, comparison of Kant with his British critics in H. J . Paton’s T he Categorical Imperative: A Study in Kant’s Moral Philosophy (Chicago: University of Chicago Press, 1948), pp. 11 0 - 1 2 . 30 31
24
Studies in the Philosophy of K an t
viz., obligatoriness constitutes the essence of morality. T o explain this hard fact of moral obligation, Kant requires a Copernican Revolution in ethics. This revolution results in the doctrine of moral autonomy— man gives the law to himself. He can do so because he exists as sovereign in the intelligible world, and, even while a subject in the world of sense, he may respect both himself and his fellowmen as pure noumena, pure rational beings. T his constraint of self by self is obligation; subjectively, it is respect for the law. Action from respect for law is the key to the universality and necessity of moral precepts, which would be variable and contingent if they depended upon particular ends, however general they might be. Universality and necessity are the marks of apriority. And just as necessity which would bind thought could come only from the mind itself, so the moral law which is a priori must be dictated by the reason which is to fulfill it. Otherwise this reason would not be free and would have to be encouraged to obedience by incentives of reward and punishment. Reason, unaffected by individual differences, gives only the uni versal form, not the particular material. From the analysis of the apriority of duty follow both the formula of morality and the chief characteristic of Kant’s system: it is subjective a priori formalism. This characterization of his ethics has often been taken lightly, as if by itself it constituted a sufficient criticism of his system. When properly interpreted, it is its truth and strength. At this point it is well to see what is not involved in the concept of the formal a priori character of duty. First, that which is a priori is not prior in time. T he moral law is not inborn, nor is it given by outer fiat. On the other hand, it is found not by induction from experience but by a critical analysis of the conditions of experience, commencing from what is implicit in any moral experience. Moral education is a process of developing the child’s mind toward per ceiving the essential features of morality. Though K ant’s ethics claims universal validity, the fact that there is no universal agree ment on ethical principles constitutes no objection to his theory. Ethical universality does not entail anthropological uniformity. Second, though all experience must conform to the a priori, no experience is adequate to its universality. Because of this, Kant can disapprove of the appeal to examples for ethical theory and yet use them in ethical training.
Studies in the Philosophy of Kant
25
Third, the a priori is not a supreme principle from which deduc tions can be made as if it were a mathematical axiom. In the Inquiry Kant defined the true function of formal principles. He says: Just as nothing follows from the primary formal principles of our judgments of truth except when primary material grounds are given, so also no particular definite obligation follows from these two [formal] rules except when indemonstrable material principles of practical knowl edge are connected with them.32
It has always been objected that the categorical imperative gives no specific moral guidance. Hegel’s objection, for instance, is that only when an institution (say, property) is universalized can the cate gorical imperative indicate a condition which a moral action must meet. Yet the function of a formula is not to supply the variables but to provide the procedure toward a solution; it is the necessary but not sufficient condition for a given solution.33 But should not the condition defined by the formula be the suffi cient condition, assuming that moral action is action done only from respect for law? T his is a difficult question, to which Kant gives no univocal answer. Yet, I think, the objection that a formal principle cannot have particular material consequences may be answered in perfect harmony with Kant’s ethics as a whole. He denies that the moral law could be derived from some one final goal of life, such as happiness.34 In questions of morality we should simply give no regard to particular motives of happiness, though they may be occaAcademy edn., II, 299. Later, in the period of the Dissertation (1770), he wrote: “ T h e supreme principles diiudicationis m oralis are, it is true, rational, but they are only principia formalia. T hey do not determine any specific purpose, but merely the moral form of every purpose; therefore, in concreto principia prima materialia occur according to this form” (Fragment 6633, Academy edn., X I X , 120; quoted from Schilpp, Kant's Precritical Ethics, pp. 105-6). 33 Broad defends Ktant from such criticisms by suggesting that the categorical im perative might better be called the “ supreme principle of categorical imperatives.” It is, he says, a second-order principle which states the necessary and sufficient con ditions to be fulfilled by any first-order principle qualifying as a categorical im perative. He compares it to one of the valid modes of the syllogism, not to a true major premise (C. D. Broad, Five Types of Ethical Theory [New York: Harcourt, Brace & Co., 1930], pp. 120-21, 123). If this view had been followed consistently by Kant, however, conflicts between duties could have been honestly admitted, and this he does not do in On a Supposed R ight To L ie from Altruistic Motives. See also William David Ross, Foundations of Ethics (Oxford: Clarendon Press, 1939), pp. 173, 313. 34 On the Saying, “ That May Be T ru e in Theory But It Does Not H old in Practice,” I (Academy edn., VIII, 279). 32
Studies in the Philosophy of Kant
26
sions of the moral problem. T h e moral problem arises as a problem of the whole man; moral law is a formula, a procedure35 for its solution. Human beings are ends in themselves and fulfill their destiny by being worthy of happiness. But the moral law, Kant affirms, is the only factor which should be regarded in making a decision, though it is obvious that the problem to be solved by the agent is not posed by moral law alone but rather by the tension between law and natural inclinations. In excluding the ends of inclinations from moral legislation, moral law does not degrade morality to a mere mechanical routine of carrying out abstract rules. Rather, it indicates the way to achieve specifically moral goals, which must be defined a priori without reference to the contingencies of experi ence and man’s partiality toward himself or others. T he formal element, implicit in ordinary morality, is separated and purified and then seen as the basis for widening the moral claims beyond the phenomenon in which they were discovered. Kant cor rectly compares his procedure to that of the chemist who analyzes a complex mixture. With pure principles, as with pure substances, procedures which have heretofore been “ empirical” can be made rational and comprehensible, and new relationships can be uncov ered which are a priori conditions of a broader experience and permit a more penetrating insight. So much, then, for the straightforward analysis to which Kant subjects the moral phenomenon. Since man is obligated by laws which flow from his own reason, the elucidation of these laws presents no great difficulty. T he distinction between the hypothetical and categorical imperatives, and between legality and morality, and the gradual process by which the categorical imperative becomes more and more concrete until finally it is a definition of the unique ends of moral action, are clearly and systematically presented in his works. Criticism of the foundations of Kant’s ethics has not been lacking. Schilpp, in his excellent analysis of Kant’s early works, has especially de veloped the “ procedural aspect” of formalism in which “ rational reflection, creative construction, and transition [are] the only method capable of coping successfully with the exigencies of human experience and with the needs of moral obligation” (Kant’s Precritical Ethics, p. 173). T h a t formalism has been quite differently interpreted, and that this interpretation has had serious historical consequences, is argued by F. S. C. Northrop, T h e M eeting of East and West (New York: T h e Macmillan Co., 1946), pp. 201, 217, et passim. 35
Studies in the P h ilo s o p h y of K a n t
27
In spite of the variety of arguments employed, they may be divided into two types— those that reject moral value as a separate species and those that accept moral value as irreducible yet localize it in a way incompatible with Kantianism. T h e doctrines which deny the specificity and irreducibility of moral value— whether they arise in Marxian, Darwinian, theological, or utilitarian philosophies— use one or both of two tactics. They deny the phenomenon of obligation as anything more than a myth, fiction, or mistaken feeling; or they accept it as a prima facie fact to be justified on their own assumptions. Against the first tactic, Kant did nothing and could do nothing except exhibit the phenomenon of morality and show that its inter pretation as a natural fact (the theory advanced by his opponents) is incompatible with their own teleological theory of nature. He points out that, if moral injunctions were part of natural economy, nature would have done better by leaving decisions to instinct. On the other hand, if the conception of nature is not one based on design and if no supranatural source of morals is postulated, the phenomenon of morality becomes an illusion, to be explained in the light of a mechanistic natural science. Kant rejects this theory for the same reason that he rejected Hume’s conclusion in science: it flies in the face of facts which should be explained, not explained away. For him, moral constraint is the starting point, which cannot be justified except by its philosophical fruits— the intelligibility of the world revealed through its analysis. Just as Aristotle waited until moral character was formed before teaching the theory of morals to the young, so Kant presupposes the acceptance, in broad outlines, of a Christian-humanistic moral attitude and speaks only to those who share it. T h e second tactic fails for the same reason that the various natural istic explanations of the logical a priori failed in the nineteenth century. In each case a speculative cosmological system is made the foundation of rational necessity, and the results of science are made fundamental to the process by which they are discovered. This is a vicious epistemological circle. T h e a posteriori cart is placed before the a priori horse. One of Kant’s greatest contributions was his dem onstration of the inevitability of skepticism, once knowledge is re garded as a natural fact on the same level with its object. Similarly,
28
Studies in the Philosophy of Kant
moral relativism and skepticism must result from ethics which derives the concept of what ought to be from that of existence as it seems to be in the light of contemporary: scientific knowledge. And this moral skepticism is equally self-refutjng, because science has its moral foundations just as it has its rational presuppositions in the narrower sense of theoretical rationality. T he criticism to which we now turn is of greater philosophical interest. Accepting moral value as an irreducible and unique fact, the German phenomenologists, the British analytical moralists, and their many fellow-workers throughout the world renounce Kant’s transcendental subjective formalism and yet seek to preserve the apriority of moral law or moral values discovered through some kind of intuition or immediate rational insight. T o evaluate this criticism, we must first review Kant’s own attitude toward the moral sense and then examine one of these countermovements. In the beginnings of modern European philosophy, Christian dualism had kept the senses in a secondary position in morals long after the importance of sense perception to scientific knowledge had been re-established. But it was inevitable that the increasing seculari zation and naturalization of thought should bring with it a reorienta tion to natural man. This change of outlook was associated with a rejection of the moral precepts involved in Christian dualism; in time of change it is not easy to separate the central elements from the peripheral ones in opposing views. T h e revolution, therefore, re sulted not in a naturalistic ethics but in an amoralism, whose great est exponent was Thomas Hobbes. After Hobbes a reaction came; the main trend of British ethics was to reconcile an ethics of natural man with Christian tradition. Within the naturalistic framework, Shaftesbury, with his concept of “ disinterestedness” of the senses, contrived an escape from amoralistic conclusions. Like Locke, he held that the senses are capable of re flecting upon themselves and thus of becoming free from domination by the object of desire to which they are normally directed. Shaftes bury says: In a creature capable of forming general notions of things, not only the outward beings which offer themselves to the sense, are the objects of the affection; but the very actions themselves, and the affections of pity, kindness, gratitude, and their contraries, being brought into the m ind by
Studies in the Philosophy of Kant
29
reflection, become objects. So that by means of this reflected sense, there arises another kind of affection toward those very affections themselves, which have already been felt, and are now become the subject of a new liking or dislike.36
Hutcheson, developing Shaftesbury’s idea, held that something absolutely good is apprehended in actions. This good, in his phil osophy, is benevolence, or that which tends to promote “ public natural happiness.” T h e perception of moral excellence is different from perception of an object of desire, and thus the sense for natural good is distinct from that for moral good.37 It is not itself the in centive to moral action (this being benevolence) but the source of approbation or disapproval: T h is moral sense, either of our own actions, or of those of others, has this in common with our other senses, that however our desire of virtue may be counterbalanced by interest, our sentiment or perception of its beauty cannot; as it certainly might be, if the only ground of our appro bation wrere views of advantage.38
In his precritical period Kant seems to have inclined toward such a view in his Observations on the Feeling of the Sublime and the Beautiful, which is composed in the manner of Shaftesbury. Kant speaks, in the Distinctness of the Fundamental Principles of Natural Theology and Morals, of “ sensations of the good” and of feeling as the “ faculty of sensing the good.” Nevertheless, in his critical writ ings, he completely rejects the concept of moral sense, though he attempts to reconstruct the concept of moral feeling.39 It is easy to see why his critical system demands this rejection. T he mind is re ceptive rather than spontaneous in assimilating sensory material, and any knowledge resulting from receptivity is a posteriori. Hence the theory of moral sense would vitiate the autonomy of reason and destroy the apriority of morals. 36 Shaftesbury, “ An Inquiry concerning Virtue” in L. A. Selby-Bigge, British Moralists (2 vols.; Oxford: Clarendon Press, 1897; reprinted in one vol. in “ Library of Liberal Arts,” No. 152 [New York: Liberal Arts Press, 1964]), I, 11. (I have mod ernized the spelling here and in the following passage.) 37 Francis Hutcheson, “ An Inquiry Concerning Moral Good and E v il” in British Moralists, I, 78, iko. •38 / bid., p. 78. 39 Metaphysics of Morals, Part II, Metaphysical Principles of Virtue, Introduc tion, V, X II.
go
Studies in the Philosophy of Kant
With the modern phenomenological approach, however, the old issue has been reopened on a level not anticipated by Kant. T h e phenomenologists claim for intuition what is specifically denied to it by Kant— an a priori insight into content (material) and not merely formal conditions. Phenomenology thus represents a CounterCopernican Revolution in philosophy, locating the a priori in the essential relationships between experiential objects grasped through immediate insight. At the same time the a priori ceases to be a mani festation of a transcendental subject; it is the basic structure of any possible object of experience. This powerful method offers a new alternative to Kant’s formalism. Hartmann agrees that if Kant’s disjunction between “ nature” and “ reason” is exhaustive, then Kant’s formalism must be accepted as the only alternative to an a posteriori naturalism. But he and other phenomenologists deny the exhaustiveness of this disjunction, assert ing that there is a realm of values into which we have an a priori insight. On this foundation, it is claimed, an ethics which is both material and a priori can be erected. It is profitless to argue, outside of a general treatise on metaphysics, concerning the relative merits of two such radically different ap proaches to philosophy. Nor is it possible to borrow a little from Kant and a little from Scheler and Hartmann to piece out an eclectic system— a method repeatedly repudiated by Kant with impatience and irony. Nevertheless, in some respects, especially in ethics, these two approaches are complementary; and, while it is out of the ques tion to make a system of them, it is important to notice how they supplement each other. T h e Kantian ethics is weak in two respects. First, the manner in which moral concern comes into experience is never satisfactorily examined. Kant often speaks of the “ fact” of obligation, and he calls the moral law the “ sole fact of pure reason.” Yet the sense in which it is a “ fact” is never made clear; it is the starting point, but the analysis of it leads away from its factuality. Second, “ monotony” arises in the Kantian ethics from the attempt to discover everywhere the same basic pattern in all actions having moral worth. Naturally, every ethical theory must seek the highest common factor, but in Kant’s writing the road back to the rich and variegated complexity of the moral phenomenon is seldom followed. T h e store of ethical phenomenology found in the Observations on the Feeling of the
Studies in the Philosophy of Kant
3i
Sublime and the Beautiful was never put to use in Kant’s mature philosophy. On the other hand, it is precisely the factuality and empirical variety of the moral phenomenon which have attracted the closest attention of Scheler, Hartmann, Ross, and others who are phenom enological or “ analytic” in their approach. Though their picture lacks the logical and metaphysical simplicity of the Kantian, it is much more subtle in its portrayal of the facts of the moral life. For this richness it must pay by an occasionally uncritical assumption of faculties and principles.40 Finally, phenomenology, with all its emphasis on the a priori, is unable to give as convincing an account as Kant does of the necessity inherent in moral imperatives. Kant states that an ethics beginning from the good necessarily leads to heteronomy and to hypothetical imperatives. T o attribute to the object (the good) a character of “ obligatoriness” in its relation to human action, as some phenomenologists do, is to make use of an ad hoc hypothesis which may well be discarded without making the phenomenon unintelligible. On the other hand, the Kantian conception of the relation of the good to obligation is simple, clear, and cogent. T he logique du cœur of the phenomenologists seems to be a poor foundation for imperious duty exalted by Kant. HUM AN N A T U R E AND FREEDOM
Kant repeatedly refers to his theory of morality as the “ ethics of intention,” and moral personality is everywhere the center of his thought. Only the good will is good; every rational being is an end in itself; humanity in man must be treated as an end and never as a means only; and our own moral perfection and the happiness of others are ends which are also duties. Kant’s theory, however, legitimizes the moral command for rational beings in general. Moral ity cannot be derived from the empirical nature of man, nor should it be applied to man alone. It lies in the essence of rationality itself and applies to all rational beings. Although the following story may be apocryphal, it represents an impatience which is often created in a reader by the elaboration and overelaboration of entities in phenomenology. It is said that, when Scheler first read Hartmann’s Ethik, he exclaimed, “ My colleague Hartmann believes he can take a stroll through the realm of values as though it were Cologne!” 40
Studies in the Philosophy of Kant
32
Man, however, is not merely a rational being; and his peculiar position in the world, as a rational being affected by sensuous needs, creates special problems for him. It requires special efforts on K ant’s part to show how man as a citizen of the sensuous world can fulfill his destiny in the intelligible world. T h e problem of human morals is thus resolvable into two questions: How can man be a member of the intelligible world? How can the demands made upon his in telligible nature be met in the world of sense? T he first is the prob lem of freedom; the second, that of the empirical character which is obligated to act morally. T he moral law is the “ sole fact of pure reason.” 41 But freedom is the “ ratio essendi of morality” and “ the keystone of the whole archi tecture of the system of pure reason.” 42 It is the central problem of Kant’s entire work and, as such, appears in various perspectives and at various strata within his philosophy. There is, first of all, freedom of choice, which in one place43 is said to be independent of transcendental freedom, the truly metaphysical concept springing from the third antinomy. Then there is the meta physical concept itself, fully developed by means of the distinction between phenomena and noumena. T his metaphysical concept of freedom is worked out problematically in the theoretical phases of the system and is asserted practically in the great ethical treatises. Finally, there is the concept of freedom which results from the faculty of judgment as the mediator between the worlds of appearance and reality. This is the latest of the three, elaborated in full in the Cri tique of Judgment. Critique of Practical Reason, Academy edn., V, 32, 48; trans. Beck, pp. 32, 48. T h e difference between the two passages cited (" F a k tu m ” “ gleichsam ein Faktum” ) is of interest. T h e former passage states that it is not an empirical fact, and the lat ter hesitates to call it a fact at all. In modern terminology Kant might call it a con struct in both passages. Alfred Hegler (Die Psychologie in Kants Ethik [Freiburg i.B.: Mohr, 1891], p. 92) writes that the moral law is Faktum “ because it shows itself as real in maxims determined by it, and constitutes an ultimate possession of con sciousness not reducible to any other,” and that it is gleichsam ein Faktum “ be cause it is not a single empirical datum [Thatsache] in consciousness, like some presentation or volition; it is not an empirical fact, but the sole fact of pure reason.” [Cf. this volume, Essay XII.] 42 Critique of Practical Reason , Academy edn., V, 3-4; trans. Beck, pp. 3-4. 43 Critique of Pure Reason, A 8o3=B 831. [I no longer accept the interpretation given here of this passage; cf. my Commentary on Kant’s Critique of Practical R e a son (Chicago: University of Chicago Press, i960), p. 190, n. 40.] 41
Studies in the P h i l o s o p h y of K a n t
33
Freedom on the first level is empirically given. It is freedom from the immediate importunities of sense. Kant is correct in saying that its reality is independent of the answer to the speculative problem of whether or not the will whose freedom is thus assured is ultimately a part of the mechanism of nature. He is not consistent, however, in his statement that this freedom is adequate to the requirements of morality as he expounds them in the remainder of his work. Never theless, this rudimentary concept of freedom helps us understand how the empirical personality can act morally. Wholly within the realm of appearance, then, there is in man a faculty which can be called empirically free— the faculty of choice.44 Choice “ is the faculty of desire so far as it is connected with the consciousness of the competence of its action to produce its object.”45 It is free to the extent that choice is determined by reason (con sidered simply as the highest faculty of the mind which in its turn is a part of nature) and is contrasted with the animal will (arbitrium brutu m ) , which is exercised without this control.46 Choice can thus be considered, within the one world of nature, to be both free and necessitated.47 Such freedom, however, is of “ limited liability.” It is not capable of freeing man from all his yesterdays. Because it leaves man finally a part of nature, it does not justify us in imputing a man’s actions to him as if he, and not nature, were the author of his works. It is free dom only in the legal sense. Freedom, Kant says elsewhere, cannot be understood psychologically; it is the stumbling block of all em piricism and cannot be salvaged if time is the mode of existence of things-in-themselves. Empirically, it is at most a partial manifesta tion of true freedom. F ree d o m consists in
o f choice
w ith
respect
to h u m a n
the ca p acity o f ch o o sin g b e tw e e n
actions as p h e n o m e n o n two o p p o s in g things, the
44 Willkür. In Kant this word does not have the implication of willfullness that it possesses in modern German, where it implies some degree of arbitrariness and irresponsibility. A b b o tt translates it as “ elective w ill,” which conveys the meaning very well but has the disadvantage of rendering a common everyday German word by a technical philosophical term. 45 “ Introduction to the Metaphysics of Morals,” I, in Part I of Metaphysics of Morals (Academy edn., VI, 2 13). 46 Critique of Pure Reason, A 534=0 562. 47 Lose Blätter, Reicke edn., II, 28.
Studies in the Philosophy of K an t
34
lawful and the unlawful. Herein man regards himself as phenomenon, but as noumenon he himself is theoretically and practically legislative for objects of choice. In this respect he is free, but he has no choice.48
These two concepts— choice without true freedom and freedom which is conformity with the law prescribed by reason— are con trasted as the negative and positive concepts of freedom.49 T h e former is freedom from something; it is arbitrary and lawless in itself, for reason as its determining ground is as yet undefined. Reason might conceivably be merely the name of a transcendent thoughtprocess directed ultimately to the satisfaction of the senses which are unable to command the faculty of choice directly. But even though such a faculty of thought might transcend the ordinary laws of psy chology, so long as its goals were in nature it would be restricted by man’s empirical character as a being of wants and needs. Hence to be free from the senses in an unqualified way requires something more than the possibility of enlightened choice between alternatives. A determining ground of choice is required which will legislate for it directly, not indirectly through its objects. T his determining ground must be independent of the entire world of sense. Only then can freedom from the world of sense become freedom to something else positively defined. If we could see freedom as an efficient cause, Kant affirms, we could also see that the practical law (reason’s legis lation) is the supreme law for choice, for it is the efficient cause issuing from rational beings. But Kant does not make the transition from the negative to the positive concept of freedom in this way, since we cannot perceive the freedom of an efficient cause.50 Rather, he seems to make a fresh start. This phase of the argument has two distinct parts— one which shows freedom, positively defined, to be possible, and one which Ibid., p. 140. 49 “ Introduction to the Metaphysics of Morals,” I, in Part I of Metaphysics of Morals. 50 In one place Kant writes as if we might make such a transition, though what he says will also bear a different interpretation. He says that the positive concept of freedom is “ the capacity of pure reason to be practical of itself. This, however, is not possible except through the subjection of the maxims of any action to the condition of the worthiness of these maxims to serve as universal law” (“ Intro duction to the Metaphysics of Morals,” I, in Part I of Metaphysics of Morals [Academy edn., VI, 213]). 48
Studies in the Philosophy of Kant
35
shows it to be real. Taken together, they constitute the transcen dental-practical doctrine of freedom universally associated with Kant’s name. T h e argument begins with the third antinomy, whose antithetical propositions, we remember, may both be true provided their respec tive scopes are defined. Thus nature as existing in time is determined under the category of causality. It is, however, possible to think (though we cannot know) noumena— things-in-themselves— which are not in time and are therefore independent of the law of nature. Hence, if man is not merely a phenomenon but also a noumenon, then he may be free as noumenon (in accordance with the thesis) without ceasing to be mechanically determined in his role as temporal phe nomenon (as the antithesis asserts). Evidence for this dual character, apart from its being suggested by the fact that we possess a priori knowledge and are thus free of the time order in knowing, is found in obligation. Obligation is charac terized by subjective necessity, the objective counterpart of which is moral law. Neither obligation nor the law can be derived from ex perience, for experience establishes no necessity. T h e moral law requires man to act from a rational principle and not from an inclina tion to an object in the sensuous world, no matter how desirable. It therefore implies that man must be capable of actions free from any inclinations and determined only by reason. Thus his absolute spon taneity as a being outside of time and his membership in a com pletely determined temporal order are jointly guaranteed without any mutual interference. These two human roles are distinguished as the rational and the empirical character or as will in contrast to choice. “ Will is the faculty of desire not with respect to action (as in the case of choice) but with respect to the determining ground of choice to an action. Of itself it has really no determining ground. It is practical reason itself so far as reason can determine choice.” 51 T o the extent that reason is autonomous (i.e., determines its own ends in a categorical impera tive instead of being merely the guardian of the senses), it constrains the empirical personality to respect and obey its law. This it does even though the law may be broken— indeed, may never have been truly fulfilled. 51
Ibid.
Studies in the Philosophy of Kant
36
[Practical freedom] presupposes that although something has not h ap pened, it ought to have happened, and that its cause, [as found] in the [field of] appearance, is not, the'refore^ so determining that it excludes a causality of our will— a causality which, independently of those natural causes, and even contrary to their force and influence, can produce some thing that is determined in the time-order in accordance with empirical laws, and which can therefore begin a series of events en tirely o f itself .52
Why man is free is a question as unanswerable as why he should be moral or why he is rational. How is man free, as a theoretical question, is likewise unanswerable. It is a speculative question, lying beyond the scope of theory. At most, speculative reason can answer the question negatively by showing how the contradiction between the causality of freedom and that of nature is only apparent. If the alleged contradiction could not be resolved, Kant confesses that free dom would have to be surrendered in favor of natural necessity. But just as the Critique of Pure Reason limited knowledge in order to make a place for faith, it limited temporal causality to give a place to causality of freedom, or— what is the same thing— it limited nature for the sake of morality. While we cannot show in any theoretical or empirical way how freedom is possible, we can prove that it is pos sible, and we can show how faith in freedom functions in our moral convictions and conduct. Man is in the world but not of it. T h is concept, tenuous as it is, is superior, even from a moral point of view, to those ancient and modern conceptions which leave man in the world as a natural being while exempting him from its laws.53 For these speculative conceptions end with an indeterminism which is of no greater moral utility and comfort than mechanism itself. Yet the theoretical questions insist upon an answer. We want an intellectual justification for faith. T o be told that one ought to have done otherwise, when a satisfactory psychological investigation con vinces us that one could not have been expected to do otherwise, seems often— as Kant admits— to “ conflict with equity.” 54 T o have acted “ otherwise” — i.e., spontaneously to have commenced a new Critique of Pure Reason, A 534=1* 562, trans. Kemp Smith. 53 In the ancient world Lucretius, and in the contemporary world those physi cists who seek comfort in the Heisenberg principle, would give man a freedom he could not manage. Scientific unpredictability would not imply moral freedom so much as a hazardous irresponsibility. 54 Critique of Practical Reason, Academy edn., V, 100; trans. Beck, p. 103. 52
Studies in the Philosophy of Kant
37
causal series— is in conflict with the Analogy of Experience,55 which treats the entire world of sense as a causal system. T he initiation of a new series cannot be interpreted as an influxus mysticus without surrendering the category of causality. Kant says that, in a given case, reason could not otherwise determine action except by making the empirical character itself different,56 and this in turn would require a change in the entire causal order so that sufficient empirical con ditions could be found for its alternative state. Hence a causally integrated totality of appearances wrould require a completely inte grated noumenal order which, in the case in point, would have had to be different from what it actually is. If these two orders are inde pendent of each other, it is impossible to see how the manifestation of the rational character could be considered a spontaneous initiation of a causal chain within the empirical world. The consequence seems to be Spinozism: nothing can be different from what it is. T h e mere possibility of evil in a world which is supposed to be the appearance of the world of reason, the source of all morality, presents a difficult problem for the theory. T he actuality of evil actions ac centuates this difficulty. If we grant that the first two Critiques sufficiently justify the possibility of moral actions, they do not explain how actions which ought to be morally good can fail to be so. The existence of evil is incompatible with a pure will which is free but without choice. Evil cannot be regarded simply as phenomenal, for then it would not be imputable as moral evil.57 It is rather a mistaken subordination of pure and empirical maxims in the empirical char acter, and the theory of freedom is still unable to justify the claim that one might have refrained from an evil action which actually was performed. These and similar difficulties are unavoidable in the second phase of Kant’s conception of freedom because the argument is still on a theoretical level. They are irresolvable for two reasons. First, the concept of freedom at this level, though distinguishing between appearance and reality, does not adequately define their relationship. It leaves morality and science each claiming its own Critique of Pure Reason, A 21 i = B 256. Ibid., A 556 = B 584. See R eligion within the Lim its of Reason Alone, trans. Theodore M. Greene and Hoyt H. Hudson (Chicago: Open Court Publishing Co., 1934), p. 17. 55 56 57
g8
Studies in the Philosophy of Kant
concept of causality, and these concepts are kept from open conflict only by a strict definition of areas of Application. But when the two kinds of causality are applied to a specific moral action or character, their claims again inevitably conflict. Nevertheless, it is one reality in its several manifestations that we need to judge in order to bring peace and not just a truce to philosophy. Until the harshness of the distinction is tempered without loss to the positive results of the two Critiques, such contradictions will occur in every specific judgment. Though the first two Critiques established an “ impassable chasm” between the two realms “ as though they were so many different worlds,” the law of freedom, Kant says, “ ought to have an influence” on the world of sense.58 T h e problems arising from the “ two-world theory” can be solved only when the relationship between the two worlds is developed affirmatively. Second, the claims of reason and those of understanding are at variance, yet each is sovereign in its own sphere. Neither is capable of extending its claims at the expense of the other. But, more sig nificantly, neither is competent to apply its own principles directly to its object. These principles are without exception formal. T h e problems, however, are not formal but material, arising from the rival but justified claims of these two faculties to the phenomenon whose true estimation is in dispute. For the adjudication of these claims, a mediating faculty is required. T his faculty is judgment, and the third Critique presents the last and most profound treatment of freedom to be found in K ant’s works. It is also the most difficult, because freedom itself is seldom the center of his attention. Rather, Kant turns primarily to consider a different, but related, problem— that of purpose. T h e Critique of Pure Reason requires the formal unity of the a priori laws of nature, but it does not establish the idea of an equally inclusive system of individual objects. Nature, Kant says, constitutes a system by its transcendental laws, but . . . there is such an infinite multitude of empirical laws and so great a heterogeneity of forms of nature . . . that the concept of a system accord ing to these empirical laws must be wholly alien to the understanding, and neither the possibility nor even less the necessity of such a whole can be conceived.59 •r>8 Critique of Ju dgm en t, Introduction, II, IX. 59 First Introduction to the Critique of Ju d g m en t, II (Academy edn., X X , 203).
Studies in the Philosophy of Kant
39
This concept of an empirical whole is nevertheless demanded as an ideal, as a regulative concept giving sense and direction to the search for interrelationships among the phenomena of nature. This inter relationship is provided by the faculty of judgment which subsumes particulars under given universais or finds universais that fit given particulars. In carrying out this function, judgment presupposes a “ formal design” in nature, i.e., an all-pervasive, comprehensible order. We are never excused from searching for a mechanical explanation of any single fact (including human actions), yet at the same time we cannot anticipate a “ Newton of a blade of grass.” T h e mechan istic theory does not lead us from the parts which it investigates to the whole of its organization. This limitation is particularly relevant to human actions, even though they are in principle, Kant says, as predictable as astronomical events. T he idea of formal design is of pre-eminent importance in the case of man, for man’s character is purposive in the sense that all parts are to be properly subordinated to the whole; moreover, man’s moral conduct understood as conformity with his own internal law tran scends mechanical causality, in which causes are external to the acting character.60 For these reasons it is proper to judge man by the regula tive idea of his design and purpose, which is moral, and we can do so without prejudice to a mechanistic explanation of nature. Because both of these ideas, purposive design and mechanism, eo William T . Jones, has emphasized this point in “ Purpose, Nature, and the Moral Law ,” in The Heritage of Kant, ed. G. T . Whitney and D. F. Bowers (Prince ton: Princeton University Press, 1939), and in Morality and Freedom in the Philosophy of Immanuel Kant (London: Oxford University Press, 1940). He denies the validity of the concept of noumenal causality and denies that the third Critique can bridge the gap between theory and practice by its attempted proof that the highest good is possible in nature. In fact, he goes further and says that this attempt is itself incompatible with the deepest insights of K ant’s system concerning the' true relation between morality and nature. He holds that the existence of the organism permits a constitutive use of the category of purpose (which is, of course, denied by Kant), and that it supplies an analogon by which we can see the supplementation to natural necessity required for the achievement of the highest moral end. Mr. Jones quite correctly sees in the organism not a natural purpose requiring a peculiarly transcendent cause but an objective analogon of morality itself. “ T h e peculiar kind of order which distinguishes organisms from other natural objects,” he says, “ is identical with the order which, we find, connects motive and act in morally good action and which distinguishes it from all other action” (The Heritage of Kant, p. 240). T h u s freedom (but not freedom of choice) is found in an order of being to which mechanistic explanation is simply irrelevant.
Studies in the Philosophy of Kant
40
function methodologically without interfering with each other, we see how the ends of freedom may be thought of as possible within the system of nature. T he existence of natural purposes in the world indicates the heuristic inadequacy of the category of mechanical causality. T he existence of organisms does not prove conclusively that there is an over-all design in the world and that man is the pur pose of creation. But it shows something very important about the constitution of the mind. T h e mind, we are made to see, because of the nature of its operations, must apply the causal concept and also supplement it with the concept of purpose. In contrasting moral necessity as dictated by duty with factual necessity as determined by nature, we contrast action as some thing which ought to be done with action as simply occurring. But, says Kant, it is because of the constitution of the mind that moral law must be represented as a command, i.e., as prescribing what ought to be done whether it is actually done or not. T he necessity of dutiful action cannot be represented as a mode of physical being, but only as an ought-to-be; but this would not be so if reason alone were legisla tive in the world of appearance. For then the world of appearance would be a perfect representation of reason’s practical conception of reality. T he idea of duty is a practical conception, an ideal which would then have as much power as it has “ manifest authority.” T h e conflict is not between two worlds but between two subjectively neces sary modes of viewing the world. T h e mind is unable to think purposively and mechanically, or practically and theoretically, at the same time. But the characteristic of the finite human mind should not be attributed to its objects.61 Design, or organization, we seem to find everywhere in nature. It is the dependence of all the parts upon all the others and hence upon the whole. In an organism, as a natural organization, every part is both cause and effect of all the others. T h e understanding, however, because its procedure is discursive, is unable to make theoretical use of this concept of mutuality. Understanding arrives at universals through particulars, thus achieving an abstract universal from which the particulars cannot again be derived. Yet they seem, in the case of an organism, to be derived from the concrete universal which is the organized whole itself. T h e categories of the understanding can61
Critique of Judgm ent, §77.
Studies in the Philosophy of Kant
4i
not be formulated as rules for this kind of synthesis. We can compre hend the observed dependence of part upon whole not through a category of purpose or wholeness— as many biologically oriented writers on Kant have proposed— but only through a regulative idea, by analogy with our own technique. As technicians, we do combine parts into wholes and order means to ends in the light of a guiding idea, a concept of the whole. By analogy, we think of a technique of nature, as though nature worked for certain ends in the light of a guiding idea, i.e., as created or as creating with a design. But in nature it is not the whole as such which is the condition of the parts (for the parts are parts of physical nature and are under the condition of time and efficient causation). It is rather the idea of the whole which conditions their relationships as we see them. Purpose, however, is the concept of an object so far as the concept contains the ground of the reality of the object itself.62 Hence the discursive nature of human understanding makes concrete constitutive insight into the organization of nature impossible and compels us to interpret design, in a merely regulative manner, as a consequence of purpose. We have, thus, only an abstract insight into nature’s structure, and our con crete insight into nature’s relationships is always constituted by the category of causation in time. But it is possible to think of another kind of understanding, one free from human limitations, an “ intuitive understanding.” Such an understanding would proceed from an intuition of the whole to its parts, thereby reversing the procedure of the human understanding, which can only construct a whole out of its parts. T his superhuman understanding would, by directly intuiting wholes, see wholes as the efficient causes of parts. Our intuition presents us with conditioned wholes which, in the order of time and efficient causation, are the effects, not the causes, of their parts. We can think abstractly of an absolute and unconditioned whole, but we cannot intuit it. As our conception of it is abstract, we comprehend it only by an analogy, in which the concept of the whole is seen as the condition of the par ticular constellation of its parts. T his is precisely what we mean by design or purpose. Therefore it lies in the nature of the human mind to think in teleological as well as in mechanical terms. The mechan ical pattern, in which the category of causality has a schema in time, 62
Ibid., Introduction, IV.
Studies in the Philosophy of K an t
42
is constitutive of the world of experience, while the teleological pro cedure, without a temporal form, is regulative of our search for the order between various causal sequences. Freedom, like purpose, is a regulative idea. It posits a design, and this design does not lie in the constitution of nature. As purpose is regulative of our theoretical judgment, so freedom is regulative of practice. Purpose and freedom are not constitutive, and the world of nature is not the world of morality. But just as the Critique of Judgment permits us to guide our thought by the idea of purpose, the Critique of Practical Reason states as the typic of the moral law the injunction: act as though the moral law could become a law of nature. T he faculty of ideas— pure reason— thus becomes practical. Moral ity and freedom are the positing of a regulative idea as an ideal to be achieved and the conduct of reason so as to achieve it. T h e regulative idea of the unconditioned condition which functions theoretically in the teleological judgment now becomes practical. T hat which is regulative always has direct relevance to the prac tical. A ll practical ideas are regulative not merely of thought but of conduct. It means little or nothing to demand that the practical be constitutive—man’s reach should exceed his grasp.63 For an intuitive understanding, design, purpose, and even freedom might be con stitutive categories, and they would not in the least conflict with the demands of mechanism. For us, there is an antinomy between free dom and efficient causation when both are taken constitutively, and this conflict can be resolved by us only by considering freedom and design to be regulative. T h e conflict holds between our diverse pro cedures, not in reality. T h e Critique of Judgment, while not breaking down the distinc tion between appearance and reality, thus points the way to a de facto justification of the judgment of freedom. Without making T h e answer to the question, “ What ought I to do?” “ consists of one presup position and one inference. T h e presupposition: It must be proved (and the two Critiques do prove) that the unconditioned is not impossible, that there is some hope of its achievement and realization, small though it be. For no man could knowingly and without reserve throw himself into something conceded impossible and indeed apodictically impossible. T h e inference: If there is the slightest hope that the Absolute can ever be realized in the intelligible world, then so act as if the maxim of your action should become, by your will, a law of nature” (Lucien Goldmann, Mensch, Gemeinschaft und Welt in der Philosophie Im m anuel Kants [Zurich and New York: Europa Verlag, 1945], pp. 173-74). 63
Studies in the Philosophy of Kant
43
freedom a category of nature, this Critique explains how the realm of nature is conceivable as harmonious, or at least compatible, with the ends of freedom. But then there must be a purpose. T he eighteenthcentury optimists had little hesitation in seeing man and his welfare, considered eudaemonistically, as this all-justifying end, for purpose was still, in principle, a constitutive concept of their cos mology. But Kant could not identify this purpose with man as a part of nature. A regulative teleology and a formal morality, however, stand together: I f things in the world, which are dependent in their existence, need a supreme cause acting toward ends, then man is the final end of creation; for without him the chain of graduated ends would not be perfectly grounded, and only in man (but in him only as subject to morality) is there unconditional legislation with respect to ends. T h is alone makes him capable of being a final end, to which all nature is teleologically subordinated.64
It is in the light of this consideration that “ ends that are also duties,” in the language of the Metaphysics of Morals, are comprehensible. They are the ends of free, and therefore moral, beings. T H E C H A R A C T E R OF MAN
Man is considered in two roles in Kant’s ethics— man as the moral ideal and man as an empirical character trying to achieve personality. Man in the first role is the archetype for man in the second role. Empirical man always has an obligation to respect personality in the former sense, to serve and promote it. T his introduces a definitely teleological element into Kant’s other wise formalistic ethics. Can this “ moral purpose” be made consistent with Kant’s assertions that moral actions are actions done without regard to ends to be achieved? An affirmative answer to this question is self-evident when it is realized that the idea of purpose is itself involved in the doctrine of autonomy. We have seen in the preceding section that purpose and freedom are two aspects of a causality which is not mechanical. Rational nature exists as an end in itself, says Kant; any other end would be heteronomous. T his end is not imposed upon the self by an alien 64 Critique of Ju dgm en t, §84 (Academy edn., V, 435-36).
Studies in the Philosophy of K an t
44
universe but is the very nature of the self. This immanent teleology of moral obligation is expressed in an early fragment: “ If there is any science man really needs,* it is the one I teach: how he may properly fill the place assigned to him in creation. From it he may learn what he must be, in order to be a man.” 65 But how different from the ideal of man is actual man! Man is a creature of needs, the satisfaction of which gives pleasure. T his pleasure is the only natural incentive to action. Consequently, all theories of human nature which place the driving force of conduct in the satisfaction of needs are ultimately forms of hedonism. More over, ethical theories based on them are heteronomous, since they view reason at most as a guide to the successful attainment of the ends of desire and not as active on its own account. Happiness as an ideal is the satisfaction of all desires. Though technical imperatives can be formulated for the achievement of pleasure arising from the satisfaction of an immediate need, happiness is a concept of such fluctuating content that no definite imperative can be derived from it. T h e moral law, however, lays down a rule without exceptions, giving commands instead of counsels. Before its majesty the senses are humiliated. This humiliation awakens respect for the law. R e spect thus functions as moral feeling. It is the subjective aspect, not the cause, of morality. It is painful as an arrest of the natural inclina tion but akin to pleasure in presenting a moral goal which every man can achieve out of the inward resources of his rational nature, ir respective of his position in the world. T h e moral law, as a universal principle integrating the diverse forces in man’s nature, is the basis of what is commonly called “ character.” T he inner form of character is virtue, or “ the moral disposition in conflict” with the wayward and chaotic inclinations. In building his character, man develops a genuine interest in his own perfection as a moral goal. T h is in terest, combined with faith, gives strength and substance to character by awakening hope of the highest good in the world. Such are the central features of Kant’s account of human nature. T w o criticisms of it are sufficiently important to require exami nation: that Kant’s theory does not give sufficient scope to human emotions and is “ false to human nature” and that Kant’s supposition 65
Fragmente aus dem Nachlasse, Academy edn., X X , 45.
Studies in the Philosophy of K an t
45
that all natural motivation is egoistic and hedonistic forces his own theory into an extreme and untenable position. It is easy to dispose of the first criticism, once we decide what is meant by being “ false to human nature.” Kant expressed his im patience with criticism of what ought to be in the light of what is when he wrote his essay On the Saying, “ That May Be True in Theory But It Does Not Hold in Practice.” In ethics it is human nature itself which is on trial. When theory does not apply to prac tice, two courses are open: either we may do as the scientist often does and add to our theory those special conditions which will close the gap between abstract theoretical and concrete empirical propo sitions or we may force the fact, so far as it is under our control, to conform to the criterion we erect. Naturally, if ethics has any norma tive function, the latter must be the procedure of the moralist. This criticism, however, might better be rephrased as follows: Kant’s theory, because it is exclusively rationalistic, is not adequate to what enlightened moral persons find worthy in human nature. Human actions and virtues have an emotional setting which is not accidental to them but constitutes a part of their essential character, and this is ignored by Kant. T he critic can find evidence that seems to justify this criticism. There is a rigoristic “ Prussian” trend in Kant’s writings and in his personality as we know it through various anecdotes. This often causes a modern reader to lose patience with the great philosopher, and it has even caused another philosopher to make an invidious comparison between “ Kantian morals” and “ human morals.” Schiller, who knew better, satirized Kant in a verse famous as a paradigm of this kind of criticism: “ G ladly I serve my friends, but, alas, from inclination, A n d often I ’m troubled because I am not good.” “ N othing else will do: You must try to despise them, A n d with revulsion do what Duty orders you.” 66
Although this picture of Kant’s ethics is incomplete, it is easy to see how it came about. Almost everyone else at that time was accen tuating feeling at the expense of reason in morality; Kant’s aim, on the contrary, was to establish reason as the exclusively legislative 66
Schiller, “ Der Philosoph.”
46
Studies in the Philosophy of Kant
moral faculty. Therefore, even when he acknowledged the contribu tion of feeling to character, it was not his purpose to emphasize it, and for obvious polemical reasons he often slighted it. Nevertheless, feeling did not cease to play a role in his ethics after 177°- Schilpp has adequately shown that in the earlier period Kant was never an uncritical disciple of the Shaftesbury-Hutcheson school,67 and it is equally true that in his later period he was not the cold thinking machine he is often pictured. Though he never admitted that feelings could generate a genuine moral disposition, or desire for happiness could function as a moral motive, Kant was quite aware of the “ syner gistic” relationships between reason and empirical character. T h e careful reader of the Critique will readily perceive this, and I shall merely cite some collateral evidence from Kant’s other writings. In the 1770’s Kant wrote to Herz: T h e supreme ground of morality must not merely allow us to reach delight; it must itself delight in the highest degree, for it is not just a speculative conception, but must have a moving force. T h o u g h it is indeed intellectual, it must have a direct relation to the basic incentives of the will.6»
During the same period Kant was formulating his own concept of moral feeling and defined “ moral incentives” as inclinations arising from feelings which are self-consistent and therefore the basis for orderly, integrated activity.69 Such feelings might be called the material causes of morality. Alone they have no moral worth, but they are the material which reason fashions into character. Later, in the Critique of Judgment, Kant describes and commends certain sentiments which attune man to morality, compares the “ moral feeling” to the aesthetic feeling, and regards the beautiful as a “ symbol” of the moral.70 T h e Metaphysics of Morals, in which man as a sensuous being is always the center of attention, does much to counteract what is perhaps one-sided in the Foundations and in the 6? K ant’s Precritical Ethics, p. 35. 68 Academy edn., X , 138. 69 Reflections 6690, 6696 (Academy edn., X I X , 134, 135). [All references to the Reflections are by number as established in the Academy edition, vols. X I V - X I X . T h e later volumes of fragments and unpublished manuscripts, vols. X X - X X I I I , are cited by page number only.] 70 Critique of Judgm en t, § §86, 29, 59, etc.
Studies in the Philosophy of K ant
47
second Critique, where the argument attempts to establish an objec tive law valid for every rational being as such.71 A long footnote in Religion within the Limits of Reason Alone supplies further evidence— if it is needed— that Kant’s refusal to place feeling in the center of ethical consideration does not mean that it can be disregarded. Schiller had objected to Kant’s rigorism and lack of emphasis on the “ graceful” attributes of morality, and Kant answered that “ grace” can come after duty but should not re place it or be mixed with it as one of the moral motives. He then says: N ow if one asks, W hat is the aesthetic character, the temperament, so to speak, of virtue, whether courageous and hence joyous or fear-ridden and dejected, an answer is hardly necessary. T h is latter slavish frame of mind can never occur without a hidden hatred of the law. A n d a heart which is h appy in the performance of its duty (not merely complacent in the recognition thereof) is a mark of genuineness in the virtuous spirit— of genuineness even in piety, which does not consist in the self-inflicted torment of a repentant sinner (a very ambiguous state of mind, which ordinarily is nothing but inward regret at having infringed upon the rules of prudence), but rather in the firm resolve to do better in the future. T h is resolve, then, encouraged by good progress, must needs beget a joyous frame of mind, without which man is never certain of having really attained a love for the good, i.e., of having incorporated it into his m axim .72
Without being a hedonist, one cannot go further. T he second objection to Kant’s psychology is that, as psychology, it is too narrowly hedonistic. In rightly denying the ethical adequacy of nonrational incentives, he came to deny to empirical motivation any value whatever— a mistake indicative of a defective theory of human desires. Kant’s psychology on this question, we admit, is often confused and occasionally wrong. T o begin with, it is obvious that Kant has not learned the lesson Butler taught Kant’s British contemporaries,73 There is an exhaustive collection of passages showing the relations between morality and happiness, duty and inclinations, in Paton, T h e Categorical Im pera tive, Appendix to ch. 3, pp. 55-57. 72 R eligion within the Lim its of Reason Alone, trans. Greene and Hudson, p. 19 n. Italics omitted. 73 There is no evidence that Kant had ever heard of Joseph Butler. 71
48
Studies in the Philosophy of Kant
namely, that inclinations are disinterested in respect to persons. In clinations are, as Butler rightly saw, aimed simply at their own satis faction, and where this primary satisfaction lies— in myself or in others— is not determined by the nature of the particular passion but by its circumstances, its direction either toward myself or toward others. The satisfaction “ belongs” to me in the sense that it is my desire which is satisfied, but this does not imply that it was a desire for my satisfaction. Rather, whatever satisfaction I get out of the action is due to the fact that my desire has met its satisfaction. T o infer that the desire itself was a desire for my satisfaction simply and immediately is to equivocate on the term “ satisfaction.” Still another equivocation is involved in the argument. “ Satisfac tion” may indicate either the fruition and fulfillment of a need, or it may mean simply pleasure or agreeableness. Now it is true that pleasure is usually associated with satisfaction in the former sense. When, however, it is rightly stated that satisfaction is the goal of a need or a desire and of the action they engender, it is often errone ously supposed that satisfaction is being taken in the second sense, and the argument for hedonism becomes a tautology. In the former sense, “ Satisfaction is the object of desire” is a definition; in the latter sense, however, it is an empirical proposition with little or no psychological evidence to support it. Yet most arguments for psychological hedonism depend upon interpreting the subject of the proposition, “ Satisfaction is the object of desire,” as equivalent to pleasure. Kant himself generally does this, and he can find no ma terial for a precept which is not thus reducible to pleasure. In order to break the bond of pleasure, therefore, he is forced entirely out of the empirical realm. Nowhere in Kant is there a clear-cut analysis of the pattern of desire. He generally argues that all principles built on empirically given interests are finally forms of the principle of self-love and one’s own happiness; but, in the light of Butler’s conclusive analysis, this argument must be rejected. However, the consequences of these errors are not such as to re quire any fundamental revision of Kant’s position as a whole. Butler’s analysis may serve to refute Hobbes, but not Kant. For K ant’s objec tion to the principle of private happiness is not that it is selfish in a disparaging sense. It is rather that this principle gives no universal precepts and can furnish no transcendental justification to the moral
Studies in the Philosophy of Kant
49
imperative. T he whole problem of selfishness lies this side of the essential moral question which Kant alone among his contemporaries was discussing. T he entire argument on altruism versus egoism is not, according to Kant, a moral problem at all; at most, it belongs to casuistry. T he same is true of the consequences of the second equivocation. For though Kant’s argument for psychological hedonism as a purely descriptive theory is invalid, his rejection of empirical motivation in morals is not predicated upon the specific character of the goal, wrhether it be pleasure or satisfaction in some broader sense. He re jects pleasure and the pursuit of happiness not because either is ordinarily immoral per se but because it lies beyond their power, as empirical, to deliver moral imperatives. Character built on them has no stability or dignity. With respect to them, man is passive, not active and free. Hence the argument does not concern their specific features; their empirical status condemns them. Such equivocations would be fatal to any ethics derived from natural man. But, since Kant’s ethics is not empirically derived, these errors do not infect the system as a whole. R A T IO N A L F A IT H
Morality stands independent in Kant’s philosophy. T he contribu tions made to it by theology are at most supplements. In establish ing the principle of morality, even the “ highest good” — faith in which implies practical-metaphysical propositions— can be com pletely disregarded.74 Nevertheless, morality has metaphysical implications and at least one presupposition. If human reason were capable of theoretical certainty in metaphysics, and were able to demonstrate the falsity of the collateral propositions of theology and metaphysics, then morality would assert its claim in vain. Reason, however, lacks this capacity. Its impotence seemed to some of Kant’s contemporaries to open the gate to irresponsible speculation. T h e consequence was enthusiasm, fanaticism (Schwärmerei), and superstition. “ Flights of genius” start from subjective sentiments which are inadequate to attain objective status even in the world of sense— let alone in the 74 See On the Saying, “ That May Be T rue in Theory But It Does Not H old in Practice,” I (Academy edn., V III, 280).
5°
Studies in the Philosophy of Kant
supersensuous realm of speculation, where, Kant says, they beat their wings in empty space.75 Reason needs to supplement its conditional knowledge with unconditionally certain knowledge; this is both theoretically76 and practically necessary. Yet to mere theory this is impossible, as the Critique of Pure Reason shows. Kant’s argument against both camps of his opponents— those who anticipated positivism and those who drew their inspiration from Romanticism and its cult of feeling— is found in his essay on What Is Orientation in Thinking? Kant shows that reason has its own needs independent of those of feeling and that it is able to satisfy them by deducing the necessity of certain a priori metaphysical pos tulates. But the crucial point of his discovery is that apriority ex tends beyond the scope of knowledge; there is a moral necessity in certain propositions which we cannot claim for them as parts of knowledge. This moral necessity is subjective in the sense that it is not based on or directed to objects of knowledge but is objective in the sense that it is not restricted by particular conditions in any sub ject. With this concept of an a priori faith, Kant is victor over dogmatic theology, which claims knowledge of the supersensuous; over the philosophy of feeling, which bases metaphysics on our subjective and contingent wants; and over skepticism, which disclaims interest in the supersensuous. T h e consequences of all three are equally in adequate for practice and untenable in theory. We have seen that if reason could proceed in a purely speculative way, metaphysics would be merely an extension of physics, and freedom would be impossible. Thus dogmatic metaphysics, developed as a rigorous science, would destroy its chief partisans, the theologians. Aware both of m an’s need for metaphysics and of the impossibility of certainty in specula tion, the philosophers of Storm and Stress, like practitioners of Lebensphilosophie, today, let the a posteriori subjective wants of men guide their reason. Kant refused to do this on theoretical grounds, and in some of the most eloquent passages of his works he points out the ruinous practical consequences of such a tendency. While rational faith is compatible with a doctrine that denies 75 Critique of Pure Reason, B 9. 76 Reason’s need to extend itself even in speculation is ultimately a practical need. We have already seen that all practical principles are regulative, not con stitutive; here is the converse of that.
Studies in the Philosophy of Kant
51
metaphysical knowledge, a faith conjoined to a metaphysics which denies the object of rational faith (as Kant holds Spinozism to do) can be only an irrational faith. This degrades reason, and Kant ex horts his opponents to consider the practical and political conse quences of their irrationalism. No duty can be acknowledged unless reason is legislative, and, if reason does not legislate, force will re place it, nullifying the Romanticist’s sublime dreams. Against the skeptics, Kant’s argument is less explicit, inasmuch as his whole philosophy is an examination of their limitations. Skep ticism is a dogmatic disbelief and suffers from all the ills of dogma tism, in addition to those brought about by failure to seek uncondi tional certainty or to justify it where it is found in both science and morals. Against all these theories, Kant puts forward rational faith— a faith which demands the same degree of assent required for theoreti cal knowledge and yet avoids the speculative claims of those who believe metaphysics to be a theoretical science. T he Supreme Being is the proper object of rational faith. We have seen already that the view that immortality is necessary to the highest good gave way to the belief that it is necessary to morality itself. A similar development occurs later in the postulate of the existence of God. T his development is in the direction of an increasing imma nence of God in the practical sphere. In the first Critique, Kant speaks of moral laws as commands associated a priori with promises and threats, an association which can occur only if there is a supreme intelligence as the moral governor of the world. And he states that to reject belief in God and immortality would overthrow the moral principles themselves.77 In the second Critique God appears as necessary to the existence of the summum bonum. God is the being that guarantees happiness in proportion to virtue; and moral laws, in whose fulfillment lies man’s worthiness to be happy, can be looked upon as divine com mands. T heir acceptance is defined as religion. Finally, a still further revision is made in the Opus postumum, in which God is identified with the moral law itself. “ God is not a being outside me, but merely Critique of Pure Reason, A 8 11 = B 839, A 828=B 856. See also Fragment 6674 (Academy edn., X I X , 130), which states that all duties disappear if there is no God, and Fragment 6858 (X IX , 181), which states that without religion there is no genuine incentive to morality. 77
Studies in the Philosophy of Kant
52
a thought in me. God is the morally practical reason legislating for itself. Therefore there is only one God in me, about me, above me.” 78 Kant was aware of the antinomic relationship between belief in a commanding, rewarding, and forgiving God and the autonomy of the moral agent. This is undoubtedly the source of his final explicit re jection of any theology but a moral one and of his making that theol ogy not so much an adjunct to his ethics as identical with it. God identified with the moral law is not the danger to moral autonomy that a God rewarding morality would be, for belief in the latter may destroy purity of motive. Since it is practical and not speculative reason which finally war rants Kant’s use of the concept of God, theology for him can have no theoretical content, and religion is only the attitude of performing all duties as divine commands. T h e judgment, “ There is a God,” is not a theoretical judgment; it is not a hypothesis in a theoretical con text. It is a practical postulate, a point of orientation. In the ordinary sense of the word “ know” we do not know that it is true,79 yet there is an a priori guaranty for it. Nor would such knowledge, were it possible, be desirable, for it would be predicated upon an extension of mere theory, of that aspect of mind which leads to amoral dogma tism. “ Inscrutable wisdom” in denying us this knowledge, says Kant, “ is not less worthy of veneration in respect to what it denies us than in what it has granted to us.” 80 In view of these postulates which are conjoined to a denial of speculative metaphysics, attempts to label Kant’s philosophy as a whole should perhaps follow Richard Kroner,81 who considers Kant Kants Opus postumum, ed. Erich Adickes (Berlin: Reuther & Reichard, 192°), p. 819. 79 Critique of Pure Reason, A 82g=B 857. 80 Critique of Practical Reason, Academy edn., V, 148; trans. Beck, p. 153. 81 See Kants Weltanschauung , chs. 1 and 4. [I would not now (in 1963) be in clined to take so extreme a view. I have, rather, come to distrust all efforts to put Kant under one rubric to the exclusion of the others. Though one form of NeoKantianism, like Kroner’s Heidelberg interpretation, does exclude other forms of Neo-Kantianism, I am not now so inclined as I once was to see Kant through NeoKantian eyes— eyes often sharper for the issues of 1881-1900 than for those of 1781-1800; cf. my reviews of Heinz Heimsoeth’s Studien zur Philosophic Immanuel Kants, in Philosophical Review, L X V I (1957), 405-9; Gottfried M artin’s Im m anuel Kant. Ontologie und Wissenschaftstheorie, in Journal of Philosophy, X L I X (1952), 424-25; and Mariano Campo’s Schizzo storico della esegesi e critica Kantiana, in Philosophy and Phenomenological Research, X X I I (1961), 132-33.] 78
Studies in the Philosophy of Kant
53
a radical voluntarist, not simply because he made the will basic (as Schopenhauer and Nietzsche were to do) but because he made the rational will the true organ of philosophy itself. While other volun tarists have developed their philosophies into theoretical apprecia tions of the wrill, Kant places the will beyond all theory. In its limited function as theoretical reason it judges nature and attempts to judge itself; as practical reason it leaves this kind of understanding behind. In its true nature it acts but it cannot be known, for all knowledge is theoretical limitation.
II
KANT’S LETTER TO MARCUS HERZ, FEBRUARY 21, 1772 Let it be said at the beginning that Immanuel Kant was not a great letter writer. In a century as celebrated for the epistolary art as for that of conversation, Kant was a master of the latter but had neither interest in nor talent for the former art. His letters are always matterof-fact, invariably polite and reserved, and seldom significant for our understanding of his philosophy or revelatory of the deeper and hidden levels of his personal being. Most of them seem to have been written reluctantly and with procrastination. T h e great man knew his faults so well that he warns his correspondents that he may not answer their as yet unreceived letters; and he accordingly apologizes in advance while at the same time begging the favor of a reply! Yet there is one group of letters which is of extraordinary interest not only to the biographer of Kant but to the student of Kant’s philosophy. This is the set of about twenty extant letters written to Marcus Herz; and outstanding within that group are the three letters written in 1771, 1772, and 1773. T hey are perhaps not great letters, of the kind that anthologists love; still the judgment I have expressed in the first paragraph about Kant’s letters in general does not do justice to them. T hey are friendly, informative, and decently wellwritten. They tell us more of Kant’s health, his moods, his habits, and the history of his thought at the beginning of work on the Critiqu e of Pure Reason than any other documents we have. T h e one Mr. r e pr in te d f r o m
T h e Ph ilosop hical F orum (Boston U niversity), X I I I (1955),
9 6 -10 2 , by k i n d permission o f the Editor. T h e letter in the tran sla tion b y M r. A r n e U n h j e m was p u b lis h e d in the same issue, a n d is r e p r in te d in the A p pen d ix
to
this v o lu m e ,
w ith
the
permission
U n h je m .
54
of
the
Editor
and
M r.
Studies in the P h i l o s o p h y of K a n t
55
Unhjem has translated is the most important from the point of view of understanding the early history of the critical philosophy. A man to whom Immanuel Kant could write as we find him writing to Marcus Herz must have been an unusual man; Kant did not lightly give his friendship, yet through all the rococo of eighteenth-century politesse the genuineness of affection and mutual respect of these two men is clearly seen. Who was Marcus Herz? He was born in Berlin in 1747 and died in 1803, the year before Kant’s death. He was a student in Königsberg from 1766 to 1770, and during the last wreek of his residence there he was the “ respond ent” or “ public defender” of the inaugural dissertation written by Kant on the occasion of his appointment to the professorship of logic and metaphysics. This was the first of many signs that have come down to us of Kant’s confidence in Herz. Herz returned to Berlin the following week with copies of Kant’s dissertation and with letters from Kant to the leading men of learning in the capital: Lambert, Sulzer, and Mendelssohn. This not only assured early and expert attention to Kant’s dissertation (which it would have received from these men anyway), but it opened doors to Herz that might have been closed and led to his brilliant career in the social and intellectual life of Berlin. He studied medicine in Halle, where he took his degree in 1774, and returned to Berlin where he opened his medical practice. By about 1776 he had begun to give public lectures on the philosophy of Kant. Among his most constant hearers was Zedlitz, the minister of education, to whom the Critique of Pure Reason was later dedi cated. In 1779 Herz married Henriette de Lemos, daughter of a leading Jewish physician in Berlin; and from that time onwards his home was a meeting place of the best minds in Berlin; literary and scientific lights revolved around the beautiful and witty Henriette, the perfect hostess, and Marcus, the popular physician and, as it were, Kant’s ambassador to Berlin. An interesting glimpse of the Herz family, with whom the Humboldt brothers studied philosophy, electricity, and Hebrew in the 1780’s, is given in Helmut de T erra’s life of Alexander von Humboldt.1 T h e letters of those who knew Herz confirm what we discern in 1 H elm ut de T e rra , H u m bo ld t: T h e Li f e and Times of Alexander von H u m boldt (New York: Alfred A. Knopf, Inc., 1955).
^6
Studies in the Philosophy of Kant
his correspondence with Kant, that he was a modest, industrious, open-hearted man with real philosophical talent. Mendelssohn wrote Kant, soon after he had met Herz, that he “ possesses a bright mind, a tender heart, a well-controlled imagination, and a certain subtlety of spirit which seems characteristic of the [Jewish] people.” 2 T his was the common judgment, both before and after he attained his high standing in Berlin. Many commented on his ugliness; Henriette recollected years after she had first met him that the “ spiritual [geistreicher) expression of his face” could not make her forget his ugliness.3 We know little of the personal relations between Kant and Herz while Herz was a student. Herz himself, in his first letter to Kant, gives all the credit to Kant for his advancement in learning, which, with his characteristic modesty, he says is as nothing compared to that of others, but is infinite when compared with his own igno rance before he knew Kant. There must have been a firm basis of friendship between master and disciple; and after Herz’s departure from Königsberg there were also other grounds for initiating the correspondence— old time’s sake alone never caused Kant to write a letter. Among the reasons I think the following were most important. First, Herz understood and sympathized with Kant’s philosophy perhaps better than anyone else during the 1770’s. This is shown by his role at the inauguration, by his own publications (on one of which, his Observations on Speculative Philosophy, Kant compliments him in this letter), and by his Berlin lectures. Kant perhaps felt the need of communicating his developing thoughts to someone who could understand them and who was enthusiastically active in spreading them. Second, we must remember that Herz was a physician, that K ant’s health was poor, and that, perhaps even more important, Kant seems to have been strongly hypochondriac. In his letters there are ex tended descriptions of his symptoms and of the treatments he has tried, and requests for advice. When he was not describing his own illnesses, he was asking for advice for his friends, or even sending his own practical advice on regimen through Herz to Mendelssohn. Herz 2 T o Kant, December 25, 1770, Academy edn., X , 103. 3 From Vorländer, Kant, der Mann und das Werk (2 vols.; Leipzig: Felix Meiner, 1924), I, 209.
Studies in the Philosophy of Kant
57
consulted his colleagues, sent counsel and the latest medical books. Kant was perceptive of matters now summed up under the name of psychosomatic medicine; he was especially concerned with the rela tion between his work and his health and how they reciprocally in fluenced each other; and he playfully admonished Herz: “ Keep up your study of the multiplicity of men’s constitutions. My own would be completely upset by a physician who wasn’t also a philosopher.” 4 Thirdly, there was the position Herz later occupied in Berlin. His relations with men like Biester, Mendelssohn, and Zedlitz made Herz a source of information about the learned world and the men in it whom Kant admired but some of whom he never met. Herz was able to use his influence in favor of others who were seeking positions, and more than one student of Kant availed himself of Herz’s inter vention. And Herz, being in contact with the best minds in Berlin, was in a position to call some of the best young men, e.g., Maimon, to Kant’s attention. Thus, there were solid bases, both in fact and in sentiment, for the long continued, though never copious, correspondence between Kant and Herz. After the 1770’s there is little of philosophical in terest in the letters Kant wrote to him, and by 1790, Herz’s medical practice was so large that he had given up his philosophical work. But there is abundant evidence of the esteem and affection the two men felt for each other even when their correspondence lapsed. I turn now to consider specifically the letter of February 21, 1772. It contains much of interest about Kant’s personal life, and it reports on a stage of his mental development that is of interest to anyone who wishes to trace the path from the Dissertation of 1770 to the Critique of 1781. T his letter, written before Kant had succeeded in formulat ing the most distinctive doctrine of the Critique , gives us a picture of his state of mind just before his “ recollection of David Hume . . . interrupted [his] dogmatic slumber and gave [his] investigations in the field of speculative philosophy a quite new direction.” 5 Even before Herz’s departure, he writes, he had been working on a treatise on ethics and aesthetics, in which the distinction between 4 Letter of 1773, Academy edn., X , 144. 5
Prolegomena to Any Future Metaphysics, Introduction (Academy edn., IV,
260; trans. L. W. Beck, “ Library of Liberal Arts,” No. 27 [New York: Liberal Arts Press, 1950], p. 8).
5«
Studies in the Philosophy of Kant
the sensible and the intellectual would be basic. In his letter of the previous year, he had mentioned the title, “ T h e Limits of Sense and Reason,” and in the letter before us he gives an outline of its major parts. T he mention of the “ Critique of Pure Reason,” later in the letter, may not denote a book planned under that title, as Kant often used these words to refer to his whole philosophic project and pro gram, even after the book called Critique of Pure Reason had been published. But while the letter leaves Kant’s literary plans in some obscurity, it makes perfectly clear that he was at this time concentrat ing on a problem of theoretical, not practical, philosophy, the solu tion of which would be propaedeutic to the metaphysics of nature and of morals that he had had in mind even before the Dissertation. T he letter describes the problem very clearly. T h e Dissertation had distinguished intellectual from sensible concepts, and understand ing from intuition. It ascribed to the latter, even when pure, the power of knowing things only as they appear, and reserved to the former the prerogative of giving us knowledge of things as they are. Kant had been criticized, as he tells us, by Lambert and Schulze for not according objective metaphysical reality to space and time; but with respect to them, he never wavered in his doctrine, which under went few changes between 1770 and 1781. Rather, he is concerned now with a problem which his critics seem to have completely over looked, but one which stands out before the modern reader (who has the inestimable advantage over these able men of having read the Critique of Pure Reason!): how can intellectual concepts refer to objects, since these concepts are not produced by the objects and the objects are not produced by our understanding? T h at they do so is the doctrine of the Dissertation, of Kant’s rationalistic colleagues, and of this letter. But Kant is rightly puzzled, apparently for the first time, as to how they can have objective validity. In spite of con fessing that the problem “ is still left in a state of obscurity,” Kant rather inconsistently says| that he has made a discovery that will permit him to answer this question and finish that part of his work within three months. This discovery was a systematic formulation of the table of cate gories. He says that he has discovered a way to “ reduce the tran scendental philosophy . . . to a fixed number of categories.” Already he is criticizing Aristotle’s list for being unsystematic; his categories,
Studies in the Philosophy of Kant
59
on the other hand, follow “ a few fundamental laws of the under standing.” (In the Critique, of course, there is only the form of judg ments which determines the number and character of the categories.) In this paragraph, Kant is reporting the first formulation of what later came to be known as the “ Metaphysical Deduction of the Cate gories,” and Kant apparently at this time assumed that it provided a solution to his problem. Of course, it is easy for us, who have the benefit of the finished Critique, to see that it was not the solution but only a clearer state ment of his problem. T h e letter to Herz seems to have been written to report that the worst of his problems had been solved; but we see that he was hardly even at their beginning. This letter is really ask ing for what is later called a transcendental deduction of the cate gories; what it reports is only the metaphysical deduction in an early and incomplete form. And when Kant is later able to provide the transcendental deduction, the question that motivated this letter is seen to be based on a misconception. T he categories do not give knowledge of things as they are, but only of objects of experience; Kant’s subsequent discovery may best be summarized by saying that he found that the question as formulated here does not admit of an answer under the supposition that a priori synthetic judgments have reference to noumena instead of phenomena. T o this discovery Kant was led by recollecting Hum e’s discussion of the limits of one intel lectual concept, causation; he generalized Hume’s problem and extended it to all categories, and he found that intuition was a necessary condition of every synthesis of concepts. T he half-skeptical, half-speculative position of the Dissertation and of this letter came to be rejected, we may suppose, during the few months after February 1772, during which Kant had hoped to finish his labors on this problem. The proper historical significance of this letter in the development of the Critique of Pure Reason has been subject to much debate. It is concerned with the issue as to whether the Critique, and especially the deduction of the categories, is a “ patchwork.” Did Kant put the Critique together out of manuscripts prepared, in what he then be lieved to be nearly final form, at various times between 1770 and 1781? Are there thus in the finished Critique various strata, represent ing, like strata of rocks in the earth, different periods of its history?
6o
Studies in the Philosophy of Kant
And if this is so, do we not have a perfect explanation of what appear to be gross inconsistencies in the finished work? T h at the Critique was composed in this way; that the apparent inconsistencies are real inconsistencies reflecting the slow development of Kant’s thinking from the point of view of the Dissertation towards views fully de veloped only in the second edition of the Critique in 1787; and, specifically, that this letter to Herz is a fragment of one of these strata whose next outcropping is in §13 of the Critique, where it is in obvious contrast to other strata originating after his “ recollection of Hume” — these are theses of the theory formulated by Vaihinger and defended and elaborated by Kemp Smith.6 T h at §13, on the contrary, represents a stage, even in 1781, in the setting of the problem of the deduction and not a solution rejected late in 1772 but revived later even though incompatible with that offered in other parts of the de duction, is the thesis of opponents of the patchwork theory.7 For both parties to this controversy, the letter to Herz is important because of the central and at least allegedly equivocal position of §13 of the Critique; the interpretation of §13 is a test case for the patchwork theory, and the letter to Herz is the most important single piece of evidence. All students of Kant, therefore, should be acquainted with this letter, and many will be properly grateful to Mr. Unhjem for his graceful and accurate translation. 6 Norman Kemp Smith, A Commentary to Kant’s ‘ Critique of Pure Reason’ (2nd edn.; London: Macmillan & Co., 1923; reprinted New York: Humanities Press, Inc., 1950), pp. 219-70. 7 For example, H. J . Paton, Kant’s Metaphysic of Experience, I, 323, 328-39; “ Is the Transcendental Deduction a Patchwork?” in Paton’s In Defence of Reason (London: Hutchinson & Co., 1951), pp. 65-90.
KANT’S THEORY OF DEFINITION i In most contemporary writings on the distinction between analytic and synthetic propositions, an analytic proposition is defined as one that follows from an explicit definition by rules of formal logic. If, as is usual, it is assumed that all definitions are nominal or stipulative, and further that all a priori propositions are analytic, it follows that the necessity of an a priori proposition is linguistic in origin and scope. T h e original distinction between analytic and synthetic proposi tions, however, was drawn by Kant, who did not make any of these three assumptions. Confusion arises through discussing, in Kantian terms, a distinction whose modern usage differs widely from that of the author of the distinction; discords are produced by Kantian tones in otherwise empiricistic harmonies. Sometimes one or more of the three doctrines mentioned above is attributed to Kant him self,1 or more often it is argued that the Kantian doctrine is impor tant and plausible only when seen as anticipating and preparing the way for the more recent doctrines. Either of these tactics keeps Kant’s r e pr in te d
, b y permission o f the Editors, from
T h e Philosop hical R ev iew ,
L X V (1956), 1 7 9 - 9 1 .
1 T h re e widely scattered specimens are: (a) “ La notion kantienne du jugement analytique semble d ’exiger que les concepts soient d ’une part absolument suscepti bles d ’une définition unique, rigoreuse et sans aucune ambiguité, et que d ’un autre côté leurs définitions soient susceptible d ’être analysées sans q u ’on aboutisse à des jugements synthétiques” (Paul T an n e ry , B u lle t in de la Société Française de P h ilo so p h ie [1903], 124); (b) “ Kant scheint bei der Einteilung der Urteile in analytisch und synthetisch von der Fiktion auszugehen, dass auch die nicht-mathe matischen Begriffe definiert werden können” (K. Marc-Wogau, Theoria, X V I I [19 5 1]» (c) “ T h e distinction . . . is easy and clear as long as we deal with merely stipulated or nominal definitions, as Kant seems to have supposed we co u ld ” (R. E. Gahringer, Journal of P h ilo so p h y , LI [1954], 435).
62
Studies in the Philosophy of Kant
own doctrines from teaching us anything important and distinctive by obscuring what was unique and original in them but has since been forgotten or neglected. My purpose here is to try to show the relationship between K ant’s own views of definition and of analytic judgment. I shall suggest that the interpretation of his analytic judgments as those based upon definitions is without historical warrant. T his raises the question whether modern disputes about the possibility of a priori synthetic propositions, in which the theory of definition plays a decisive role in the formation of criteria for analyticity, are really discussions of the problem to which Kant devoted the first Critique. II
T o define, according to Kant, means to present the complete con cept of a thing within its limits and in its primary or original char acter. A complete concept is one with a sufficiency of clear predicates for the entire concept to be distinct; and the predicates stated are primary or original in the sense that they are not derived from other predicates included in the definition. T h e predicates must, in other words, be primitive and coordinate; no derivative and subordinate predicates are admissible in a definition, for otherwise the definition would require proof.2 If a definition does incorrectly contain deriva tive predicates— properties instead of essentialia— it is lacking in precision. Definition is a “ sufficiently distinct and precise concept (conceptus rei adaequatus in minimis terminis, complete determinatus).” 3 T he definition of “ definition” that Kant gives here leads him to deny the name “ definition” to many sentences commonly so called. It is reached partly by an analysis of usage, and partly by a decision which makes the concept more precise: “ There are definitions of concepts which we already have but which are not correctly named. In these cases, it is not that the meaning of a word is analyzed, but that a concept, which we already possess, is analyzed; and then it must be particularly shown what name properly expresses it.” 4 2 See Critique of Pure Reason, A 7 2 7 = 6 755 and note. 3 Logik. Ein Handbuch zu Vorlesungen, §99 (Academy edn., IX , 140). H ere after cited as Lectures on Logic. 4 Reflection 3003 (Academy edn., X V I, 610).
Studies in the P h i l o s o p h y of K a n t
63
Kant distinguishes two major and independent divisions of defini tions: into analytic and synthetic, and into nominal and real. A definition is analytic if it is of a given concept; synthetic if of a concept made or synthesized by the definition itself.5 T h e former makes a concept distinct, the latter makes a distinct concept.6 Under each of these major divisions, there is a subdivision: the concept de fined may be given or made a priori or a posteriori.7 An analytic definition states the original analytic predicates of the thing defined. An analytic predicate is a partial concept of a thing actually thought in the concept of the definiendum.s Thus an analytic definition is an analytic judgment containing no subordinate predicates. A synthetic definition, however, contains synthetic predi cates, predicates whose union first establishes a distinct concept of the definiendum. T h e other major division is between nominal and real definition. Kant does not draw this distinction as one between the definition of a word and the definition of a thing; because he regarded the con cept, rather than thing or word, as the definiendum he was prevented from using this formula of the distinction. Rather the difference lies in the content of the definiens and in the methodological function of the two kinds of definition. A nominal definition states the logical essence of the concept of the thing, or serves merely to distinguish this thing from others. If it does only the latter, it is called a diagnos tic definition in contrast to a definition stating essential primitive predicates.9 T h e logical essence, stated in nominal definition, is the original primitive concept of all the essentialia ;10 the diagnostic defi nition may state only the irreducible minimum of some easily recognized attributes or properties, sufficient as a criterion in a dichotomous classification by a pass-fail test. A real definition not only puts one word in place of others, but the definiens contains a clear mark by which the object can be. recognized and by virtue of which the defined concept is shown to have “ objective reality”— by which it is shown that there is a defined 5 Lectures on Logic, §100. 6 Ibid., In troduction, V I I I (Academ y edn., I X , 63); see also Reflection 2929.
7 Lectures on Logic, §101. 8 Ibid., A c a d e m y edn., I X , 59.
9 Reflections 2994, 3003.
10 Lectures on Logic, Introduction, V III (Academy edn., IX , 61).
64
Studies in the Philosophy of Kant
thing.11 (The diagnostic definition does this, but not by stating the diagnostic symotom as an essentia of the thing.) Real definition, therefore, is a part and not merely a tool of knowledge. Real defini tion states the real essence constituted by real predicates, not merely by logical predicates included (“ already thought in” ) the concept of the subject. A synthetic predicate is a determination (Bestimmung ) not con tained in the subject-concept but enlarging it; it is not found by analysis. It determines a thing, not merely its concept. A n y th in g w e please can be m ade to serve as a lo g ical p red icate ; the subject can even be p red icated o f itself; for lo gic abstracts fro m a ll co n tent. B u t a determ ining p red icate is a p red icate w h ich is a d d e d to the con cep t o f the sub ject an d en larges it. C o n se q u e n tly it m ust n o t be a lre a d y co n tain ed in the co n cep t.12
A real definition, therefore, is always a synthetic judgment, even though the real definition, as definition, may be analytical and is analytical if the concept is given.13 Real predicates are never arbitrarily synthesized into a logical product called the essence; in every case the determinations are not purely conceptual but intuitive representations. General logic is concerned only with the logical essence or predicates; or, rather, in abstracting from all contents it treats determinations as if they were logical predicates. But knowledge of things requires knowledge of and through determinations, not merely the mouthing of their names, and this knowledge is knowledge of the real possibility of the object through a specific determination as both its ratio essendi and ratio cognoscendi.14 We find the logical essence by reflecting on the predicates which constitute or are made to constitute the nominal definition; for real essence, we seek data from experience or intuition 11 Critique of Pure Reason, A 241-42 n. 12 Critique of Pure Reason, A 5g8=B 626, trans. Kemp Smith; cf. Reflection 405513 Reflections 2955, 2994. 14 Lectures on Logic, Introduction, V III (Academy edn., IX , 61). Ratio essendi is, of course, to be understood not as having a bearing on the thing itself, corre sponding to the “ real essence” in Locke; for Kant, like Locke, admits ignorance of that. But ratio essendi may be applied also to the object of knowledge; and when its ratio cognoscendi and ratio essendi in part coincide, there is a priori knowledge.
Studies in the Philosophy of Kant
65
to determine whether and under what condition the object is really possible.15 This difficult and obscure matter is involved in the distinction between general and transcendental logic, and it cannot be made intelligible within the limits usually imposed on discussions of definition in formal logic. Kant is saying that in a real definition we do not merely equate a word with a logical product of arbitrarily chosen logical predicates, but we make at least a problematical existential judgment and state the conditions under which this judgment could be verified so that the d e fin ie n d u m will be seen to have “ objective reference.” There must be, in the d e f n i e n s , some determination or compound of determinations that can be “ cashed” in possible sensible (intuitive) experience. Its absence is the reason wrhy all definitions in speculative metaphysics are only nominal. Its specific epistemological character is the reason also why general logic does not deal with (or at least does not distinguish) real defini tions, since general logic disregards the transcendental difference between a predicate and a determination; and neglect of this differ ence is, finally, the reason why logic, when used as an organon in metaphysics, develops into dialectic. T he notion of real definition is not only excluded from general logic by Kant (though he dealt with it in his Lectures on Logic , which far exceeded the bounds he set up around the field of general logic), but is challenged on other grounds by most modern writers who reject the ontological distinction between essence and property.16 They ad mit, in any specific case, the distinction between an essential and an accidental definition, though on pragmatic not ontological grounds. Kant, in accordance with a tradition going back at least to the Port
ia Critique of Pure Reason,
A 2 i8 = B 265. Kant insists on the distinction be tween the two meaning^ of possibility as early as the Einzig möglicher Beweisgrund . . . (Academy edn., II, 77-78), the most important point always being that existence is not a logical predicate. T here are many things logically possible that are not really possible, because the nonconceptual condition that would show them to exist is not possible. T h u s “ a two-sided plane figure” is logically but not really possible, while a “ two-sided triangle” is not logically possible (Critique, A 2 2 i= B 268). T h e only kind of possibility subject to formal definition is logical (Critique, A 244=B 302). See Richard Robinson, Definition (New York: Oxford University Press, 195 °), PP- *54 - 5 5 -
Studies in the Philosophy of Kant
66
R o y a l L o g i c , 1 1 uses the distinction between nominal and real defini
tion to designate this other, quite different, distinction: a real définition is one from which other properties can be derived, while a nominal definition suffices only for “ comparisons” and not for “ deri vations.” Thus “ T he circle is a curved line all of whose parts can be made to coincide” is described by Kant as a nominal definition despite the fact that it prescribes an applicable test; he means that it is a definition that contains a predicate already derived, not an essentia; but instead of pointing this out, he calls it nominal.18 Having now set up the major divisions, I turn to the specific types of definitions resulting from the two independent divisions. T hey may be most easily seen from the following table. «
N O M IN A L
REAL
Analytic
Synthetic
L O G IC A L D E F IN IT IO N
D E C L A R A T IO N
A p rio ri E x p o sitio n
A po steriori D escrip tio n
A p rio ri C o n stru ctio n
A posterio ri In v e n tio n
1) A n a ly t ic n o m in a l d efin itio n . Kant says little about this, and even that little is confusing. Because it is in any case of small im portance to our inquiry, I shall not undertake to examine the various confusing statements he makes, but merely list the passages for the interested reader.19 2) Syn thetic n o m in a l d efin itio n . Such a definition is a stipulation or a “ declaration” of an intended usage, the concept being created by the definition. Since they are not determined by experience or by analysis of a given concept, Kant says that such definitions are a priori synthetic, not realizing, perhaps, the inappropriateness of this adjective to what is not a proposition or judgment proper.20 3) A n a ly t ic real d efin itio n . A definition of this type states the defining predicates of a given concept known to have objective validity, and it contains the synthetic predicate ( B e s tim m u n g ) which Arnauld, Port-Royal Logic, Part I, ch. 12. 1 8 Reflection 2916; see also Reflection 2995. He does, however, point out the real infirmity of this definition in the Critique, A 732=!* 760. 19 Lectures on Logic, §106, Note 2; Reflections 2918, 2931, 2963, 3004. 20 Reflection 3007. Such a definition— in the case under discussion, it happens to be Kant’s definition of analytic judgment— cannot be in error. See Vber eine Entdeckung . . . , Academy edn., V III, 232. 17
Studies in the Philosophy of Kant
67
gives the defined concept this objective reference. Nevertheless, upon investigation it turns out that any attempt to state such a definition fails to meet the formal requirements of definition, with respect either to completeness or precision. If the concept is given a priori, we cannot be sure that we have a complete analysis of it into its coordinate predicates. A concept given a priori may include “ many obscure representations, which we overlook in our analysis, although we are constantly making use of them in our application of the concept.” Therefore the complete ness of a proffered definition is never more than probable, and rather than call such an indefinite analysis by the name “ definition,” Kant calls it an “ exposition.” 21 If the concept is given a posteriori, its analysis suffers from the same infirmity mentioned above in discussing the definition of an a priori concept. Such a concept has no precise and complete analysis, for the concept itself is not a fixed union of predicates. It is variable, depending upon the scope of the experience we classify under it. Kant in one place says that it cannot even be nominally defined.22 A statement of the attributes and properties of a thing meant by an empirical concept is at most a description, which is not held to rules of precision and completeness; description provides many truths which serve as the “ material for definition,” 23 the definition itself being only an ideal. 4) Synthetic real definition. It is obvious from the very name what falls here: such a definition must not only make a concept, but must show its real possibility by including the Bestimmung which is its ratio essendi and cognoscendi. If the synthesis is of pure concepts, the real determination must be a character of pure intuition; if of empirical concepts, the real determination must be an empirical intuition. T h e synthesis of pure concepts is a construction. Construction is the presentation of a concept through the spontaneous production of its correspond ing and verifying intuition. Concepts, if pure, can have an a priori representation only in pure intuition; and such representation is definition as this occurs in mathematics. If the concept is empirical in Critique of Pure Reason, A 728-29=!* 756-57, trans. Kemp Smith. 22 Reflection 2992. 23 Lectures on Logic, §105, Note 3 (Academy edn., IX , 143). 21
Studies in the Philosophy of Kant
68
its components, we have the presentation of an actual empirical in tuition not through the productive imagination alone but through a change effected in the real world. A definition of such a concept may be genetic, telling us how to make a corresponding object,24 and the devising of the object is proof that the concept has real objective possibility and is not chimerical. Kant calls such a defini tion (as of a ship’s chronometer) a “ declaration of a project” 25 or an “ exposition of appearances.” 26 Since “ exposition” and “ declaration” are both used in other senses, I have called this, in the table, “ inven tion.” In mathematics, we make a concept by synthesis. “ T h e mathe matician in his definitions says, Sic volo, sic jubeo .” 27 But in spite of the modern sound of this statement, mathematical definitions for Kant are real, not nominal. Mathematical entities are not arbitrary logical products of compatible logical predicates; the concepts have objective validity (in pure intuition) shown through the presenta tion of the corresponding determination. If the presentation is a product of the productive imagination, the construction is called schematic or pure, as of a figure (no matter how roughly drawn) used in a geometrical proof. Such a figure is not used empirically, and the actual drawing of it is not a part of the science of mathematics but belongs to art. Kant calls the empirically made sketch “ technical construction” 28 and, indeed, it is like the “ invention” of any empiri cal object. Mathematics is the only science able to construct its con cepts a priori, and only by construction can we achieve completeness and precision in knowledge. Therefore mathematics is the only science which contains proper and strict definitions.29 Kant often speaks of synthetic definitions, including mathematical definitions, as willkürlich. T he word willkürlich , ordinarily trans lated as “ arbitrary,” does not, however, suggest the caprice sometimes understood in the word “ arbitrary” ; “ arbitrary” does not mean 24
See Reflection 3001.
Critique of Pure Reason, A 729=!* 757. Lectures on Logic, §102. 27 Reflection 2930 (Academy edn., X V I, 579); see also Inquiry on the Distinct ness of the Principles of Natural Theology and Morals (hereafter referred to as Prize Essay), Academy ecUj., II, §1 (trans. L. W. Beck, in Kant’s Critique of Prac tical Reason and other Writings in Moral Philosophy, p. 262). 28 über eine Entdeckung, Academy edn., V III, 192 n. 29 Critique of Pure Reason, A 729=B 757. 25 26
Studies in the P h ilo s o p h y of K a n t
69
“ random.” Arbitrariness, as it is now commonly interpreted, is not a feature of mathematical knowledge as Kant interprets it; mathe matical concepts are limited by the fixed conditions of intuition, just as empirical concepts are synthesized under the limits imposed by the actual content and order of empirical data. Kant contrasts willkürlich with empirisch, not, I think, with notwendig .30
III
I shall now consider the role that definitions play in the progress of knowledge, as this is described by Kant. T h e search for definitions of empirical concepts is justified by the technical demands for communication in relatively unambiguous language. We need to “ fix” the meaning of a concept from time to time, and we do so by nominal definition or declaration. Such definitions, if made too soon or especially if taken too seriously as a part rather than as an instrument of knowledge, distort inquiry by permitting logical analysis to usurp the place of empirical ampli fication. “ What useful purpose,” Kant asks, “ could be served by de fining an empirical concept, such, for instance, as that of water? When we speak of water and its properties, we do not stop short at what is thought in the word ‘water’ but proceed to experiments.”31 Description suffices; definition which aims at being more than nominal is a useless presumption. Turning from empirical to rational knowledge, Kant insists upon a sharp distinction between the methods proper to mathematics and those of philosophy. T h e mathematician begins with definitions and proceeds by a synthetic method (involving constructions) to his conclusions; his definitions cannot be false, and their only fault may be lack of precision, which is progressively corrected.32 T h e philos opher, on the other hand, must begin with concepts already given to him, though confusedly and without sufficient determinateness. T h e thing meant is not intuitively clear in the sign, as in the con cepts of mathematics,33 all of which are subject to construction in 30 31 32 33 278.
Lectures on Logic, §103, Note. Critique of Pure Reason, A 728=!} 756, trans. K e m p Smith. Reflection 2979. See Critique of Pure Reason, A 7 3 4 = 6 762; Prize Essay, A c a d e m y edn., II,
70
Studies in the Philosophy of Kant
intuition. T he symbols, such as a set of points representing a number, have their meaning “ on their face” ; whereas the philosopher must use his symbols only as poor representations of richer concepts. These he must analyze in order to compare their segregated characteristics with those originally intended by a ready-made concept used to render unanalyzed experience intelligible. A definition reached by synthesis in philosophy could only by accident be a definition of a concept which originally posed the philosophical problem to us. In mathematics there are few unanalyzable concepts, and they can be used with assurance according to explicit rules without any need for analysis. Analyses of concepts, if made at all, belong to the philosophy of mathematics rather than to mathematics itself. In philosophy, on the contrary, there are many unanalyzable and indefinable concepts, but we do not begin our work with them. We discover what they are only by the analysis of given concepts, which are not entirely clear and distinct. Thus (if he is fortunate) the philosopher ends where the mathematician begins, to wit, with indefinable elementary concepts and definitions of the concepts given in the beginning. Definitions in philosophy, therefore, are not the conditions of knowledge; they are what we hope to conclude with, not the raw material with which we begin. From these textual inquiries, we can conclude that definition does not play the dominant role in Kant’s philosophy that it does in later theories of analytic judgment. In only one field, mathe matics, does Kant admit strict definitions, and in mathematics it is possible to decide indubitably what is analytic and what is synthetic. In empirical knowledge, definition is only loose and informal, and we should expect what we do find, namely that decision on the character of specific judgments is variable and without great impor tance. It is a priori judgments outside of mathematics that Kant is chiefly concerned to establish, and of their concepts definition is impossible. Yet it is with respect to these judgments that it is of fundamental importance to distinguish the apriority of formal logic (analytic) from the apriority of transcendental logic (synthetic). Definition is not essential to certainty in knowledge. Quite apart from Kant’s belief that not all a priori knowledge is analytic, he does not even assert that analytic judgments are necessarily conse
Studies in the Philosophy of Kant
quences of definitions. Though he indicates34 that analytic judgments are deducible from definitions, this statement occurs in the reply to Eberhard, in a context supplied by his opponent; it is not his char acteristic way of stating the nature of analytic judgments. Definition would be a sufficient, but is not a necessary, condition for analytic judgments; we may have a priori knowledge of undefined concepts provided we can either exhibit the concept in pure intuition (sche matize it to give a basis for synthetic judgments) or give a partial analysis of the concept.35 And in three places,36 at least, Kant de scribes the way logical certainty in knowledge is gained, showing clearly that definition is given a secondary role. He tells us that we begin by analyzing concepts, expressing the analyses in analytic judgments, and only then organize these analytic judgments into definitions. Even so, definition requires a completeness and precision that is often an unattainable ideal; yet its absence does not jeopardize the analytic judgments already made. The Critique of Pure Reason is supposed to answer the question, How are synthetic judgments a priori possible? But if it is not pos sible to decide objectively whether a given judgment is synthetical or analytical, the entire Critique seems to be wasted effort. Can we make synthetic judgments and know that they will, as it were, remain synthetic while we examine their apriority? Or do not definitions grow and so extend their sway that a judgment once known only empirically can, under better definitions, come to be logically neces sary? Can we not agree37 that a “ synthetic a priori judgment” is a judgment with an ambiguous term, and that when we remove the ambiguity by definition we either remove the apriority or the syntheticity? This presupposes that analytic judgments are determined by definitions, and it at least suggests that definitions are arbitrary in such a manner that we have a choice as to whether the judgment in question will be made analytic or synthetic. Expressed in other über eine Entdeckung, Academy edn., V III, 229. Critique of Pure Reason, A 731= 1* 759; Prize Essay, Academy edn., II, 285; Lectures on Logic, §109, Note (Academy edn., IX , 145). 36 Prize Essay, Academy edn., II, 282; The False Subtlety of the Four Syllogistic Figures, Academy edn., II, 6 1; Prolegomena, §2 c 3 (Academy edn., IV, 273). 3 " With H. W. Chapman, Mind, N.S., L X I (1952), 391. 34 35
Studies in the Philosophy of Kant
72
ways, this is one of the oldest and probably the most common of all criticisms of Kant’s theory. T he difference which he thought was fundamental seems to be a subjective, shifting distinction, dependent upon how much one knows at a given time, and how one formulates what one knows. Very early in his use of the distinction Kant seems to have anticipated this objection38 though he gave no answer at that time, and for many years used the distinction as though com pletely oblivious of the objection. He does not seem to have realized its full force until he prepared his reply to Eberhard. Even then, in the published reply he does not come to grips with the problem; but in the working paper prepared under his direction by Schultz there is a passage39 which deals with the shifting of the line between the two types of judgments by the modification of definition. T h e passage is obscure, but I will try to describe what I think Kant would have said had he put it into shape for publication. Kant invites his opponent to add any attributes he wishes to a concept, so that whatever it is he wishes to prove he can prove by deduction, i.e., analytically. But then Kant asks him: How did you come to include in the concept precisely those attributes you needed in order to render previously synthetic judgments analytic? He cannot reply that he is giving a definition of the concept unless he can show that he is obeying the rules of definition in formal logic. T hat is, he must be able to show that the newly introduced attributes are logically independent of the old, yet invariably attached to the subject in experience, so that the conjunction of the old and new attributes has the same denotation as the original concept. A nar rower denotation will not do, for that means that a new concept has been introduced, not that an old one has been defined. Now in order to know the identity of the old and new denotation, he must know the connection of the independent attributes before stating them in a new definition; he must know this synthetically, for if they are analytically related the rule concerning the precision of definition is broken. Hence, definitions devised for the purpose of rendering synthetic judgments analytic are not real definitions, or in making them we must already know with certainty the synthetic judgment 38 39
Reflection 3928, dating from the late 1760’s. Rezension von Eberhards Magazin, Academy edn., X X , 408-9.
Studies in the Philosophy of Kant
73
they were designed to establish as analytic. If they are not real but only nominal definitions, then the problem of synthetic a priori knowledge (which Kant calls the metaphysical problem) is not touched by this exercise in logic.40 IV
In contrast with the views mentioned at the beginning of this paper, sometimes erroneously attributed to Kant, we have found that Kant’s views on the relation between definition and analytic judgment are as follows. While a judgment logically implied by a definition is analytic, analytical judgments are not necessarily or even usually known or justified by deduction from definitions. Ana lytic judgments are made by analysis of concepts which need not first be established by definition. Definition is a late stage in the progress of knowledge, being preceded by the analysis of given con cepts, expressed in analytic judgments. Because definition is a sec ondary and more or less adventitious element in Kant’s theory of the criteria of analytic judgment, the view that synthetic proposi tions can be rendered analytic by a change in definition is foreign to the distinction as Kant established and used it, and does not contribute to a solution of his problem of justifying a priori judg ments whose necessity is not that of formal logic. I have translated this passage and discussed the issue of the variability of the synthetic-analytic decision in “ Can K ant’s Synthetic Judgments Be Made Analytic?” [the following essay]. 40
IV CAN KANT'S SYNTHETIC JUDGMENTS BE MADE ANALYTIC? In the 1760’s, when Kant had first drawn his distinction between analytic and synthetic judgments, he made the following note: “ If one had the entire concept of which the notions of the subject and predicate are compars, synthetic judgments would change into ana lytic. It is a question of how much arbitrariness there is.” 1 T his ques tion has been asked repeatedly since that time, and the clear and unmistakable trend of the answers has been that the decision whether a specific judgment is analytic or synthetic is arbitrary or at least is dependent upon variable conditions of how much the judger knows about the subject of the judgment and on his arbitrary decision of the choice and formula of his definitions. In recent discussions of the distinction, analytic judgments are those that follow from explicit definitions by the rules of logic; and definitions are nominal or stipulative, to some degree arbitrary. If it is further argued, as is often done, that all a priori judgments are analytic, it follows that the distinction between a priori and a posteriori is likewise a shifting, arbitrary distinction. Kant, who first asked the question, seems to have decided very early that the line of demarcation between these two types of judgment was not variable or arbitrary. T h e purpose of this essay is to inquire into the reasons for his decision and to indicate some of its implica tions for his philosophy as a whole. A N A L Y T IC
AND
SY N T H E T IC
JUDGM ENTS
Judgment, for Kant, is a synthesis of representations, having objec tive validity. T h e synthesis must be in accord with some objective, r e pr in te d
, w ith the permission o f the Editors, from
( 1955 ), 168-81 . 1 Reflection 3928 (Academy edn., X V II, 350).
74
Kant-Studien, X L V I I
Studies in the Philosophy of Kant
75
normative rule, and not merely illustrate some contingent law of association. A representation, functioning in the synthesis of judg ment, is not just a brute given mental content, but is a mark of an object, its meaning fixed by a rule. Abstraction from the given com plexity of representations in consciousness, and the generalization that a particular kind of representation is the mark of a particular kind of object, are necessary in converting raw representations into marks which can be manipulated in knowing.2 Concepts are such marks functioning in knowledge; they are representations under an analytical (abstractive) unity through which they are discursive and not merely given sense contents. As concepts, they are not given; they are made concepts by being involved in a special attitude of in tention and the interpretation of data. All that we directly have of an object is such marks. Our original consciousness is a congeries of raw materials for concepts, and the business of consciousness is to refine and organize these representations, assigning to some of them the role of subjects and to others that of predicates in judgments which are their objectively valid syntheses;3 only as predicates of possible judgments do Vorstellungen serve as concepts, and only as containing representations under themselves do concepts refer to objects.4 Besides the analytical unity by which hie et nunc representations are made to serve as marks under a discursive concept (e.g., this quale at this time is seen as an example of a specific quality also instanced in another quale at another time), in order that there be judgment there must also be a synthetical unity through which the concepts (and their corresponding representations) are referred to the same object. This object may not be given at all, or if given it is given as only a still further complex of representations which refer to “ the same object” only by virtue of some precedent synthetical unity. T he syn thetical unity, which is a form and not a content of experience, is not given, but is prescribed to experience by a rule that requires a com mon focus of meaning of the several concepts that appear in a judg ment; if one such object is not meant by the various concepts, the synthesis of the concepts is a comparison, a setting of them side by 2 Reflection 2881. 3 See Reflections 3920, 4634. 4 Critique of Pure Reason, A 69=B 94.
Studies in the Philosophy of Kant
76
side, and not a judgment. This common object is called by Kant X, and the rule of synthetical unity means that the terms in a judg ment (concepts, derived through the analytical unity of representa tions), such as A and B, must be regarded as marks of X . Then through A and B we know X , and the cognition of X through A or B is a concept of X .5 A and B are, epistemologically, predicates of X, but one of them is made to serve as the logical subject and the other as the logical predicate. T he one called subject is directly related to X, the one called predicate is indirectly related to X in the judgment, though its occurrence in experience may be direct evidence of the existence of X (usually it is the wider concept, and is applied to a specific X only through the mediation of the subject concept).6 Thus, to summarize and make specific: when X is known through two con cepts related to each other in a synthetical unity, then a judgment whose form is given by a category or rule of this synthetical unity is established. If the rule is, for instance, the category of inherence and subsistence, the judgment reads, “ There is an X such that X is A and X is B .” If B is related to A directly by being included as a part of its con notation, so that “ X is A ” implies logically “ X is B ,” the judgment is analytic. In an analytic judgment, reference to X is otiose, and we say simply, “ All A is B ” where A and B are “ partial concepts” of X , and B is a constitutive part of A. But “ All A is B ” is an elliptical expres sion since A is a complex concept containing B. Fully expanded, therefore, the analytic judgment is the tautology, “ All A.B is B .” When B is a concept of X because it is a nota notae of X , i.e., a mark or constituent of A, we can speak of the judgment as one in which the certainty of the connection of subject and predicate is “ through identity.” 7 If the identity is explicit, the judgment is inconsequential. T h e important case is the one in which the identity is implicit, so that its explication “ widens our knowledge formaliter ” though not materialiter. B may be “ covertly contained in the con cept” 8 and not thought “ so distinctly and with the same (full) conReflection 3920. 6 The Progress of Metaphysics, Academy edn., X X , 274. 7 Lectures on Logic, §36 (Academy edn., IX , 11 1 ) . Kant objects to calling them identical judgments, however; see Über eine Entdeckung, Academy edn., V III, 244. 8 Critique of Pure Reason, A 7 = B 10. 5
Studies in the Philosophy of Kant
y>j
sciousness” as A . 9 It is an “ analytic attribute” of A contained in it and elicited from it by logical analysis.10 But it is essential that it be “ con tained in” A , so that the judgment is explicative, not ampliative, and independent of further experience of the X of which both A and B are concepts. Now if the decision on analyticity of a specific judgment could be based on a definition of the subject, it would be easy enough to determine whether the judgment is analytic. But Kant rejects this procedure, because he holds that “ definability” is a stricter condition than “ analyzability,” and that we can therefore make analytical judgments with concepts we cannot define. It is, in fact, through organizing analytic judgments that we gradually approach to defini tion,11 which is the end, not the beginning, of knowledge. Since Kant has so restricted the scope and value of definition, these statements about the inclusion of one concept in another are exceed ingly obscure. It seems that, without a stated definition, they can be understood in part only psychologically or phenomenologically. Speaking for the phenomenological interpretation is the emphasis upon what is “ actually thought” in the subject; speaking for a logical interpretation is the fact that analytic attributes may be uncovered and brought to light only by sustained inquiry, and are not present, in any phenomenological sense, in the thought of the concept of the subject. If we investigate each phrase in these passages, the possible con fusion of the two meanings is not removed. For instance, “ contained in” (enthalten in) was a logical term used by Kant’s contemporaries to describe predicates belonging to all individuals denoted by a concept.12 But Kant obviously does not mean it only in a logical # Prolegomena, §2 a.
Über eine Entdeckung, Academy edn., V III, 228-29. 1 1 Prolegomena, §2 c 3; see also Lectures on Logic, §109, Note; Prize Essay, Academy edn., II, 282; False Subtlety of the Four Syllogistic Figures, Academy 10
edn., II, 61. I have studied the relation between K ant’s theory of definition and the distinction between the analytic and synthetic in some detail in “ K ant’s Theory of Definition” [the preceding essay]. 12 Hans Vaihinger, Commentar zu Kants Kritik der reinen Vernunft (2 vols.; Stuttgart: W. Spemann, 1881), I, 258. “ Contained in ” is contrasted with “ contained under” (Reflection 3043). T h e latter, used in describing synthetic judgments, seems to mean for Kant what Vaihinger says was commonly meant by “ contained in.” See also Reflections 2896, 2902.
Studies in the Philosophy of Kant
78
sense, for then synthetic attributes would be contained in the subject concept, which he denies; “ contained in” seems to have reference to the subjective intension, and thus to have at least psychological over tones. But the words “ actually thought in the concept of the subject” are elsewhere given a strictly logical meaning, since Kant says that what is really thought in a concept is “ nothing other than its defini tion.” 13 I think we have to suspect here a fundamental failure on K ant’s part to distinguish the logical from the phenomenological aspects of thought. Where definitions or fairly complete analyses are available, he thinks of the distinction between analytic and synthetic judgment as logical; where they are not, but are rather the objects of search, he has recourse to a phenomenological criterion, by virtue of which he seeks definitions through analysis of what, in the plainest sense, is “ actually thought” in a concept or even “ contained in” a complex experience subject to subsequent analysis.14 While we cannot speak of two definitions of the analytic, and can1 at most say that the analytic has both a logical and a phenomenologi cal dimension, we can discern two criteria for analytic judgment. Kant, in apparent disregard of their differences, uses first one and then the other as it suits his purposes, perhaps in the conviction that their answers will in any specific case be the same. i) T he logical criterion of analytic judgment is its conformity to the law of contradiction, a necessary condition of any judgment and a necessary and sufficient condition for an analytic judgment. T h e test is applied as follows: substitute in a judgment synonyms for synonyms, or an analysis or definition of the subject concept for the subject itself. Then the contradictory of this judgment will infringe the law of contradiction if the original judgment is analytic. And 13
Critique of Pure Reason, A 7 i8 = B 746.
A recent paper by Robert S. Hartman, “ Analytic and Synthetic as Categories of Inquiry,” in Perspectives in Philosophy (Ohio State University, 1953), pp. 55-57, has the special merit of singling out the two kinds of analyticity, one of which it calls definitional and the other expositional, and distinguishing both from “ ana lytic” in the sense of descriptive of what is “ contained in ” an experience of an empirical object. H artm an’s paper presents very clearly the processes by which analytic judgments lead to definitions, and definitions then establish a new and stricter criterion of analyticity. Another study of the process by which an analytic judgment may become synthetic is K. Sternberg, “ Über die Unterscheidung von analytischen und synthetischen Urteilen,” Kant-Studien, X X X I (1926), 17 1-2 0 1. 14
Studies in the Philosophy of Kant
79
as the contradictory of a self-contradictory proposition is necessary, the original judgment is necessary. In applying the logical criterion, a definition in the strict sense is not required, for it is from the analytic judgments in informal exposi tion that we first gain the definition. All that is needed is a partial analysis of the subject concept. T he absence of definition may at most prevent only the decision that some specific judgment is not an alytic,15 for what is mentioned as the predicate may be an unnoticed analytic attribute that we would have noticed had we possessed a full definition. But no criterion is infallible; even given a strict definition, the pertinancy of a specific attribute as analytical may be a discovery of the most difficult and surprising kind. It is in such cases that there will be the greatest divergence between decisions made on this and those made on the phenomenological criterion. 2) T he phenomenological criterion is the issue of an inspection of what is found introspectively to be really thought in the concept of the subject. Though we have seen that what is “ really thought” is said to be a definition, and that the mention of predicates not thought “ with the same (full) consciousness” suggests a very wide range of predicates that might pass the logical but fail the phenomenological test, still it is clear that Kant was not free from a psychologizing, introspective tendency in his decisions on what is analytic and what is synthetic. T h e Port-Royal Logic 16 demanded “ moderate attention” to see whether the predicate is “ truly contained in the idea of the subject,” and not a completely articulated logical system as a criterion for this decision; the same kind of “ moderate attention” seems to provide a criterion for Kant. He repeatedly asks himself and the reader what he thinks when he thinks a particular concept, and though undoubtedly one may think much, by casual association, which is not “ contained in the concept,” what he does not think is not included in the content of the concept. Just as he has previously distinguished between what is contained in and what is contained under a concept, so also he distinguishes what “ lies in” a concept and what “ belongs to” it.17 There seems to 9
See Lectures on Logic, §109, Note. Part IV, ch. 6. Critique of Pure Reason, A 718 = 1} 746; but cf. Lectures on Logic, Introduc tion, V III (Academy edn., IX , 61) where attributes belong to the essence, so far as they are derived from it. 15 16 17
Studies in the Philosophy of Kant
8o
be here a tacit distinction between two kinds of concepts, one being a concept of a highly refined analytical or abstractive unity, subject to strict definition, and the other being a looser complex of represen tations, more or less loosely held together and expandable through the accretion of new experience or subject to restriction in content through the supervention of a definition.18 I now turn, for the space of one paragraph, to Kant’s description of synthetic judgments, after which I shall come back to these two cri teria of analytic judgment. T h e following material is essential for evaluating the issues raised by the two criteria. B may be related to A indirectly by virtue of the fact that both are predicates of the same X . Then the concept A does not include the concept of B as a part of its logical essence, and to relate them to each other in judgment requires reference to the X of which each is a partial concept. There are three kinds of X which serve to mediate between A and B. (1) X may be a schema of an object in general (of a thing, cause, etc.). (2) X may be a determinable intuition of space or time or both, which A and B both refer to and make determinate. (3) X may be a datum or concretum of experience, “ the complete experience of the object which I think through the concept A .” 19 In the former two cases, the judgment will be valid regardless of the empirical content of the concepts, and in the first case there is estab lished the kind of judgment which appears in “ metaphysics as science.” Failure to provide a schema without the conditions of space and time and to put the thing in itself in the role of the X makes synthetic judgments impossible except of objects of possible experience. T h e second is the situation with respect to mathematical judgments, where X is a construction. In the third alternative, the judgment is a posteriori. But in each case it is a synthetic judgment, since the predicate is not found by analysis of the logical subject. If X is (as is actually the case) a subjective condition for the synthesis of T h e confusion between these two meanings of “ concept” has been discussed by W. Koppelmann, “ Kants Lehre vom analytischen Urteil,” Philosophische Monats hefte, X X I (1885), 6 5 -10 1; and by H. Ritzel, “ Über analytische Urteile,” Jahrbuch f. Philosophie u. phänomenologische Forschung, III (1916), see especially pp. 2 6 176, 324. T h e full significance of it, as representing the interpenetration of two stages of inquiry dominated respectively by the analytic and the synthetic method, is ably worked out by Hartman, “ Analytic and Synthetic as Categories of Inquiry.” 19 Critique of Pure Reason, A 8; omitted in B. 18
Studies in the Philosophy of Kant
81
A and B, the resulting synthetic judgment is, in the transcendental sense, only subjectively valid; though we can say still that the predi cate is a part of the real and not of the nominal essence. In the same sense, an analytic judgment is objectively and even transcendently valid, not being restricted to the conditions of synthesis placed upon the X .20 From this account of the origin of synthetic judgment, and from the two criteria mentioned above, it is clear that the distinction between analytic and synthetic judgment is not one of formal logic, for formal logic abstracts from the meaning of all terms. V A R I A B I L I T Y O F T H E D IS T IN C T IO N
Eberhard interpreted an analytic judgment as one the predicate of which is an essentia of the subject, and a synthetic judgment as one whose predicate is an attribute derived from an essentia. But Kant denies that this is his meaning, for he holds that “ derived from” is equivocal. If the attribute is derived by logical analysis, the judg ment is indeed analytic whether we knew that the attribute was “ contained in” the subject concept or not; but there are other attributes, synthetic attributes (Bestimmungen), that are not con tained in the logical essence, even though they might be associated with it in our minds, e.g., as weight with body. They are derived not by logical analysis but by construction or exhibition of a correspond ing intuitive object. From such an experience the attribute can as it were be read off, though it is not a nota notae of the subject concept but a nota of the real object. It is this kind of synthetic predicate which is a part of the ratio essendi of the object, and it gives the concept of the subject and all its judgments whatever objective validity they have. Though Eberhard was a mediocre thinker, much of whose argu ment is vitiated by being based upon patent misunderstandings of Kant, he did nevertheless ask a difficult and important question, “ How do we decide what is ‘actually thought’ in a concept?” Unless a definite and plausible criterion can be given that is exempt from the vagaries of the phenomenological criterion and of the logical criterion when Kant attempts to employ it unarmed with definitions, 20 Reflection 3950.
82
Studies in the Philosophy of Kant
then an important member of the structure of his philosophy must be given up. Modern writers, reacting against both psychologism and phenomenology, wanting a behavioral rather than an introspectional criterion if a significant logicaLcriterion cannot be given, have directed their main attack on the possibility of maintaining the distinction, in any particular instance, without a complement of definitions. Rather than considering the views of those who give up or relativize the distinction for the reasons just mentioned, however, it will be more profitable to consider the views of a critic who admits a sharp distinction between analytic and synthetic, yet who does not base it on the test of nominal or stipulative definition. A critic this close to Kant is likely to be more instructive, at this juncture, than one more radically opposed to Kant. T h e criticism I shall consider is that by C. I. Lewis, which is in part an infinitely improved version of some debating points raised by Eberhard. Kant’s cognizance of these argu ments, admittedly in a more primitive form, makes a study of them especially worthwhile for an understanding of Kant himself. Lewis argues as follows. T h e notion of a necessary but nonanalytic proposition such as “ Every event has a cause” is based on an equivo cation. For “ event,” as a concept which does not contain “ having a cause” as a part of its meaning, is not the same as the concept of “ event” which does contain the concept “ having a cause.” Part of Kant’s argument is based on the former and simpler concept, and here Kant rightly infers that the proposition is synthetic. But the argument that the proposition is a priori is based on the second, richer, concept. We can, according to Kant, think without contradic tion an uncaused event; hence the relation expressed in the judg ment is synthetic; but we cannot imagine, represent, or know an event as objective without relating it to another event by a rule of causa tion; hence the judgment is known a priori. T he equivocation is that “ event” in the second case means “ phe nomenal event in objective space and time,” while in the first case it is not so restricted. If this restriction is made explicit, however, the rela tion between the restricted concepts is seen to be analytic. T h e second Analogy of Experience seems to be synthetic only because the word “ event” is not usually given the restricted meaning. T h e term needs to be fixed by definition before one can pronounce the judgment to
Studies in the Philosophy of Kant
83
be analytic or synthetic; and in defining it, we must be sure to include in its meaning everything needed to determine the objective applica bility of the term in question: . . Anything which is essential to the temporal character of an event must be included in the adequate concept of it as a temporal event. . . . A definition which does not logically entail all characters essential to what is defined, is faulty.” 21 Kant’s reply to this kind of criticism, as it appeared in its first crude form, or rather Schultz’s reply written under Kant’s supervision, makes two responses. 1) T w o different propositions, one of which is analytic and one synthetic, may be expressed by the same sentence, for the same word in the sentences may refer to two different concepts, one narrower and one broader. 2) Closely related to this is the assertion of the “ fixity” of a concept. A concept cannot be arbitrarily widened through the accumulation of information. It can be replaced by another called by the same name; but of any given concept it can be decided what is implicit in it to be explicated in analytical judgment and what does not lie in it at all. When one changes a definition, which may change the status of many judgments, the judgments are changed not merely in status but in meaning and validity. Definitions should not, therefore, be arbitrarily changed; a new one must pass the same kind of test of “ realness” that the old one originally passed and later failed, if it is not to be merely stipulation without objective reference. We cannot convert empirical knowledge into a priori knowledge simply by refining our language: L e t o n e p u t in to the con cep t o f the sub ject ju st so m an y attribu tes that the p red ic ate w h ich o n e wishes to p ro ve o f the subject can be d erived fro m its con cep t m erely by the law o f co n trad iction . T h e critical p h i lo so p h y p erm its h im to m ake this k in d o f a n alytic ju d g m e n t, b u t raises a q u estio n a b o u t the con cep t o f the subject itself. It asks: how d id you com e to in c lu d e in this con cep t the d ifferen t attribu tes so that it [now a n a ly tic a lly ] en tails syn thetic proposition s? First p rove the o b jective re a lity o f y o u r concept, i.e., first p rove that an y one o f its attribu tes re ally b e lo n g s to a possible o b ject, an d w h en you h ave d o n e that, then prove 21 C. I. Lewis, An Analysis of Knowledge and Valuation (La Salle: Open Court Publishing Co., 1947), pp. 161-62. I have given a fuller exposition of Lewis’ views (without discussion of the point raised here) in “ Die Kantkritik von C. I. Lewis und der analytischen Schule,” Kant-Studien, X L V (1954), 3-20.
84
Studies in the Philosophy of Kant that the other attributes belong to the same thing that the first one belongs to, without themselves belonging to the first attribute. T h e whole ques tion of how much or how little the concept of the subject is to contain has not the least bearing on the metaphysical question: How are synthetic a priori judgments possible? It belongs merely in the logical theory of definition. And the theory of definition without doubt requires that one not introduce more attributes into a definition than are necessary to dis tinguish the defined thing from all others. Hence [in a good definition] one excludes those attributes of which one can demand a proof whether and on what grounds they belong to the former attributes [that are in cluded].22
Put in our own words, Kant is saying that a definition which will change a synthetic into an analytic judgment must be either nominal or real. If nominal, it does not in the least affect the cognitive status of the original judgment; while it may make the original sentence formally analytic, it does not give to the knowledge it expresses any logical or epistemic necessity it previously lacked.23 And if the defi nition is a real one, we must know the necessary conjunction of independent, coordinate attributes in order to make it; and this con junction is precisely what was stated in the synthetic judgment whose status is now being disputed. All that is effected by such a procedure, we might say, is that the locus of a priori synthesis is shifted. I N D E F I N A B I L I T Y O F T H E C A T E G O R IE S
Thus far I have considered only Kant’s explicit answers to the criticism that the analytic-synthetic distinction is variable. I now examine Kant’s reply in its general philosophical bearings. I have already mentioned that there are in Kant’s writings two quite different species of concept. In one case, like that of “ water,” the word is “ more properly to be regarded as merely a designation 22 Rezension von Eberhard’s Magazin, Academy edn., X X , 408-9. 23 In this, Lewis is in agreement with Kant. In criticizing these who identify the a priori and the analytic, and then define the analytic in terms of linguistic rules or procedures, Lewis writes: “ If implications of conceptions of this sort should be well worked out, it must appear that they are fatal to the thesis that what is a priori coincides with what is analytic; since the notion that what may be known true without recourse to sense experience, is relative to vocabulary or dependent on conventions of procedure, is not credible” (Analysis of Knowledge and Valuation, p. 36).
Studies in the Philosophy of Kant
85
than as a concept of the thing,” 24 and its meaning does vary with experience. In the other, the concept is fixed either by definition, or fixed because it is a pure concept which, while not subject to defini tion, is not subject to revision by the accumulation of experience. In the latter case, Kant believed that a fixed decision could be made concerning what was and what was not included in it, even at a time before a stated definition had been reached. T h e rationalistic tradi tion in which Kant wrote fixed many of the most important concepts by “ implicit” definition and common use or by nominal definitions that had become well established.25 Thus Kant could confidently decide that a given proposition is analytic without the necessity of referring to a “ rule book” of stipulative definitions. We, in a more conventionalistic period, are usually puzzled by some of his decisions, and can only feel that Kant and his contemporaries were committing what Whitehead called the “ fallacy of the perfect dictionary” — when the dictionary could not, in principle, exist for Kant at all. But the more important point is that the concepts with which Kant is most concerned, viz., the categories, are not fixed by definition and need not be fixed in this way. They are fixed because, as pure, they are not susceptible to experiental modification. Let us consider what Kant was attempting to do with these con cepts. It had been shown by Hume that they could not be given objective validity by definition, and though Kant might have given a richer, more determinate definition to such a concept as cause, a still more extended Humean argument would have been fatal again to its claims to objective validity. Definition and proof of objective validity are not the same except in mathematics; which, for quite peculiar reasons, does not have to meet the Humean type of criticism. Assuming a broader definition, a proof of the objective validity of its analytic consequences is still called for if Hume’s criticisms of the rational structure of empirical knowledge are to be met. Given the broader definition, of course, antecedently synthetic judgments be come analytic. So long as the definitional component is expanded ad lib, any a priori judgment can be shown to be analytic. But apriority 24 25
Critique of Pure Reason, A 728=!} 756.
See J . H. Hyslop, “ K an t’s Treatm ent of Analytic and Synthetic Judgm ent,” The Monist, X I I I (1903), 3 3 1 - 5 1 , which emphasizes Cartesian and Newtonian con clusions as they “ infected” the concepts Kant used.
86
Studies in the Philosophy of Kant
is not dependent upon this kind of analyticity; the analyticity of such a judgment is not a condition of its apriority but a subsequent, factitious addendum to it. T hat is, there must be recognition of some special dignity of function of a specific proposition that makes it worthwhile to devise a language in which it will be necessary; but the linguistic necessity is established subsequent to this recognition. Kant did not simply suppose that causality had objective meaning; he tried to show that it did, and in doing so he found that he had to add to the concept of sufficient reason determinations which neither Hume nor the rationalists had suspected; he had to give a new inter pretation to “ possible experience” as the mediator between the terms of such a judgment. T o have suppressed this interpretation for the sake of a formal definition of cause which would render the second Analogy of Experience analytical would have distorted the whole procedure of the critical philosophy, and would have left unanswered the reiterated question, how can this judgment, based on definition, be valid objectively? Kant thought that real definitions should come at the end of in quiry, not at the beginning. One might expect, therefore, that the contribution of the Critique of Pure Reason might have been seen as a new set of definitions subsequent to which a priori judgments previously called synthetic would now be called analytic. Why did Kant not see his work in this way, but obstinately regarded the Analogies as synthetic judgments— in spite of the fact that he might have seen the logical classification as tentative, dependent upon the richness of the concepts? There were several reasons why Kant did not do this. Among them was his respect for tradition; more important was his recognition that Hume’s objections to the rational foundations of empirical knowl edge could not be met by new definitions. And a still more funda mental reason is to be found in his repeated denials of the definability of the categories: the definitions which some might think would serve for this reduction of all a priori knowledge to analytic knowledge cannot be given. Definitions, however elaborated, are still conceptual relations; but what is needed is some way to get a concept into rela tion with an object, and to do it in an a priori fashion. Concepts alone, however richly furnished with predicates, do not establish contact
Studies in the Philosophy of Kant
87
with things; only intuition can provide this contact. We can indeed conceptualize and name the requisite intuition; but in doing this, we treat it like a universal concept, and as such it fails to establish the objective reference. It always leaves open the question: does this complex universal apply? T h e category, whether it can be defined or not, must be schematized— must be provided, in Lewis’ terminology, with a sense meaning as well as a linguistic meaning. Kant is profuse in his definitions of pure categories, but these definitions are nom inal.26 Schematizing a category is very different from defining it: T h e r e is som eth in g strange a n d even n onsensical in the n o tio n that there sh o u ld be a con cep t w h ich m ust h ave a m e a n in g b u t w hich can n ot be defined. B u t the categories are in a u n iq u e position, fo r o n ly by virtu e o f the gen eral condition of sensibility can they h ave a definite m ea n in g an d re la tio n to an object. T h i s co n d itio n , how ever, is om itted in the pu re category, fo r this can co n tain o n ly the lo gical fu n ctio n o f b rin g in g the m a n ifo ld u n d e r a co n cep t,27
without specifying the concept or the condition of its application to a specific manifold. No philosopher has emphasized more than Kant the fundamental difference between sense and understanding while at the same time asserting their complementary function. This fundamental differ ence is essential here. It is not the concept of an intuitive condition, which might be added to a concept or included in its definition, that gives full meaning to the category; it is the c o n d itio n of sensibility itself,28 the condition of its actual use in specific circumstances ac cording to rule. T his is a transcendental addendum, a real predicate, a synthetic predicate, a B e s tim m u n g , an element in the ratio essendi as well as the ratio cognoscendi. It is not just another attribute with out which the definition is “ inadequate.” Make the added condition a conceptual amendment to the definition, and the entire question is 26 27 28
Critique of Pure Reason, A 244=1} 302. Ibid., A 244-45; omitted from B. Italics supplied.
T h e difference between a concept of an intuitive condition and the intuitive condition itself is formally like that between the concept of existence and existence itself. K an t’s criticism of the ontological argument, mulatis mutandis, could be used here against the view (expressed by Lewis, Analysis of Knowledge and Valua tion, p. 162, middle paragraph) that the concept of space suffices, if we assume, with Kant, that mathematics is knowledge of something real.
88
Studies in the Philosophy of Kant
postponed: we would still have to ask, “ How does this concept have a priori objective application?” 29 Because Kant does admit definitions, in the strictest sense, only in the field of mathematics,30 it is easy to admit a sharp distinction be tween analytic and synthetic judgments here; in fact, mathematical definition has been taken as establishing the paradigm of the analytic-synthetic distinction.31 Granting the sharpness of the dis tinction between analytic and synthetic here, most competent critics of Kant are in agreement that he was in error in saying that mathe matical judgments are synthetic. It is said that what kept him from seeing that they are analytic was the lack of adequate mathematical definitions, definitions not available until much later. Professor Lewis characteristically writes: “ It would be ungrateful and unjust to blame Kant for not foreseeing that, from genuinely adequate mathematical definitions, the theorems of mathematics might be deducible.” 32 Obviously, deducible from definition and analytic are here regarded as equivalent notions. This however, as we have amply seen, is not what Kant meant by “ analytic.” In the Prolegomena he wrote: . . . as it was found that the conclusions of mathematicians all proceed according to the law of contradiction . . . men persuaded themselves that the fundamental principles were known from the same law. T h is was a great mistake, for a synthetical proposition can indeed be established [eingesehen] by the law of contradiction, but only by presupposing an other synthetical proposition from which it follows, never by that law alone.33
From this we see the following: (1) Mathematical theorems may be synthetic even if proved by the law of contradiction, i.e., by strictly There is still another argument in the Critique (A 245; omitted from B) against the definability of categories, to wit, that such definitions are circular. I do not think the argument is valid; but inasmuch as it applies, if at all, to the pure as well as to the schematized categories, it is not relevant to our purposes here. 30 Critique of Pure Reason, A 729 = 6 758. 31 “ Kant scheint bei der Einteilung der Urteile in analytisch und synthetisch von der Fiktion auszugehen, dass auch die nichtmathematischen Begriffe definiert werden kônnen” (K. Marc-Wogau, “ Kants Lehre vom analytischen U rteil,” Theoria, X V II [1951], 150). 32 Lewis, Analysis of Knowledge and Valuation, p. 162. 33 Prolegomena, §2 c 2 (Academy edn., IV, 268; trans. Beck, p. 15 )=Critique of Pure Reason, B 14. 29
Studies in the Philosophy of Kant
89
logical procedure. Deducibility is not a sufficient condition for analyticity. T o be analytic, in Kant’s meaning, a proposition would have to be proven by the law of contradiction alone, i.e., its contradictory would have to be ^//-contradictory; but in mathematical proof by strict logic, the contradictory of the proposition contradicts some other assumed propositions. (2) A proposition will be called synthetic if among its premises is a synthetic proposition, such as an axiom, or a mathematical definition, i.e., a synthetical definition which can be exhibited in a construction. (3) Mathematical axioms (fundamental principles) are synthetic since they are not established by the analysis of a given concept, but only by the intuitive construction of the con cept, which will show the necessary presence of attributes not in cluded in a logical definition of the subject.34 T h e theorems, therefore, can be called synthetic even though they are strictly (analytically, in modern usage) demonstrable. T he famous discussion of the example, “ 7 + 5 = 12,” two paragraphs later, is quite independent of the grounds given in the quotation for calling the theorems synthetic. It is, in fact, inconsistent with it. In the quo tation, Kant is conceding that a theorem does follow from premises by strict logic; whatever may be the nature of the premises, the in ternal structure of the proof is logical. But in the discussion of “ 7 + 5” Kant is arguing that a theorem does not follow logically even from synthetic axioms, but that intuitive construction enters into the theorem itself and its proof. These two theses— that an intuitive synthetic element is present in the primitive propositions, and that an intuitive synthetic process is present in demonstration— are inde pendent of each other. Because a mathematical judgment is often synthetical by the phenomenological criterion, Kant seems to have supposed that there were good logical reasons for calling it synthetic. Of these two theses, ohly the first is of any moment in the epistemology (not the methodology) of mathematical knowledge, but it is only the second of the theses that could be corrected by the use of what Lewis calls “ genuinely adequate mathematical definitions.” T h e real dispute between Kant and his critics is not whether the theorems are analytic in the sense of being strictly deducible, and not whether they should be called analytic now when it is admitted that Prize Essay, Academy edn., II, 277; Über eine Entdeckung, Academy edn., V III, 229 -31; Critique of Pure Reason, A 730=1} 758. 34
90
Studies in the Philosophy of Kant
they are deducible from definitions, but whether there are any prim itive propositions which are synthetic and intuitive. Kant is arguing that the axioms cannot be analytic, both because they must establish a connection between concepts, just as definitions do, and because they must establish a connection that can be exhibited in intuition. And this is what is denied by the modern critic of Kant. I think Kant is obviously right in saying that there cannot be a system of nothing but analytic propositions; there must be some com plexes to analyze, and these must be stated synthetically. But if the postulates are not analytic, this does not mean that they are synthetic propositions, i.e., synthetic statements expressing truths. A stipula tion can “ establish” synthetic relations, but it does not thereby qual ify as a proposition. If it be assumed that mathematics is a game, then the analytic-synthetic distinction is of no importance in discussing the postulates, because the premises are not propositions at all but are only stipulations or propositional functions.35 Kant did not espouse the game theory. Mathematics was for him objective knowledge. T hat is why he regarded the axioms as propo sitions, not proposals. Were they mere relations of ideas, in Hum e’s sense, they could be made as “ adequate” as one wished, yet the ques tion of how they could be objectively valid would remain untouched. But for Kant, real mathematical definitions are possible, because the definition creates the object. T his sounds like stipulation again; but the object is not an arbitrary logical product of subjectively chosen independent properties. T o define a mathematical concept is to pre scribe rules for its construction in space and time. Such a definition is a synthetical proposition, because the spatial determination of the figure is not a logical consequence of the concept but is a real condi tion of its application. T he real property is joined to the logical prop erties synthetically, not analytically. Objections to Kant’s views of mathematics, therefore, cannot be 35 Kant says that mathematical definitions are willkürlich, which is usually translated as “ arbitrary.” But the connotation of “ random” present in “ arbitrary” is not present in K ant’s word “ arbitrary,” for Kant makes the antonym of “ arbi trary” not “ necessary” but “ empirical” {Lectures on Logic, §103, Note). Willkürlich has reference to the volitional character of a synthetic definition, a rule for the synthesis of a concept; but a mathematical concept is synthesized only under given conditions of intuition, and is therefore not arbitrary in the modern sense of this word.
Studies in the Philosophy of Kant
9i
removed merely by the substitution of more adequate sets of defini tions and postulates, as if being a better mathematician would have corrected K ant’s philosophy of mathematics. T h e syntheticity of m a th em a tica l k n o w le d g e in K a n t is not a co nseq u en ce o f the in a d e quacy o f his definitions. It is an essential feature of his entire theory
of mathematical knowledge, by which the identity of mathematics and logic was denied. Mathematical knowledge in his view of the world has objective reference, and this is obtained not through definition but through intuition and construction. His mathematical definitions are real; what is deduced from them may be, in modern but not Kantian terminology, analytic propositions. But the proposi tions admitted as theorems by Kant are not like the analytic proposi tions of modern mathematics or the relations of ideas of Hume, for they have a necessary relation to experience through the synthetic, intuitive character of the definitions and axioms. Even propositions which Kant admits are analytic belong to mathematics only if they can be exhibited in intuition.36 Whatever improvements in Kant’s definitions might have been introduced for the sake of making the theorems analytic in his sense would have cost a high price in setting mathematics apart from the discussion of the conditions of possible experience. And had they been seen as analytic, Kant’s long and deep concern with mathematics would not have positively contributed to his interpretation of the problems of empirical knowledge. For Kant saw in mathematics a clue to the objectivity of all a priori knowl edge, both analytic and what he considered to be synthetic. This is indeed the sense of the Copernican Revolution: even empirical ob jects are constructions; and their necessary conditions are geo metrical. Had Kant radically sundered mathematical knowledge from the intuitive a priori structures of empirical knowledge, as he criticizes Hume for doing,37 both would have been rendered unin telligible to him. T h e question is thereby raised whether, in intro ducing modern amendments into Kant’s theory of mathematics (per haps for the purpose of “ saving what is essential in the critical philos ophy” ), we do not at the same time overlook or destroy everything distinctive in his theory of empirical knowledge. 36 37
Critique of Pure Reason , B 17 —Prolegomena, §2 c 2. Prolegomena, §2 c 2; Critique of Practical Reason, Academy edn., V, 51.
V ON THE META-SEMANTICS OF THE PROBLEM OF THE SYNTHETIC A PRIORI This note is concerned with the semantics (in a wide sense) involved in discussion of the question: “ Are there any synthetic a priori truths?” I am not here concerned with the meanings of the truths that are claimed or denied to be examples of the synthetic a priori, and least of all am I concerned with the decision as to whether any specific proposition is synthetic and a priori. T h a t semantic and other linguistic changes may alter the position of a given sentence in the synthetic-analytic dichotomy is well known; it is likewise granted that the development and progressive articulation of a body of knowledge can alter the attribution of necessity to a given statement. From these now well-known facts it follows that the decision as to whether a given sentence is analytic or synthetic, or a given truth is a priori or a posteriori, is not the comparatively simple question that it seemed to be when, for instance, Kant could confidently decide that “ All bodies are extended” is analytic while “ A ll bodies have weight” is synthetic. In the dispute about specific judgments and the kind of meaning and truth they possess, however, there has been a pervasive change in the meanings of some of the terms used in the controversy. Here I shall discuss the changes in the meaning of the terms “ analytic” and “ synthetic” themselves. Adopting by analogy a somewhat widespread practice in recent discussions, I may be per mitted to refer to this discussion as belonging in the “ meta-semantics” of the problem of deciding whether any given truth is a priori and synthetic. While the problem of distinguishing what were later to be called analytic judgments from synthetic judgments begins before Kant, r e pr in te d
,
by permission of the Editor, from Mind, N.S., LXVI (1957),
228-32. 92
Studies in the Philosophy of Kant
93
these names are his own original technical invention. Unlike the dis tinction between the a priori and the a posteriori, which he inherited and modified in few or no essentials while providing a radically new theory about the a priori, the distinction between analytic and syn thetic judgments is, historically, what he said it was. In his contro versy with Eberhard, in fact, he claimed the right of ownership in them, claiming in effect that “ analytic” had to mean what he said it meant. No doubt he said different things at different times; the criteria he set up would not be applied in the same way by other thinkers; indeed, he did not himself seem to make unequivocal u?e of them. But all this can be left aside at the moment, as bearing on points of only historical interest, while we concentrate on one central point he was making throughout his discussions. This essential point was that we can classify judgments in two ways: (a) according to their epistemic status, as necessary or contingent; and (b) according to some aspect of them which belongs to the judgments themselves without respect to their relation to a system of other judgments, the rest of experience, or our sense of assurance of their truth. T he first gives the classification into a priori and a posteriori judgments; the latter gives the classification of judgments as analytic or synthetic. T he former is an epistemological distinction, the latter a logical distinc tion. If we recognize that there are two principles of division, we are prevented from solving the problem of whether there are a priori synthetic judgments by deduction from the definitions of the four kinds of judgments. No real issue is solved by defining “ a priori” in such a way that all a priori judgments must be analytic. Whether there are such judgments is a question of fact. Undoubtedly the ques tion of fact appeared in a very different light before the invention of non-Euclidean geometries and non-Newtonian physics. Discussion of these facts and not tampering with the definitions has been the fruitful way in which many of Kant’s specific decisions concerning the status of single judgments have been settled. But when the ques tion is discussed in greater generality, I think I detect a tendency to modify Kant’s definitions in such a way that the general issue is de cided against the possibility of a priori synthetic judgments. It is this kind of argument which seems to me to be perverse. I am not here pleading for Kant’s definitions, least of all for his
94
Studies in the Philosophy of Kant
definition of “ analytic.” But I do suggest that any definitions, how ever drawn, ought to take account of the fact that there is at least a prima facie difference, which ought to'be reflected in new definitions, between logical necessity and some other kind of necessity which I have called “ epistemological.” And if the new definitions are to be of Kant’s terms “ analytic” and “ a priori,” they should take into ac count the historico-linguistic fact that “ analytic” is a logical or a linguistic concept, and that “ a priori” is an epistemological concept. Any set of definitions or considerations which obliterates this distinc tion begs the question of whether all necessary truths are logically necessary, and they arouse suspicion of perfectly valid arguments, based on fact, which might be given to show that logical necessity is, in fact, the only kind of necessity in knowledge. Prima facie, there are four species of propositions or sentences which are abstractly possible: (1) analytic a priori, (2) analytic a posteriori, (3) synthetic a priori, (4) synthetic a posteriori. While it might follow from Kant’s definitions that the class (2) has no mem bers, that any sentence which is analytic is necessarily true (if true at all), and that a proposition which is a posteriori is synthetic, it does not follow from Kant’s definitions, or from any other definitions which recognize that two principles of division are abstractly pos sible, that there are not statements of kind (3). I shall not recite the details of the history of the long controversy, which began in his own lifetime and which has played so important a part in the development of mathematics, philosophy of science, ethics, and general philosophy since that time. I merely point out that, from about 1920 to about 1950, most competent original thinkers (especially in English-speaking countries) had concluded (3) to be an empty class, and that at least towards the beginning of the period this conclusion had been based on meticulous examination of the kinds and limits of the necessity claimed for a priori knowledge in mathematics, ethics, and science. One of the chief writers on the problem wrote correctly in 1946 that the question of whether there were a priori synthetic judgments was “ a dead, or nearly dead, issue in philosophy.” It is hard to be confident that discussion of the issue towards the end of the period and since then has always been based on factual considerations of the semantics and syntactics of the truths under dis
Studies in the Philosophy of Kant
95
cussion. With the triumph of the point of view that there are no synthetic a priori truths, there occurred, almost inevitably, a subtle meta-semantic change. Analytic a priori was seen as the only species of the a priori, and therefore “ analytic a priori” seemed to be a re dundancy. Since it was the analyticity of a statement that rendered it independent of experience, either in origin or in confirmation, the decisive word was “ analytic,” and “ analytic” came to function as a synonym for “ necessary,” which had previously been regarded as synonymous with “ a priori.” At least they were ascribed the same denotation, and since the criterion of necessity came to be regarded as exclusively logical or linguistic, nothing essential was omitted in saying simply “ analytic” instead of “ analytic a priori.” And few people in the main current of recent philosophy talked about “ a priori” at all. With the collapse of the two distinctions into one, the debate on the old issue of whether there are any synthetic a priori judgments became singularly difficult. For if someone said, “ p is necessary” p would not be granted as a case in point until it was shown that it was logically or linguistically necessary; so p came into the discussion only on the condition of its analyticity. And if it could not be shown to be analytic, it was either ruled out as being not necessary, or some what vague rules (e.g., “ grammar of color words” ) were promised which would show, at some future time, that if necessary it is analytic. Much of this kind of argument was based on the vagueness of terms in p that could be fixed only by definitions which would arbitrarily insure the inclusion of the proposition, or its exclusion, as desired. T h e tactics of such dispute belong to the semantics of the problem, and will not be recounted here. But there was, no doubt, another vagueness which could be fixed by arbitrary definition; this was the equivocality of the word “ analytic,” and discussion of this does be long to the meta-semantics of the controversy about the status of the disputed truths. Whenever any term which once had a narrow range of application is made to apply to things which did not fall within its narrow range, if there is not to be an irrevocable obliteration of some important dis tinctions it is necessary that subspecies be found within the new range of the term. If “ analytic” is allowed to take over the entire denotation of “ a priori” (when the range of “ a priori” is properly restricted so
96
Studies in the Philosophy of Kant
that it contains all statements thought to be necessary, by any chosen criterion different from the classical or established criteria of logical [or linguistic] necessity), then we m&y reasonably expect that the subspecies of a priori, which were the analytic and the synthetic, will reappear at least transiently as species of the new genus of analytic, the genus I call “ analytic2.” T his is reasonable unless we can say that the subdivision of the a priori into species, by Kant, had nothing even prima facie to recommend it; and it is obvious that there are at least prima facie differences (whatever we may think of Kant’s treatment) between the necessity or alleged necessity of “ p. q. D p ,” “ No unmarried man is married,” “ No bachelor is married,” “ Noth ing red all over is blue,” “ 7 + 5 = 12,” a principle of causation, and “ F = m a ” — all of which have been called analytic because they are or are thought to be, for various reasons and certainly in various senses, necessary. We do find what I have said we might reasonably expect to find. Let us take a theory of the analytic which holds that a statement is analytic to the extent or degree to which it will be held impregnable against revision by experience. (Other recent formulations, such as the theory that a proposition is analytic to the extent or degree to which we feel shock when it is denied, would serve equally well, provided only that they do not restrict the range of “ analytic” to explicit tautologies— a stringency rare in recent discussions.) Then there are two ways in which a proposition may be found to be ana lytic (i.e., analytic2): (1) by inspection of the sentence itself, if it is logically or linguistically true; and (2) by investigation of its role in an organized body of experience we call knowledge. It may well be that we can apply the first test to sentences which are not explicit tautologies, and in this event we have to take a short step to another proposition, such as a law of logic or to a statement in an actual dictionary of standard usage. But even if we do so, we look to rules (or to empirical propositions functioning as rules, as in referring to a dictionary), and we are not likely to become engaged in an arduous search into the actual articulations of an established and internally systematized body of knowledge and precepts of procedure. T hat there may be some sentences the test of which we do not know whether to describe as (1) or as (2) does not alter the fact that there
Studies in the Philosophy of Kant
97
are some instances when we do apply and know clearly that we are applying one or the other test. T h e first, which I shall call the “ microscopic” test, is formally like the Kantian test for analytic (analytic^, though it may be vastly improved beyond his somewhat naive notion that such a judgment is found to be true by an analysis (in a literal sense) of the subject concept. But other statements, such as a principle of causation or even “ F = ma,” are found to be analytic2 by a “ macroscopic” test in which we do not inspect the state ment itself or analyze its subject concept, but analyze (in some non literal sense1) the context of organized knowledge of which it may be an essential structural member. Now by this test statements may well turn out to be analytic2, i.e., necessary in some sense, which need not be analytic!. But what is “ analytic2” ? It is not defined by the procedure which defined “ analytic/’ but simply by reference to the statements’ neces sity (in an epistemological or methodological, and no doubt watereddown, sense which would not perhaps in some cases have been recognized by Kant as constituting necessity at all). Here “ analytic2” means necessary in some sense, and not exclusively in the logical or microscopic sense (“ analytic/’). But then “ analytic2” is formally equivalent to the old “ a priori.” Or, better, it is a substitute for the old “ a priori,” differing from it, if at all, only in the somewhat more modest claims made for the exclusiveness and irrevocability of the decision that a given statement is analytic2 but not analytic!— a modesty and tentativeness singularly absent in most earlier decisions about whether a statement was a priori. Here we observe the meta-semantic shift to which I wish to call attention. “ Analytic” comes to mean the same as “ a priori” with all the vagueness or precision of the latter. But this meaning of “ ana lytic” is not the classical meaning, and is indicated as “ analytic2.” Then there are under it two species, the microscopic and the macro scopic analytic. T h e former is the logical or linguistic analytic (ana lytic^, and the latter means any one or more of the following: “ ana lytic truths which are factual or ‘real,’ ” the “ functionally a priori,” On the literal and non-literal meanings of “ analysis,” and the consequent con fusion in the meanings of “ analytic,” see the acute remark of Richard Robinson, Definition, pp. 188-89. 1
98
Studies in the Philosophy of Kant
the “ methodological a priori,” the “ hypothetically necessary,” the “ material a priori,” the “ regulative a priori” — and the “ synthetic a priori.” * x Now just the same reason that one might now say that a proposi tion like a principle of causation is (macroscopically) analytic in a particular language or body of science led Kant to say that it is a priori synthetic, viz., that it makes experience of objects possible, and is not known to be true either by reference to empirical facts or by inspection of its logical form but by reference to its contribution to the “ possibility of experience.” In fact, I suspect not only that there are the same reasons for saying both; I suspect also that, apart from the more modest claims made for any specific statement claimed to be analytic2 (and indeed analytic^, the same thing is being said. Plus ga change. . . .
VI REMARKS ON THE DISTINCTION BETWEEN ANALYTIC AND SYNTHETIC Messrs. W ild and Cobitz have undertaken to fight a two-front war on the issues depending on the distinction between analytic and synthetic. One group of their opponents consists of those who have developed, largely from Kant but even more from Hume, a func tionalistic, syntactic, or positivistic theory of the a priori according to which all a priori propositions are analytic in one or both of the senses they find in the word. T h e other opponent is Kant himself. I take it that the former may be interested in defending themselves. I propose to examine some of the strictures Messrs. W ild and Cobitz put on Kant and to try to show that if the issues are more carefully stated, more of the Kantian epistemology can be saved than they think possible or desirable. I shall discuss (1) their criticism of the Kantian concept of “ in cludes” as the relation in an analytic judgment; (2) their “ epistemological distinction” between analytic and synthetic; and (3) their charges that the Kantian epistemology is phenomenologically in adequate. 1) T h e meaning of “ includes” in K a n t ’s definition of analytic judgment. Messrs. Wild and Cobitz (p. 651) give as the logical defi nition of analytic: “ A proposition is analytic if the subject in some sense ‘includes’ the predicate.” Similarly Kant: “ Analytic judgments express nothing in the predicate but what has been already actually r e p r i n t e d , by permission o f the Editor, from
Philosophy and P h e n o m e n o
logical Research, I X (1949), 720-27. T h i s article was w ritten in r e p ly to the article by J o h n W i l d a n d J. L. C o b itz, “ O n the D is tin c tio n b e tw e e n the A n a l y t i c a n d the S y n th e tic ,” Philosophy and P h en om en o log ica l Research, V I I I (1948), 6 5 1 - 6 7 . Messrs. W i l d a n d C o b i t z resp on d ed to this article in the same jo u rn a l, I X
(1949), 728-30.
99
lOO
Studies in the Philosophy of Kant
thought in the concept of the subject, though not so distinctly or with the same (full) consciousness” ;1 if “ the predicate B belongs to the subject A as something that is contained covertly in the concept A ,” the judgment is analytic.2 T he “ in some sense” in their definition causes Messrs. W ild and Cobitz their misgivings. T he only clear meaning they can attach to the word “ includes” lies in its spatial connotation. But we shall see that the spatial connotation makes it possible for them to reduce Kant’s account to nonsense. Take the proposition, “ S includes P ,” which is ex hypothesi ana lytic. In Euler circles, S must be the larger, for the whole is greater than the part. But when Kant says, “ 5 includes P ,” he means that P is included in the concept of 5 and not that all the members of P are included among the members of S. Kant is using the word “ includes” in an intensional sense. He means that P is the genus of S, its inten sion is less than the intension of S, and its intension can be found by the analysis of the intension of 5. Even if this were not clear from his statement about the concept of A instead of simply A, it would be clear from the procedure described for establishing the analyticity of a judgment. T he supreme prin ciple of analytic judgment is the law of contradiction. T h e contradic tory of an analytic judgment is self-contradictory; and the contradic tory of a self-contradictory judgment is necessary. But there is no contradiction in saying “ S does not include P ,” even if it is true that S does include P, provided that the relation of inclusion is extensional. There is self-contradiction only if P is included in the concept of S, as Kant holds. T o diagram in Euler circles the Kantian reading of the proposition “ S includes P ” would require that the P-circle be the larger. Therefore the spatial analogy makes the Kantian usage ap pear paradoxical, i.e., makes the smaller seem to include the larger. Surely then the “ usual” interpretation of inclusion is not the one suitable for Kant. Messrs. Wild and Cobitz ask: “ In what sense is the subject term in the proposition 7 + 5 = 12 logically wider than the predicate term?” T he answer is, in no sense. T h e judgment is, according to Kant, synthetical though necessary; and our authors are led to ask 1 Prolegomena, §2 a (Academy edn., IV, 266; trans. Beck, p. 14). 2 Critique of Pure Reason, B 10.
Studies in the Philosophy of Kant
101
this question because, in the preceding sentence, they have tacitly equated necessary with analytic, in accordance with their “ epistemological distinction.” But then it might be asked in return, in what sense is the predicate logically wider than the subject term? Its intension is different from that of the subject, as they admit, though 12 is not (as might be ex pected from my foregoing argument) the genus of which 7 + 5 is a species. T o define a synthetic judgment it is not necessary to stipulate that P shall include S; it suffices merely to stipulate that 5 does not include P. Extensionally, 7 + 5 and 12 are equivalent, but the inten sions of the two terms are different, and that of one is not included in that of the other. Messrs. Wild and Cobitz argue for this conclu sion with more cogency than Kant did. T heir reading of inclusion in extensional terms leads them to another paradox which intensional interpretation would avoid, “ ab is b” as a paradigm for K ant’s “ body is extended” does not mean ab is a larger whole equated with one of its parts, b. ab is not a plus b (i.e., the things that are either a or b or both); ab therefore is not ex tensionally wider than b. ab means the things that are both a and b (i.e., in our present illustration, the things which are extended and which also contain the specific difference between occupied and un occupied space); ab is therefore extensionally narrower than b, or at least not wider. In extensional interpretation, ab < b, but b < a + b. Even without accepting the universal validity of the principle of in verse variation of extension and intension, it follows that if ab < b, then the intension of b is at least not more than ab, that it might be smaller, and that it might be included in the intension of ab. Hence on this specific point I cannot find Kant guilty of the ambi guities and paradoxes imputed to him. More potent objections have been made to his distinction in view of the changing intensions with the growth of knowledge, so that a synthetic judgment may become analytic when the predicate, in the course of scientific development, comes to be “ really thought” in the subject. Inasmuch as these ob jections imply alternative necessities, they would, of course, be re jected by our authors. 2) The “ epistemological distinction” between analytic and syn thetic. T h e authors write (p. 651): “ A proposition is analytic [in the epistemological sense] if it is certified and validated a priori solely
Studies in the Philosophy of Kant
102
by reference to the meaning of the proposition, synthetic if it is veri fied a posteriori from empirical fact.” As they rightly point out, this is not a Kantian definition, because synthetic judgments are not, for Kant, necessarily a posteriori. They are, furthermore, correct in maintaining as a matter of historical fact that this distinction has been merged with the logical distinction, so that in present-day logic and epistemology it is usually thought that a proposition analytic in one sense is also analytic in the other. As it is not self-evident that this is true, Kant in his day and Messrs. Wild and Cobitz in our own do well in pointing out this fact. It is worthwhile noting that the proposed definition has no source in Kant’s philosophy. Neither the Kantian a priori nor the Kantian synthetic is defined in terms of its origin; the a priori and the analytic are not defined as that which results from an analysis of meaning without reference to experience. Even analytic judgments presup pose a prior synthesis. T h e a priori is that truth whose warrantability is greater than can be justified by induction, and it includes the ana lytic; the analytic is that truth whose contradictory is self-contradic tory. While it is obvious that an analytic judgment is a priori, it is false to conclude that an a priori judgment is necessarily analytic; it is taken simply as a matter of fact that there are a priori synthetic judgments. Kant writes that all empirical judgments are synthetic;3 this statement is careless and inexact. But from the definition given by Messrs. Wild and Cobitz, it follows that all synthetical judgments are empirical, and this is false for Kant. Certainly Kant’s use of synthetic as having one meaning in tran scendental logic, another in general logic, and still another in meth odology is unfortunate. In view of his confusing terminology, it would be well to fix the meanings more exactly, as follows: A. Judgments considered with respect to logical form i) Logically analytic judgments— judgments in which P is found by analysis of the concept 5; explicative judgments; judgments testable by the law of contradiction.4 ii) Logically synthetic judgments— judgments not logically analytic. 3
Prolegomena, §2.
T h e equivalence of these three marks should be shown a priori, but Kant does not attempt to show it. 4
Studies in the Philosophy of Kant
103
B. Judgments considered with respect to their content i) Logically empirical judgments— judgments whose subject is a concept derived from experience. ii) Pure judgments— judgments not logically empirical (Critique, B 3). C. Judgments considered with respect to evidence i) A priori judgments— judgments whose evidence exceeds that derived from induction. ii) A posteriori judgments—judgments whose evidence equals that derived from induction. D. Judgments considered with respect to origin i) Epistemologically analytic judgments—judgments derived by abstraction, and always presupposing a synthesis (Critique, B 103). ii) Epistemologically synthetic judgments—judgments derived from the conjunction of representations. N.B.: (i) and (ii) are not contradictories. a) Epistemologically empirical judgments—judgments re sulting from a synthesis of empirical intuitions. b ) Epistemologically pure judgments—judgments resulting from a synthesis of pure intuitions. T h e most important theses, to whose defense the Critique of Pure Reason is dedicated, are: I. All judgments are epistemologically synthetic.5 II. A priori judgments may be logically synthetic, logically em pirical, and epistemologically empirical. III. Logically analytic judgments must be a priori. IV. Logically analytic judgments may be logically and epistemo logically empirical. V. A posteriori judgments are logically synthetic and, if true, log ically and epistemologically empirical. I l l and V are obvious. For Kant, the most important is II. II is defended by showing that I is true and that all epistemologically empirical judgments presuppose epistemologically pure judgments. T h e chief problem of a positivistic epistemology, as I see it, is the 5
Critique of Pure Reason , B 130.
104
Studies in the Philosophy of Kant
validation of IV, because it, with III, implies that a priori judgments apply to experience. But the problem is difficult for the positivist because he does not, as Kant did, argue that epistemologically em pirical judgments presuppose the objective validity of epistemolog ically pure judgments; rather, the epistemologically pure judgment is applied to a contingent organization or chaos of contents, which it may or may not fit.6 If we try to discover the basic objection which Messrs. W ild and Cobitz make to the Kantian epistemology, I believe we shall find it directed to I and to have little to do with the Kantian a priori as such. This I shall try to demonstrate in the next section. At the moment it suffices to point out that in the Kantian conception of knowledge, all knowledge is judgment, whereas in the phenomeno logical conception, knowledge has more the character of vision. Hence the issue between Kant and his present critics really turns on a basic difference in the interpretation of knowledge in general, and only incidentally on a dispute about the a priori. 3) T h e p h e n o m e n o lo g ic a l in a d e q u a c y o f the K a n t ia n theory o f the a p rio ri. T he general question is: Does the epistemology of the a priori account for the facts of knowledge? a) Clarification of the question. T h e argument against the phe nomenological adequacy of the theory is based on the fact that, for Kant, a priori synthetic judgments are synthetic in origin, though synthetic in a different sense from the one defined on page 651. Synthetic, in Kant, refers, as an epistemological term, to “ the act of bringing together different representations and of grasping their manifold in one cognition.” 7 Hence the central question is not now whether synthesis can or cannot produce higher warrantability than induction. It is only whether knowledge, either a priori or a poste riori, can be regarded as the product of a synthesis. Therefore the form that the argument takes might be more clearly indicated if the question were phrased as follows: Is the epistemological theory of synthesis able to account for the facts of knowledge? T h e three facts singled out as test cases are: the character of direct apprehen6 Cf. the quotation from Professor Lewis given on p. 664 of the paper by W ild and Cobitz. 7 Critique of Pure Reason, B 103.
Studies in the Philosophy of Kant
105
sion, the phenomenon of intention, and the phenomenon of neces^ sity. b) T h e p h e n o m e n o n of direct a p p re h e n s io n . T he first trait of knowledge which, it is said, a theory of knowledge must fit is the character of direct apprehension. I pass over the a rg u m e n tu m ad ig n o ra n tia m that “ no alteration in any material object has ever been observed to result from its being truly known,” and the p etitio in assuming the “ original phenomenon of knowing by which some thing is apprehended as it is objectively with no new production whatsoever” (p. 658). I do so because I have dealt at length else where8 with an elaborate argument that Kant has no account to give of empirical realism and the phenomenon by virtue of which it is plausible; whether that attempted refutation is valid or not, at least I cannot improve upon it here. But I cannot pass over the second paragraph of page 658 in which our authors attribute to Kantianism the equation of knowing with making. T h e synthetic act is not for Kant noetic, though it is a condi tion of knowledge. “ Synthesis in general . . . is the mere effect of imagination, a blind but indispensable function of the mind, with out which we would have absolutely no knowledge, but of which we are only seldom conscious.” 9 If knowing and making are equated, the paradoxes pointed out do surely arise; but they are not equated by Kant. What does making accomplish for Kant? “ We know the object if we have effected a synthetic unity in the manifold of intuitions.” 10 Knowledge is a systematic whole of “ compared and conjoined repre sentations.” 11 Synthesis produces a unity of consciousness and a systematic connection of phenomena into representation of an object and guarantees a conformity between them, for, as the transcendental deduction shows against the subjectivists, the one is not possible with out the other. c) T h e p h e n o m e n o n o f in t e n tio n a lit y . Does the epistemology of 8 “ Nicolai H artm ann’s Criticism of Kant’s Theory of Knowledge” [see Essay V III, pp. 132-44, below]. 9 Critique of Pure Reason, B 104. 10 Ibid., A 105. 11 Ibid., A 97.
io6
Studies in the Philosophy of Kant
synthesis account for the fact of intentionality? Much of this argu ment (pp. 659-63) is directed against a'fiat-theory of apriority, which would be as uncongenial to Kant as t,o his present opponents. But, Messrs. Wild and Cobitz argue, every noetic element, including even definitions, is intentional, while the a priori if it arises from a synthe sis cannot be intentional, for, if it were intentional, “ it would have to be viewed [by Kant] as a bare datum and could not be distin guished from the so-called a posteriori” (p. 656). Here again the kernel of the argument is found in the alleged synthetic origin of knowledge and not in any peculiarity of the a priori as such. Does the fact that the order is imposed by the mind mean that there can be no intentional reference? Though it has been almost universally assumed or argued that Kantianism and intentionalism are incompatible,12 I must confess that I do not see any necessary incompatibility, though this may be a confession of obtuseness rather than an argumentum ad rem. If the Kantian synthesis gener ated contents, or if it proceeded without contents and yet produced knowledge, it would be easy to see. T h e Dialectic shows both to be impossible. Furthermore, Kant’s refusal to consider pure mathe matics as knowledge indicates the necessity of a referential factor in his conception of knowledge; he says mathematical concepts per se are not cognitions “ except insofar as it is presupposed that there are things which can be given to us only in accordance with the form of pure sensuous intuition.” 13 Synthesis which gives knowledge presupposes something that is carried over into the form which the synthesis generates and applies. T h a t something is phenomenal content; unless the mind of man were like the mind of God, it could not be the thing in itself. If the only product of synthesis were psychological complexes of subjective representations or associated ideas, it is again easy to see that there could be no veridical phenomenon of intention, and Kant would be in the same position as Hume. Kant realized that, at least in the second edition, and the subjective deduction was weak precisely because it was subjective. But the synthesis accords with an objective norm which defines objective reality, even though it has no Platonic 12 T h e most noteworthy exception I know is Gunther Jacob y’s Allgemeine Ontologie der Wirklichkeit (Halle: Niemeyer Verlag, 1925). 13 Critique of Pure Reason, B 147.
Studies in the Philosophy of Kant
status. T h e understanding relates intuitive representations to a “ something as an object of sensuous intuition,” though this some thing, the transcendental object, is not known, i.e., is not itself a phenomenal representation.14 Appearances (phenomena) are distin guished from “ appearances of appearances,” i.e., subjective con tents.15 Appearances always have “ two sides” 16 and the first repre sents the second. Similar defenses of objective reference, in which the mechanism is nonphenomenological but the results are inten tional, are found in the Second Analogy and in the Refutation of Idealism. d) The phenomenon of necessity. Does the epistemology of synthe sis account for the factor of necessity in knowing? Here I agree with Messrs. Wild and Cobitz that the positivists de-psychologizing of Kant has left them with “ no necessity at all,” (p. 664) because a neces sity which has alternatives is a contingency. In order to avoid the errors of both, they fall back upon “ the natures of things which con strain our reason to recognize their necessary implications” (p. 666). Actually if the Kantian synthesis is compatible with intentionality, I can find no argument in the paper that apriority and necessity per se are incompatible with their alleged synthetic origin. Since Messrs. W ild and Cobitz do not make a bill of particulars against the Kant ian explanation of necessity, holding only that the positivistic ac count is “ no doubt in many ways an improvement on the Kantian view” (p. 664), I shall not undertake a detailed defense. I need only to point out that the positivists, who argue against Kant that he made things necessary which are in fact contingent (e.g., the Newtonian laws), seem to me to have touched perhaps a more crucial point in Kant’s philosophy than those who accept eternal verities in the manner of the classical rationalists, Platonists, scholastics, or phenomenologists. Ibid., A 250. Opus postumum, Academy edn., X X I I , 339, 363-65. 1 6 Critique of Pure Reason, B 55. 14 15
vii LEWIS’ KANTIANISM1 i T w o diametrically opposed criticisms of Kant’s theory of knowl edge are distinguished by divergent attitudes towards the Copernican Revolution. T h e first criticism is that Kant, through this Revo lution, facing the question of how there can be a priori synthetic knowledge, gave an answer which destroyed the objectivity of cogni tive judgments generally. T h e argument of those who put forth this criticism is that Kant makes “ knowing” a kind of “ making,” so that the essence of knowledge as the grasp of the antecedent fact and independent factuality is misrepresented. Critics who make this objection assert the possibility of a priori synthetic judgments; but since they hold that knowing is not a manufacture of objects out of the data of sense, they do not limit the scope of these judgments, as Kant did, to things in experience. T ru e judgments, they say, are t h is p a p e r
was w ritten for, a n d w ill be i n c lu d e d in, T h e P h ilosop hy o f C. I.
Lewis, edited b y P a u l A r t h u r S c h ilp p in the series, T h e L i b r a r y o f L i v i n g P h ilosop h y, to be p u b lish e d b y the O p e n C o u r t P u b l i s h i n g C o m p a n y ; it is p r e p r in te d here by k i n d permission o f Professor S c h ilp p a n d Dr. E u g e n e Freem an, Editor, T h e O p e n C o u r t P u b l i s h i n g C o m p a n y . T h e O p e n C o u r t P u b l is h in g C o m p a n y retains all rights to this essay.
i T h is essay is based in part upon, and includes a few paraphrases and q u o ta tions from, three earlier studies of mine: “ Die Kantkritik von C. I. Lewis und der analytischen Schule,” Kant-Studien, X L V (1954), 3-20; “ Can K a n t’s Synthetic Judgments be Made Analytic?” Kant-Studien, X L V I I (1956), 168-81; “ K a n t ’s T h e o r y of Definition,” Philosophical Review , L X V (1956), 179-91. [T h e latter two articles appear in this volume.] I am grateful to the editors of these journals for permission to use some material that first appeared in their pages. In this article I have not dealt with the relation between Lew is’ and K a n t’s theories o f value, but w ould call attention to “ C. I. Lewis: Hedonistic Ethics on a K an tian M od el,” by Mary Mothersill, Philosophical Studies, V (1954), 81-88. 108
Studies in the Philosophy of Kant
109
about an independent and metaphysically real object, and whether these objects are sensuously given or not is only an accident. On how such judgments are possible, they differ among themselves, accord ing to whether they are phenomenologists, or scholastic, American, or British realists; but they are at one in opposing the subjectivism they find inherent in Kant’s apriorism. Those who make this kind of attack on Kant I shall classify, in spite of their diversity, as the “ realist critics” of Kant. T h e second objection comes from those philosophers who see in K ant’s work an unsuccessful attempt to give knowledge an objective and perhaps metaphysical ground that it does not in fact possess. In their opinion, the Copernican Revolution is not radical enough. In Kant’s work, this Revolution ends with a series of allegedly necessary synthetic principles which are presupposed in all possible knowl edge. But while Kant denied to such principles the transcendently metaphysical status that realists say they have, he did assert their uniqueness, certainty, and necessity. This “ universal and necessary” character is what the other critics of Kant deny. They wish to push the Copernican Revolution still further, and in doing so they wel come, as a first step, Kant’s denial of what would universally be re garded, for good or ill, as transcendent metaphysics. But they see in his own “ immanent metaphysics” (“ metaphysics as science” ) still too much metaphysical dogmatism in the form of a single, uniquely necessary set of eternal presuppositions. They deny the validity of Kant’s claim to give a transcendental deduction of categories neces sary to all possible knowledge of phenomena, just as the realist critics deny its validity with respect to the repudiation of knowledge of noumena. While claiming that Kant was correct in insisting upon the subject’s activity in the construction of phenomenal objects, they deny that his rules for such construction are necessary and unique. They find pragmatic variables entering into a situation which Kant believed was logically— in some sense of the word “ logically” — fixed. Philosophers who criticize Kant in this way are indebted to him and to Hume for the key to their own epistemological edifices. But indebtedness of this kind often contributes to bitterness in criticism. Though Hum e’s psychologism usually has merely been rapped on its knuckles, the Kantian philosophy has received some of the most
1
io
Studies in the Philosophy of Kant
thorough-going censures from positivists, pragmatists, conventional ists, language analysts, and sociologists of knowledge, all of whom join Kant in regarding physical.objects as some sort of construction. In spite of their divergences among themselves, for our present pur poses I shall call them all by the same name, viz., the “ analyst critics” of Kant. It is noteworthy that the realist and the analyst critics of Kant are closer to him than they are to each other. Each has learned something important from Kant, and but little or nothing from the other. This is all the more remarkable when elements of each type of criticism are found in the writings of one man. II
Clarence Irving Lewis has seemed to many to be a leader of the latter movement of anti-Kantianism— a leader in part because of the originality and cogency of his own constructive work, and in part because unlike most analyst critics he has come to his critical position with a sympathy for and a detailed knowledge of Kant’s philosophy that many of his fellow analyst critics seem never to have had. He speaks, in his autobiographical essay, of his youthful admiration for Kant and of the unmistakable “ evidence of Kant in [his] think ing.” 2 Yet a critic he is; an apocryphal story points up his ambivalent position: “ I am a Kantian,” he is reported unreliably to have said, “ who disagrees with every sentence of the Critique of Pure Reason.” But ambivalent or not, M ind and the World Order can be seen as a locus classicus for analytical criticism of Kant. T h e conceptual pragmatism espoused in that book emphasizes the freely creative act by which complexes found in experience are supplemented by ex pectations and interpreted as objects. These complexes, spread out in time and attended with anticipations, are the real things of experi ence, because they meet our freely chosen but pragmatically justified criteria for the ascription of the status of reality. T h e concepts are, as it were, the constitutional law for a world we create from the flux of experience. They are a priori because they are not given in experience or dictated by experience, while they regulate the 2 Logic and Pragmatism,” in Contemporary American Philosophy, ed. G. P. Adams and W. P. Montague (New York: T h e Macmillan Co., 1930), II, 3 1-3 2 .
Studies in the Philosophy of Kant
1 1
1
admission of experience to the status of evidence for or against some object. In Kantian language, they are like categories which inform raw experience and make it experience in the sense of knowledge of objects. At most, experience may suggest them to a pragmatically interested being or advise him to modify them; but experience cannot refute the principles employing them. While Kant sharply distinguishes between forms of intuition, which govern our experience, and the categories that govern experi ence in the sense in which he identifies experience with “ knowledge of objects,” Lewis denies the “ transcendental machinery” by which categories seem to be brought to bear upon the data of intuition; and he rejects out of hand the notion of any a priori regulation of experience in the former sense. For the transcendental machinery, often (though not by Lewis) misinterpreted as psychological, Lewis substitutes a pragmatic machinery that gives the categories the “ character of fiat” and of “ deliberate choice.” T he principle of sub stance, for instance, is not an existential statement that there is a permaneht in perception, as in Kant’s first Analogy of Experience. It is rather a freely chosen decision not to ascribe substantiality ex cept to those chains of experience which exhibit the analytical con sequences of our freely chosen definition of substance as that which is permanent amidst change. Such definition is not a command to experience; it is a claim which experience may or may not honor, but which it cannot abrogate. Such a principle of the permanence of substance is true “ no matter what.” A ll that my experience might force me to do is to admit sadly and with disappointment that I never seem to come across any substances; but it might suggest to me that my system of concepts might be a little less of an “ unearthly ballet of bloodless categories” if I modified my definition of substance. Only categories that do find continuous and repeated application can render knowledge effective in the control of our future passages of experience. M ind and the World Order ends with a kind of challenge to ex perience. In the last chapter, Lewis, having surrendered any govern ance of experience by a priori factors, says to the flux of experience: Be as chaotic as you will, I shall still be able to interpret you with some categories that will make intelligent anticipation possible. But he cannot venture to say what categories will serve.
1 12
Studies in the Philosophy of Kant III
It is easy to see why M ind and the World Order made Lewis a representative thinker among the analyst critics of Kant. Yet the book was not quite typical of the analytical movement as a whole. What was most distinctive in that book was the pragmatic origin and variability of the categories. But there were elements of realism, in various senses of the word, which hardly conformed to the typical positivistic model of a relativistic theory of categories, a theory without any perhaps metaphysical acknowledgement of the hard coerciveness of independent factuality. An Analysis of Knowledge and Valuation brings these divergences from the typical analysts’ position to light, and shows Lewis to be in many respects closer to the camp of the realist critics. If we think of orthodox Kantians (if there were any today) involved in a twofront war with, say, Ayer to the left and the ghost of Prichard to the right, we must think of Lewis as fighting against both the left wing and Kant, using some of the weapons of the realist critics against both, and yet so cautious and meager in his commitments to realism that the realists hardly can claim him as an ally. T h e tactical picture is therefore very complicated. But Lewis’ main effort is clearly di rected against the other analysts, for he regards the battle against Kant, on the central point, to have already ended with victory by those who deny that there is any a priori synthetic knowledge.3 Yet it was a Pyrrhic victory, for in it the conventionalism of the linguistic a priori made it incapable of doing the job assigned to the a priori elements of knowledge. A view that the analytic is equivalent to the a priori but is dependent upon linguistic conventions cannot, Lewis maintains, justify the epistemic function of the a priori, which is characteristic of knowledge independent of sense experiences. It explains the analytic by divesting it of that object-relatedness which alone can make cognitions true.4 Yet Lewis believes that the results of recent developments have clearly shown that all a priori knowledge is analytic, and that the Kantian question is a “ dead issue” in philos ophy.5 It was incumbent upon him, therefore, to develop a realistic 3 4 5
Lewis, Analysis of Knowledge and Valuation, pp. 35, 158. Ibid., pp. 36, 147. Ibid., p. 158.
Studies in the Philosophy of Kant
theory of the analytic in which the analytic is equivalent to the a priori and in which a priori knowledge really is knowledge and not just a decision to talk in a certain way, no matter what. Such a theory presents, however, only a minimum concession to the realists. I pointed out that armed with a realistic theory of the a priori, these critics of Kant had extended the scope of knowledge into speculative metaphysics. Because of Lewis’ theory of meaning, however, in which the meaning of a proposition resides in its veri fiable consequences in future passages of experience, nothing in the way of knowledge of objects absolutely transcending experience is gained by Lewis’ rejection of the typical analytical doctrine con cerning a priori knowledge. It is for this reason that Lewis is hardly a bona fide member of the realist school of critics of Kant. IV
Lewis’ criticism of the analyst critics is that they are not able to give a plausible account of the epistemic function of analytic judg ments. They make linguistic meaning basic, with the result that “ analytic” means “ determined by language rules” ; but “ determined by language rules” does not mean “ known independently of experi ence,” and it is. the latter which is characteristic of a priori knowl edge. Against Kant, on the other hand, Lewis argues that the necessary connections between intensions are not synthetic but analytic con nections. What can be known independently of experience is what follows by analysis of the meanings involved in our interpretation of the given, though not necessarily by analysis of the meanings actually entertained in consciousness. Sense meaning is the criterion in mind by which decision on the applicability of a term to a part of experience is to be determined, and this criterion may not be found merely by a cursory inspection of what is included in the sense meaning as actually entertained. The applicability of a complex concept to a part of experience requires a step-wise process. If the decision to apply the term “ square” presupposes the decision to apply the term “ rectangle,” then “ every square is a rectangle” is an analytic proposition. While “ every square is a rectangle” would certainly have been
Studies in the Philosophy of Kant
ii4
regarded by Kant as analytic, there are judgments that Kant held to be synthetic yet necessary; and this is what Lewis denies. Lewis joins with Kant in denying that there is an intellectual intuition of necessary but not analytic relations Between concepts or meanings. But Lewis affirms what Kant denies, to wit, that all epistemically necessary connections are logical connections; even if not formally logical, they are still connections between meanings. Kant’s theory on this point, of course, is that pure concepts refer to necessary schemata in time; that there is an intuitional or sensible element in even pure schemata which is necessary, determined not by the con cept but by the structure of time itself as a given; hence that meanings can be got from schemata that were not put into the schemata by the concept. T o the extent that the schema has its own a priori structure, what follows from it by a kind of phenomenological analysis is not restricted to what follows from the concept itself by a logical analysis. Lewis suggests that his sense meanings are schemata in the Kantian sense. The Kantian schema is the condition of the application of an intellectual concept to a sensuous material; but given the concept itself, the schema is not in any sense a logical part of it. One can think with the concept without knowing or using its schema; what the absence of the schema limits is not logical analysis, but appli cation to experience, and hence knowledge. Lewis’ schemata are very different from Kant’s; but on the basis of his conception of the schemata, which he does not sufficiently distinguish from Kant’s, Lewis brings a devastating criticism against the Kantian theory. The criticism is this. The schema must be included logically in the concept itself as a criterion of its application. It is not, as Kant believed, a “ third thing,” neither wholly conceptual nor wholly sensuous, which bridges the gap between concept and data. Accord ingly, if the schema as the sense meaning is not included in the con cept and stated in its real definition, it appears that a judgment using the unschematized concept is synthetic; on the other hand, the judgment expressed in the same sentence but using the schematized concept is a priori. This is the error with which Lewis charges Kant, and he says that Kant’s error is based upon a “ failure of analysis.” 6 I shall examine this alleged failure of analysis by reference to Lewis’ criticism of the second Analogy of Experience. Lewis argues 6 Ibid., p. 163.
Studies in the Philosophy of Kant
as follows. The notion of a necessary but nonanalytic proposition such as “ Every event has a cause” is based on an equivocation. “ Event” as a concept which does not contain “ having a cause” as a part of its intension is not the same as “ event” of which it is said to be necessarily true that it does have a cause. “ Event” in the second case means “ phenomenal event in one objective space and time,” while in the former case it is not so restricted, and denotes, inter alia, per ceptual or conscious events as well as perceived events. The judgment that every event has a cause is synthetic only when “ event” has the thinner meaning; but it is a priori only when “ event” has the richer meaning. It appears to be a priori synthetic only because it is erro neously believed that “ ‘event” means the same thing in both sen tences. A term needs to be fixed before one can pronounce a judgment in which it figures to be analytic or synthetic. In defining its meaning, Lewis says, we must include in its meaning everything needed in order to determine its applicability to experience; if we do not do so, the definition is faulty, and arguments based upon the definition are fallacious.7 None of the elaborate and almost infinite variety of criticisms of the Analogy offered during the past century and a half approaches in elegant simplicity this attack by Lewis. Yet its full evaluation requires a detailed consideration of matters in the philosophies of both Kant and Lewis which do not have the straightforward char acter of this perspicuous refutation, and which have been left in considerable obscurity by both philosophers. T o these more obscure matters it is necessary to turn now. V
It would perhaps have been more appropriate for Lewis to have accused Kant of a “ failure of synthesis.” For the essential question is not: What can be got from a concept or meaning by analysis? but: How are meanings constituted, so that analysis can get at the neces sary truths employing them? In Kantian language, analysis presup poses synthesis; and if we are to replace allegedly a priori synthetic judgments with analytic judgments, we must pari passu replace 7 Ibid., pp. 161-62.
116
Studies in the Philosophy of Kant
thinner concepts with richer ones—we must, as it were, provide a priori synthetic concepts to take the place of a priori synthetic judg ments. Kant says that the Critique exists fdr the study of the synthesis of knowledge, not for the analysis of concepts. Let us forget for the moment that Kant is concerned with establishing the possibility of a priori synthetic judgments; let us grant for the moment the validity of the Lewis criticism of the Second Analogy. But if we do so, only the locus of the problem is changed; to speak the classical language, we could say that we now have to do with a problem in the logic of concepts instead of in the logic of judgments. So long as both philos ophers believe that there is a rule by which later events are deter mined by earlier ones—as they do— the problem itself remains essentially unchanged. And Kant’s answer to it is fairly clear. The relation of a concept to an experience which falls under it is not a logical relation found by the analysis of the concept. It is not our conception of the conditions of its application which is essential in the constitution of actual knowledge, but it is the actual conditions of application themselves. This condition is the givenness of data in a variety of fixed temporal patterns, which Kant has argued from the beginning are not adequately defined except ostensively. T o take a simpler example than that of cause, we do not know how to apply the concept “ red” by knowing its “ analysis” (if it has one). Stating in the “ analysis” of the concept that “ ‘Red’ applies to red things” does not in the least help us to find out whether there are any red things or not. Similarly, “ ‘Cause’ applies to that without which an event would not occur” is a conceptual analysis (a poor one) of cause; but “ ‘Cause’ applies to an event in one objective time if there is another event which regularly follows it” does not help us in the application of “ cause” unless we are given some directions for deciding what is meant by “ follows.” And in Kant’s view, the mean ing of “ follows,” like the ostensive meaning of “ red,” cannot be found just by analyzing the connotation of a concept. It is for this reason that so much of the Kantian philosophy stands or falls with the success or failure of what is often regarded as its weakest part, the Transcendental Aesthetic. But unless there is an a priori content of experience (e.g., necessary spread-outness in space and time) to which some concepts have an intuitively necessary appli
Studies in the Philosophy of Kant
1*7
cation, a controlling factor in the constitution of a priori synthetic concepts is missing. Without intuition, the concepts are empty no matter how many concepts of intuitions are included in them. Of this Kant is firmly convinced, so much so that he even suggests that the concepts and forms of intuition may have a common root. While Lewis has written of the social origin and the pragmatic determinants in the constitution of concepts, throughout his work there is a Platonizing assumption of the directly given and privately inspectable ready-made meanings,8 and in his later book he “ take[s] reality as we find it already disclosed and conveyed in our commonsense meanings.” 9 The concepts we analyze arise through neither induction nor definition nor stipulation. Not through induction, because without them experience can neither be classified nor queried. Not through real definition, because real definition is an analysis of them and presupposes them.10 Not through stipulation or nominal definition, because meanings so established are merely linguistic. Sometimes definitions seem to be reports on what we mean or expect or intend; sometimes they seem to be reports on what the things themselves mean in terms of eventual experiences. But in any case they are fixed, and from the fixing of them there follows a hard and fast dichotomy of analytical consequences and accidental accompaniments of their application. Expressions for the meanings may be changed, and it is worthwhile to establish a lan guage in which what does follow from them analytically will be linguistically necessary; but we must recognize in advance what con nections of meaning, however expressed, are in fact necessary before it is worthwhile developing a language in which the sentences re porting on these connections are linguistically necessary. Lewis’ expressed purpose in Analysis of Knowledge and Valuation was not to give a phenomenological, transcendental, or genetic ac count of the structure and growth of knowledge, but to discover by analysis the criteria of the validity of knowledge we already have.11 It is not fair to criticize a man for writing the book he did write and 8 C. I. Lewis, M ind and the World Order (New York: Charles Scribner’s Sons, 1929) pp. 70-88. (All citations, however, refer to the page numbers in the Dover edition, New York, n.d.) 9 Analysis of Knowledge and Valuation, p. 22; cf. Academy edn., X , 353 n. 10 Analysis of Knowledge and Valuation, pp. 112 , 130. 1 1 Ibid., p. 22.
118
Studies in the Philosophy of Kant
not another one. But it is fair to point out that one may fairly inquire concerning what must have gone on before to provide us with these wonderful meanings which make a priori synthetic judgments otiose, inasmuch as they can be replaced with analytic judgments which say the same thing because they contain more complex concepts. VI We are now ready to return to the criticism of the Second Analogy. There are two passages in Kant relevant to this criticism, written nearly thirty years apart and having quite different implications. The first, an early fragment,12 suggests that the distinction between the analytic and the synthetic is arbitrary, depending solely upon the amount of knowledge we have of the analysis of the concepts. T he implications of this modern-sounding suggestion seem never to have been explored by Kant. Had they been worked out, many of the examples of so-called a priori synthetic judgments in Kant’s major works would never have been given, and the essential points which they are meant to illustrate would not have been hidden by Kant’s poor examples. I sometimes wonder, however, if Kant himself would have discovered what I shall try to show is his essential point, had he held to this early and all too easy belief. Be that as it may, when Kant was challenged late in life by Eberhard, who held precisely this gradualistic view of the analytic-synthetic distinction, the essen tial doctrine of the Critique had been established in Kant’s mind, and he was then prepared to give an answer to Eberhard which pro vides us with a key to the answer that I think he would have made to Lewis’ attack on the Second Analogy. Eberhard, like Lewis, believed that there were no synthetic a priori judgments, and that redefinition of the terms would show this to be the case. Kant replied13 by inviting his opponent to add any attributes he wishes to a concept, so that whatever he wishes to prove can be proved analytically. But then Kant asks him: How did you come to include in the concept precisely those attributes needed in order to render analytic what was previously said in a synthetic judgment? He cannot reply that he is giving a definition of the 12 Reflection 3928 (Academy edn., X V II, 35). 13 Academy edn., X X , 408-9.
Studies in the Philosophy of Kant
concept, unless he can show that in doing so he is obeying the rules of proper definition. That is, he must be able to show both that the newly added attribute is independent of those already included, and invariably present in the entities denoted by the original definiendum, so that the conjunction of the old and the new attributes has the same denotation as the original definiendum. A narrower denota tion w7ill not do, for that means that a new concept (with the same name) has been introduced by definition. In order to know the identity of denotations, he must know the connection of the inde pendent attributes before stating them in the new definition; and he must know this not by analysis, else the rule is infringed that the attributes stated in the definition must be independent. Hence definitions devised for the purpose of rendering synthetic judgments analytic are not real definitions, or in making them we must already know with certainty the synthetic judgment they were designed to translate into an analytic judgment. If they are nominal definitions, they do not do the job; if they are real definitions, they are (or in clude) the synthetic truths in question. In fact, Kant elsewhere states that real definitions— definitions which refer to the sensuous condi tions of their applications— when possible at all are synthetic a priori propositions.14 When possible at all: but Kant argues repeatedly that definitions of the basic philosophical concepts or categories are not possible. By a definition is meant a complete listing of essential and independent attributes; yet we are able to make analytic judgments without first possessing a definition.15 By a real definition, he means a definition that refers to a real predicate, a Bestimmung, which when found will show that the definition does have application. The Bestimmung, however, is not a predicate contained in the definition; it is not even the concept of a condition of application; it is the condition of application itself. A real definition is a true synthetic judgment, having objective validity. When dealing with categories, Kant gives nominal definitions of them; but these nominal definitions (e.g., “ Substance is that which is subject but not predicate” ) do not guide Reflections 2955, 2994 (Academy edn., X V I, 586, 606-7). 15 T he False Subtlety of the Four Syllogistic Figures, Academy edn., II, 61; Inquiry into the Distinctness of the Principles of Natural Theology and Morals, Academy edn., II, 282; Prolegomena, §4 (Academy edn., IV, 273). 14
120
Studies in the Philosophy of Kant
us in finding substances within experience. T o show the objective validity of the categories, we must cite the necessary condition of their application, and this is always some mode of sensibility, e.g., a schema. >> The categories with which Kant was concerned could not be given objective validity (application to experience of objects) by definition; this is known to every empiricist since Hume. However many conditions of application are formulated as predicates and put into definitions, a more extended Humean criticism of the enriched concept would again show that a necessary semantic bridge between conception and object had not been erected. Assuming a broader definition, a proof that the concept is objectively valid because the objective condition is included in the definition would still fall before the Humean criticism. By enrichment of the concept, we would make more judgments analytic, without giving any one of them a claim to objective validity. Apriority is not dependent upon such factitious analyticity; there must be recognition of some special dignity of function which makes formulating an a priori judgment in an explicitly analytic manner worthwhile.16 This special dignity of function for Kant lies in its making experience of objects possible. Now Kant, apparently unlike Lewis, did not tacitly assume the necessary temporal features of experience, nor did he assume the direct givenness of the causal relation. He tried to show that causality was inherent in the order of objective time, and in doing so he found that he had to add to the concept of sufficient reason not further predicates, but quite other conditions that neither Hume nor the rationalists had suspected. He found that he had to refer to “ possible experience” as the enabling ground of its application to any actual experience. T o have suppressed this interpretation for the sake of a merely conceptual enrichment of the concept, so that the Second Analogy could be made analytic, would have distorted the whole procedure of the critical philosophy, and would have only deferred the Humean question: How can this causal judgment, with a definitionally enriched set of terms, be valid of all possible experi ence of objects? T o answer this question, it does not suffice to con struct or inspect concepts; it is necessary to construct an object in 16
Cf. Analysis of Knowledge and Valuation, p. 167.
Studies in the Philosophy of Kant
12 1
series under the condition of sensible givenness of the data, when the order of the data may not correspond to the order to events. While Lewis calls sense meaning a schema, a part of the intension of the concept which must be included in the definition of it, I do not believe that his theory meets the Humean question even where Lewis believes that there is such necessary objective validity. In order to show this, I shall first examine the way in which Lewis would handle a concept that admittedly does not have necessary objective validity. Let us take the concept of red. If “ red” is to be applied to some experience, we can state conceptually some of the conditions of its application; we can state the conditions under which we would normally expect to produce a red sensation; we might even say that we should not apply the term “ red” except to things which look red. (The latter is, of course, an oversimplification; but the overly simple element in this illustration does not have any bearing upon the point at issue here.) That the concept does have objective application depends upon the fact or is equivalent to the fact that some things do look red; and this is of course not entailed or included in our “ definition” of red. For this reason, no one has ever said that the concept of red has necessary objective validity; red is not a category. A statement about red is a claim upon experi ence that may or may not be honored; it is not a fiat that experience must obey. Most of the concepts that Lewis deals with are of this kind, and are not categories in the Kantian sense of being necessarily applicable to every experience. Lewis’ categories are high-level ab stract concepts such as “ physical object” or “ Euclidean triangle,” which experience may or may not provide examples for. With respect to these concepts, therefore, neither Lewis nor Kant would wish to meet the Humean attack; Lewis joins in it.17 Only one monitory note must be added: it is sometimes thought that the Kantian cate gories are like the concept of red in our example, and that all prin ciples therefore are, in Hume's sense, “ relations of ideas” and all applications of them hypothetical. But implicit in the process of testing concepts for application to experience there are hidden other categories which, if Lewis’ philos ophy is to work, do have a necessary objective validity, do involve 1 7 Ibid., p. 353 and note.
Studies in the Philosophy of Kant
122
a real anticipation of experience. This validity and anticipatory function cannot be got at analytically merely by including a concept of the experiential condition of application in the definition of the category. You cannot build the condition of application into such a category by logical synthesis, without falling into an ontological argument. As Kant says,18 it is the sensuous condition itself, a neces sary given, which provides the categories with validity for experience. The concept of the sensuous condition will not serve. Such categories are necessarily associated with a sensuous schema, and therefore they cannot fail to have objective reference. The categories I have just described formally and epistemically resemble the Kantian categories. I make no defense of the specific list of categories Kant gave; but his list does have the virtue of not being rhapsodic and episodic. Lewis has correctly remarked that “ the easily recognized categories do not cover the easily recognized distinctions,” 19 and we do not find anywhere in Lewis’ writings a list of the concepts of which he ought to have said that their appli cation to experience is not contingent upon the content of experi ence. I will suggest four such categories or principles which function in Lewis in much the same way the categories function in Kant. The if . . . then of terminating judgments and the serial character of the verificatory experience and procedures are categorical in the sense of being illustrated in every possible experience. They are not cate gorical by virtue of any definition of time or of the relation of real connection, which experience might or might not illustrate. Then there are two principles, the Principle of Induction and the principle that “ mnemic presentation constitutes a prima facie probability of past actuality,” which are “ ingredient in, and together with the certainties of given experience, are constitutive of our sense of that reality which we cannot fail to acknowledge, unless we would repudi ate all thought and action and every significance of living. T hat we cannot do.” 20 Neither of these principles is an analytic proposition. Yet they are both a priori in the two senses attached to this word by Kant, Critique of Pure Reason, A 244-45; omitted from B. 19 “ Logic and Pragmatism,” p. 47. 20 Analysis of Knoxvledge and Valuation, p. 362. 18
Studies in the Philosophy of Kant
1 23
and not merely in the one sense in which the a priori is characterized by Lewis. That which is a priori for Lewis is irrefutable by experience because any experience which would refute it is ruled irrelevant by the classification of experience into evidence and non-evidence by the meaning itself. This is characteristic of the analytic a priori. But the principles that I am considering at the moment are a priori in the Kantian sense, a sense neglected or denied but still used by Lewis. This is the characteristic that the concepts or principles are necessary in application to experience, and therefore universally applicable to experience, be its content what it may. The justifica tion or deduction of the a priori that has both of these characteristics cannot be that of the analytic a priori. The justification of the prin ciples by Lewis is precisely of the kind that Kant gave of his prin ciples, to wit, in conjunction with the given they make experience as knowledge of objects possible, whatever be the generic traits of the objects. It is easy to see that defining “ mnemic presentation” in such a way that it would logically imply any probable propositions about the past would accomplish nothing. But to speak of these prin ciples as having a pragmatic vindication rather than a logically analytic justification should not obscure the fact that the mode of justification employed is of the kind that Kant called a transcendental deduction. Lewis is almost alone among analytical philosophers in the im portance he attaches to these uniquely fundamental principles. T o say that there are principles or concepts which must be applicable to every experience is to say something the pragmatic analysts are not in the habit of saying. Such a priori presuppositions cannot be rela tivized by linguistic analysis, for they do not arise by analysis of given conceptual meanings or by conventional stipulation of what we shall decide to mean by “ induction” or “ hypothetical connection” or “ mnemic presentation.” Their application is not to just those parts of experience selected by us because of harmony with them. They are principles for the interpretation and anticipation of experience itself; Kantian apriority and not analytic apriority is their mark. The if . . . then of terminating judgments and the seriality of verificatory procedures are constitutive of experience somewhat like the Kantian categories and forms of intuition respectively; the principle of induction and the principle by which memory is justified are
124
Studies in the Philosophy of Kant
regulative very much like the dynamical principles of the Critique.21 For the elaboration and defense of these a priori presuppositions, the Critique— though every sertfence in it should prove wrong—has given Lewis a model for the justification of principles as conditions of possible experience and knowledge of a common objective world. It is these presuppositions that are the mark of Kant, albeit faded, that is still discernible in Lewis’ epistemology. 21 Not to be confused with the regulative Ideas. Cf. Critique of Pure Reason, A 178=1} 220-21; A i8 o =B 223.
V lll
NICOLAI HARTMANN’S CRITICISM OF KANT’S THEORY OF KNOWLEDGE' Early in life, Nicolai Hartmann was a representative of logicistic Neo-Kantianism, but as early as 1915 he began his apostasy from the tenets of Marburg. So strong was the Kantian influence and so strong has it remained, however, that opposing Kant was no easy task for Hartmann as it had been, unfortunately, for those realists who did not think Kant worth the trouble of attentive reading. Thus, instead of ignoring Kant and instead of attacking him outright and from the ground up, it was Hartmann’s difficult task to say precisely where he thought Kant was in error. It would not be far wrong to say that every one of his books is, in a degree, an Auseinandersetzung with the Kantian philosophy. He has given an outline of a book on Kant “ which must be written, and which I see as a task for our time, but which I myself shall not write.” 2 And no one has been in a better position than Hartmann to give such directions, not only because of his unsurpassed learning in the letter of Kant’s text and his old sympathy with the critical philosophy, but even more because he does not fall under the ban by which all metaphysicians have been “solemnly and legally suspended from their occupations”3 until they can answer the question as to how synthetical judgments a priori are possible— for Hartmann, almost unique among modern realists, has attempted to meet this question. r e pr in te d
, w ith the permission o f the Editor, from
Philosophy and P h e
nom enological Research, I I (1942), 472-500. 1 It was the a u th o r’s privilege, as Julius Rosenwald Fellow, to be a m ember of Professor H artm a n n ’s seminar on the Critique of Pure Reason at the University of Berlin, 1937-38. 2 Nicolai H artmann, “ Diesseits von Idealismus und Realismus,” Kant-Studien, X X I X (1924), 161. Hereafter referred to as “ Diesseits.” 3 Prolegomena, §5 (Academy edn., IV, 278; trans. Beck, p. 25). 125
Studies in the Philosophy of Kant
That Hartmann has discovered many weak points in Kant’s philos ophy cannot be denied. That he has cogently argued for the inevi tability of many of those difficulties which have been pointed out by others, by showing how they are rooted in Kantian fundamentals and are not due, as many have hoped, to mere carelessness of expression— neither can that be denied. But no exposition of these points is needed; it would be indeed presumptuous for another writer to undertake to defend Hartmann or even to give another exposition of his views when Hartmann himself is so convincing in his own argu ment and so well able to deal with all who disagree with him. And if the writer mentions Hartmann’s specific arguments only to disagree with them, that is no sign of lack of sympathy with or respect for those positions which are not mentioned. T H E P R O B L E M O F S T A N D P O IN T S
The subtitle of Hartmann’s Kantgesellschaft lecture of 1922, “ Diesseits von Idealismus und Realismus,” reads, “ A Contribution to the Differentiation of the Historical and Transhistorical in the Kan tian Philosophy.” This indicates accurately the form of Hartmann’s treatment; it is not critical in the sense that it marks a complete renunciation of Kantianism, but rather in that Hartmann is inter ested in distilling out of Kantianism that which is of lasting value and thus free from the limitations of Kant’s standpoint in history. The “ historical” and the “ transhistorical” in this study refer to the same elements in the works of a philosopher which Hartmann has distinguished respectively as the “ systematic” and the “ aporetic” moments in philosophical construction. “ Whatever in philosophy is transhistorical,” he says, “ must necessarily be above all stand points.” 4 “ Arbitrary presuppositions,” such as idealism and dogmatic realism, are always standpunktlich.5 The standpoint, as such, always determines a system, because its acceptance by a philosopher means that all problems will be dealt with from a previously assumed perspective, which will sometimes involve distortion or force the philosopher to do violence to a particular problem so that its solution 4 “ Diesseits,” p. 162. 5 Ibid., p. 165.
Studies in the Philosophy of Kant
127
can be brought into a harmonious whole whose structure has been decided upon in advance. Thus when Hartmann distinguishes be tween the historical and the transhistorical in Kant, he means to distinguish between the systematic and the aporetic, or between the constructional and the “ phenomenological” (in his very broad usage of this latter term). With reference to Kant himself, that means that he will search out those parts of his philosophy which are determined by Kant’s “ arbitrary assumption” of transcendental idealism, reject them because they are “ historical” and systematic, leaving the rest of the philosophy as a result of the presuppositionless aporetic or phenomenological method which bases itself solely on the temporary “ standpoint” which the specific and single problem and the present phenomena afford. This resulting philosophy will be a fund of phenomena which have no specific reference in themselves to their function in Kant’s actual systematization.6 But there are several questionable points in this ideal. First, a formal question may be asked. Can the aporetic method proceed without presuppositions? T o be sure, there is a difference in method which is important; Hegel and Kant are sufficient proofs of this in themselves. But is the ideal of presuppositionlessness attainable, and even, we may ask, is it meaningful? As to its attainability, perhaps only a history of the further refinements of phenomenology can decide. No one can claim, I think, that it has been attained, and Hartmann does not assert categorically that it has. One has to read only the first sentence of Hartmann’s Metaphysik der Erkenntnis to see his presuppositions: “ The following investigations proceed from the conception (Aufjassung) that. . . .” It is well to have such suppositions on page 1, but that does not make them any less of a presupposition. In Hartmann’s case, this “ presupposition” is that knowledge is a knowledge of something independent, and further on he argues that his “ standpoint” is not necessarily realism simply be cause it makes this supposition. The belief is not a schlechtweg 6 Here is a noteworthy resemblance between the evaluations by Hartmann and by Ledger Wood (“ T h e Dialectical Method,” in The Heritage of Kant, ed. G. T . Whitney and D. F. Bowers [Princeton: Princeton University Press, 1939])- Whether Hartmann specifically has influenced Wood I do not know; but it is a fact that many of H artm ann’s opinions may be illustrated equally well by quotations from his own works or from W ood’s essay.
128
Studies in the Philosophy of Kant
hingenommenes Vorurteil but only a vorgefundener Ausgangspunkt which will be held only as long as it is adequate to the problem.7 The expressed readiness to give up the conception does not render it less of a presupposition, but it is only a confession of such a philosophical seriousness as a systematist like Hegel might make. In a word, it is difficult or impossible to see how, without a stand point of some kind, any question could arise from a conflict of phe nomena, how an Aporie could be acknowledged, or how any situation could be recognized as problematical. For why should the world “ make sense” ? Without presuppositions that it does, it is the most obvious thing that, prima facie, it does not. Only the general pre supposition that a philosophy is possible, that the wonder and curiosity of the lover of wisdom can in some way be at least partially satisfied, makes it possible for us to leave even for a moment the rhapsodic phenomena and to hunt for their “ meaning.” And to go a little further, there is a more specific and peculiar pre supposition. (I call it “ peculiar” because it is less generally held than the first, which is the a priori beginning of all philosophizing.) It is that a systematic approach to philosophy must do violence to some of the problems it touches. Granted that there may be no such thing as a consistent and true system of philosophy, it does not follow that we should reject systematicity altogether, nor that we should, as Hartmann says we should, relegate systematic thinking to the end of the endless task; the antidote for bad systems is better systems. Hart mann is proud of his aporetic method because it does not look “ longingly for results.” 8 But this wise caution against enthusiastic haste in metaphysical speculation which would lead to some precon ceived and uncriticized goal amounts to an actual fear of solution of problems in Hartmann’s own thinking. T o take some examples: Hartmann holds that all of Kant’s antinomies are unresolvable be cause they involve a relation of phenomena to noumena, and this relation is “ irrational.” “ All attempts to solve the antinomies are merely perspectival (standpunktlich) and artificial.” 9 And when idealism has an answer to the question of the relation of the subject 7 Nicolai Hartmann, Grundziige einer Metaphysik der Erkenntnis (2nd edn.; Berlin: de Gruyter & Co., 1925), p. 193. Hereafter referred to as Metaphysik. 8 “ Diesseits,” p. 164. 9 Metaphysik , p. 233.
Studies in the Philosophy of Kant
129
to the object in a priori knowledge, Hartm ann fears it because of its lack of mystery, its ease.
Idealism weakens the meaning of the thesis [that principles of the object are identical with those of the subject] since it takes the object into the subject, even though only into the transcendental. The identity of cate gories is then almost self-evident. But if one lets the object stand as real beyond the subject, even the transcendental, the whole burden of the difficulty falls to [the principle of] the identity of categories to bind the subject and object together as they are in a priori knowledge. That the identity of categories can do this constitutes its strength.10 T o object to a philosophical principle because it is “almost selfevident” is almost unique in the history of philosophy. Prima facie it would appear that H artm ann avoids the “almost self-evident” principle in order to show the strength of another interpretation, but this strength consists in overcoming difficulties avoided by Kant’s own expression.11 Again, H artm ann’s fear of solution leads him to say,
Ontology makes the attempt to assume the absolute minimum of hypoth esis for the consideration of the philosophic problem. The criterion of its legitimacy lies in the question, how well does it succeed in taking up the irrational loose ends of problems into its insight and putting them in the place in the connection of the entire system, into which, in spite of their irresolvability it does introduce them? A closed rational system can never satisfy this demand. But an open system, which closes itself only beyond the limits of knowability, has within itself room for this possi bility.^ 10 Ibid., p. 187. Here the concept of “ identity of categories” is introduced only as a part of a typical example of H artm ann’s argumentation. It will be discussed in its own right in the fourth section, “ T h e Identity of Principles,” below. 1 1 T h a t Kant avoids them will be shown later. Hartmann refers later to the ob jectivity of the Metaphysik as having been an ultima ratio to be defended in other ways. But the defense, as we shall see, is only of “ gnoseological being,” of the object as different from the subject in knowledge, and not of “ ontological being,” i.e., of real being without any dependence on the knowledge situation itself. 12 Metaphysik, p. 306. “ Ontology,” in this sense, is largely a project, not a product. It is the “ neutral” treatment of subject and object, and conceivably of mere appearance, as forms of being. As such a project, “ It follows, as a matter of principle, only the problem, and allows only that to stand in the picture of the world which is demanded by the problem. It moves as little as possible away from the natural standpoint, and it preserves its point of view as far as it is valid. . . . Its
i go
Studies in the Philosophy of Kant
Examples could be multiplied indefinitely to show Hartmann's fear that solution means rationalization,13 A more material objection can be rilade to Hartmann’s own inter pretation of what the Kantian “ system” is. It is not too much, I think, to say that many of the views and emphases Hartmann attributes to Kant are largely Hartmann’s own presuppositions which he is now for the first time rejecting. That is to say, the “ Kantian system” Hartmann rejects is a Neo-Kantian system. Hartmann is, no doubt, quite correct in saying that there is in Kant a “ decisive preponder ance” of the aporetic over the systematic tendency.14 After this is admitted, what is it that remains to be attacked? It is Kant’s “ sys tematic moment” of idealism, which is merely a historically condi tioned vestige of rationalism. But it is significant that this is the moment which is decisively preponderant in Marburg Neo-Kantianism. When Hartmann denies the “ systematic consistency” of the Kantian philosophy for the admitted presence of the thing itself, ought not his criticism perhaps be directed more at his own teachers than at Kant himself? For have we any reason to believe that critical idealism is the proper name for Kant’s ultimate philosophy? Mar burg Neo-Kantianism could not, to be sure, admit the thing in itself without falling victim to Hartmann’s attack, but that is because it was professedly an ultimate idealism. Hartmann sees in the thing in itself the “ pre-eminently critical concept” in Kant15— but Kant had no very different opinion of it. Kant had a transcendental idealresults must bear the stamp of some standpoint: its methods can remain this side of all standpoints and pre-decisions” (Metaphysik, pp. 179, 180). T h u s we may say that ontology, as understood in the passage quoted, is the “ aporetic standpoint” itself, if such a term is permitted. Though the text is not always unambiguous in this point, it seems necessary to separate ontology in this sense from H artm ann’s own substantive ontology, which is the system or science of the principles of being and which is largely realistic. 13 It is well to point out here the meanings of rationalization which may be con fused in this attitude of H artm ann’s: it may mean systematization (which is ac ceptable at the end, if it is possible); a renunciation of the irrational in its various senses (whose recognition as such constitutes the metaphysical problem of knowl edge as Hartmann sees it, as he, by implication, describes his work as “ eine Phäno menologie der Erkenntnis als Wesensanalyse des Metaphysischen im Erkenntnis phänomen” [Metaphysik, p. 36]); or a false systematization, a “ standpunctual” act of violence in an almost psychoanalytical sense. But that all systematic procedures are bad rationalizations in the last sense seems to be assumed by Hartmann w ith out any grounds. 14 “ Diesseits,” p. 165. 15 Metaphysik, p. 220.
Studies in the Philosophy of Kant
ism, but it did not extend to the “ noumenon in the negative under standing” ; nevertheless, the latter is still there. Does that then mean, as Hartmann seems to believe, that Kant was not systematic, but aporetic? I f Kant’s ultimate systematic attempt was in the direction of idealism, we should have to agree with Hartmann that in this concept the problem triumphed over the system. But we have no right to say that Kant’s ultimate system was transcendental or critical idealism. T o be sure his “ refutation of idealism” does not affect his transcendental idealism, but the immediately following chapters on the distinction between noumena and phenomena would hardly have been written by a metaphysical idealist. Hartmann himself recognizes the possibility of a “ systematic realism” in Kant.16 Nevertheless, this is precisely the form of criticism which a member of the Marburg school would be expected to make if he ever rejected “ Kantianism.” We should be very sure what it is which suffers from the attack. The critical philosophy as a whole does not; those parts of it which may be described as idealistic may— though this is a matter we shall have to discuss in greater detail later; but only a Kantianism which stops at the beginning of the dialectic, or a NeoKantianism which is wholly transcendental and idealistic will be utterly destroyed by Hartmann’s argument concerning the necessary position of the “ realistic” elements in Kant, if it prove successful. One more question of “ standpoint” needs to be considered before we go into the details of Hartmann’s criticism. Every age reads history and the historically significant in terms of its own needs and interests. When we separate the historical from the lasting in a philosophy, are we not in danger of merely distinguishing between the “ old fash ioned” and the “ new fashioned,” even though the latter will be just as “ historical” in its way as the former was? Adolf Seelbach17 raises this question, and in the symposium of the Kant Society in 1931 Emil Utitz implied that he thought Hartmann’s shift to realism was zeitbedingt. Hartmann’s answer to his critic on that occasion is worthy of study: I would regard it as misleading to see the problem of reality itself as merely an expression of a passing time. It is a very old basic problem of philosophy, which periodically again and again comes to the fore. And 1 § Metaphysik, pp. 148-49. 1 7 Adolf Seelbach, Nicolai Hartmann’s Kantkritik (Berlin: Pan Verlag, 1933).
132
Studies in the Philosophy of Kant
rigorously taken, all and every philosophy must deal with it. .. . From this standpoint [and not from the standpoint of fashion] the investiga tions I have presented take their origin'. T h ey should not serve a “ move ment” but a timeless demand.18 ' ^
When one goes into Hartmann’s discussion of the problem of reality and its evidence, one easily sees how it is a “ timeless demand” —a demand that philosophy not be absurd on its very face. For the “ timeless demand” for reality is so little speculative, and so indis criminate in regard to the reality which will satisfy it, that the sort of reality Hartmann provides is of little or no distinctive use in the Wendung zum Realismus.19 Ontology, the assertion of being, “ has a standpunctual indifference to idealism and realism, and the same distance from them even though it is not equally visible in the two cases.” 20 The ontology which is assumed as ultima ratio in the Metaphysik der Erkenntnis and argued for in the symposium of 1931 as part of the shift to realism is wholly negativistic; there must be being, but what it is the argument does not and cannot show: “ Being is the common sphere in which subject and object stand over against each other.” 21 E M P I R I C A L R E A L IS M
It is from the proof of the givenness of the real that we must begin our detailed discussion of Hartmann’s strictures on Kant. Hartmann does not directly charge Kant with failure to account for empirical reality, and for that reason it might seem, at first, as if the argument in “ Zum Problem der Realitätsgegebenheit” would be irrelevant to our problem. Perhaps ultimately it is; but strictly speaking this entire paper by Hartmann seems to me to be irrelevant to his own position too. Nevertheless he thinks his realism is strengthened by the argument, and if his position is strengthened, Kant’s more “ ex travagant” metaphysics can be attacked by implication. Only when we see what Hartmann attempts to prove and how, can we say how significant it is for Kant’s position. Nicolai Hartmann, “ Zum Problem der Realitätsgegebenheit,” in “ Philoso phische Vorträge veröffentlicht von der Kantgesellschaft,” No. 32 (Berlin: Pan Verlag, 1931), p. 96. Hereafter referred to as “ Problem.” See Metaphysik, p. 138. 20 Ibid., p. 300. 21 Ibid., p. 138.
Studies in the Philosophy of Kant
133
For the establishment of realism it does not suffice merely to throw the responsibility for proof upon the shoulders of the skeptic and “ idealist,” for as Hartmann said in his reply to Moritz Geiger, the grounds for the displacement of the burden of proof must be shown forth in some phenomenon, and the exhibition of this phenomenon is itself productive of evidence for the givenness of reality.22 This is of course Kant’s own procedure against the idealists in his Refutation of Idealism, but it is easily seen that this refutation does not really meet Berkeley’s problem at all. Kant accepts the “ reality phenom enon” of empirical realism just as much as Hartmann does, though Hartmann does have objections to this assumption on the part of Kant. Hartmann holds to an a priori certainty of empirical reality, and he says, K a n t satisfied h im se lf in sh ow in g how em p irica l realism was co m p atib le w ith his idealism as one o f its corollaries. B u t if he had seen it as a p r io r i, then precisely acco rd in g to his ow n con cep t o f a p rio rity it w o u ld have been necessary to dem on strate, in ad d itio n to this co m p a tib ility , also the co n d itio n s o f its u n ive rsality an d necessity. F o r its claim to u n iversal v a lid ity is p erh a p s n a iv e ly obvious, but n ot p h ilo so p h ically obvious. I t is the precise p ro b le m .23
Let us see, then, why Kant does not carry through an a priori justi fication of the assumption of empirical reality. It is safe to say he never thought of it at all, since the givenness of reality is not a formal a priori Sachverhalt. Therefore “ according to his own conception of apriority,” which was formal justifiability, he is not required to “ deduce” the belief in empirical givenness; and if the mere givenness as such of reality were a priori for him, needless to say his conception of apriority would have had to be different; he would have had to be just a phenomenologist. T o make this requirement of Kant with his own conception of apriority would be to require him to use an ontological argument to deduce givenness. Hartmann had begun his Metaphysik der Erkenntnis, as we have seen, with the supposition of reality. There it was taken as an ultima ratio, as he himself later says. His argument here, in the “ Zum Prob lem der Realitatsgegebenheit,” is supposed to bear this assumption out and justify it. The argument moves from “ emotional-transcend 22 “ Problem,” p. 85. 23 Metaphysik, p. 132.
134
Studies in the Philosophy of Kant
ent acts,” phenomena which are so constituted that they “ tran scend their own phenomenal character,” or phenomena which have the marks of self-transcendence within themselves; but it merely establishes the fact that the concrete jnan is in a real environment. And this has never been denied. He tells us nothing about the reality which confronts man, except that it is as real as he. This failure to determine its nature is a consequence of the nature of the phenom ena to which he appeals. T h is mode of givenness of reality is not only different from that of knowledge, but occurs independently of it. Being affected by something does not wait for the occurrence as such to be known. It is immediately there. Of course knowledge of the situation can follow it, but it need not. What it really “ was” which happened to me or affected me can remain hid den to me.24
But the that and the what do not appear equally visible: “ The weight of the givenness of reality in an act is the greater the more indissolubly the reality of the object is connected with the actphenomenon.” 25 In other words, the more concrete and immediate the experience of reality, the less we know about it and the more certain we are that it is. Cognition containing a theoretical element has less certainty in the more it asserts than a mute Erleben has in its mere acknowledgment that something has happened somewhere. The only ground which we have for supposing that which we do think to be real is that we can apparently have emotional and cognitive acts related to the same thing, in other words, that there is one reality and not two.26 But why, we may ask, so great a devotion to the emotional“ Problem,” p. 18. 25 Ibid., p. 30. 26 T h e interpretation at this time does not bring it out, but here we have a Zwei-Instanzen System as criterion of truth as this has been presented in H art mann’s earlier works. In a word, if two “ ideas” of a thing agree they may both be wrong; but if we have two “ sources” of experience and their deliverances agree, then their agreement in diversity is a criterion of their dependability. Hartmann uses in his other works the different sources of apriority and aposteriority as the two instances, but if in any way he could separate emotion and cognition, the argu ment could be employed very usefully here. But in this lecture, more than in any of his other writings, Hartmann is close to a kind of Lebensphilosophie or anthro pology, and he emphasizes the errors introduced into the theory of knowledge by the mistake of isolating knowledge from other life-functions. For this reason, the fact that we erfahren and erleben the “ same world” cannot be used as a criterion of its actuality. 24
Studies in the Philosophy of Kant
»35
transcendent acts? That Hartmann’s phenomenological analysis of them is subtle and thoroughly accurate no one denies, but why does it occur here? In order to show both the strength and the weakness of his position, let us compare what he says about these acts with what might be or has been as truly said about cognition. “ This con frontation [Betroffensein] is thoroughly real and is experienced as real. And because it is a confrontation with ‘something,’ immediately behind it must stand the happening [Widerfahrnis] itself, by which the subject is touched and as the real thing which is experienced in the act.” 27 T o be sure, if the experience of having something befall you is accurately described in these words, then there must be the thing there which does as a matter of fact befall you. But in knowl edge the argument can be phrased in precisely the same way, and with as easy and obvious a phenomenological justification. Let us say: “ Cognition is thoroughly real, and in every experience it is known as real. And since it is a cognition of something, so there must be immediately behind it the ‘cognized’ itself, which the subject cognizes, and which is the real thing which in knowledge the subject does as a matter of fact cognize.” The only way to avoid this analogy is to say that, in Lockean terms, it is the idea which is cognized while the thing may be felt (as reminiscent of the function of intensity and “ sensible knowledge” in Locke); but Hartmann of course explicitly rejects this anti-phenomenological construction (as for example in his sharp attack on Leo Polak). Further: “ In this occurrence, self obtrusion, oppression [Zustossen, Sichaufdrängen, Bedrängen], the happening [Widerfahrnis] shows a weight of reality which the subject cannot defend itself from. This reality is in a sense given; and against this, skeptical and idealistic attack must remain mute.” 28 But Berkeley’s argument for reality is precisely the same: I am acting in putting myself into a position to experience (I open my eyes) but what I experience cannot be my product and is not affected by my desire or dislike of its own nature. “ Whatever more there is—as that I perceive such a particular smell, or any smell at all— this is inde pendent of my will, and therein I am altogether passive.” 29 That Berkeley thinks of it not as having an altogether independent ex27 “ Problem,” p. 16. 28 ibid., p. 16. 29 George Berkeley, Three Dialogues between Hylas and Philonous , ed. Colin M. Turbayne, “ T h e Library of Liberal Arts/* No. 39 (New York: T h e Liberal Arts Press, Inc., 1954), Dial. I, p. 37.
ig6
Studies in the Philosophy of Kant
istence (in his sense, matter and in Hartmann’s, Ansichsein) does not affect either the cogency of the argument or its relevance to Hart mann’s position, for both are interested here not in the ontological status of the thing but solely in its existential position with reference to the concrete and particular subject. And for both of them it is equally real and equally evident that it is real; but for Hartmann to infer Ansichsein and ultimately realism from it is a subreption, essentially like Dr. Johnson’s refutation of Berkeley, and no more justified. But it is further argued that Erfahren in the broad sense of L eb en serfahrung is in its own nature “ the act-correlate of the happening [Widerfahrnis]” and it is immediately known as such. “ I have not spoken of subsequently breaking through the subjective limit, either in transcendent acts or anywhere else.” And in cognition this phe nomenological description also holds. For in the phenomena of knowing, we do not experience the object as “ idea” or “ modification of mind,” but without saying anything about the ultimate dignity of the object, we have in cognition an experience of “ objects.” If we do not, “ object” is merely a flatus vocis.30 When the givenness of reality is on such a firm foundation in our theoretical experience, where the object is not only given as real but also as definite, why go to another form of argument for reality? It is a practical matter. “ In experiencing [in this broad sense] or suf fering, it would be senseless to understand the happening [Widerfahrnis] as merely supported by the act [aktgetragen], standing and falling with it, and thus as unreal.” 31 But this would be no more nonsensical (and no one has ever supposed that solipsism or even skepticism was practical) than to do the same thing with the act of “ Problem,” p. 16. For this reason, it seems to me one may rightly suspect some of Hartm ann’s phenomenological descriptions of knowledge. Especially note worthy here are the descriptions of the Satz des Bewusstseins and the B ild des Objects im Subjekt (.Metaphysik, Part I, sec. 2, ch. 5). T h e fact is that one of the great apories of his system arises from the conflict of this Satz and intention. T h e antinomy of consciousness (p. 6), the conflict between “ knowing consciousness” (Satz der Erkenntnis) and “ knowing consciousness” (Satz des Bewusstseins) may appear to many to be artificial, since it is based on the statement, “ Zum Wesen des Bewusstseins gehört es, dass es nie etwas anderes als seine eigenen Inhalte zu fassen bekommt, nie aus seine Sphäre heraustreten kann” (p. 59). T h e antinomy, which Hartmann overcomes only at great metaphysical expense, seems to me to be phenomenologically ill founded. 31 “ Problem,” pp. 30-31. 30
Studies in the Philosophy of Kant
»37
observation. If one does make this supposition, “ One does not at all understand the act, does not comprehend its phenomenon, and has playfully mistaken the seriousness of human life.” But is the serious ness of human life more directly or obviously threatened by the failure of this defense than it would be if a solipsism were intellec tually founded? For to attack this argument, like the defenses against intellectual solipsism, is still possible, if one cares to go on. Skep ticism can . . . c e rta in ly ag ain an d ag ain v a lid a te the vie w that h ap p e n in g s, obstruc tions, fates [Widerfahrnisse, Widerstände} Schicksale] are o n ly self-created forces. B u t such stereotyped a rg u m e n ta tio n becom es less a n d less sig n ifican t [wesenloser] the stro n ger the p h e n o m e n a l transcen den ce o f the act becom es. I t is pu sh ed m ore a n d m ore in to em p ty abstraction. A n d so it exh au sts itself, in that it becom es objectless [sic!].32
But is this last Potenzierung of the skeptical argument any more arbitrary and “ objectless” than the first one was, since the very first attack of the hypothetical skeptic is an attack on act-transcendence in its most naïve and outstanding form? C’est le premier pas qui coûte. In fact, though it must be admitted that all arguments against skepticism of this variety are futile so long as the skeptic makes his stereotyped reply, still the single judgments have a greater “ objec tivity” or “ possibility of meeting the object” if emotional tran scendence is forgotten. Moritz Geiger, in fairly close agreement with Hartmann, argued that since illusion could be discovered only by a comparative study within experience, experience as a whole could not be illusory, such a statement would be not only false but meaning less. But Hartmann doubts the reality-evidence of these judgments which are clarified and abstract enough to be tested in this way, and it is true, though neither Geiger nor Hartmann argues in this way, that Geiger’s disproof of the possible illusoriness of all experience will hold whether knowledge is a Realitätzeugnis as they believe or only a Sinneinheit.33 32 Ibid., p. 31. 33 But one may ask what is the criterion in Hartmann’s scheme? Every emotional act, except perhaps “ moods,” claims a certain transcendence and objectivity. But what of pathological cases? And Hartmann explicitly refuses to base an argument for the existence of God upon the mere emotional conviction. “ In regard to the re
138
Studies in the Philosophy of Kant
Empirical realism then can be neither dialectically proved nor refuted. But why is a proof needed? Helmut Kuhn pointed out that every form of realistic argument grows out of a polemical situation since it is always directed at some specific form of Realitatsbezweiflung. Hartmann’s argument is directed against the “ pseudological” form which characterizes ancient skepticism’s recognition of the pos sibility of deception but it neglects the more modern critical limita tions. Opposing purely empirical skepticism, Hartmann’s attack can lead only to empirical realism. But that this empirical Ansichsein (which is not denied by idealism) can hold its own against Kant’s or even Berkeley’s attack— that possibility is not even touched by Hart mann. Thus we may say, if the realism Hartmann is here arguing for were metaphysical and not merely phenomenological, it would be systematic, conditioned by a standpoint in the stream of history as Utitz and Seelbach say, and therefore it would according to Hart mann lack weight. But realism of this kind is not systematic, and it is to be sure one of the timeless demands we may make on any philos opher or sane man. But then it does not touch Kant, even by implica tion. The idealists were quick to realize this. During the symposium Willy Moog said that a phenomenological analysis of the acts could be built into an idealistic as well as into a realistic system, and thus Hartmann’s polemic did not touch really critical idealism.34 Hart mann’s answer is revealing in that it shows how far he had gone since he expressed the opinion that Sein was a neutral category including subject and object and therefore jenseits idealism and realism, which is his position in the Metaphysik der Erkenntnis. “ Such a true ideal ism would have to draw, however, also the further consequence, and allow the ‘acknowledged’ reality of the world, and indeed of the empirical world, to hold as existing in itself,” but he doubts if it could be called “ true idealism” if it drew this consequence.35 But Hartmann ligious experience, I do not indeed mean that from it one could directly ‘prove’ the existence of God, though perhaps that in it the reality of God could be immediately given. But even this only if there is a religious experience in the sense of a wirkliches Erfasstsein of a higher power . . .” (“ Problem,” p. 88). But this does us no good at all, for every such experience claims to be a ivirkliches Erfassen. 34 “ Problem,” p. 30. 35 Ibid., p. 88.
Studies in the Philosophy of Kant
139
is here confusing his own useful concepts of gnoseological and on tological Ansichsein; and from the first, which is proven, we cannot infer the second without begging the entire question. Hartmann goes even further in his attack on the idealists who would allow this much of “ realism” in their system, and thus he by implication attacks Kant.36 Against Liebert, who like Hegel thought of the true idealism as being the only true realism, and who sought a synthesis beyond perspectival distinctions (as opposed to Hartmann’s epoche this side of idealism and realism), Hartmann replied: T h e opposition of matter and spirit is not a contradiction, as that of views from a chosen standpoint is; they may easily be united in one world view. Therefore there is no need for a “ standing beyond them,” even if they are both fully autonomous; nor is there any need for a dialectical synthesis. Both of them are rather included in the same reality. And who ever is aware of that stands already on the foundations of realism.37
And against Heinz Heimsoeth, he continues, “ The great idealists have all of them only sublated one reality for the sake of another.” This is a noteworthy argument, meaning that all philosophers are realists. But if this is the case, realism is a wholly nondescript philos ophy, a tautology. The ontology of realism can only be the idea of an ontology in general. If every idealistic attack on the metaphysical Ansichsein of objects, beginning with Berkeley, is to be met by this realistic supererogation, we may paraphrase Hartmann’s own argu ment against skepticism: such realism is always a formal possibility, but this stereotyped argumentation becomes less and less significant the more idealistic that “ being” is whose “ reality” is asserted. The argument is analogous to, and as futile as, Parmenides’ strictures on talking about non-being. It is unfortunate that Hartmann thought what he said was an attack on anyone, let alone Kant. T h e writer must confess that he regrets having had to spend so much time discussing a problem which he considers ultimately irrelevant to Kant criticism. But here Hartmann, who never accuses Kant of neglecting empirical realities, im plies that this phenomenological analysis is opposed to Kant’s system. It is this belief of H artm ann’s which has necessitated this already too long discussion. 37 “ Problem,” p. 89. T h us obviously the Wendung zum Realismus may be sup posed as having taken place after the Metaphysik der Erkenntnis in which Sein was “ equally distant” from the hypostatizations of idealism and realism. Are we then to suppose that “ realism” is proven by Realitdtsgegebenheit? 36
140
Studies in the Philosophy of Kant T H E SU PR EM E P R IN C IP L E OF S Y N T H E T I C A L JU D G M E N T S
We are ready to turn to specific points in Hartmann’s explicit attack on Kant’s idealism, now that we have seen idealism in general, and Kant’s transcendental idealism in particular to be unaffected by any “ proof” of empirical givenness. The question is, can those specific parts of Kant’s philosophy which are called idealistic be considered so peculiarly idealistic as to be affected by Hartmann’s philosophy of being? I shall make no futile attempt to defend Kant’s idealism by deprecating the importance of the thing in itself; Hartmann, I think, is quite correct in his censures of those who, like his Marburg teach ers, tried to explain the noumenon as a Grenzbegriff whereas Kant only argued that the concept of the noumenon is a Grenzbegriff. The major problem lies in Hartmann’s attack on the supreme prin ciple of synthetical judgment, or rather on the idealistic interpreta tion Kant gives to this principle. Here again I shall have to neglect many important points in which I think Hartmann is correct: his construction of the principle “ A is not A ” as a complement to the principle of analytical judgments. “ A is not non-A” ; his emphasis upon the nonjudgmental character of much a priori knowledge, etc. These points are important, but they are more by way of supple ments to Kant’s own theory and do not materially affect the main points in his theory. That the conditions of experience should be at the same time con ditions of the possibility of objects of experience is a statement which can be read in many ways. In order to avoid them, Kant has another and more careful statement of the principle, “ Any object stands under the necessary conditions of a synthetical unity of the manifold of intuition in a possible experience.” 38 However, we may interpret Kant to mean that the term “ possibility of experience” is to be pre ferred to “ possible experience,” and at B 195 we read, “ T h e possi bility of experience is thus that which gives all our knowings a priori objective reality.” 39 Possible experience has its own laws, the laws of Critique of Pure Reason, B 197. 39 It is odd that Hartmann does not emphasize the preferability, in K an t’s own thinking, of “ conditions of possibility of experience” to “ conditions of possible experience,” since it seems to bring Kant closer to his way of thinking; this subtle point has been elucidated by Paton, Kant’s Metaphysic of Experience, II, 90. 38
Studies in the Philosophy of Kant
141
its own possibility, and these laws are the necessary conditions of any actual knowledge; but it would be a mistake to interpret them, in this section, as merely the formal conditions of experience; nothing in this section is asserted specifically about a priori synthetical judg ments, and what is said is applicable to all synthetical judgments provided they be true. Because its laws are the same as the laws of our actual (true) experience as empirical subjects, and because its formed content is the empirical reality we actually experience, we have the possibility of the coexistence of empirical reality and transcendental ideality if grounds for the latter can be found. It does not prove that we have this coexistence actually; it merely means that, so long as we consider objects as objects of experience, the transcendental ideality of the conditions would not prejudice the empirical reality of their specifications in actual experience. Because the possibilities of experience do not extend to the things in themselves as they are (or, in Hartmann, since they do not cover the entire Ansichsein of the object), we do not have a priori knowledge, or indeed any knowledge, of what is not under these conditions. Because there is no higher principle which unites the possibility of experience with the possibility of the existence of things in themselves, phenomena, says Kant, do not contain a priori determinations of things in them selves. Thus we have in Kant a three-level system of the relation of the subject to its counterpart: empirical realism— the objectivity of this or that spatial object and of this or that psychological subject; corresponding, in Hart mann, to gnoseologisches Ansichsein, Realitatsgegebenheit. transcendental idealism— the ideality of the phenomenal object in a possible experience; the thesis that the conditions of the empirically real are, in part, conditions which underlie the possi bility of experience; and according to Hartmann, the “ dogma” that these principles are transcendentally subjective or ideal. metaphysical dualism— the speculative but unavoidable assump tion of the ultimate duality of the transcendental subject and the thing in itself which “ affects” it, a dogma which is denied by NeoKantianism in favor of the transcendental subject (idealism) and modified by Hartmann in favor of the supreme sphere of Being (ontologism).
142
Studies in the Philosophy of Kant
If the discussion in the preceding section has any value at all, it shows that the first level does not properly concern us now. We must restrict our attention to the second and third. The third concerns us first. Kant assumes the ultimate duality not only to “ limit knowledge in order to make way for faith,” but also to provide for the material contents of experience which he did not believe were capable of an a priori justification. That existence is not dialectically derivative is an opinion shared by Hartmann, and it is not only made on empirical grounds (the failure of science to deduce particulars), but also in order to avoid an ontological argument, Hartmann preferring the former ground and Kant perhaps the latter. Hartmann’s monism is an objection to this dualism from the other side, however, since he holds the common principles of the subject and the object are com mon principles of Being, and since the subject is, it must be under these conditions. Several objectionable features of this doctrine can be seen immediately. We have already pointed out the poverty of this concept of Being; if it were a real concept, and not just a limit, we should run into certain of the difficulties I shall point out in regard to the second level. If it is not a real concept, not a “ punctual unity,” but a catchall for all principles, I can see no legitimate objection one may raise against it except that it is as empty as Parmenides’ Being. But let us forget for the moment Being in general and consider the being of the subject. Hartmann is quite clear that Kant did not mean the transcendental consciousness to be a hypostasized existent,40 yet he just as often argues against it as an unnecessary Being. He says, T h e vehicle o f a p p e a rin g m ust h ave reality, else there co u ld be no such th in g as a p p e a ra n c e as such. T h u s a “ th in g in itse lf” w o u ld be m ad e the fo u n d a tio n o f the system, a n d in d eed its kn ow ab len ess w o u ld be reco g nized as a m atter o f p rin c ip le , though K a n t has d eclared the things in them selves to be u n k n o w a b le .41
Obviously if “ thing in itself” is to be interpreted as equivalent to Being, Hartmann is no longer talking Kant’s language and the objec tion does not meet the issue. On the other hand, if they are not the same, and they certainly were not for Kant, Hartmann’s argument likewise fails. 40 See Kant’s Letter to Marcus Herz, February 7772 [see below, p. 233]; and
Metaphysik, p. 146. 41 Metaphysik, p. 150.
Studies in the Philosophy of Kant
143
T he notion of Being has the disadvantage of allowing us to think that we have made a beginning towards categorization, though this is a delusion for it leaves “ all cows equally black.” Thus no material objection can be made to Kant or to any other system on the ground that its highest principle is not Being; the only objection which can be legitimately made is on a “ lower level” and concerns what this being is. Hartmann’s ontology is, as we have seen, negativistic: “ The possibility of ontology remains, independent of the adequacy of its concepts.” 42 Any attempt to qualify Being seems to be either too specific or not to have sufficient evidence behind it. Hartmann’s par ticular objection to idealism is that it does not allow Seinsprinzipien which are not Erkenntnisprinzipien . 4 3 We must consider Hartmann’s objections to idealism’s equation of these two. T o do so will lead us to the detailed study of the second level we mentioned above. A fundamental difference lies here. Kant says that the principles of actual experience must be like the principles of possible experi ence, that is principles from which the a priori character of actual experience can be derived, but not principles of any ultimate Ansichsein. Hartmann, on the other hand, says it is sufficient if some of the principles on the two sides of the knowledge-situation, em bedded in a common sphere of Being, be the same, so that objective Ansichsein is at least under the partial domination of principles which also hold for the subject as a particular Seinsgebilde. Thus Hartmann, in effect, eliminates the second level altogether, regarding the empirical subject-object situation as metaphysically derivative from the peculiar nature of the being of the subject, placing over this duality a monism of common ontological principles which are neu tral as regards this distinction between subject and object. The degree of identity of principles on both sides is the degree to which knowl edge can be extended; beyond that limit there is ineradicable ignorance. But the principles which give the mind a necessary iso morphism with empirical existence are no more adequately described as principles of experience than as principles of being. The subject überhaupt is a fiction, a prejudice Kant has taken from his standpoint to idealism. T h e sub ject in g e n e ra l is a p u re fiction o f a stan d p o in t, an d w ith it idealism falls in to the atavism o f the d o gm atic systems it has struggled
42 Ibid., p. 280. 43 Ibid., p. 362.
144
Studies in the Philosophy of Kant
against, in which the fiction of the intellectus infinitus, archetypus, or divinus had played the same role. And so far as all these Potenzierungen of the intellect, whether hypostasized oi not, nevertheless are finally con structed only by analogy to the only intellect known to us men, so there is hidden in it, as probably in all epistemological idealism, a vestige of an old anthropomorphism.44 H artm ann’s argument against this conception of Kant, constituting the central theses of his own epistemology, may be summarized in three theses: (a) Principles which are to provide for knowledge of objects are neither objective nor subjective in any definite sense, though they may appear almost indiscriminately on the subject and object sides in knowing. ( b ) Since these principles are not products of the ratio, that they should be rational or knowable does not lie in their essence, and a fortiori the a priori need not be known a priori. (c) T he object side in knowledge is homogeneous; that is we do not know a mere phenomenon, with the being which has Ansichsein re maining heterogeneous with it and necessarily unknowable. W hat we know of an object is that part of it which is under principles which our knowing mind is under, and more specifically that aspect of this part which is not too “distant” from the subject and which lies within its own psychological or metaphysical limits of actualization. T H E I D E N T I T Y O F P R I N C IP L E S
“Only an identity of principles can bridge the duality of subject and object without abolishing them; and therefore it is indifferent whether these principles are primarily principles of the object or of the subject, or stand neutrally above both of them.”45 “T h a t which is philosophically valuable in the intellectus infinitus is not the in tellectus but the infinitum. It is not clear why this incontestable infinitum, which goes beyond the human ratio , should be attached to an intellect.”46 Thus argues H artm ann against the subject in general. At times he seems willing to allow it as a possible interpreta tion, as a m atter of personal belief, but to be neither refutable nor provable and therefore to be a “transgression against the critical “ Diesseits,” p. 17 1. Metaphysik, p. 151 46 Ibid., p. 301.
44 45
Studies in the Philosophy of Kant
145
minimum of metaphysics.” At other times he objects to Kant’s formu lation of the supreme principle as not even adequate to its own task; for example, apparently forgetting that Kant's principle is restricted to “ objects of e x p e r i e n c e he asserts that it stands “ in obvious con tradiction to his theory of ideas” since the latter points out the divergence between subject and object conditions.47 But for the sake of argument let us grant all of Hartmann’s restrictions of the use of this principle in Kant, and then let us face the issue in its simplest form. Hartmann says, and I take this to be the central disharmony between him and Kant, “ Principles of knowledge . . . are ontic [seiende] principles of knowledge, which is possible only if knowledge itself has a being, is a kind of Being.” 48 What are the ontic principles which thus cover knowing and being known as well as unknowable Being? Hartmann does not say that the principles of Being, or our knowledge of them, are wholly empty; it is only the unity of all principles about which ontology can say nothing.49 In fact, a very great deal can be said about Being in ex tension; and we do know that some of the principles or conditions of knowledge are conditions or principles of being known. They are ontic principles because knowledge is a being. They are not “ ideal.” They are shared first of all by all subjects as peculiar kinds of beings (the root of intersubjective apriority) and by some aspects of all subject-beings and some aspects of all object-beings (the root of transcendent apriority). This part of Hartmann’s ontology is not negativistic, since we know a great deal about these beings we call knowers and some of their principles, or at least our conceptions of them. We have then to ask this question: Does the ontological assertion of the primacy of Being to knowing hold its own when we come to specific and pregnant beings we call subjects? In other words, do the characteristics which make Being the fundamental and ultimate, even if intensionally empty, concept obtain also for particular things which we call specific beings (Seinsgebilde as contrasted with Sein)} Is it not rather that general ontological characterizations retire to the background as soon as the emphasis is placed upon the Gebilde which 47 48 49
Ibid., p. 350. Ibid., p. 253. Ibid., p. 310.
146
Studies in the Philosophy of Kant
are, or which have Being? And do we not find some Gebilde which are beings in the ultimate and empty sense, but not beings in any pregnant sense of the word? T he last two questions must, I think, be answered affirmatively as far at least^as the knowing subject is con cerned.50 Kant’s entire philosophy is a denial that the subject and its knowl edge are beings under laws characteristic of things which are known “ to be.” Mind, the root of knowing, is not a thing among things or a being among beings; it is the known-ness of being (Bewusst-sein). This known-ness is the condition of the assertion of being in any pregnant modal sense. Being as modally pregnant is a Sinneinheit, even though as a noumenon in negative understanding it is more than this, and that despite the fact that a science (ontology) of this empty Being is, for Kant, impossible. The transcendental possibility of experience has a being, but its modality is empty, since modality has a meaning only through relation of an assertion to other possible experiences of the same thing; the significance of possible experience is to be found in its relation to actualized experience, in which being is known, and not to empty Being itself. The objection may be made, of course, that knowers are existents; this is, in fact, the meaning of Hartmann’s own “ Copernican Revolution.” 51 But this raises the issue of the status of knowledge in a world of existence or Being, and we mean here the world of Being in a preg nant, assertible, and identifiable sense. Hartmann’s solution to the problem as to how a knowing subject can be and find its principles in a Being-world raises more questions than its answers. It leads unavoidably to a confusion between psychological categories (Beingprinciples in a pregnant sense) and logical or epistemological cate gories (which are Being-principles only in an empty Parmenidean or indeterminate sense). In the midst of phenomenal Being we are in troduced to “ transcausal determinations” whose being is wholly irrational and can never be understood; these transcausal relations lend to logic and ontology a spurious effectiveness in their own right. T o be sure this is an error of post-Kantian idealism too, but it is not an error in Kant since he held to a sharp distinction of spheres, with See my “ T h e Psychophysical as a Pseudo-Problem,” Journal of Philosophy, X X X V I I (1940), 5 6 1-7 1, where this point is argued in detail. 51 Metaphysik, p. 277. 50
Studies in the Philosophy of Kant
147
relations to each other obtaining between them only as wholes, not as an interpenetration of their “ effects.” The error is introduced by Fichte and is present in Hartmann; what Hartmann finds as the Plotinian and “ chorismic” errors in Kant’s philosophy, that the categories are transcendent and their “ application” to experience must be justified, shows itself to be one of the strongest points in Kant’s doctrine since it sets up limits to various types of explana tion.52 Hartmann avoids psychologism and anthropologism by not inter preting the knowledge principles as primarily principles of subjectbeings, though the intersubjective a priori is psychologically con ceived and all the a posteriori is existential and anthropological. The only way to find out whether a principle which is given with a priori evidence is transcendently a priori, that is, also a principle of objects, is to test it perhaps empirically so as to eliminate prejudices which are also a priori, but only subjectively so. The principles which are eliminated are not illusory and of no metaphysical significance; they are rather metaphysical principles of Being which apply only to some things (some or all subject-beings) and tell us much about them, rather than what they purport to assert. Here in this differentiation Hartmann is merely using an ontologi cal language to describe what Plato discovered and Kant repeated: namely, that the a priori is a principle of knowing and of the known, whereas the empirically general characteristics of the empirical ego do not have this universality and objectivity. Hartmann is introduc ing into the negativistic sphere of empty Being all of the distinctions Kant discovered, so that Being in the pregnant sense (the only sense in which Kant used the word) is in effect distinguished from the necessary conditions of its own knownness, which have being only in the negativistic but, to be sure, imprescriptible sense. And having once made this distinction, Hartmann goes on further in a Kantian way, distinguishing the knowable from the unknowable in the object by reference to the functions of the subject, admitting that the antin omies may be dependent upon discrepancies between subject52 How badly such a principle is needed by Hartmann is shown in his wholly un-Kantian attempt to solve the third antinomy in his Ethik (2nd edn.; Berlin: de Gruyter & Co., 1935), pp. 590-95; trans. Stanton Coit, Ethics (3 vols.; London: Allen & Unwin Ltd., 1932), III, ch. 17.
148
Studies in the Philosophy of Kant
conditions and world-conditions in the sphere of pregnant Being and making the world as a whole (mere Sein) antinomial in its own intrinsic nature.53 T o summarize: Kant and Hartmann differ primarily only in the names they give to the principles at the second level; all of the dis tinctions pointed out by the former are admitted by the latter. Kant, however, considers being as a phenomenal, i.e., a pregnant, determi nation, and therefore inapplicable to the conditions of its own known ness. Hartmann uses Being and principles of Being in a much broader sense so that it applies to both. But since the boundaries between transcendental conditions and phenomenal Being coincide with that between Being in an empty and in a pregnant modal sense, there is no real point at issue. All of the serious problems in this part of Kant’s philosophy— turning on the point of how actual experience is related to possible experience or to the possibility of experience— are left where they were by Hartmann. Transcausal determinations, categorical fundamental relations, and psychophysical fundamental relations do not aid us in solving the problem of the relation of the phenomenal to the supersensible, or of pregnant to tautological Being. T H E R A T I O N A L I T Y O F C A T E G O R IE S
The second major ground on which Kant is criticized is his belief in the rationality of the principles of knowledge. By rationality here is not meant logical deducibility, for everyone is clear on the point that the ultimate principles even of logic are not “ deducible,” and Kant admits that the characters of space and time are not “ logical,” else they would be concepts and not intuitions. By rationality in this case Hartmann means knowableness, and if the a priori forms are not ideal, i.e., not functions or productions of the ratio, Hartmann thinks they need not be known at all, and certainly their knowableness is not an a priori character of them, even though they must be known a priori if they are to be known at all. Some of them, he thinks, may be irrational in a preeminent sense, that is, both alogical and un knowable. It is one of the strong points of Hartmann’s writing that he does not confuse these meanings of the irrational. Nicolai Hartmann, “ Wie ist kritische Ontologie überhaupt möglich?” in Festschrift fü r Paul Natorp (Berlin: de Gruyter & Co., 1924), pp. 174-75. 53
Studies in the Philosophy of Kant
149
Let us examine Kant’s theory of the status of the formal a priori elements in experience. For there to be an apriority, Kant requires that there be two characteristics: a certain trait or structure must be pervasive in experience and known to be such, and second it must be a condition of experience so that the occurrence of experience how ever characterized or categorized would be impossible without it. These two features (illustrated, for example, in the third paragraph of the metaphysical exposition of space, B 38-39) must coincide, and this is the distinctive feature of the Kantian a priori. The tran scendental method of Kant is to take the former (as found in his “ phenomenological” description of space and time, the Aristotelian table of judgments, the axioms of natural science, etc.) and to make a critical regression upon their conditions. T o make this clearer, let us consider space as an example: Kant says spatiality is a universal characteristic of all external experience, and he says so on grounds of his phenomenological (to use Hartmann’s own description) report of the experience which he imaginatively varies to see what happens if spatiality were abolished; and also on the basis of his dogmatic assertion of the validity of geometry (as in the analytical or regressive method in the Prolegomena). This is space as an intuited form of experience; it is known a priori as a feature of experience as we have it, though it is not asserted to be a form to be intuited in all possible experience of a being otherwise constituted for experiencing and not to be a form necessarily present in things apart from all experiencing. Kant is here asserting an absolute minimum of formal apriority; if it were originally with the weight only of a de facto generalization from experience, it is nevertheless taken as a criterion of our sane human experience, and if it failed, experience as the stuff analyzed by Kant disappears; thus it becomes a kind of presupposition for future experience, giving the futurity which is a necessary even if not suffi cient condition of apriority. Now Kant asks how we can assert this apriority which is an absolute minimum, or, to put it in another way, why is space a priori? He might have said space is valid of objects in themselves and in some way it migrates into experience, or he might have said it is an a posteriori human generalization; but in addition to the reasons (largely “ phenomenological” ) he had for not doing so, there are deeper reasons lying in his transcendental method. They do not follow solely from the impossibility of an a priori
150
Studies in the Philosophy of Kant
knowledge of things in themselves nor from the impossibility of the Humean denial of a priori synthetical knowledge; they are dictated by Kant’s desire for a metaphysical minimum, his desire to state what he finds in experience as only 'an experiential necessity. Kant says this intuited form can be said a priori to characterize all our experience, but at the same time to be known definitely as character izing only this experience, if it is at the same time a form of intuition. Thus central to the Kantian argument is a principle which is formally like the supreme principle of synthetical judgments: the intuited forms are likewise the forms of intuition. Mutatis mutandis, this is Kant’s assertion of the identity of categories, but materially it is far different from Hartmann’s “ minimal” reading of this supreme principle. If the transcendental condition, without which experience as it is constituted would be impossible, were not identical with the conditions which are necessary for the objects as experienced, we should have no grounds for connecting a priori necessity with universality. T o be sure in Kant these condi tions are not wholly rational; but they are necessarily knowable. Hartmann attacks Kant here at almost every point. a) Erkenntnisgesetz is not the same thing as Gesetzerkenntnis; the former is the prius of knowledge, the latter the posterius . 5 4 This is not a significant objection to Kant, however. The a priori is not first in order of time or obviousness; and if it asserts more than this, it erects metaphysical (ontological, but at least non-epistemological) principles which are constructs or hypotheses to account for that which is known directly. Hartmann’s grounds for doing this lie only in his interpretation of experience as a kind of being, which concep tion we have already discussed. And interpreted most narrowly, the statement is a truism. Kant would completely agree with the follow ing assertion of the means of knowing conditions of knowledge: “ Conditions are certainly, if at all, known with a priori evidence, though this evidence is not immediate, but mediated through that regressus [Rückgang resp. Rückschluss].” 55 b) The concepts we make of those conditions underlying the pos sibility of experience are and remain concepts of categories; they are 54 55
Metaphysik, p. 350. Ibid., p. 253.
V
Studies in the Philosophy of Kant
not themselves categories. “ Knowledge of principles is related to the principle which is in the same way that knowledge of objects is re lated to the transcendent object.” 56 “ Concepts of categories are nothing but attempts to comprehend a category or to progressively formulate it. Between them and the categories themselves there is the same tension of the progress of knowledge as obtains between the knowledge of the object and the object itself.” 57 These statements are dependent again upon Hartmann’s basic ontological theory, that the principles of knowledge are ontic principles of Being which are shared by the object in itself and by the knower as a kind of being. T o be sure, then, they are not correctly labeled psychological or cosmological principles or conditions of knowing an object; but formally, ontological and gnoseological formulation of these principles is like the existential and psychological. That is to say, there are certain partially or wholly unknowable conditions which are partially or wholly identical for the knowing-being and for the known-being. We do not know the principle as it is or the object which it dominates; we know only that aspect of the principle which we intuit in that object. This principle is indicated or sug gested, and on account of its compulsiveness it is inferred to domi nate also our experiencing the object, since experience is also a kind of being under universal ontological laws. Thus the principles of the a priori structure (and for Hartmann, also content) of knowledge are not specifically subjective or objective, but neutral; still objective (realistic) or subjective (idealistic) interpretations could equally well account for the phenomenon of apriority. They would, however, assert more than we know about the a priori than we can know a priori, which is merely that there is an identity of principles on the two sides of the knowledge situation. Nevertheless, because Hart mann thinks this identity extends also to the real object (thing in itself), he is asserting more than experience justifies, for the partial identity of principles is, at the very least, possible, a partial identity of the appearance and our experience, in the duality we directly know. (Whether this is justified will be discussed in the last section, on the distinction between appearance and reality.) 56 57
Ibid., p. 2 K1 . Ibid.
152
Studies in the Philosophy of Kant
c)
K ant’s theory of the transcendental consciousness is involved in his identification of the conditions of experience with the necessary features found in the experience; but H artm ann says that there is an identity of subjective and objective cbnditions in Being, though these conditions appear only in a disguised form in a priori evidence in experience. Thus to Hartm ann it appears that K ant’s transcendental consciousness as the “source” of these principles is merely a hypoth esis, a metaphysical speculation unnecessary to Kant's own justifica tion of the a priori.58 If the interpretation of Kant is correct whereby these conditions must be identical in the three phases but have no relevance (as they cannot if the distinction between appearance and reality, however interpreted, is valid) to things in themselves, then this is not a hypothesis but, as Kant thought, the sole sufficient condi tion of experience. Thus H artm ann’s apriority is always merely hypothetical since it is a construction of being beyond experience, whereas K ant’s apodicticity is found in a regressus upon conditions, found as it were beneath experience; Kant’s apodicticity is therefore not hypothetical when regarded only in terms of that experience based upon it, and not in terms of a neutral or objective ontology. T o infer characters of being from experience (and by this being one must mean a modally pregnant being, another being than experience itself), is to construct a hypothesis which goes beyond the fact of experience. Thus, though it can be said that Kant’s transcendental idealism is an ontological hypothesis, it is a hypothesis giving the minimum conditions of an actual experience. Hartmann's assertion of being is, since it contains nonexperiential determination of both experience as a being and the being experienced, in contrast with it, speculative and dogmatic. In a word, Hartm ann hypostasizes Being as necessary in a way in which
it is here that Wood’s criticism most nearly approaches Hartmann’s. “ The critical epistemology is not incompatible with the transcendental idealism, yet Kant's argument from the one to the other is certainly inconclusive. I am con vinced that the really significant insights of the critical epistemology could be pre served even if the whole transcendental paraphernalia were either abolished altogether or else translated into psychological terms. Kant’s metaphysical idealism —for it is that despite Kant’s professed repudiation of all dogmatic metaphysics— is, in relation to his critical rationalism, a mere supplementary hypothesis which has to compete on terms of equality with the rival hypothesis of realism” (Ledger Wood, “ The Transcendental Method,” in The Heritage of Kant, p. 18). For “ psy chological” read “ ontological,” and the agreement is perfect. 58
Studies in the Philosophy of Kant
J53
he admits Kant does not have to hypostasize the transcendental subject as a being in any pregnant sense.59 Kant does not hold there may be no unknowable principles of Ansichsein, nor does he hold there may be no unknown aspects of the twelve categories and two forms. All that he denies is that there may be unknown principles which are a priori for experience as it is con stituted. (We know this from his assertion of the exhaustiveness of his analysis.) But his theory in its essentials would not be adversely af fected, as Ledger Wood has shown, by regarding the list of categories as merely a de facto analysis just as his statement of the forms of intuitions was. All that is essential is the assertion that every neces sary principle in experience corresponds to a necessary condition of experience; as to the manner of discovery we need make no restric tions. T H E T H I N G IN I T S E L F
Finally, we come to the last decisive problem, that of the status of the thing in itself. Though Kant expresses himself sometimes loosely in regard to it, for purposes of evaluating Hartmann’s criticism sev eral easily documented theses may be taken as expressive of Kant’s views: 1) The thing in itself is that which “ affects” the senses to produce the “ presentations of the things which are unknown in their nature” and these presentations are phenomena.00 2) What we do know is the phenomenon which is under condi tions which do not hold for the things in themselves but only for our own experiencing; therefore it is not only possible but also necessary that there be a priori knowledge of appearance.61 3) But the thing in itself, not falling under these conditions, is not known a priori.62 4) Furthermore, there is no resemblance between the quality of See Seelbach, Nicolai Hartmann’s Kantritik, p. 14. On the hypostasis of Being as a necessary fact, see my review of Weiss’s “ Reality,” in Philosophy and Phenomenological Research,! (1940), 114 -19 . 60 Critique of Pure Reason, B 164. 61 Ibid., A 29. 62 ibid., B 56. 59
154
Studies in the Philosophy of Kant
the phenomenon and that “ character” of the thing in itself which “ excites” the specific sensation in me; in fact, such a conception is meaningless, and no a posteriori knowledge of the thing in itself is possible, or at least no a posteriori knowledge of it as it is in itself is possible.63 5) Nevertheless, the thing in itself is real, because it is logically required that the appearance be of something real. It is the “ other side of appearance.” 64 The first of these theses requires little attention. It has long been realized that Kant committed a subreption in applying causality beyond possible experience. Furthermore, it is also generally recog nized that a causal theory of knowledge is inadequate, regardless of the status of the entities supposed to be concerned in knowledge; and in this first thesis, and also in part in the fourth, Kant has not com pletely resisted the seduction of Locke’s analysis. Hartmann is rightly insistent upon the “ transcausal” character of the knowledge relation. Moreover, there is agreement as regards the fourth thesis in so far as Hartmann thinks of sensuous material as a system of symbols of that which is not sensuous65 and Kant holds that an “ empirical ideality” can be attributed to the a posteriori sense contents.66 They agree that there is no “ migration” of the characters of things in them selves into sensation, and that there is no resemblance between a sensation and its correlate. Hartmann holds, however, that there may be an “ intersubjective apriority” of the sense material; such a belief is not expressed by Kant, but there is no fundamental reason why he might not accept that view. Nevertheless, Hartmann does argue that there may be a posteriori knowledge of the thing in itself, for all true knowledge is of an Ansichsein; therefore their disagreement here is a mere corollary of their divergent views concerning the second thesis. Similarly, little discussion of the last thesis is required. In both systems, appearance is of a real thing which is not limited to the sphere of experience. For Hartmann, the Ansichsein includes the “ objectified” and the "transobjective” as continuous parts, with the Prolegomena, §13, n. 2. 64 Critique of Pure Reason, B 55; A 391. Hartmann would, of course, say “ phenomenologically required.” 65 Metaphysik, pp. 382-87. 66 Critique of Pure Reason, B 52-53. 63
Studies in the Philosophy of Kant
1 55
boundary between them being shifted in the direction of the latter in the progress of knowledge. For Kant, however, this boundary is fixed, being that between the phenomenal world and the noumenal. The progress of knowledge occurs wholly within the former, with the transcendental object a mere x which is to be progressively character ized. The thing in itself, in its own nature, does not appear, if we mean by that that it becomes phenomenal. It is at most the sub stratum of experience. It is here that Hartmann makes an illicit criticism. He says, “ If the thing in itself is the partially appearing behind the appearance, and in its unknown and infinite remainder but the continuation of the known, then it would be an astonishing coincidence if the accidental limits of possible experience were the same as the limits of the validity of the categories in it.” 67 But for Kant there is not this continuity which Hartmann hypothesizes and which he finds only on the basis of his original Auffassung68 and because of his empty assertion of the community of all being. There fore the limits of the former are not accidental for Kant, and their coincidence with the limits of the categories is not astonishing. In fact, they are essentially related as two different expressions of the basic heterogeneity between appearance and that which is (ambigu ously) said to “ appear” in it. Hartmann, however, is here arguing from his own thesis of the ontic continuity which Kant denies in the second thesis. But the force of Hartmann’s thesis of homogeneity is considerably weakened by one consideration. We have already seen how the concept of being is unable to effect any actual synthesis or categorization of “ everything which in any sense is,” and the spurious unity of Being is disrupted by the various tensions which Hartmann finds within it. Here he asserts that the unknowable aspects or parts of the object depend, for their unknowableness, not on any intrinsic peculiarity they have, but solely on the incapacity of the subject-being or the knowledge rela tion to “ bridge the gnoseological distance” from the subject, and he admits that there is an undetermined point dependent only upon the nature of the being of the subject, and not on any intrinsic peculiarity of a specifically “ unknowable” being, beyond which this Metaphysik, p. 227. 68 Cf. above, p. 132. 67
156
Studies in the Philosophy of Kant
boundary cannot be pushed. Thus the. world of Being itself develops the heterogeneities he had decried in Kant’s separation of appearance and reality.69 Once again, the looseness of the unification through the concept of being is obvious when Hartmann comes to the problem of the an tinomies. It was the discovery of the antinomies which led Kant to abandon rational ontology after he had written the dissertation of 1770. Kant’s attempt to solve the antinomies does not fail to do justice to the Aporie itself; it is an attempt not to deny the antinomial character of the different aspects of experience, but rather to use these to show the distinction between phenomenally appearing and ontologically being. Hartmann, in denying the ultimacy of the distinction, is faced with the same kind of problem (with a different coefficient, as it were) which Berkeley would have had to struggle with had he known of the antinomial character of his metaphysical assertions. That is to say, when the distinction between appearing and merely being is regarded as accidental, then esse, whether in terpreted as percipi or Ansichsein, itself becomes antinomial. Thus Hartmann develops, as the only alternative to Kant’s distinction which will not rationalize the antinomies away, the highly specula tive and uneconomical theory that “ being itself is disharmonious, and conflict is the form of its being. In this case we must assume that the principle of contradiction does not actually, or at least only con ditionally, obtain for the sphere of the ontic real.” 70 Here, with the logical law as a form dominating only the ratio, the distinction be tween appearance and reality is surreptitiously introduced even after being has been made antinomial and chaotic. We have now seen that the divergences with reference to each of these theses, where they do exist, depend upon Hartmann’s rejection of the second and third of Kant’s views which I have listed at the T h e distinction between appearance and reality returns in Hartmann with almost the same force as is in the Critique of Judgm ent, §77, “ Von der Eigentüm lichkeit des menschlichen Verstandes. . . .” For it is merely a peculiarity of know ing, says Hartmann, that it takes its object from a perspective, and “ Appearance is the objective structure of consciousness differentiated from the object” (Meta physik, p. 225). 70 “ Wie ist kritische Ontologie überhaupt möglich?” pp. 174-75. T h ou gh it can not be dogmatically asserted that this is H artm ann’s final view, nevertheless at that time (1924) it was the only alternative to K an t’s theory which he gave more than a mere nod to. 69
Studies in the Philosophy of Kant
157
beginning of this section, or at least that part of them which says that the a priori conditions of phenomena are not conditions of things in themselves. But this thesis is absolutely central for Kant, and its rejection is decisive for Hartmann. Each philosopher gives an essentially similar picture of the fact of phenomenation; what Kant cannot do and what Hartmann attempts in spite of the ac knowledged antinomies is to generalize beyond these phenomena as kinds of being to a general ontic continuity. This ontic continuity is not, however, as we have seen, necessary for accounting for the realityphenomenon or for empirical realism.71 In Kant, the thing in itself provides only for aposteriority; but to construct an a priori theory which is to apply to it is to assert more about it than we know, and indeed more than we need to know in order to understand the phenomena. Therefore Hartmann’s ontology, which tries to do this, is speculative and, in Kant’s sense, dogmatic. 7i Kant is quite insistent upon the adequacy of his concept of object to satisfy all the phenomenological demands, and Hartmann concedes this in admitting K ant’s empirical realism. T h e fact that it is not a priori does not, as we have seen, touch any relevant point here.
IX ON HENRY MARGENAU’S KANTIANISM i Professor Margenau, in his recent T h e Nature of Physical R e a lity 1 has made a significant contribution to the study of the methods and philosophical interpretation of the results of modern physics. Were I writing a review of this book, I should certainly devote most of my space to an appreciation instead of a criticism of it, for what I con sider good and valuable in it far outweighs the things about which I still have doubts. But I wish rather to discuss two weaknesses I find in it, with the hope that my discussion will occasion a further elabora tion and defense of Professor Margenau’s well-considered views. In entering upon such a discussion, I find myself somewhat em barrassed by my polemical desire to occupy two prima facie divergent perspectives— Kantianism and a more straightforward physical real ism. I have defended in previous writings a generally realistic, as opposed to a Kantian, phenomenalistic, or positivistic view of scientific methodology, though I have lasting sympathy for the point of view represented by Kant and real enthusiasm for many things Margenau and the Neo-Kantians represent. But while on the one hand Professor Margenau is opposed to the kind of realism I have defended, it does not seem to me that he has gone far enough in the direction of Kantianism, as I understand it, to gain strength from some of its more important insights. Consequently I feel that Pro fessor Margenau is subject to more criticism from my own point of view than would have been the case if he had been more thoroughly r e pr in te d
, b y permission o f the Editor, from
P h ilosophy and P h e n o m e n o
logical Research, X I (1951), 5 6 8 -7 3 . Professor M a r g e n a u r e p lie d in the same jo u rn al, pp. 5 7 4 -7 8 .
1 H enry Margenau, T h e Nature of Physical Reality (New York: M cG raw -H ill Book Company, Inc., 1950). 158
Studies in the Philosophy of Kant
*59
Kantian. I feel that he has given up some of the more valuable tenets of realism without thereupon availing himself of the great ad vantages which would flow to his position from a slightly different attitude toward Kant and the Marburg school. Inasmuch as he has already been criticized on the positive basis of his Neo-Kantianism,2 it is perhaps wise for me to indicate the way in which, I think, he has failed to exploit the full resources of the Kantian philosophy. (I shall, of course, leave out of account those features of Kant’s philosophy which were consequences merely of the historical situation in physics and mathematics at the end of the eighteenth century; “ back to Kant” does not now mean, and has never meant, going back to them.) The essential feature of the Kantian philosophy, as I see it, is to find within experience those conditions of its structure without which it would not be experience—objective, intersubjective, having fea tures of universality and necessity; that these conditions are not empirical facts for either physics, psychology, or history, but are epistemological principles continuously but never completely con firmed by experience and yet of such a universal and necessary nature that they would be confirmed by any imaginable (human) experience even if the empirical facts were different. These conditions are ex pressed in propositions which, in Professor Lewis’ words, we take to be true “ no matter what.” While Professor Margenau does not assert his “ metaphysical requirements on constructs” to be eternal, fixed, and absolutely necessary, and does intimate a theory as to how they are developed through a long history of analysis of our experience, his “ requirements” are far closer to the Kantian theory of the a priori than to more modern functionalistic and relativistic theories of the nature of the a priori. Yet there is one aspect of all the Kantian principles to which Pro fessor Margenau refuses to have recourse. I refer to the role of possible experience as regulative in the interpretation of actual experience, wherein the test for “ objective existence” is to be found in the con formity of an alleged object of experience, starting as a more con tingent representation, to the conditions of a possible experience in general. For Kantian idealism, “ ideal knowledge” is the locus and 2 Adolf Grunbaum, “ Realism and Neo-Kantianism in Professor Margenau’s Philosophy of Quantum Mechanics,” Philosophy of Science, X V II (1950), 26-34.
i6o
Studies in the Philosophy of Kant
test of (empirical) reality; Kantianism is the source of coherence theory of truth and of ascribed existence. Such a principle is explicitly rejected by Professor Margenau.in his refusal to count the desideratum of convergence of concepts or constructs as one of the metaphysical requirements that a valid construct must meet. Professor Margenau explicitly argues against giving this status and dignity to the ideal of convergence of constructs— an ideal or principle necessary for there to be a common reality as the common object of diverse concepts. He does so because he says there is no evidence of any necessary con vergence of a group of successively developed concepts (p. 459). But there is no coercive empirical evidence for the metaphysical require ments he does acknowledge; they are rather conditions for empirical evidence in general. Thus I cannot see why Professor Margenau derogates the role of convergence as only historically important and does not permit it a position among the requirements. Perhaps of all the distinctions introduced by Kant, the one between the regulative and the constitutive is hardest to maintain. Before his time, the regulative was generally regarded as constitutive, thus giv ing rise to “ dialectic.” Since his time, the converse has generally been held, and even the categories and principles are now usually regarded not as constitutive of experience but as regulative of the ascription of reality. But Professor Margenau seems to preserve the distinction, and only grudgingly, if at all, he admits convergence to a more per fect possible experience as perhaps regulative; but it is not a require ment that all constructs must meet before they are to be accepted into the corpus of science or the body of nature. It cannot be such a requirement, of course, because the decision on admission has to be made now before we have the possible experience. But it is because this possible experience functions as an ideal that some constructs are later banished. But surely a “ good construct” is one which we be lieve has a continuous and future status in the corpus of the science. Professor Margenau is, I believe, left with only actual experience as the measure of “ reality,” and this results in a historical relativism or subjectivism of scientific concepts. He consistently concludes from this that physical reality changes with our knowledge of it; in fact, this proposition is almost taken as a logical identity. The consequence is that the date of a judgment becomes a factor in its truth, and the distinction between sentences and propositions is abrogated. T o take
Studies in the Philosophy of Kant
161
his example, it would not have been true in 1918 that the neutron existed, while it is true after 1932 that it does exist. Now if his praise of a “ changing reality” is justified, it would mean that the neutron first came into existence in 1932 when its “ construct” became a “ verifact.” But in spite of his liking a changing, progressive reality (instead of a more or less permanent world revealed only piecemeal), he does not believe that the neutron came into existence in 1932. For, he says, in 1950 it is true that the neutron did exist in 1918, though the same proposition (viz., “ Now, in 1918, the neutron exists” ) would not have been true (pp. 293-94). A failure to distinguish between warrantable assertibility and truth—a distinction requiring possible experience to be regulative of actual experience and, per corollary, convergence of constructs as a test of varifactuality— leads to this paradox. But it is questionable whether Professor Margenau has any right to say, “ After a construct is validated, it must be said to have been real before it was formed” (p. 294). His argument for this certainly very plausible statement is based on his “ requirement of permanence and stability of constructs.” But the kind of permanence and stability which is required in the defense of this postulate (Chapter V, §5) seems to be nothing more than semantic identity (cf. pp. 89-90), while the argument for the retroactive truth of a true statement (Chapter XV, §1, 2) depends upon the assumption that an established con struct must have temporal endurance. The first does not entail the second. Since Professor Margenau would accept the verifactual status of the neutron in 1950 and hold this acceptance to be a correction of, or to have some other prerogative over, the statement true in 1918 that neutrons do not exist, it seems to me that he is appealing to “ better experience” even if he is unwilling, with Kant, to postulate perfect (possible) experience as a goal to which we assume all verifacts of 1950 will converge. For if we knew— which we do not of course know except in a very abstract and general sort of way— that the verifacts accepted now will be rejected later by restrospective judgments (such as “ There were, in 1950, some mesons for which there was no evi dence” ), then we should have no justification for the open-minded tentativeness (in Professor Margenau's language perhaps a “ Milquetoastian” attitude) that is one of the most striking and valuable
162
Studies in the Philosophy of Kant
features of the scientific approach to things. There are always results we have not yet got; and one of the requirements we always make for a present verifact is that it should hold its own upon further inquiry and in further experience. If'this fact, though admitted, had been taken more seriously, it would have led, I believe, to a different view about “ changing reality.” It would have been a view more in accord with both common sense and Kantianism— to be sure an unusual confirmation of one by the other. But what Professor Margenau seems to me to be doing is to permit present experience to function as the criterion of past truth-claims without properly emphasizing that present truth-claims are to be judged in the light of later ones. The same argument which leads to the former will lead to the latter and, if extended, will, I believe, lead to an acknowledgment of the regulative role of possible experi ence and its rules of convergence of all accepted constructs as a condition of their becoming verifacts. Professor Margenau writes: It is easy to succumb to the temptation of distinguishing at the outset between the permanence of physical entities and the [im-] permanence of theories about them, saying for example that the entities themselves are not affected by the vicissitudes of theories. Such an assertion amounts to granting the breakdown of the epistemological approach to reality (p.
295)-
I submit, on the contrary, that the “ epistemological approach to reality” cannot be saved from skepticism if “ possible experience” is not accorded an epistemological prerogative over actual experience. And this would require taking the principle of convergence as one of the “ metaphysical requirements on constructs.” II
It is possible to discuss Professor Margenau’s relation to realism in Kantian terms, though this is not the most natural way to approach the problem. But there is in Kant a very important strain of tran scendental realism as a guarantee of the legitimate claims of empiri cism. The thing in itself provides in some way the empirical content or material of our experience, according to Kant, and if the thing in itself is denied, there is no way for him to avoid a thoroughgoing
Studies in the Philosophy of Kant
163
rationalism; the history of idealism after Kant shows many examples of this dogmatic and rationalistic consequence of challenging Kant’s realism. Now I do not wish even to intimate that Professor Margenau’s objections to realism and his denial of the thing in itself lead him to rationalism; he is too much a practicing scientist to derogate to that extent the role of mere matter of fact. But it is nevertheless difficult for me to see what is the role he ascribes to matters of fact, and even more difficult for me to see what can be the status and source of the given. Since Mr. Griinbaum has dealt with this aspect of Professor Margenau’s epistemology, I shall turn to another but closely related problem. But I must express my hope that Professor Margenau will apply his epistemological views to an inquiry into the problems of psychophysics; whether one could agree with them or not—and I do not prejudge the matter— they would be extremely instructive. The problem I wish to discuss is that of the element of vagueness present in actual things but lacking in constructs. It seems to me that the “ growing point” in scientific systems lies in an area of vagueness in our concepts, wherein we acknowledge that the object of the con cept has properties which are not included in the definition of the concept, and that the function of further investigation is to determine that vagueness, make it explicit, lead to a decision among possible accidents of the object of the concept, and produce if needed a new definition. The “ existential surplus of meaning” or what I have called the “ potentiality” of the object of a concept seems to me to be neces sary if we are to understand the dynamics of the growth of a science. And it seems to me to be neglected in positivisms, and either denied or left unexplained in phenomenalisms which identify clearly de fined and confirmed constructs with the allegedly actual objects of the physical world. Kant’s thing in itself would provide for this element of ambiguous potentiality were it not for his decision that experience cannot con verge upon it. It is questionable if Professor Margenau has any way of providing for it, though he does seem to acknowledge it (e.g., foot note to p. 230). He does admit that physical objects do not constitute the whole of reality; but I should guess that he would use the same constructionistic approach for the establishment of other species of reality, so that their status would fail to permit them to fulfill the
164
Studies in the Philosophy of Kant
role of vaguely defined and slowly approached realities which are reconstructed rather than constructed by concepts. It is interesting to observe, in conclusion, that Kant, in order to avoid skepticism, appealed to the structure of possible experience, yet to avoid dogmatic rationalism he limited the metaphysical competence of even possible experience and did not identify its veri fied constructs with reality. It seems to me that Professor Margenau, in his otherwise splendid book, has failed to follow Kant in these two only prima facie opposed directions, and has consequently placed himself in danger of falling into two opposed errors from which Kant saved himself by his precarious balance between rationalism and empiricism and between phenomenalism and realism. Professor Margenau has not actually fallen into dogmatism or skepticism; but I feel that he is saved from one or the other by his common sense more than by the logic of his argument. This discussion is not written in any captious spirit, but with the sole hope that Professor Margenau will elaborate and successfully defend the points of view I have questioned.
X
SIR DAVID ROSS ON DUTY AND PURPOSE IN KANT Kant’s F oundations of the Metaphysics of Morals has recently at tracted the attention of the philosophical community in a most sur prising degree. In about twenty years there have been four new English translations, three commentaries,1 at least two inexpensive students’ editions, and a rapidly growing periodical literature that cites it almost as though it were the only work on ethics we have from Kant. K an tphilolog en have again taken up, in the manner of Vaihinger, the microscopic study of the text. W ith the exception of the Critique of P ure Reason, no other work by Kant— indeed no other work in German philosophy— seems to be subjected to such intense scrutiny as this slim volume. K a n t ’s Ethical Theory 2 is interesting for two reasons. First it is a sensitive and not unsympathetic interpretation of Kant. Its brevity recommends it as a real help in the understanding of Kant. Second, and more important, it is interesting because it is by Sir David Ross. Readers of his other works can easily guess at the ways and points in which he agrees and disagrees with Kant; and here they will find, in a polemical context, discussion of the specific character of duties and of the conflicts between prima facie duties— subjects in which Kant r e pr in te d
, b y permission o f the Editor, from
Philosophy and P h e n o m e n o
logical Research, X V I (1955), 9 8 -10 7 . 1 [Since this was written, still another com m entary on the Foundations has been p ublished, viz., A . R . C. D u n c a n , Practical Reason and Morality. A Study of
Im m anuel K an t’s Foundations for the Metaphysics of Morals (London: T h o m a s Nelson and Sons, Ltd., 1957), and I reviewed it in T h e Philosophical Review, L X V I I I (1959), 400-2.] 2 W i l l i a m D a v i d Ross, K an t’s Ethical Theory: A Commentary on the Grundle-
gung zur Metaphysik der Sitten (Oxford: C la re n d o n Press, 1954), p. 96. 165
i66
Studies in the Philosophy of Kant
is weakest and Ross is strongest.3 But here for the first time they will find him dealing with ipsissima verba of Kant and not merely using common knowledge of Kant 'in the exposition of his own moral philosophy. In this paper I shall restrict myself to comments on Ross’s criticism of Kant’s views of purpose and inclination in their relation to duty. More specifically I shall consider only the accuracy of his interpreta tion and try to avoid assuming an adjudicative role in re formalism vs. intuitionism. Where Kant and Ross really disagree, I shall be silent. But in a few places where they seem to Sir David to disagree, I think I can point out ways in which his views on Kant are incorrect or at least are expressed with such brevity that many readers of the commentary will regrettably find certain common misapprehensions strengthened by Sir David’s authority. PU RPO SE AND D U T Y
“ What makes us ascribe moral value to an action in itself is rather what it is meant to produce than what it in fact and perhaps by mere accident produces. Kant’s intuition here is profoundly original. .. . It is this, rather than the subtleties of his theory, that makes his greatness.” 4 But on page 19 we are told that it is Kant’s opinion that “ An action done from a sense of duty derives its moral worth not from the purpose which is meant to be attained by it, but from the maxim by which it is determined.” T he second is, and is em phasized by Ross to be, the authentic Kantian doctrine,5 though it is apparently not difficult to read some passages as evidence that the former was Kant’s view. But from the correct interpretation, given 3 Surprisingly, the argument concerning the infinite regress in the alleged duty to act out of a sense of duty (Ross, Foundations of Ethics, pp. 116-24) is omitted. For discussions of this argument, which, if valid, is far more devastating than any thing contained in this commentary, see H . J . Paton, The Categorical Imperative, p. 118 ; A. C. Ewing, The Definition of the Good (New York: T h e Macmillan Co., 1 9 4 7 )» P- l 4 2> and A. T . Shillinglaw, “ Dr. Ross’s Criticism of K an t’s T heory of Duty,” Mind, N.S., X L I I (1933), 2 13 -16 . 4 Kant’s Ethical Theory, p. 12. 5 Foundations of the Metaphysics of Morals, Academy edn., IV, 399; trans. L. W. Beck, “ Library of Liberal Arts,” No. 1 1 3 (New York: Liberal Arts Press, Inc., *9 5 9 )» P- *5 - Hereafter referred to as Foundations.
Studies in the Philosophy of Kant
167
in the second of the sentences quoted, it is easy to draw a very wrong conclusion, to wit, that a moral action has no purpose. Accordingly, Sir David writes that Kant requires us to choose be tween describing a good action as done from duty and as done with a purpose, and that “ Kant must be wrong . . . in saying that we must act from a principle and not from a purpose.” 6 In Foundations of Ethics, the same distinction is expressed as follows: “ Kant . . . main tains that an action is good only when it is done not from a motive but from a maxim.” 7 There is a sense, I suppose, in which it is correct to say that Kant tries to establish such a disjunction, for it is respect for the law and not desire for a consequence, however good, that makes an act morally good. But this way of saying it is at least as likely to mislead as anything in Kant’s Foundations itself (which is saying a good deal), for it suggests that Kant makes us choose be tween an action which has a purpose and an action which conforms to and is motivated by a moral principle, and that we ought to choose the latter. This is not what Kant is saying, but the reader of this commentary will not be likely to see the way in which purpose and duty coexist in moral decision and action. And to avoid the common interpretation according to which there is a radical antagonism between purpose and principle, this coexistence of principle and purpose is what needs to be pointed out. Every action, properly so called, is guided by a maxim. In this, action differs from mere reflex or blind instinct. A maxim is a rule to do certain things in certain circumstances. These circumstances always have a bearing on the attainment of some purpose; otherwise there is no occasion for acting at all. Desire and purpose are there fore the material of the maxim; the maxim is the form in which purpose becomes operative in consciousness and determinative of the will. Impulse and desire do not directly determine volition, but will is practical reason, i.e., impulse guided or thwarted by con sciousness of some rule. 6 Kant’s Ethical Theory, pp. 20, 22. But Ross rightly adds: “ We shall see later that Kant admits that the seeking of an end is one aspect of the nature of good action,” but his “ later” refers only to the second formula of the categorical impera tive. I shall insist that it is present from the beginning, in the first formula and even in Sec. I. 7 Foundations of Ethics, p. 293.
i68
Studies in the Philosophy of Kant
Actually, every sensible and intelligent action is under two maxims. For instance, I want to build a hous^ for my dog. My maxim is not merely to use a hammer and saw (a technical imperative), but that, whatever my purpose may be, as a prudent man to use the most effi cient means to the satisfaction of my interest. Only the conformity of the former maxim to the latter makes it proper for me to choose a saw instead of a pocket knife; I cannot say, in general, that a saw is better than a knife and “ use a saw” and “ build a dog house” certainly can not be “ deduced” from the counsel of prudence, the pragmatic imperative. The counsel of prudence as a hypothetical imperative is as “ empty,” by Kant’s own admission (though in fact he adduces a different reason), as the categorical imperative has been found to be by his critics. In each case, there is purpose as well as principle. It is perhaps natural to emphasize the purpose in the hypothetical imperatives and the principle in the categorical; but it is wrong to forget that action under each imperative implicitly contains both. The peculiarity of the moral good is that it is not derived from the goodness of the ends which the will as a matter of fact intends. Rather, the moral maxim itself determines the acceptability of the ends the will may have; hence the moral imperative is independent of the specific teleological content of the specific maxims subsumed under it. The goodness of the good will is not— and here Ross and Kant really disagree— dependent on the goodness of the ends it intends; though Kant does hold that it always does intend something that is good in some sense, and Sir David does hold that the desire to do one’s duty because it is one’s duty is the highest of all motives. T he good ness of the good will, for Kant, lies solely in the fact that every maxim it enunciates and follows must be under, and adopted for the sake of subsumption under, the supreme condition that the maxim be objectively valid (for all rational beings) regardless of the (human, or even merely private) purposes it may subserve. In morality, every material maxim for the attainment of purposes must be subsumed under another maxim by which certain ones of them are chosen and others rejected. The supreme maxim of the good will, therefore, is not: “ Act on some maxim of creating this or that good situation,” but “ Act only on a maxim that you could also will to be a universal law. . . .”
Studies in the P h ilo s o p h y o f K a n t
169
There are thus two maxims even in moral action: (a) the maxim of an action as a means to certain consequences— this expresses the intention of the act and depends upon specific conditions known only through experience; and (b ) the maxim that any maxim of type (a) should also be objectively valid. T h e restrictive condition of (b ) over (a) is duty, which is absent if maxims of type (a) and (b) agree “ by nature,” as in the holy will for which purpose and the morally good necessarily coincide. Desire for an end and knowledge of empiri cal fact together are the subjective ground of maxims of type (a); rev erence for law is the subjective ground for maxim (b), since the action, if done “ from duty . . . must be determined by the formal principle of volition as such [maxim (6)] since every material principle has been withdrawn from it.” 8 In the context, however, it is clear that though the material prin ciple has been “ withdrawn” from the determination of the action and its moral worth, its presence in a maxim of type (a), in cases like that of sympathetic feeling, may obscure the presence of maxim (b). Even where maxim (b ) is salient, however, a maxim of type (a) does not cease to be the material upon which the practical reason, through maxim (b ), exercises control. If this were not so, maxim (b ) would, in principle at least, determine our specific duties in specific cases without the cooperation of intent and knowledge of the world summed up in maxims of type (a). This presence of both types of maxims is not as clear in the examples Kant gives as one would per haps wish, for these examples were devised precisely to bring out the formal element, the maxim (b ), which is usually overlooked;9 but there are plenty of explicit statements, when the occasion demands them, that all moral actions have both rational form and teleological content.10 T h e content, however, need not necessarily be inclination or a maxim of inclination. A unique, and highly controversial, feature of Kant’s ethics is his theory that a maxim of type (a) can be as it were %Foundations, Academ y edn., IV, 400; trans. Beck, p. 16. [A more complete analysis of this situation is given below (pp. 180-81) and in my Commentary on K a n t ’s C ritiq u e of Practical Reason, pp. 85-88 and 119-20.] 9 Paton writes: “ Kant does not forget [that moral action has a purpose]. He e x pects his readers to rem em ber” ( T h e Categorical Imperative, p. 75). 10 Fou nd ations , Academ y edn., IV, 436; Reflection 6633 (Academy edn., X I X , 120).
170
Studies in the Philosophy of Kant
generated by maxim (b) and empirical knowledge but without the presence of an inclination. Thus Kant interprets the command to love our neighbor not as a command to have an inclination11 but a command to act with a certain end in view, the creation of good for others. The end may not be presented ready-made by inclination, so that the pure practical reason can merely decide its status; it may, as it were, be presented by the categorical imperative, maxim (b), itself, as a worthy end even if one is not inclined towards it. But an action devoted to this end is a purposive action, the end being a “ necessary [not contingent] object of the faculty of desire.” 12 There are thus ends that are also duties. Now when Kant does explicitly acknowledge the purposive ele ment in action, Ross seems to chide him for it; but at the same time I think he misinterprets the way Kant introduces purpose into the evaluation of an act. In the discussion of “ imperfect duties” under the formula of “ the law of nature,” Sir David writes: “ Here there is a frank appeal to purposes, i.e., to desires of his own. . . . Here again the appeal is to results, and in this case to definitely hedonistic results; benevolence is justified by an appeal to farsighted selfishness.” 13 This is indeed the usual interpretation of Kant’s discussion of duties of benevolence and of developing one’s talents, and it is on grounds of such an interpretation that Mill claimed Kant as a forerunner. Yet in the context of the formula, “ Act as if by your will. . . ,” it ought to be clear that Kant is not suggesting that the judgment of actions depends upon consequences, or that the motive is a desire for my own satisfactions. For the consequences and satisfactions would be actual only under a condition that all agree is merely a fiction. Kant is rather asking: could you will that a system of nature should subsist in which there would be no development of talents and mutual aid? Only if we actually could by volition establish laws of nature could we have occasion to make such a prudential calculation as indicated by: Would I be happy in such a world? Kant is rather arguing that we would not, even if we could, will that there should be a world in which the will would be systematically thwarted. T he particular consequences as they bear upon my desires— and in this 1 1 Foundations, Academy edn., IV, 399; Kant’s Ethical Theory, p. 15. 12 Critique of Practical Reason, Academy edn., V, 58; trans. Beck, p. 60. 13 Kant’s Ethical Theory, pp. 46-47; on Foundations, Academy edn., IV, 423.
Studies in the Philosophy of Kant
171
formula there is no implication that there are such consequences— are not the grounds for my rejecting the actions in question. IN C L IN A T IO N AND D U T Y
Let us return for a moment to the situation developed in consider ing the relation between the two maxims in moral action. We saw there that though every moral action has a purpose, the purpose need not be one which we are induced to pursue by the allurements of desire. The sense of duty, applied with empirical knowledge of human nature, and the world, can project its own ends. Most men are fortunately so disposed, however, that the maxim (b) does not have to generate maxims of type (a), and they only have to undertake the surely arduous enough duty of judging and controlling the inclina tions, at least some of which are beneficent in character and intent. When they are beneficent, they need not be eradicated, but only directed. Hence Sir David rightly rejects Schiller’s famous parody on the antagonism between duty and inclination, the implication of which is that an inclination directed towards a goal makes the ensuing action evil. But Sir David nevertheless charges Kant with “ definite error” on three points; and if Kant did commit these errors, it is hard to see on what ground Schiller’s parody can be rejected. 1) “ It can be maintained that it is possible to have a direct inclina tion to do a certain act and yet do it purely from a sense of duty. . . . We may want to do a thing, and yet do it simply because it is our duty.” 14 I do not know how one can say “ purely” and “ simply” in such a case. There seems to be an open contradiction in saying: I have two motives, A and B ; each would lead me to do action C; I do per form C, but I do so purely and simply from motive A alone. How can there be two motives pointing in the same direction, but only one of them actually effective in the determination of the action? And if anyone claimed that he did an act “ purely” and “ simply” out of a sense of duty but admitted at the same time that he had an inclination to do it, I think Kant would say this was the merest cant. Nevertheless Kant would agree in principle with the point Sir 14
Kant’s Ethical Theory, p. 17.
172
S t u d i e s in th e P h i l o s o p h y o f K a n t
David is making, if this be interpreted to mean that there could be a motive to do what is right because it is right, coexisting with an inclination to do it, though he would be very reluctant to grant that such a case could actually occur. But even so, Kant, would not say that such a will would have the experience of duty, which is felt precisely because the inclinations are wayward. Kant might agree that an action dictated by respect for the law could, in principle, be performed by a being whose inclinations alone would have sufficed to produce the same external act, and that such an action (granting both conditions of respect and inclination) would be as good then as if performed under the less favorable but far more typical circum stances of a less fortunate disposition. Kant was constitutionally suspicious of “ beautiful souls,” so he does not take his illustrations from them; and had he done so, his illustrations would not so clearly have brought out the role of duty, for which his examples were devised.15 2) “ Kant tends to assume that the value of the motivation of an act is to be got by striking the average of the worths of the cooperating motives.” T h at is to say, “ the presence of a morally indifferent motive . . . along with the sense of duty reduces the moral value of an act. . . .” 16 But previously he has said, “ Strictly [a kindly natural dis position] makes [acts of altruism] (according to Kant) not less good, but less purely actions done from a sense of duty.” 17 T his is a tautology, but if one goes further and asserts, in the manner of the first of this pair of sentences, that the less purely moral acts are for Kant less moral acts, it is hard to see how one can criticize Schiller. It must be admitted that there are places in which Kant seems to go further. And it is these places which may have aroused Schiller. For instance, Kant writes of the man who tears himself, unsolicited by inclination, out of dead insensibility to act only from duty and with 15 Foundations, Academ y edn., IV, 397, 11. 19-21. It w ou ld indeed have been
advisable for Kant to follow his precept in the Critique of Pure Reason and avoid examples. Far more attention has been given by most expositors and critics to the details of his examples than to the principles they are devised to exem plify. If something is left out of an example, it does not necessarily mean that it is denied in the reality. 16 K an t’s Ethical Theory, p. 17; Foundations of Ethics, pp. 305, 325. 17 Kant’s Ethical Theory, p. 14.
S t u d i e s in t h e P h i l o s o p h y o f K a n t
»73
out (he could have said “ against” ) inclination: “ then for the first time his action has genuine moral worth.” 18 Following the intention of all the examples— which is to show the good will under the form of duty, i.e., under “ subjective hindrances [which] . . . bring it out by contrast and make it shine all the brighter,” 19 Kant may have meant that then for the first time we may know, or have good reason to believe, that the action was done from duty and that it therefore has moral worth.” 20 Certainly this interpretation is consistent with his general view of the relation of good will to inclination and with the purport of the examples; it is not inconsistent, I think, with the word ing of his conclusion. (With regard to the latter, so far as I can see, Kant quite arbitrarily includes or omits words meaning “ genuine” in the conclusions of the four examples; in two instances Kant says roughly “ only then does the action have moral worth” and in two other instances roughly “ only then does it have genuine moral worth.” ) Such emendation, by reading, “ only then do we know the act to have genuine moral worth,” is, of course, hazardous; but I believe it is justified. But one might suggest, in the manner of Nicolai Hartmann, that here we have a case in which any “ unidimensional” grading of moral values is too simple. “ Genuine moral value,” as Kant uses the term here, can not mean something better than just “ moral value” ; there isn’t any moral value that is not “ genuine.” Rather, it might differ not in “ goodness” but in Hartmann’s dimension of “ nobility.” 21 It is perhaps such multidimensionality of moral values which makes Ross profoundly correct when he asserts, as against the Kantian theory, that there is not one characteristic by virtue of which all right acts are right. 3) T h e third charge brought against Kant is that inasmuch as no motive except a reverence for law has any moral worth, Kant thinks that all inclinations are on the same (amoral) level. In evaluating this, it is best simply to compare the passages concerning the man who acts altruistically from a sympathetic motive. 18 Foundations, Academ y edn., IV, 399: trans. Beck, p. 14. 19 Ibid., p. 397, 1. 10; see also p. 425, 11. 26-31 (trans. Beck, pp. 13, 43). 20 T h i s im port of the examples is recognized by Ross, Kant's Ethical Theory,
p. 16. 21 H artm ann, Ethics, trans. Stanton Coit, II, 201-3.
S t u d i e s in t h e P h i l o s o p h y o f K a n t
»74
Kant
Ross
‘how ever
“ I say that, h o w ever proper, h o w
proper, h ow ever am iable it may -be . . . is on a level w ith other in clin a tions. . . . It deserves praise and e n couragem ent, but not esteem.’ . . .
ever am iable it may be, that k i n d o f a ctio n has no true m oral worth. It
“T h i s
motive,
he
says,
W e cannot agree with K a n t that the inclination to make oth er people h a p p y is ‘on a level with other in clinations’22 and, if valuable, is so because it tends to act to the pu blic advantage.” K a n t ’s E t h i c a l T h e o r y , p. 18; the ellipsis w ith in the q u o ta tion from K a n t is due to Ross.
is on a level w ith oth er inclinations, such as the in clin ation to honor, which, if f o r tu n a te ly d ir e c t e d to xohat in fact accords w ith d u ty a n d is g en e ra lly u s e f u l a n d th u s h o n o r a b le , deserve praise and en co u rage
m ent bu t not esteem.” F o u n d a t i o n s , A cad em y added.
edn.,
IV ,
398;
italics
It must be said that Kant’s sentence is not clear, since he seems to be putting an action “ on a level” with an inclination; properly, the second sentence should have read: “ It is on a level with actions done from other inclinations. . . .” But what he does not say is that all inclinations are on the same level, but only that altruism is on a level with other inclinations which issue in legally correct acts. T h a t altruism has a greater tendency to issue in such actions cannot be denied, and for that reason, if for no other, it should be cultivated. And in the light of other passages, we can say that Kant did encour age such “ adopted” virtues, though he does not budge an inch from his doctrine that no inclination has any moral value in a strict sense. Thus— I restrict myself to one citation from the F ound ations whose whole purpose is to direct attention away from the peculiar ities of human nature and towards the purely rational element in morality— Kant says that some dispositions “ seem to constitute a part of the inner worth of the person . . . [but] they are far from being good without qualification.” 23 Even Sir David himself does not argue 22 In Foundations of Ethics, p. 306, Sir David ascribed to Kant an even more extreme view which entirely omits the qualification italicized in the right hand column: “ Plainly great violence is done to what we really thin k,” he says, “ w hen we are asked to believe that ordinary kindness when not dictated by the sense of duty is no better than cruelty.” A view more like that of the commentary is in Foundations of Ethics, pp. 205-6. 23 Foundations, Academy edn., IV, 394, 11. 5-8; trans. Beck, p. 10. H ere he is speaking of Stoic virtues, but his remark easily applies to philan throp ic feelings as well.
S t u d i e s in th e P h i l o s o p h y o f K a n t
175
that acts issuing from such dispositions are right without qualifica tion; they are only prima facie right, while actions issuing immedi ately from a sense of duty are always (subjectively) right. Sir David says, “ Contrary to this [view ascribed to Kant], it may perhaps be held that in such natural benevolence, in its freedom from selfish interests, we have the germ out of which conscientious ness develops, and something which, like conscientiousness, has moral value.24 Kant does deny what is suggested in the last clause;25 but there is nothing in the first clause which he would deny. Certainly warmhearted fellow feeling was not Kant’s dominant trait, and he does not appreciate it in others as much as Stoic characteristics of independence, moderation, self-control, coolness, reserve, propriety, fortitude, and the like. But none of these has genuine moral worth in the sense that the good will has it, though they are nonetheless con stituents of the good life and the emotional ground upon which moral education must work;26 without them, duty is only a voice crying in a wilderness.27 I doubt, therefore, if there is a real issue here between Kant and his commentator, except that Sir David uses “ moral value” in a wider sense. Were this intended as a general review of Sir David’s little book, it would be unfair to concentrate upon about 10 pages in the first part, and to neglect the remainder of the book, in which I find much to agree with. But c’ est le premier pas qui coute; after one gets inside Kant’s ethical system, the going is fairly easy. T h e first section of the Foundations, the entrance hall to the system, as it were, certainly does not make the mansion itself seem commodious or even habitable. 24 K an t’s Ethical Theory, p. 18. 25 “ Even an inclination to do that which accords with duty (e.g., beneficent acts)
can at most facilitate the effectiveness of moral maxims but not produce them ” (Critique of Practical Reason, A cadem y edn., V, 118; trans. Beck, p. 122). In the same work, p. 75 (Beck, p. 78), he seems to go further and deny even this role to the inclinations: “ In the subject there is no antecedent feeling tending to morality.” B u t this remark is made when he is trying to bring out the peculiarities of a feeling of reverence for the law, which is not an antecedent incentive to morality but “ is morality itself.” 26 Critique of Practical Reason, Academy edn., V, 88. In Part II, the “ M ethod ology,” however, this is played down in order to bring out the unique features o f K a n t’s educational program. 27 Metaphysics of Morals, Part II, Metaphysical Principles of Virtue, §35; for the same view in earlier works, see Schilpp, K an t’s Precritical Ethics, p. 53.
S t u d i e s in t h e P h i l o s o p h y o f K a n t
A t least this is the way it appears if one does not read between the lines ahd constantly remember that everything “ anthropological” has been denied admission and that whatever of the empirical is there is admitted only for the sake 6f bringing out by contrast what is not empirical. A commentator on Kant, or a teacher expounding Kant to students, should never forget that this work is only the groundwork of a metaphysics of morals. T h a t it is such a groundwork is the reason why external criticism and criticism on grounds of adequacy should always take into account the other books in which Kant had no such restricted purpose. It is because Sir David tends to permit the reader to forget that Kant imposed these restrictions on his presentation that I find the opening pages of his commentary so unsatisfactory.
XI APODICTIC IMPERATIVES R E LA TIO N
AND
M O D A L IT Y
O F IM P E R A TIV E S
At the end of §1 of the Critique of Practical Reason there is an ex traordinary statement in which Kant distinguishes between the practical rightness and the categoricalness of an imperative, and indicates that the former might be independent of the latter. This is astonishing, because the reader has been led, in the Foundations, to believe that by establishing a categorical imperative a moral imperative is established. But Kant, in discussing the example, “ Never make a deceitful promise,” says: “ If, now, it is found that this rule is practically right, it is a law, because it is a categorical imperative.” Though Kant says in the Foundations that there is only one cate gorical imperative, from time to time he calls specific imperatives to do certain actions categorical imperatives. For. instance, in the Metaphysics of Morals, “ Obey authority” is called a categorical im perative.1 I do not know of any instance in which Kant calls an imper ative “ categorical” and does not regard it as “ practically right.” But there are indeed imperatives which are formally categorical and yet not “ practically right.” There are likewise hypothetical imperatives which are morally binding and hence practically right. “ Shut the door” is formally categorical, while “ If you are married, remain faithful to your spouse” is formally hypothetical. Presumably it would be possible to show that the latter is “ practically right” and is t h i s p a p e r was first p u b lish e d in the
p r in te d from Kant-Studien, X L I X
1 Metaphysics VI, 3 7 0 -
Festschrift for H . J. P ato n , a n d is re
(1957), 7-24, by permission o f the Editors.
of Morals, Part I, Jurisprudence, Conclusion
177
(Academy edn.,
1 78
S t u d i e s in t h e P h i l o s o p h y o f K a n t
in some sense a law.2 And certainly the former is not “ practically right” in the sense required if we ^re to regard it as a law; it is at most technically practical and not morally practical.3 It is clear from these examples' that the term “ categorical” in Kant’s ordinary usage of “ categorical imperative” does more than refer to the form of the imperative. And it is worth noting that the distinction among imperatives, when first introduced in the F o u n d a tions, is not based on a difference in form of the imperatives them selves, but on their ways of commanding: “ All imperatives command either hypothetically or categorically.” 4 T h e distinction between hypothetical and categorical imperatives, which seems to be a formal distinction, is derivative from this prior distinction, which is modal. What is involved in being “ practically right” is not the category of relation, but the category of modality. T h e best name for what Kant called the “ categorical imperative” and identified with the “ imperative of morality” would seem to be “ apodictic imperative.” In the Foundations and occasionally in the second Critique, terms of modality are used to distinguish among the several imperatives. T h e technical imperative is called problematic, the pragmatic is called assertoric, and the categorical is called apodictic. But in gen eral the modal division is neglected in favor of the formal. Whereas Paton expresses a certain distrust of the modal division,5 the purpose of this paper is to justify its use by drawing attention to certain con siderations which are neglected as a result of Kant’s insistence upon the formal or relational distinction. T h e “ Table of the Categories of Freedom” in the C ritiq u e of Prac tical Reason is regrettably of no help to us in deciding on the pro priety of the term “ apodictic imperative.” This T ab le is, I may venture to say, one of Kant’s less successful productions in the archi tectonic genre, and it is hard to make heads or tails of it, especially 2 Though
Kant explicitly distinguishes “ la w ” from “ im perative” and from “ rule,” he continually uses these words interchangeably, even in §7 of the Critique of Practical Reason which professes, in its title, to give as it were the official state ment of the “ law ” but in fact states only an imperative. 3 “ Technically practical” and “ morally practical” are distinguished in Critique of Judgment, Introduction, I. 4 Foundations, Academ y edn., IV, 414; trans. Beck, p. 31. In Critique of Practical Reason, Academy edn., V, 11 (trans. Beck, p. 11), K ant speaks of “ problematic, assertoric, and apodictic grounds of determination.” 5 T h e Categorical Imperative, p. 115.
S t u d i e s in t h e P h i l o s o p h y o f K a n t
»79
in its division of moments under the dynamical categories.6 T h e dis tinction between hypothetical and categorical does not appear under “ Relation.” Under “ Modality,” corresponding to the moments of necessity and contingency there is “ Perfect and Imperfect Duty,” while “ D uty” itself appears as the member corresponding to the assertoric judgment of existence. According to this Table, an “ apodictic imperative” would be an imperative of strict or perfect duty, commanding without exception a definite action (e.g., paying a debt); such imperatives belong to jurisprudence, not to ethics, because they command a specific action without concern with the motive.7 T h e contrary of an apodictic imperative is not a prohibi tion8 but the imperative of imperfect duty. Such an imperative is not a command to do a certain action, but to act on a certain motive, with room left for the determination of the occasion and the specific kind and direction of the action. Imperatives of imperfect duty belong only to ethics and not to jurisprudence. In this essay I propose to use the name “ apodictic imperative” for all moral imperatives whether they allow latitudo 9 in action or not. Imperatives which are not apodictic or necessary will be treated as imperatives which carry no moral weight, according to Kant’s esti mate. T h ey correspond to the first category of modality in the Table in the C ritiqu e of Practical Reason and not to the second, which already introduces the concept of duty. But they correspond, in their two species, to the problematic and assertoric modalities in the Table of Judgments of the first Critique. There are two textual justifications for this, apart from the nega tive justification of the obscurity of the T able itself. T h e first is that moral imperatives, whether strict or loose, are necessary— the strict necessitating a specific action, the loose necessitating a general maxim and rendering it a law which yet permits “ variety in the 6 It seems to have puzzled K a n t’s contemporaries as much as it puzzles me. See the letter from Schütz to Kant, June 23, 1788 (Academy edn., X , 541). [I subse quently tried to make sense o f it in my Commentary on K an t’s Critique of Prac tical Reason, pp. 144-54.] 7 Metaphysics of Morals, Part II, Metaphysical Principles of Virtue, Introduc tion, X V II I (Academy edn., VI, 410). 8 See R ichard Barber, “ T w o Logics of Modality [in Kant],” Tulane Studies in Philosophy, III (1954), 46. « Metaphysics of Morals, Part II, Metaphysical Principles of Virtue, Introduc tion, VII (Academy edn., VI, 390).
i 8o
S t u d i e s in th e P h i l o s o p h y o f K a n t
rule.” 10 As necessary, they are products of reason, which (in its logical use) is the faculty of apodictic knowledge.11 And, more importantly, we have seen that Kant himself on occasion calls them apodictic imperatives. A N A LYSIS O F H Y P O T H E T I C A L IM P E R A T IV E S OF T H E F IR ST T Y P E
A hypothetical (problematic or assertoric) imperative is one which holds for any rational being under the condition that this being has a certain end or purpose which is believed to be the effect of the action he is told to execute, and which states this condition as the restrictive condition on the validity of the imperative. We shall call these imperatives conditional imperatives or hypothetical imperatives of the first type, and under them are included two modal types: asser toric and the problematic imperatives. A full analysis of such an imperative would make the following components explicit: a) In the protasis:
1) A conative element: an incentive, impulse, or interest in something (B ), which is the purpose of the commanded ac tion; an object or state of affairs the representation of which is one of the causes of its existence through the relation of the representation to the overt action of the subject; expressed in the imperative as “ If I want B ” or “ Because I want B . ” 2) A cognitive element: knowledge of the causal relation be tween the action commanded (A) and the purpose or end (B); equivalent under condition (1), to “ A is a means to B . ” 3) A logical form: “ If I fully will the effect [B ], I must also will the action [A ] necessary to it.” 12 b) T h e apodosis: “ Do A ” In the technical imperative or rule of skill, the fully expanded form would be: “ (1) If you will B , (2) because A is a necessary condi tion of B, then by (3) do A . ” Here the problematic element lies in the first conditional: “ if you will B. . . .” I f A is really commanded under 10 Critique of Practical Reason, §1, Remark. 11 Critique of Pure Reason, A 75= B 100 n. 12 Foundations, Academy edn., IV, 417; trans. Beck, p. 35.
S t u d i e s in th e P h i l o s o p h y o f K a n t
181
this condition, (2) must here be assertoric.13 In the pragmatic im perative or counsel of prudence, the full exposition would be: “ Be cause you will B (£=happiness), (2) if A is a necessary condition of B, then by (3) do A . ” Here the problematic element lies not in the ante cedent conation, since a rational being having any desires necessarily desires his happiness;14 it lies in our irremediable uncertainty as to what is the empirical content of B and whether A will in fact lead to it. Kant tells us that the possibility of these imperatives is easy to es tablish, because they are, in what concerns the will, analytical.15 This is correct,16 but it requires a little attention to see what it means and why it is true. It is not entirely clear in what sense an imperative can be either analytic or synthetic. Imperatives are not judgments with a subject and predicate, and so do not fall under Kant’s explicit division of the types of judgment. And though for any imperative a set of indica tive (but not necessarily factual)17 statements may be formulated un der which the imperative could be derived by a syntactical change of mood, in the case of hypothetical imperatives of the first type, some of the corresponding judgments are themselves hypothetical, and hence likewise fall outside Kant’s official rubric for distinguishing analytic from synthetic judgments. It is not entirely pedantic to 13 T h e r e is here no imperative to do A unless A is known or believed to be a
means to B. A purely problematic protasis could not be the basis for an actual command. Kant in the First Introduction to the Critique of Judgment (Academy edn., X X , 200 n.) withdraws the name “ problematic im perative” as involving a self-contradiction. B u t the “ mixed problematic imperative,” one of whose premises is assertoric, does not involve any self-contradiction in its name. T h e present anal ysis shows that each of the hypothetical imperatives of the first type may have a problem atic and an assertoric antecedent, and that they differ only in the locus of each modality. 14 Critique of Practical Reason, Academy edn., V, 25; Foundations, Academy edn., IV, 415. 15 Foundations, A cadem y edn., IV, 417. 16 [In my Commentary on K an t’s Critique of Practical Reason, pp. 86-87, I modified the position taken here. I there argued that “ in w hat concerns the w ill” both types of imperative are analytical, but in what concerns the understanding, i.e., the content in the choice of means to an end, neither is analytical. T h e meaning of “ analytical” as applied to an imperative requires a fuller analysis than I gave it in this paper, but I attempted such an analysis in the Commentary.] 17 T h u s , “ Do A ” corresponds to the indicative, “ You ought to do A , ” though the two sentences have different roles in moral discourse. “ Do A ” would not be “ practically rig h t” unless “ You o ught to do A ” were true.
182
S tu d ie s in t h e P h i l o s o p h y o f K a n t
insist on these points; but the minor infelicities of expression should not permit us to overlook the important point Kant is insisting upon, even though he does it in not entirely suitable language. I shall try to say, therefore, what I think he means. If we take the statements of the conditions under which the apodosis is regarded as binding and phrase them in declarative statements, e.g., “ Under conditions C a rational being will choose A , ” then the proposition “ A rational being will choose A ” will follow analytically from this and the proposition, “ A rational being is under conditions C .” By “ following analytically” I mean that a denial of the conclu sion will contradict a premise. Thus: “ If a rational being fully wills B and knows A to be a necessary condition of B, and if (3) to will fully an end includes willing the means known to be necessary to this end, then the rational being wills A , ” is analytic in the broad sense in which Kant here uses the word “ analytic,” since to deny the apodosis is to contradict the protasis. But it is not analytic in the strict Kantian sense, since an analytic proposition for Kant is one whose denial is self-contradictory, and the mere concept of rational being does not analytically “ contain” the concept of willing A . But it is analytic, even in the broad sense, only in what concerns the will. T h a t A is necessary to B is not known analytically, and that the rational being in fact wills B (except perhaps when / ^ h a p p i ness) is likewise not known analytically. All that is known analytically is (3), and this concerns the form of volition and not the contents willed. (3) is not actually a premise,18 but a rule of practical inference, specifying the relation between the variables A and B. Kant should say, therefore, not that the hypothetical imperatives of the first type are analytic, but that their formal principle, (3), is analytic. “ If you wish to have bread, devise a mill” is not, fortunately for most of us, in any sense analytic, but is only a disguised form of the empirical statement, “ Mills grind flour, and flour is needed to make bread.” I shall call (3) the formal principle of hypothetical imperatives of the first type.19 It is indeed analytical, and it alone concerns the will 18 See R. M. Hare, T h e Language of Morals (London: O xford University Press,
P- 4 8 . 19 By analogy to B ro ad ’s terminology in his discussion of the categorical im p e ra tive, it might be called “ the supreme principle of conditional imperatives.” See C. D. Broad, Five Types of Ethical T h e o iy , pp. 120-21, 123. I call it a principle of conditional, not of hypothetical imperatives, for we shall see that there are h y p o thetical imperatives which do not fall under this principle.
S t u d i e s in t h e P h i l o s o p h y o f K a n t
'95
Hence one source of Kant’s rigorism, as in the essay on the alleged right to lie. In this essay I do not propose to examine the question of Kant’s rigorism directly, or to discuss further, in Kantian terms, the im portant distinction between prima facie obligations, qua-obligations, and real obligations. My purpose here is only to point out the man ner in which he could have distinguished hypothetical assertoric and problematic imperatives, which he called hypothetical imperatives simpliciter, and hypothetical apodictic imperatives, which he did not discuss at all. In judgments in hypothetical form, we may distinguish between two relations of antecedent and consequent. First, there are judg ments in which this relation is real and synthetical: “ If there are clouds, it will rain.” Second, there are judgments in which the rela tion is analytic and logical: “ If a man is a bachelor, he has no wife.” T h e former can be roughly translated into categorical judgments that are synthetic: “ Cloudy days are rainy days,” and the latter can be strictly translated into analytic categorical judgments: “ Bache lors are unmarried men.” Similarly, there are two types of relation between antecedent and consequent in hypothetical imperatives. T h e relation is different in “ If you want to keep your feet dry, wear your rubbers,” and in “ If you make a promise, keep it.” Unlike hypothetical judgments of both types, which can be translated into categorical judgments, the former imperative cannot be translated into a categorical imperative having the same meaning and scope of relevance. It does not say, “ Wear your rubbers,” or, if it says this it does so only to those who do not wish to wet their feet. T h e hypothetical imperative of the second type becomes, through moral rule with the admission of an exception to a moral rule. He does not recog nize that the difference between a rule with an exception and a grammatically categorical rule is often purely a verbal one” (T h e Theory of Good and Evil [2 vols.; London: O xford University Press, 1924], I, 116). Mr. Hare (Language of Morals, pp. 50-54) discusses in this connection the difference between “ occasional exceptions” and “ classes of exceptions.” T h e latter can be specified in the rule w ithout affecting its apodicticity, and Mr. Hare says that when so specified the rule is “ not looser than it was before, but stricter” (p. 53). [Since this paper was published, Mr. Singer has elaborated and defended his view in his Generalization in Ethics (New York: Alfred A. Knopf, Inc., 1961).]
196
S tu d ie s in th e P h i l o s o p h y o f K a n t
only verbal change, “ Keep your promises.” Now while this impera tive is not strictly universal, since it contains “ your,” its address is not as specific as that of the imperative “ Wear your rubbers.” 41 If the imperative, “ Wear your rubbers” is obeyed, it is not obeyed by anyone because it is a general imperative; the reasons why it is obeyed may be stated only in an expanded protasis of the correspond ing hypothetical. In this respect, “ If you make a promise, keep it” differs from “ If you would preserve your good name, keep your promise.” T h e categorical imperative corresponding to the former of these commands is an apodictic imperative under the formula of a categorical imperative; the categorical imperative enunciated in the same sentence (i.e., “ Keep your promise” ) but corresponding to the hypothetical of the first type (“ If you would keep your good name . . .” ) is not apodictic, does not have the scope of the apodictic, and expresses no necessity. T h e hypothetical imperative such as “ If you wish to keep your library intact, write your name in your books” has as its antecedent a factual condition, which is related to the consequent only through an empirical causal relation. T h e declarative sentence reporting obedience to the consequent, i.e., “ You do in fact write your name in your books,” reports a fact believed to be a causal condition of the end entertained in the antecedent. T h e fact that a man desires to keep his library intact does not create an obligation to write his name in the book. (He may indeed say, “ T o keep my books from being stolen, I have been obliged to put my name in each volume,” but he does not say, “ I have had an obligation to put my name in my books.” Wants are not the sort of things that create obligations, though to satisfy them we may be “ obliged” to do certain other things. “ T o be obliged to put my name in each book” merely calls attention to the causal necessity or efficacy of an irksome means to a desired end; but “ to have an obligation to put the University book plate in each book” calls attention to a qua-obligation of the U n i versity Librarian. T o be obliged to do something is not, etymology to the contrary notwithstanding, the same as to have an obligation to do it.) T h e obedience to the imperative’s consequent where the antecedent
Mr. Hare (Language of Morals, pp. 187-90), recognizing that no imperative in ordinary English is strictly universal, proposes an “ enriched imperative m ood,” and in this there w ould be a universal imperative corresponding to “ Keep your promise” but not one corresponding to “ W ear your rubbers.” 41
S t u d i e s in t h e P h i l o s o p h y of K a n t
»97
is obligation-creating is not a means to a state of affairs desiderated or mentioned problematically in the antecedent. T h e affirming of the consequent by obedience to the imperative is not a means to the accomplishment of the state of affairs envisaged in the antecedent of the hypothetical imperative of the second type; the promise has been made, and we do not say that “ fulfilling a promise” is a means to the obedience to the imperative— it is the obedience. And certainly ful filling a promise is not a means to the making of a promise and the obligation to fulfill a promise is not dependent upon whether the fulfilling of the promise is in fact a means to certain purposes enter tained in the original making of the promise. In all these respects, the relation of antecedent to consequent dif fers in the two types of hypothetical imperative. Mr. Jackson aptly sums up the differences by speaking of the apodosis of hypothetical imperatives of the second type as being “ derivatively practically necessary” while the apodosis in those of the first type, which com mand a means, he calls “ derivatively theoretically necessary.” 42 Only moral knowledge is required for the practical derivation of apodosis from an obligation-creating protasis, while theoretical knowledge of the world is required for learning what one must do in order to keep his feet dry or to preserve his library. Given the imperative, “ If you want to keep your feet dry, wear rubbers,” one may intelligently ask “ Why?” and expect an informa tive answer: “ Because the ground is wet, because rubber is imperme able to water. . . .” But, in Mr. Nowell-Smith’s idiom, it is “ logically odd” to ask, “ Why should I keep my promise?” as a response to “ If you make a promise, keep it.” 43 T h e protasis here seems to be a sufficient condition for the apodosis even though it may not in fact be sufficient since the apodosis may be only a prima facie or quaobligation. But if the question is asked, “ I know I made a promise, but why should I keep it?” the question seems to be one that could not be answered by new information on the same level as the fact that a promise has been made;44 the only appropriate answer to this question would be an exposition of an ethical theory, in which 42 Reginald Jackson, “ K a n t’s Distinction between Categorical and Hypothetical
Imperatives/’ Proceedings of the Aristotelian Society, 1942-43, 143. 43 Kant w ould have regarded it as “ logically odd.” See Metaphysics of Morals, Part I, Jurisprudence, §19 (Academy edn., VI, 273). 44 New information on the facts would be relevant only in the decision as to whether the prima facie or qua-obligation is really binding.
i 98
S tu d ie s in t h e P h i l o s o p h y o f K a n t
“ promise-keeping” would be put under a more general theoretical rubric, etc. Additional reasons to justify this particular apodosis might, in fact, tend to weaken the awareness of the moral connection between promise-making and promise-keeping;45 and I think this is the principal reason why Kant preferred examples of obligations in which any additional factual information might tend to obscure and impugn the purity of the moral connection between making a prom ise and keeping it.46 I do not say that “ You have made a promise” is a sufficient reason for “ Keep your promise.” Whether I am in fact obligated to keep a promise may depend upon much besides the fact that I have made it, and other facts as well as other obligation-creatings may be relevant to the decision.47 But that I desire to keep my feet dry is never a sufficient reason, and nobody ever believed it was a sufficient reason, for wearing rubbers; and even if I state the sufficient condition, in cluding all relevant facts, the connection between them and the hypothetical imperative. “ If you wish to keep your feet dry, wear your rubbers” is not like that in “ If you make a promise, keep it.” A full theory of apodictic hypothetical imperatives is not to be found in Kant. But it is not difficult to see what some of its features would be. Such imperatives would be first-order imperatives justified by the formula of what Kant called the categorical imperative, but which we have found to be a formula of an apodictic imperative regardless of its grammatical form. This formula would be a criterion of the admissibility of various kinds of conditions into the protases of the hypothetical imperative that is apodictic. It would tell us what conditions could legitimately and what conditions could not legit imately be included without destroying the apodicticity. T h e formula would make it possible for us to answer the Kantian ques tion, “ Would I be willing for all men to obey the imperative, ‘Do A ’?” in the form, “ Would I be willing for all men under condition C 45 Foundations, Academy edn., IV, 394. 46 I have discussed K a n t’s examples from this point of view in “ Sir D avid Ross
on Duty and Purpose in K a n t” [see Essay X, pp. 172-73]. 47 T h a t such considerations are permitted by Kant, how they are evaluated, and what they are in K a n t’s work have been exhaustively explored by Paton, “ A n Alleged R ig h t to Lie. A Problem in Kantian Ethics,” Kant-Studien, X L V (1954), 190-203; and by Singer, “ T h e Categorical Imperative.” [See also Mary J. G re g o r’s T h e Laws of Freedom; A Study of K an t’s M etho d of Applying the Categorical Imperative in the Metaphysik der Sitten (Oxford: Basil Blackwell, 1963).]
S t u d i e s in th e P h i l o s o p h y o f K a n t
»99
to obey the imperative, ‘If C, do A ’?” w ithout trivialization, since the range of C could be determined by examining the effect of the various possible C ’s upon the apodicticity of “ Do A .” 48 It might permit us to show why, for instance, “ If you make a promise and find it incon venient to fulfill it, don’t fulfill it,” is not an apodictic imperative, while “ If you make a promise and find that an innocent person will die if you fulfill it, don’t fulfill it” might be an apodictic imperative. Though a full theory of such imperatives is not developed any where in Kant, the Metaphysics of Morals is filled with clues as to how it could be done; and such a theory is a desirable addendum to the theory that seems to require moral rigidity and is a desirable founda tion for his casuistry. T h e apparent incompatibility of such hypo thetical imperatives with Kant’s best known ethical principle van ishes when wre realize that “ categorical” was not the most suitable adjective for Kant to use in describing the moral imperative. See W . I. Matson, “ Kant as a Casuist,” Journal of Philosophy, LI (1954), 855-60. [Mr. Singer, in his Generalization in Ethics, has skillfully spelled out the restrictions expressed in the words without trivialization; cf. especially ch. 4.] 48
X ll
THE FACT OF REASON: AN ESSAY ON JUSTIFICATION IN ETHICS IN TE R N A L AN D
EXTERNAL
Q U E ST IO N S
By the term “ external question,” which I borrow from Carnap,1 I refer to a question which concerns the collective justification of all t h is p a p e r
, u n d e r the title: “ D as F a k t u m der V e r n u n f t : Z u r R e c h tfe r tig u n g s -
p r o b le m a tik in der E t h i k ,” was read to the m e e t i n g o f the K a n t G esellschaft in B o n n on J u ly 25, i960, a n d was p u b lis h e d in G e r m a n in K a n t-S tu d ien , L I I (i960), 2 7 1 -8 2 . T h i s E n g lish translation is p u b lis h e d w ith the p erm is sion o f the E d ito r o f the Kant-Studien.
1 R u d o lf Carnap, “ Empiricism, Semantics, and O ntology,” R evue internationale de philosophic, IV (1950), 20-40. “ W e must distinguish,” Carnap says, “ two kinds of questions of existence: first, questions of the existence of certain entities . . . within the framework; we call them internal questions; and second, questions con cerning the existence or reality of the framework itself, called external questions” (pp. 21-22). Althou gh Carnap was concerned with existence only of such “ entities” as numbers, things, properties, classes, and propositions, his distinction can easily be applied to “ entities” like values and duties. Carnap tries to show that external questions have no epistemological meaning and are thus pseudo-questions. Such external questions show only that we arbitrarily assume or reject “ the forms of expression for the framework in question.” Opposed to this is the thesis of this essay: if the framework is itself a value-system (system of value judgments, im pera tives, or obligations), answers to internal questions leave no elbowroom for a rb i trary anwers to the external question. T h e form of the internal answer determines the abstract form of the external answer; the internal question itself implies a dis tinct kind of answer to the external question. In this respect, theoretical and practical questions are different. If Carnap is correct in the problem o f theoretical questions, one cannot infer from that that his theory can be simply applied to prac tical questions. I do not here raise the question of w hether his solution to the theo retical question is correct. T h e conclusions reached in this paper have a marked resemblance to those of C. I. Lewis in his T h e Ground and Nature of the R ig h t (New York: C olu m b ia University Press, 1955), though the form of the argum ent is very different. I owe important suggestions concerning the concept of “ presupposition” to Dr. Patricia A. Crawford; see her “ K a n t’s T h e o r y of Philosophical Proof,” Kant-Studien, LIII (1962), 257-68. I wish to thank Professor Johannes-Erich Heyde of Berlin for a terminological suggestion, which I have followed, in the analysis of the expres sion “ fact of reason.” 200
S t u d i e s in t h e P h i l o s o p h y o f K a n t
201
the true propositions contained in any particular realm of discourse. T h e question of the justification of a judgment within the particular realm of discourse will be called an “ internal question.” T h e problem of justification arises from the belief that no internal question is fully answered until the external question is answered, since the answer to any internal question is hypothetical until the premises are established. T h e difficulty of the problem arises from the fact that the norms for judging answers to the internal questions are never sufficient for judging answers to the external question. I shall illustrate the two kinds of questions in logic and in ethics, though the paper as a whole will be concerned only with the question in ethics. An internal question in logic is, for instance, “ W hy is the syllogism Ztaroco valid?” T h e answer is given by adducing the princip iu m de om ni et nullo, which is a principle within logic. But if one asks, “ W hy is this principle valid?” it seems that no answer can be given. Without this principle, one cannot reason logically at all, and therefore any argument designed to show that it is valid makes use of it on the assumption that it is valid; thus it involves a petitio principii. T h e dictum is the standard for validity of the argument, not the argument a support for the validity of the dictum. Similarly in ethics. An internal question is, for instance, “ Why ought one to tell the truth?” Various answers to this question are given in various ethical theories, but the form of the answer is always the same, viz., by taking recourse to some higher principle within the theory such as, in Kant’s theory, the categorical imperative. But one who asked, “ Why ought I to tell the truth?” and who is told, “ Because the categorical imperative demands it,” might then say: “ I know that, according to the categorical imperative, I ought to tell the truth; but I want to know why I ought to heed the categorical imperative.” Thus is raised the external question in ethical theory. And it seems to be a question that cannot be answered. An answer would have to be either a factual or a value judgment. If a factual judgment, the answer will not satisfy because we know that no factual statement implies a value judgment. If a value judgment, the original question was not truly an external question, for now the same question can be asked of it. Kant was aware of the distinction between internal and external questions in ethics. In the Foundations the internal question is dis
202
S tu d ie s in th e P h i l o s o p h y o f K a n t
cussed in the analytical parts, Sections I and II. These Sections give the formula of the moral law and the structure of his ethical system, and thus provide the principles by which every internal ethical ques tion is to be answered. But Kant repeatedly reminds us that nothing in those Sections shows that duty is not “ a vain delusion and chimerical notion.” 2 T o show that duty is a concept with an object, and not merely a concept within a neat logical system, requires the syn thetical use of pure practical reason, and a shift from the analytical to the synthetical method of the T h ird Section. (A similar division, between exposition and deduction, is present, though not so clearly, in the second Critique.) Though Kant’s answer to the internal questions of ethics is clear, it is not at all clear what his answer to the external question is. T h e transcendental deduction of the principle of pure practical reason is, I think, the most obscure part of Kant’s ethical theory. It is the purpose of this paper not just to expound what I think Kant’s answer to the external question is, but first to explore some possible ways of dealing with the external question and then to examine Kant’s method of answering it in the light of its possible alternatives. This preliminary exploration is the more necessary in that the authority of Kant has been invoked by defenders of two dramatically opposite treatments of the problem. T H E IN T U IT IO N IST IC ANSW ER T O T H E E X T E R N A L Q U E STIO N
Taking the external question in its highest generality, “ W hy ought I to be moral?” the intuitionist answers that one can just see that one ought to be moral, and that is all there is to it. One cannot go farther than that in either logic or ethics: there are clear and distinct self-evident principles of right thinking, and there are clear and distinct self-evident principles of right action, and these are re spectively the fundamental principles of logic and of ethics. While these truths are not, of course, seen with one’s physical eyes, the act of apprehending them is more like seeing something than it is like thinking and inferring. T o call it “ seeing” is to make use of a metaphor, and all the other words which name it likewise seem to 2 Foundations, Academy edn., IV, 402; trans. Beck, p. 18.
S t u d i e s in t h e P h i l o s o p h y o f K a n t
203
be metaphorical also: A nschauen, intueor, “ apprehend,” “ grasp,” “ see by the lum en naturale,” etc. I shall here use the word “ intuition” in two senses, which I shall distinguish as follows. An intuition in the real sense is a direct apprehension of an objective truth, value, or state of affairs. An intuition in the putative sense is an act which seems to be a real intuition but which in fact may or may not be. Now it follows from the meaning of intuition in the real sense that if there is an intuition in the real sense of an answer to the external question, ethics is in fact perfectly justified and an otherwise infinite regress of questions and answers is avoided. But I submit that all we have in the way of data are intuitions in the putative sense, that con tradictory judgments of intuitions in the putative sense exist, and that there is no valid intuitive means by which real intuitions can be dis tinguished from putative intuitions. I do not wish to deny that putative intuitions are common experi ences; nor do I wish to deny that some putative intuitions are neces sary to provide a kind of Archimedean platform on which to stand while inquiry is under way. What I do wish to deny is that putative intuitions are self-guaranteeing. For though we would not call a mental act an “ intuition” unless we believed that it was in fact revelatory, a putative intuition will seem to be revelatory and will provide an Archimidean platform for inquiry into whether it is in fact a real intuition or not. T h e necessity that there be an intuition does not guarantee the necessity of any specific putative intuition. T o say that in any well-ordered set of propositions for which truth is claimed there is at least one that must be accepted intuitively, if not arbitrarily, is only to acknowledge that not every proposition in the set can be proved. But to say this is not to decide which propositions in the set— or indeed whether any proposition in the set instead of its con tradictory— is intuitively certain, where “ certainty” is found in a real instead of in a merely putative intuition. Methodologically, to claim intuitive certainty for any proposition means only: 1) I believe it so strongly that I do not entertain any alternatives. 2) I cannot prove it by deducing it from any proposition about which I have a higher degree of certitude or conviction, so that it seems to be self-guaranteeing as seeing seems to be self-guaranteeing.
20 4
S tu d ie s in t h e P h i l o s o p h y o f K a n t
3) I must use it in the proof of other propositions which I believe but which I do not claim to be intuitively true. Each of these conditions is essential, but each is fully met by a propo sition which is only putatively intuition. T h e third of these condi tions, however, is fatal to any attempt to claim an intuitive certainty as the answer to an external question, i.e., as the answer to the ques tion of the justification of ethics. T h e reason is this. Accepting a given proposition as intuitively true — even calling it a putatively intuitive proposition— requires the prior acceptance of the truth of the other propositions for which intui tive certainty even in the putative sense is not claimed (condition 3). And the real intuitive certainty of the proposition (not just its puta tive intuitional status) requires the actual truth (and not just the be lief in the truth) of the propositions which answer internal questions; for if they are false, then by modus tollens the putatively intuitive proposition is false and therefore cannot be a really intuitive propo sition. Hence the assertion of the really intuitive truth— assuming we could in fact assert it— is not an additional warrant for the truth of the propositions internal to the system. T h a t is, answering the ex ternal question in the manner of the intuitionists— even supposing real intuition to be possible— would not justify the collection of answers to the internal questions, but would rather presuppose some independent warrant for them, and so on ad infinitum.
T H E P R A G M A T IC AN SW ER T O T H E E X T E R N A L Q U E STIO N
Another procedure in answering the external question is to show that the external and internal questions are so different that the con cept of an “ answer” is ambiguous. T o answer an internal question means to cite a good reason for a yes or no answer. But there is no answer of this shape with a good reason for an external question. Hence the quest for justification is not only in fact futile, but it is misguided from the beginning. It is like asking, “ W hat letter of the alphabet comes before A?” and not being satisfied with the answer, “ None,” because “ none” is not a letter. Suppose someone asks, “ Why should I tell the truth?” and I answer,
S t u d i e s in t h e P h i l o s o p h y o f K a n t
205
“ Because you ought not to do actions that will probably make you unhappy.” T h e other person may ask me why he should not do this, and I can reply by citing some other more general principle such as “ You should respect your own person.” For each of my answers, he can respond with another “ Why?” Perhaps in exasperation I finally tell him: “ At least you will grant that ‘You ought to do your duty.’ This is an analytical proposition. It would be self-contradictory to say you ought not do your duty. Here, then, is my justification for telling the truth. You granted that each of my answers to your in ternal questions was correct, but inadequate; and now I have showed you that the first premise of my first prosyllogism is an analytical proposition; therefore the conclusion ‘You ought to tell the truth’ is logically necessary.” This, however, is not a justification. It would be a justification only if an axiological equivalent of an ontological argument were valid. Kant showed that the concept of a perfect being may analy tically imply the concept of an existing being, but that the concept of an existing being does not imply the actual existence of the being. We can show, similarly, that if there is anything that one ought to do, it must be his duty to do it; but this hypothetical and analytical judgment does not in the least imply that there is in fact anything that one ought to do. Kant knew this. Granting this failure to find a validation (justificatio cognitionis), an effort is made to find what Feigl3 has called a vindication (justi ficatio actionis). A vindication is a pragmatic warrant for acting in a certain way (here, giving a set of answers to a set of specific internal questions) without claiming that we know the answers to be cogni tively correct, i.e., without claiming that the answers to the internal questions can be known to be true to objective fact or metaphysical essences. I fear that the belief that validation is impossible may have led the author of this distinction to think that vindication is a substitute for it. But it is not, as I shall try now to show. T o say that an ethical system cannot be validated may mean one or two things. It may mean that we cannot, in principle or in fact, H erbert Feigl, “ Validation and Vindication,” in Readings in Ethical Theory, ed. W ilfred Sellars and John Hospers (New York: Appleton-Century-Crofts, Inc., 1952), pp. 667-80; “ D e principiis non est disputandum. . . ?” in M. Black, P h ilo sophical Analysis (New York: Prentice-Hall, Inc., 1963). 3
206
S t u d i e s in t h e P h i l o s o p h y o f K a n t
give a justification for it. If Feigl is correct, the principle “ a is a” cannot be justified because we cannot give a justification for it. I shall call such a proposition participially u njustified ; this term neutrally describes the status of the proposition in question. But to say that a proposition cannot be justified may mean that it is in fact unjustified, i.e., wrong. Again if Feigl is correct, the principle “ a is not a” cannot be justified, but this proposition I shall describe as pejoratively unjustified.
Now if we accept the statement, “ Ethics as a whole is not vali dated,” then faut de m ieux we may accept vindication for it, inter preting the sentence to mean that ethics is not participially justified. But there is not the slightest reason why we must interpret the sen tence merely participially. Maybe we cannot justify ethics because it is unjustified in the pejorative sense. But— and here is the fatal difficulty— a pejoratively unjustified system may be vindicated just as well as a merely participially unjustified system can be vindicated. Success in use is no guarantee of the rightness of what succeeds, espe cially when it is a question of the rightness of the test of success itself. If an external validation for an ethical system could be given, it would give that system a unique standing, for if one system is so validated, no rival can be validated. But there is no evidence that only one system can be vindicated; in fact, it is probably false. Vindication is not criterially unique, therefore, and provides no substitute justifi cation for the answer sought along the route of validation.
N EW F O R M U L A T IO N OF T H E E X T E R N A L Q U E STIO N
T h e external question, “ Why ought I to be moral?” seems to be logically like the internal question, “ Why should I go out in the rain?” But it is not. T h e internal question admits of two answers: (1) “ You should go out, because you have an engagement” or the like; and (2) “ You should not.” T h e external question is quite unlike the internal question, inasmuch as it permits at most only one kind of answer, “ You ought to embrace this ethical system because . . . , ” and its difficulty is due to the fact that we cannot find an answer of this kind, since the because-clause must be a value judgment and the whole answer is then seen as an answer to an internal question. And if one gave an answer of the second kind, viz., “ You should not be
S t u d i e s in t h e P h i l o s o p h y of K a n t
207
moral” or “ You should reject all ethical systems,” this is still a moral judgment, though one that is self-contradictory and hence pejora tively unjustified. There is another way in which the two questions differ. T h e in ternal question asks for information or advice; the external question does not. Suppose someone asks me if it is right for him to do action X , and I reply, “ No, you ought to do your duty.” If he asks why he should, I shall not be able to give him any reasons other than reasons for doing the action n o n - X . T h a t is, the reasons for being moral and for doing one’s duty, which he asked for, are not different from the reasons for doing the act which is dutiful, which I have already told him. It is not that there is one set of reasons for being moral, which we do not know, and another set of reasons for doing the moral act, n o n - X , which wre may know and have already stated. Perhaps the terminology of “ external and internal questions” has misled us; the questions now seem t o l o c o e l o different and hardly to fall under the same genus. Let p , q, r, s be a series of internal ques tions, in which q is a question about the decision of p , r a question about the decision of q, and so on. T h e so-called external question should not be regarded as the last member of this series, Z, which we approach step-wise. It is not an additional question like the others, but a question about the questions. Question p has two answers: (a) “ You should do it” or (n o n - a ) “ You should not do it.” These possible answers to p constitute the tautology (a or n o n - a ) , and a tautology materially implies any true proposition. Thus “ Either you should go out to keep an engage ment, or you should not go out to keep an engagement” entails that the Rhine is a swift river and that Bonn is a lovely city. Among the propositions implied by (a or n o n - a ) , however, there are some which are necessary for a decision to be made between a and non-a.4 Such propositions I shall call a p r e s u p p o s i t i o n . Granting a situation of choice, which I shall symbolize as (a or n o n - a ) , we can ask: What propositions are entailed by the choice of a, and what propositions are entailed by the choice of n o n - a ? These propositions constitute answers to the internal question, and if we choose a we are commit ting ourselves to the system of ethical principles which justify this internal choice; similarly, if we choose n o n - a we are committed to an 4 T h is will be modified in the second paragraph following this.
208
S t u d i e s in t h e P h i l o s o p h y o f K a n t
alternative system of answers to internal questions. But if we ask merely what propositions are entailed by and required for a decision between a and non-a, we are' asking for the presuppositions which would make a choice between them possible. These presuppositions will be at least part of the answer to the external question, and will constitute the justification for an ethical system about which we shall not be able to ask the same question we could ask about the system of internal answers based on a, or about the system of internal answers based upon non-a. Suppose, for instance, a is the utilitarian decision in some problem of choice, and non-a is the answer that would be given by an egoist. A system of propositions entailed by a will be the utilitarian system; the system based on non-a might be the egoistic system. About the choice between these systems we might ask the external question, and we have seen how hard it is to answer the question of this kind: after all, why should I consider the interests of others? But the sys tem of propositions entailed by and required for a decision with respect to the choice (a or non-a) will not be subject to this kind of question, because it is independent of which of the contradictories we choose. It depends solely upon there being a choice between them. There are, I think, two possible presuppositions which would pro vide a decision-procedure for (a or non-a). T h e first is like a throw of a coin: heads I do a, tails I do non-a. I know of no way in which anyone could reject this as a valid decision-procedure provided it were itself chosen by, say, tossing a coin and not as a result of argu ment. But immediately the proponent of such a tychistic ethics attempted to argue for it, on whatever grounds, by saying that this was the right way to choose, I think we should have him.5 For here there would be a choice formally like that between a and non-a, and he is citing reasons which seem to be valid presuppositions for a choice between deciding to do a or, contrariwise, to do non-a, so that the question of the justification of either would present insurmount able problems like those we found in the intuitionistic and pragmatistic answers. What I am suggesting is this: that immediately one gives any rea son for a choice, there is at work one presupposition which isTTieast part of the answer to the external question of ethics. This presup5 See C. I. Lewis, T h e Ground and Nature of the R ight, p. 86.
S t u d i e s in t h e P h i l o s o p h y o f K a n t
209
position, I submit, is what Kant called “ the sole fact of pure reason.” T o an examination of this fact I now turn. T H E F A C T OF PU R E R E A SO N — F IR ST IN T E R P R E T A T IO N
Eight times in the C r i t i q u e o f P r a c t i c a l R e a s o n Kant uses the ex pression “ fact of pure reason” or “ fact of reason.” It refers to the ultimate datum of ethics, but what this ultimate datum is, is identi fied in various ways. T h e denotata of the term are given as follows: 1) Consciousness of the moral law (Academy edn., V, 31), 2) Consciousness of freedom of the will (p. 42), 3) T h e law (pp. 31,47, 91), 4) Autonomy in the principle of morality (p. 42), 5) An inevitable determination of the will by the mere conception of the law (p. 55), 6) T h e actual case of an action presupposing unconditional causal ity (p. 104). In other works there is a similar variety of denotations.6 Sometimes Kant speaks of the “ fact of reason” and sometimes of the “ fact of pure reason,” and once of the “ fact of moral-practical reason.” Sometimes he speaks of “ fact” s i m p l i c i t e r , and sometimes of “ fact, as it were.” 7 Because of this variety of terminology, we cannot take Kant’s i p s i s s i m a v e r b a and expect to find a single and unique interpretation for them. A t least there seem to be two different kinds of facts referred to. First, there are the statements that the fact of reason consists in the direct consciousness of something— that is, the fact is the con sciousness of the law (1) or of freedom (2). But we have been told elsewhere, and in such a way that leaves no doubt as to Kant’s mean ing, that we are n o t conscious of freedom, but only of the moral law 6 T h e fact is: (a) freedom (Critique of Judgment, §91 [Academy edn., V, 468]); (b) the law of freedom (Metaphysics of Morals, Part I, Jurisprudence, §6 [Academy edn., VI, 252]); (c) that the law is in us (Vorarbeiten zur Tugendlehre, Academy edn., X X I I I ,
378);
(d) the categorical imperative (Opus postumum, Academy edn., X X I , 21). (a) corresponds most closely to (5) and (6); (b) to (3) and (4); (c) and (d) to (1). 7 T h e many variations in his terminology are explored in my Commentary on K a n t’s Critique of Practical Reason, ch. 10, §2, especially p. 166 nn.
210
S tu d ie s in t h e P h i l o s o p h y o f K a n t
which is its ratio cognoscendi.8 So let us say that, at least, the con sciousness of the moral law is a fact, and that through this fact we become conscious (though not intuitively conscious) of freedom. But we have also been told repeatedly that we do not have indubitable evidence that the concept of duty is valid and that moral ideas are not vain and chimerical.9 Consequently the indubitable fact that we are conscious of the moral law does not mean that the moral law is itself indubitably valid. Second, there are statements that what we are conscious of is the fact. We are told, in the other passages, that it is a fact of reason that there is a law, that this is a law of autonomy, that this law can determine the will. These are the facts which must be real if morality (at least Kant’s conception of it) is to be justified. This division of meanings of “ fact” is perfectly analogous to the division we formulated in the second section, “ T h e Intuitionistic Answer to the External Question,” between putative and real intui tions. Consciousness of the law is a putative intuition; and if this consciousness is indeed cognitively veridical, then the law is the object of a real intuition. Hence Kant’s theory of the “ fact of pure reason” seems to authorize us to classify him as an intuitionist with respect to the external ques tion; and if my argument in the second section of this essay is correct, we shall have to declare his justification of morality to be inadequate. T H E F A C T OF REASON AN D T H E F A C T FO R REASON
In spite of Kant's thinking of the moral law as something directly given to consciousness in the unique experience of respect, I do not think that it is correct to regard him as an intuitionist with respect to the problem of justification. An intuitionist is a kind of Platonist; for him, there is a given Idea or essence, set over against a receptive consciousness which accepts it without mediation or error. But for Kant, the metaphysics of the moral law is not Platonic; the moral law is a creation of reason, and he suggests that it is “ merely the self-consciousness of a pure practical reason and thus identical with the positive concept of freedom.” 10 8 Critique of Practical Reason, Academ y edn., V, 4 n. 9 Foundations, Academy edn., IV, 402. 10 Critique of Practical Reason, Academy edn., V, 29; trans. Beck, p. 29.
S t u d i e s in t h e P h i l o s o p h y o f K a n t
211
Since I say Kant was not an intuitionist, it is incumbent upon me to find a non-intuitionistic interpretation of the expression “ fact of reason.” T h e “ of” in this expression may be a subjective or an expli cative genitive. Usually it has been interpreted as the former. Here tofore we have interpreted it in the former way to mean a fact known by pure reason as its object, m odo dirccto (reason’s fact). This is what the fact would be for a genuine intuitionist. But if it is a fact of this kind for Kant, it differs from all other facts (which, for him, must be given in a spatio-temporal intuition), and he would indeed be justified in referring to it as “ a fact, as it were.” T h e “ of” in “ fact of reason,” however, may be an explicative genitive, and will then indicate the fact that there is pure reason , a fact known by reason reflexively and not intuitively or modo directo. I shall call this “ the fact of pure reason,” while the intuitionistic interpretation of Kant would make us refer only to “ the fact for pure reason.” If Kant’s expressions do mean that he was talking of what I have called the reason and not the fact for reason, then he is im mune to the attack I made upon the intuitionists in the second section. And the fact of reason, as I interpret it, is prima facie inde pendent of the specific moral construction put upon it by Kant. T hat is, I am not arguing, for instance, that the fact of pure reason shows that we should not tell a lie. Rather, I am insisting only that there is, in willing, a principle that is rational, whether the willing be moral or immoral, prudent or foolish.11 If volition appears to be inde pendent of desires and to be morally unconditional, then there must be an unconditionally rational principle for it. T h e phenomenon of moral obligation implies the fact of reason— the fact that pure reason can be practical— even though in any particular case the judgment of obligation may be wrong. I am not therefore concerned here with the deductive elaboration of Kant’s inferences, but with their premises, not with the so to speak metaphysical exposition of the moral principle, but with its transcendental ground. 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. T h e moral law, which expresses nothing but the autonomy of reason,12 11 See my Commentary on K an t’s Critique of Practical Reason, pp. 84, 109.
12 Critique of Practical Reason, Academy edn., V, 33.
212
S tu d ie s in t h e P h i l o s o p h y o f K a n t «
can be a fact for pure reason only because it is the expression of the fact of pure reason, i.e., of the fact that pure reason can be practical. T hat is why_the moral law is the sole fact for pure reason; for it expresses only the sole fact of reason. T h e fact for reason, which we know m odo directo, is a creation of the fact of reason as legislative. We discover the law not as a given ordinance of God or in an essen tial structure of a realm of values, for then it would have to be given intuitively. No, we discover the law within volition itself; “ If pure reason is actually practical, it will show its reality and that of its concepts in action.” 13 It is not a fact (fact for reason) which we first discover by a kind of moral intuition and to which we then attempt to conform in our willing. If this seems to be a slippery argument, let us turn to the moral phenomenon itself. A moral principle is not binding upon a person who is ignorant of it or of the conditions necessary for rational choice according to it. On the other hand, if a person believes that the rational principle is valid for him, then it is in so far forth valid for him. This is true whether the specific principle he follows (e.g., the imperative to do a) expresses a claim that is materially valid or not. For only a being aware of the claim of some principle on his choice between a and non-a could even make a mistake about what he ought to do; if the principle is not acknowledged, there is nothing that he ought (formally) to do. Without a principle there is no possibility of doing right, but also no possibility of doing wrong. T o argue against the relevance of some principle is to appeal to norma tive grounds against the application of normative principles. T h e rejection of a moral principle is itself a moral act, which can beeperformed only by a rational being. Hence Kant says: “ Every being that cannot act otherwise than under the Idea of freedom is thereby really free in a practical respect.” 14 Now “ acting under the Idea of freedom” is the same as acting in cognizance of the moral law; hence a person who is conscious of moral constraint (i.e., has a putative intuition of his duty) shows that his reason is practical (even if it is ineffective), i.e., shows the fact of pure reason. If Kant is correct, a pure practical reason can produce only one fact for reason, to wit, the moral law. But I claim only this: even though 13 Ibid., p. 3; trans. Beck, p. 3.
14 Foundations, Academy edn., IV, 448 n.; trans. Beck, p. 66 n.
Studies in the P h i l o s o p h y of K a n t
21 3
every specific moral claim that a person acknowledges might be actually invalid (i.e., he might always think that he ought to approve of a when, in fact he ought to approve of non-a), the adjudication of the conflict between a and non-a evidences the fact of reason. In this unhappy case, only the derivation of the fact for reason from the fact of reason would be erroneous. But deliberation and responsibil ity for the choice show that the person will not, in this instance, be able to justify settling the issue between a and non-a by tossing a coin, but only by appealing to the standard of reason. T o recapitulate: moral consciousness or consciousness of duty is an undisputed fact. Prima facie, it does not justify the assertion that duty is a valid concept, for prima facie it seems to be an intuition of value, modo directo, which may be wrong. But the moral law is not only the object of a putative intuition; it is not given to us in the first instance as a fact for consciousness, modo directo. T h e moral law expresses nothing but the law-giving of reason itself, and this legis lation is implicit in every rational choice, even though it might happen that every specific law or rule given by an imperfect human reason should in fact be wrong. We might conceivably be wrong in every decision we make, but the decision, if made under the guidance of reason, is a fact or act of reason; and no other act is. (And if Kant is correct, this means that the decision is morally right, for the only fact for reason is the moral law.) This validity may be called external rather than internal. Internal validity means the rightness of the choice between a and non-a ; ex ternal validity means the rightness of choosing between a and non-a by some rationally chosen standard. T o deny this is to make a state ment that is pejoratively and not merely participially unjustified. In this paper, I do not wish to argue, as Kant did, that Z which is the presupposition of (