Mind Association

Kant's Antithesis of Dogmatism and Criticism

Author(s): Arthur O. Lovejoy
Source: Mind, New Series, Vol. 15, No. 58 (Apr., 1906), pp. 191-214
Published by: Oxford University Press on behalf of the Mind Association
Stable URL: http://www.jstor.org/stable/2248557
Accessed: 20-09-2016 15:32 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

Mind Association, Oxford University Press are collaborating with JSTOR to digitize, preserve and
extend access to Mind

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
IV.-KANT'S ANTITHESIS OF DOGMATISM AND
CRITICISM.

BY ARTHUR 0. LOVEJOY.

KANT'S distinction between two sharply contrasted types of

philosophical method, designated by the names of 'dog-
matism' and ' criticism' has come to be one of the accepted
rubrics in the current treatment of the history of philosophy.
It is commonly supposed to correspond to actual historical
differences that are both definite and important. For this
celebrated antithesis not only sums up in a single phrase
what Kant conceived to be the most important and dis-
tinctive feature of his own doctrine; it also, as Kant sets it
forth, contains several plain implications as to matters of
historical fact which concern much more than his own doc-
trine: implications, namely, as to the real character of the
philosophical procedure of his predecessors, especially of
Leibniz and Wolff; as to the mneasure of his own divergence
from them, in his conception of the nature and scope of the
ultimate criteria of truth; and as to the degree of essential
novelty and originality that can be claimed for the Kantian
system. In particular, the customary assumption of the
validity of the antithesis -has brought about that many
more or less instructed persons carry about with them, as
their one firmly fixed philosophical idea, the persuasion that
between the method of philosophy in vogue before Kant's
day, and that in vogue since, there is a great gulf fixed. I
propose here to show that the antithesis in question is
seriously misleading, and to point out certain misrepresenta-
tions of historic facts, certain exaggerations and certain am-
biguities in Kant's way of presenting it, and certain errors in
those ideas about the criteria of truth employed in meta-
physical reasoning, upon which the antithesis is based.
Behind this simple and popular and apparently convincing-
scheme for classifying historic tendencies lies a very con-
siderable, though not uncharacteristic, confusion in Kant's.
thinking. He was able to make so sharp the antithesis be-
tween his own and earlier systems only because he had been

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 192 ARTHUR 0. LOVEJOY:

guilty of three rather singular oversights. In the first place,

he had failed to hold clearly in mind what the doctrine of
philosophical method held by his immediate predecessors
had actually been. In the second place, he had failed to
consider duly and define clearly the relation of his own doc-
trine to a familiar principle which has been the main and
the generally accepted instrument of mnetaphysical reasoning,
as well before his day as after it. And in the third place, he
had forgotten-and most of his expositors have ever since
overlooked-the fact that, of the special arguments in which
he worked out his own 'critical ' method, the most linportant
and ostensibly the most 'critical '-namely, the argument
about causality embodied in the Second Analogy of Experience
-was merely an elaboration of an argument already employed
by Kant's 'dogmatic' predecessor, Wolff, from whom he
seems unconsciously to have borrowed it.' These considera-
tions will, I hope, throw an appreciable, and, if that be con-
ceivable, a somewhat new illumination upon the limitations
of Kant's thought and upon its historical connexions. Now
that the world, and especially the Germanic part of it, has
ceased celebrating-not without a certain measure of un-
philosophical Schwdrrnerei - the centenary of the worthy
Koenigsberger's death, nothing, I take it, could be more
fittilng than that students of the history of thought should
undertake a new inquest into the real originality, and the real
value, of this central idea of the thinker who is credited,
above all, with giving to philosophy a new and definitive
method. The outcome of the inquiry will, I think, tend
somewhat to qualify the conventional high estimate of Kant's
importance, so far as the methodological question is con-
cerned. Prof. William James has, in a public address, laid
hands on our father Parmenides with some boldness: "I
believe," he says, " that Kant bequeaths to us not one
single conception which is both indispensable to philosophy
and which philosophy either did not possess before him, or
was not destined inevitably to acquire after him, through
the growth of men's reflexion upon the hypotheses by which
science interprets nature. The true line of philosophic pro-
gress lies, in short, not so much through Kant as round him,
to the point where we now stand. Philosophy can perfectly
well outflank him, and build herself up into an adequate
fulness by prolonging more directly the older English linies."
This is rather more than I care here to maintain; it is not

I This third point cannot be dealt with within the limits of this paper.
The proof of it is offered in a forthcomrling article by the present writer
in the Archiv fiur Geschichte der Philosophie.

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 KANT'S ANTITHESIS OF DOGMATISM AND CRITICISM. 193

so much from Kant's positive contentions as from his ne-

gations and his unfair reflexions upon his precursors, that
this paper will offer reasons for disseniting. Yet it is true, I
think, that our present somewhat minute inquiry will make
it evident that 'the tendency of Kant's influence has been to
obscure or disguise the proper and the historic method of
metaphysical reasoning, and to lead to the assumption of a
breach of continuity in the secular working out of philo-
sophical problems, where none really exists.
What Kant means by his antithesis is, of course, sufficiently
clear. The distinction between doginatism and criticism is
represented as radical; and it has a perfectly definite technical
character. A dogmatist, in the special sense, is a philosopher
who deliberately goes about making synthetic judgments a
priori, without first pausing to ask himself whether, or how,
such judgmnents are logically possible. That is the definition
of the creature; and of the motive which leads him to be
such as he is, Kant gives a plain account in various passages
of the Tractscendental Dialectic and of the Prolegomena. The
weakness in human nature which causes men to construct
dogmatic systems of philosophy is the passion for complete-
ness in the conception of things and in the explanation of
their conditions, the disposition "to find for every condi-
tiorned an unconditioned and so to complete the unity of
knowledge ". The speculative metaphysician, in short, is the
man who suffers from an uncontrolled craving for the abso-
lute and ultimate. The reader of Kant cannot fail to derive
the impression that the philosophical ambitions of earlier
thinkers, and especially of the school in which Kant himself
had been brought up, had been characterised by a general
ignoring of limits, by a naive assumption of the adequacy
of the powers of human reason, and by a careless neglect to
define the criteria of truth with either proper narrowness
or proper explicitness. Kant's whole picture of the motives
and the achievement of all constructive pre-Kantian meta-
physicians is the picture of the motives and the achievement
of an Icarus.
Now this picture is entirely in keeping with the view about
metaphysics commonly held by the mnan in the street, and
even by many natural scientists and some philosophers; but
to any one who will take the trouble to understand the inner
springs of the historical movement of metaphysical reflexion,
it must seem an unintelligent caricature. The primary and
exigent philosophic passion, as the history of philosophy
displays it, is not the passion for completeness in the con-
ception of the world of experience and of the conditions of

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 194 ARTHUR 0. LOVEJOY:

it, but the passion for consistency and coherency in that

conception. The greater philosophers have been distin-
guished, not by a more insatiable lu1st for the synthesis of
abstruse ideas, but by a superior talent for the analysis of
the common ideas used in all thinking. Philosophy begins,
no doubt, in the case of a Thales or any similar pioneer, with
a demand for completeness and ultimateness of explanation
of given facts; but-if I may repeat what I have already
remarked elsewhere-philosophy once begun is kept going
-and endlessly changing by the continual fresh discovery of
latent inconsistencies in the accepted interpretations of ex-
perience. Every man is more or less disturbed at any
intimations of the presence of such inconsistencies in his
body of apparently settled opinions; and the philosopher, as
he is usually exemplified in history, is merely the inan who
is peculiarly alert to this situation, and peculiarly-and,
perhaps, unnecessarily - uncomfortable at finding himself
in it. Often enough, no doubt, it has been some religious
,or practical need that has set him to looking for the contra-
,dictions in the current opinions; but in so far as his revision
of those opinions has been philosophical, the motive that
shaped it has been the demand for consistency in the ex-
plication of the conceptual necessities implicit in the facts
of experience. We are not called upon here, however, to
review the whole history of metaphysics; and it is therefore
not possible to point out in detail how, at nearly all the
great steps of philosophical transition, the method, both of
metaphysical criticism and of metaphysical construction, has
consisted in some special application of the two logical
maxims involved in the principle of contradiction, in its
broader meaning: that which, after the completest analysis of
the ideas involved, implies the coinherence, in a single subject, of
concepts that the mind is incapable of combining in thought, can-
not be real, and that proposition of which the opposite is, in the
sense just specified, inconceivable or self-contradictory, must be
true.
It suffices for our present purpose to recall the manifest
fact that Kant's more immediate 'dogmatic' predecessors,
Leibniz, Wolff and Baumgarten, so far from constructing
their metaphysical edifices without first examining the founda-
tions, had been entirely explicit in naming the criterion of a
yriori knowledge, through the use of which metaphysics was
to be possible; 1 and had limited that criterion to the single
I In his criticism of Kant Eberhard declared that " die Leibnitzische
E'hilosophie ebensowohl eine Vernunftkritik enthiilte, wie die Kantische;
Adenn sie grunde ihren Dogmatismus auf eine genaue Zergliederung der

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 KANT S ANTITHESIS OF DOGMATISM AND CRITICISM. 195

principle of contradiction-or, to give it a name that better

covers their full meaning, the principle of the compossibility of
concepts. There had, moreover, during the period just before
the appearance of the critical philosophy, beenl taking place a
progressive clearing up of the significance and scope of this
principle, and an especially determined effort to elucidate the
whole question of the nature of a priori judgments. Since
this was the case it will be of use to inquire, first, just what
Leibniz and Wolff meant by the principle of contradiction,
especially in its relation to the question of "the possibility
of synthetic judgments a priori"; and, second, to ascertain
precisely what Kant's own views were about the validity and
the reach of that principle. We shall find the results of such
an inquiry somewhat surprising.
The principle of sufficient reason is still sometimes repre-
sented as holding in the Leibnitian system an independent
place, virtually co-ordinate with the principle of contradiction.
But in reality Leibniz fully recognised that this principle can
apodictically establish no conclusions a priori; and Wolff
either treated it as a special case of the principle of contra-
diction, or else attempted to justify it by a peculiar mode of
argument, of which Kant would have been obliged to recog-
nise the legitimacy on strictly 'critical ' grounds. With the
exception, then, of this last-mentioned argument, which can-
not be dealt with here, Leibniz and Wolff rested the whole
possibility of demonstrative proof or disproof in metaphysics
upon the principle of contradiction, and professed to have no
metempirical theoretical knowledge save such as could be
gained through the use of that principle. Wolff takes pains
to point out that even the Cartesian cogito ergo sum is nothing
more than a special case of the sort of logical necessity which
the principle establishes.
What, then, did these philosophers understand by the
Grund des Widerspruches ?
We must remember, to begin with, that Leibniz-a fact
of which both Mr. Bertrand Russell and M. Couturat have
made much, in their recent expositions of his doctrine-
liked to call all judgments analytical, contingent ones no less
than necessary; this because of the familiar truth of formal
logic that the connotation of the predicate of a proposition is
always embraced within the connotation of the subject. As
M. Couturat points out, the whole metaphysics of Leibniz

Erkenntnissvermdi6gen ". Upon this Kant remarked that, if such was the
case then, indeed, "there is no dogmatism in that philosophy, in the
sense in which our Kritik always employs the word " (Reply to Eberhard,
ite Abscbn).

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
196 ARTHUR 0. LOVEJOY:

follows from the logical principle: " Dans toute verite, uni--
verselle ou singuli'ere, necessaire ou contingente, le predicat
est contenu dans le sujet". But while Leibniz thus makes
all judgments analytical in a sense, it is also true that he fully
recognises a distinction corresponding to Kant's distinction
between analytic and synthetic judgments. Some predicates,
are contained in their subjects essentially and inextricably, so,
that, lacking those predicates, the subject would cease to be
itself, would become an inconceivability; in other words, the
concept, as a whole, is made up of a complex of attributes,
some one or more of which are inconceivable without the-
others. In other cases the inclusion of the given predicate
with the several other predicates that make up the essence-
(or definition) of the subject is purely accidental; the several
attributes do not coinhere of necessity. In the latter cases,
the connexion of predicate with subject can only be known a
posteriori; the fornmer class constitute the field where a priori
reasoning is legitimate, and where necessary and eternal
truths are to be looked for. It is, of course, true that Leib--
niz called propositions of this class identical propositions,.
their distinguishing mark being that their opposites involved
self-contradiction. But by identity Leibniz did not really in--
tend to mean mere tautology, and by contradiction he does
not always signify mnerely verbal contradiction. It was, in-
deed, difficult for hin to make out how he could mean any-
thing else; but he was, none the less, firmly persuaded that
the Gruind des Widerspruches is no empty and sterile maxim,
but the fruitful source of important insights; and he especially
(if not always very successfully) exerted hinmself to prove its.
positive utility. Thus he remarks (Gerhardt vii., p. 299)
that though there may seem to be nothing but a cocoysmus-
inutilis in identical propositions, yet levi mutattione utilia inde
axiomnata nascuntur. This, to be sure, is a somewhat na&f way-
of putting it, and the examples which follow are hardly con--
vinicing; but it all shows Leibniz's unwillinciness to take his,
principle in its strict and narrow sense. There are, more--
over, he maintains, two distinct kinds of judgment included
within its range of application: (1) identical judgments, of'
which the opposite is formally self-contradictory, e.g., A is A;-
(2) judgments virtualiter identicae, of which the opposite can
be seen to involve contradiction only per terminorum intellec--
turm et resolutionem-that is, only by an examination of the
whole implicit connotation of the terms involved, showing-
that the two notions are 'incompossible'. It is chiefly these
latter, or " virtually identical," judgments which constitute
the substance of our demonstrative knowledge, and especially-

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 KANT S ANTITHESIS OF DOGMATISM AND CRITICISM. 197

of mnetaphysics. Demonstration consists in just the process

of conceptual analysis whereby the implied content or mean-
ing of two concepts is thus brought out with suLch definite-
ness as to make clear to the mind the impossibility of thinking
them together, and therefore the necessity of thinking each
of them and the opposite of the other together. "Virtually
identical" judgments, to be sure, always bring us back ad
identicas formales sive expressas, that is, to propositions in-
volving two "simple" concepts, the necessary coinherence
of which, in the given proposition, is mnanifest because the
opposite of that proposition involves an explicit and literal
contradiction: "Manifestum est omnes propositiones neces-
sarias . . . ad primas veritates revocari possunt, ita ut ap-
pareat oppositum implicare contradictionein et cum identica
aliqua sive prima veritate pugnare" (Gerhardt vii., p. 300).
There is still, it is true, in all this a singular inconsistency,
or failure in clear thinking, on the part of Leibniz, which
Mr. Bertrand Russell has already noted (Philosophy of Leibniz,
p. 18). If the prime veritates are literally " identical " propo-
sitions-A is A, B is not non-B-and if other necessary pro-
positions are demonstrable only in so far as they are reducible
(by the process of definition of the concepts involved) to such
_prime veritates, then nothing but identical propositions can
be really demonstrable at all-the principle of contradiction
can never help us to prove anything more thain that a thing
is itself and is not anything not-itself. And again, to note
another aspect of Leibniz's confusion-if all demnonstration
consists in definition, and if definition is the analysis of
complex concepts into truly simple ones-then, once more,
between these simple concepts there can be no relations
except those of identity and (mere) difference, and therefore
no possibility of demonstrating any synthetic truths. For it
is not clear how a genuinely simple concept can contain
within itself the necessity for the coinherence with itself of
other and distinct concepts, or even aniy special or preferential
repugnancies towards any particular other concepts. If the
fundamental concepts were really simple in their connota-
tive content, they would be a sort of logical atomls, or window-
less monads, capable, perhaps, of entering casually and con-
tingently into any kind of intellectual combination, but not
necessitated by their own nature to enter into (or to refuse)
any. And so we. could discover no really instructive logical
relations between distinct ideas, could find in a given subject
no necessary coinherence with any predicate except its own.
simple and unanalysable self. And so, once more, no truly
synthetic judgments a priori would be possible anywhere,
14

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
198 ARTHUR 0. LOVEJOY:

even in mathematics; and metaphysics, possessing no mneans

of demonstration except the Principle of Contradiction, would
indeed be condemned to perpetual sterility.
We must, then, recognise this blunder in Leibniz's treat-
ment of the Principle of Contradiction. By describing the
judgm-lents based upon that principle as " identical "-by re-
garding the relation between subject and predicate in such
judgrments as purely analytical, and then further treating
both concepts as ultimately "simple" notions-he undeni-
ably destroys by implication the possibility of constructive
metaphysics; and by his confused thinking upon the point
he is, unquestionably, largely responsible for Kant's aberra-
tions in the matter of the distinction between synthetical
and analytical judgments. But on the other hand, one
must repeat that Leibniz intended no such result. He fully
meant his Principle of Contradiction to be a positive and
constructive principle; and if he habitually employs the sort
of language that I have quoted, implying that there are no
necessary relations between any two distinct simple concepts
except that of bare non-identity, it is also true that there is
an essential and frequently reiterated point in the Leibnitian
system which implies exactly the contrary. This is to be seen
in Leibniz's doctrine of definition. Definition consists in form-
ing a complex idea by the conjunction, in a single meaning,
of several simple ideas-or of less complex ideas which are
ultimately resolvable into simrple ones. The contrary process,
as we have already seen, is analysis-the taking apart of the
complex which definition puts together, and so the eventual
discovery of the simple conceptual elements of which it is
composed. Leibniz, now, on the one hand, always insists
that the analysis of a definition must bring us to such simple
and indefinable concepts; but on the other hand, he con-
stantly insists that definition is not an arbitrary process but
is always (when legitimately performed) limited from the
outset by the requirement that the notions united shall be
"compossible," compatible with one another.' But these
two contentions taken together are equivalent to the asser-
tion that there may subsist, even between ultimately simple
and indefinable concepts, relations of incompatibility. For
if there were no such relations of ultimate incomnpatibility,
definitions would be arbitrary, and any conjunction o-t positive

1 Gerhardt iii., 443: " Les definitions ne sont point arbitraires, et on

ne peut point former les idees comme l'on veut. Car il faut que ces
idees qu'on pretend former soyent veritables, c'est-A-dire, possibles, et
que les ingrediens qu'on y met soyent compatibles entre eux." So fre-
quently elsewhere.

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 KANT'S ANTITHESIS OF DOGMATISM AND CRITICISM. 199

simple concepts into a single notion would be possible and

legitimate. In such a case, though you would not be en-
titled to say that a triangle is a non-triangle, there would be
nothing to prevent your saying that a triangle is a parallelo-
gram. For if some one urged that a triangle is definable as
a three-sided figure while a parallelogram has four sides,
and that three-sidedness and four-sidedness are incompatible,
one would have to ask: How, unless "simple" concepts
have relations of fixed incompatibility inter se, can I know
that what has three sides cannot have four sides ? It could
not be merely because the two concepts are two, i.e., merely
distinct; for three-sidedness and whiteness are also distinct,
yet no objection arises when I define a figure as a white
triangle. If there are incompatibilities anywhere in defini-
tions, they must inhere in the original elements, the primary
concepts, of which the definition is inade up; and if there
were, contrariwise, no such original incompatibilities-self-
evident, not capable of being demonstrated by further de-
finition, and not to be confused with mere non-identity-
everything under heaven in the way of a definition would'be
permissible. In asserting, then, that definitions are not
arbitrary, Leibniz plainly points to the affirmation of ultimate
repugnancies to coinherence between distinct and positive
concepts; and thus to the affirmation that synthetic relations
of incompossibility, on the one hand, of necessary coexistence,
on the other-are to be found, in some cases, between the
ultimate and irreducible ideas upon which our thinking finally
depends.'
As regards ILeibniz, then, we must sum up the case thus:
He sought to give definiteness and self-evidence to the method
to be used in philosophical construction, by reducing the
grounds of all a priori reasoning to a single principle. But
in his treatment of this principle he fell into a twofold and
inconsistent doctrine. On the one hand, in the pursuiit of
extreme rigour and simplicity in the definition of his ultimate
criterion of truth, he tended to reduce it to the useless and
trivial principle of the mere self-identity of concepts. But
on the other hand, in another part of his imperfectly con-
catenated system, he plainly implied that there exist syn-
thetic relations of compatibility and incompatibility between
several distinct concepts, and that by the elucidation of
these relations, important and significant truths a priori may
be discovered. And in the form of this open choice between

1 This point has already been clearly set forth by Mr. Russell (Philo-
sophy of Leibiz, ,loc. cit.).

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
200 ARTHUR 0. LOVEJOY:

unharmonisable alternatives Leibniz left the problem of

inetaphysical method to his successors.
What has rarely been remarked is that those successors-
the inmediate predecessors of Kant-took up the problemu
and gave a reasonably clear answer to it; that to the funda-
mental Leibnitian principle they gave an enlarged but ani
explicitly defined meaning, corresponding to the second and
sounder of the alternatives suggested by Leibniz; that be-
tween this enlarged meaning and the narrower one from
which Leibniz had not fully distinguished it, they drew an
unequivocal distinction, corresponding to Kant's distinction
between synthetic and analytic judgments; that they thus
answered in advance Kant's question: " How are synthetic
judgments a priori possible? "-and answered it in a sense
to which Kant could not consistently have made objection.
It was, of course, Wolff who accomplished this advance
upon the position of Leibniz; - Baumgarten, and Kant's
contemporary and critic, Eberhard, did little more than re-
arrange or expand Wolff's ideas. What Wolff did was, in
the first place, distinctly to abandon the theory that the only
thing that you can logically discover a priori about any con-
cept is that it must always mean whatever it happens to
mean. On the conitrary, he remarks in an instructive pas-
sage of the Horae Stbsecivae (1730, vol. i, p. 154), we ought
to recognise that there are such things as notiones fcecundc,
" pregnant concepts "-of which the peculiarity is that they
contain determinationes rei, per quas cetera quce rei conveniunt,
colligi possunt; in other words, thev are such that ex iis, que
in iisdem continentur, certa ratiocinandi lege colligunttu alia qjtce
in iisdem non continentur. It is because it is a fact of psycho-
logy that there are. such pregnant notions, that we are able
to frame a numnber of propositions quarum prcedicatis positis,
ponuntur alia ejlusdem rei prcedicata. Here, then, we find
Wolff asserting that some ideas contain within theinselves-
inseparable from, yet not expressed in, their formually defined
essence-n-ecessary implications as to their relations of co-
predicability or incompatibility with other ideas; and that
meanings are thus organically interconnected.
Now in the technical logical and metaphysical treatises
of Wolff and his school, the theory here implied as to the
grounds of the legitimacy of synthetic judgments a priori is
set forth with great fulness and definiteness-and copious
iteration. In summarising the doctrine I shall quote in part
from the Acroasis Logica of Baumgarten, which, while follow-
ing Wolff's expressions without significant deviation, is some-
times more concise and compact. Logic, says Baumgarteni,

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 KANT'S ANTITHESIS OF DOGMATISM AND CRITICISM. 201

taken in the strictest sense, is precisely the science which

shows how knowledge a priori is to be had (scientta cognoscendi
a priori). Now " whatever is known a priori is known ex
internis veritatis characteribus"; by this is meant, in modern
language, that the presence of a priori knowledge is to be
veriied essentially by a psychological fact, namely, the actual
inability of the mind to divorce certain predicates from one
another. There are two kinds of universal judgment, and
only two, which are characterised by this inward compulsive-
ness or necessity. Whereas in all other judgments predication
is justified only a posteriori, since the subject is, as such, found
to be thinkable as a single coherent notion without the im-
plication of the given predicate-in these two cases predication
is necessary, and therefore valid universally and a prtori, for
the reason that the denial of the predicate carries with it for
the mind the disappearance of the subject as a consistently
definable and thinkable notion.' These two sorts of judgment
a priori are distinguished from one another with respect to
the manner in which the inseparability of predicate from
subject is grounded. The first class is that of identical pro-
positions, in which the predicate is merely the whole (essentia)
or a part (essentiale) of the attributes included in the defini-
tion of the subject. Such propositions are, evidently enough,
axiomatic, that is to say, necessary yet indemonstrable truths;
but they are purely tautological and add nothing to our know-
ledge. The second class of axioms is the one which is made
possible by the existence of notiones ftecunde, and yields us,
therefore, pregnant and instructive truths. Its logical char-
acter consists in this, that the predicate in such proposi-
tions is a " property " (attributum, Eigenschaft) of the subject.
Whereas in the merely identical proposition, or definition,
the several predicates are merely put together by the mind
to make the chosen meaning, and do not co-determine the
presence of one another, two " properties," on the other
hand, stick together as inseparable "joint-determinations" of
any subject into the definition of which either of them is
introduced-even though the framer of the definition may
have meant to admit only one of them into his proposed
meaning.2 The test of a " property," once more, lies in the
1 " NotUe entis sunt vel necessariae et immutabiles, quibus sublatis tolle-
retur ens; vel contingentes et mutabiles, quae possunt adesse et abesse
salvo ente. Ad priores essentia, essentialia et attributa pertinent." From
Acroasis Loqica in Toellner's edition (1773), ? 171. The preceding citations
are from other sections of the same.
2 Wolff, Phlilosoph1ia rationalis sive Logica, 1728, ?? 64, 65: " Ea qure
constanter enti cuidam insunt, quorum tamen unum per alterum non
deterinilatur essentialia appello. . . . Ea qine constanter insunt, sed per
essentialia simul determinantur, attributa dico."

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
202 ARTHUR 0. LOVEJOY:

subjective necessity of thinking the predicates together; no

other test can be either had or imagined.'
Wolff, then, on the basis of this distinction between judg-
ments a priori per essentialia and judgments per attributta, adds
to the enumeration of the axioms (underlying mathematics
and metaphysics) that are purely identical, another class, not
clearly recognised by LeibDiz of a 'pregnant ' or synthetical
character. "Si definitum sumitur ut subjectum, et de eo
praTdicatur quidpiam, quod notis ad definitionem spectantibus,
in ejus notione animo praesente, indivulso nexu cohaeret: pro-
positio axioma est " (Philosophia rationalis sive logica, ? 273).
In view of all this we may well ask, as Eberhard asked,
to Kant's great irritation, shortly after the promulgation of
the critical philosophy: How can one say that the so-called
'dogmatists' neglected to undertake a preliminary Vernunft-
krittk, or neglected to distinguish between analytic and syn-
thetic judgments a priori, or neglected to show the logical
justification of judgments of the latter sort? And if one
accepts the principle upon which this last-nentioned justifi-
cation rests, how is one entitled to reproach those who pro-
fessed to found their metaphysics upon no other principle,
for uncritical dogmatists ?
These questions bring us to the essential issue upon the
settlement of which must depend one's estimate of Kant's
genuine originality and historic importance as an epistemol-
ogist. Metaphysical methodology had culminated just before
his time in an explicit limitation of all a priori philosophising
within the range where the testing of propositions by the
principle of contradiction,2 in an enlarged sense, is possible,
and in an equally explicit indication of just what that enlarged
sense was to mean. It is, therefore, impossible for any one
'Wolff puts it thus in his German logical treatise: " Wenn nun zwei
Gedanken so beschaffen sind, dass der andere notwendig statt findet, wenn
man den ersten heget, oder dass ich mir das andere notwendig gedenken
muss, wenn ich von einem Dinge das erste gedenke, indem namlich durch
das erstere das andere mit determiniret wird; so stimmen die Gedanken
mit einander uiberein. Kann ich aber das andere von einem Dinge
unm6glich gedenken, wenn ich mir das erste von ihm gedenke: so streiten
meine Gedanken wider einander, oder einer widerspricht dem andern "
(Verninftige Gedanken von den Krffften des menschlichen Verstandes, 6te Auil.,
1731).
2I refer both forms of the a priori judgment, in the Wolffian classifica-
tion, to the principle of contradiction, although the phraseology of
different members of the school appears to be inconsistent upon this
point. Eberhard, for example, prefers to label judgments a priori per
attributa as examples of the " principle of sufficient reason ". The ques-
tion, however, is mierely one of nomenclature; and Wolff himself cer-
tainly appears to me to intend the subsumption of both forms of a piori
reasoning under the principle of contradiction.

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 KANT S ANTITHESIS OF DOGMATISM AND CRITICISM. 203

to judge Kant intelligently, or to know where he stands his-

torically, who has not determined the precise attitude which
Kant took towards this Wolflian theory of knowledge. There
are three possible positions which Kant might have assumed,
in order to maintain, in some fashion, his charge of 'dog-
matism' against his forerunners; and it is essential to de-
termine which of the three he adopted-or whether he
wavered between two or more of therm. He might, in the
first place, have held that the principle of contradiction, even
in its narrowest Leibnitian sense, is not applicable to reality
as such, but only to the phenomena of experience; in this
case the contrast between the 'dogmatic' systems and the
Kantian criticism would undeniably be radical. Kant might,
again, have held that the principle is valid a priori for all
Being, noumenal as well as phenomenal, but only in so far
as it is purely analytical; that judgments based, not upon
formal tautology or formal contradiction, but upon relations
of coinherence or incompatibility between distinct and posi-
tive ideas are (as Wolff's language had suggested) synthetical;
and that as synthetical, such judgments can have no a priori
validity for either phenomena or noumena. In this case, also,
he would have presented a fairly definite and significant,
though somewhat paradoxical, antagonism to the method of
Wolff and his school. Or, finally, Kant might have held
that the principle is, in its fullest sense, valid a priori; that
a judgment based upon the discovery of implicit incompati-
bilities between the subject and predicate of the contrary
proposition, no less than one based upon explicit and verbal
inconsistencies between them, is legitimate, and applicable
to reality uAberhaupt; but that certain of the special conclu-
sions concerning God, the soul, etc., which his predecessors
had sought to justify by appealing to the principle, are not
really inferrible by means of it. In this last case Kant
could not properly have charged his predecessors with a
radical error of method; he could fairly have charged them
only with errors of formal logic, in the application of a valid
method to certain special problems. And he could not in
consistency have maintained that an a priori science of mneta-
physics is impossible; for he would virtually have admitted
that metaphysical conclusions are 'possible wherever the prin-
ciple of contradiction can be carried-wherever incompati-
bilities or necessary implications between ultimate concepts
can be made out. Such are the issues, in the interpretation
and criticism of Kant, which depend upon our answer to
the question: Which of these three positions did he actually
assume ?

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
204 ARTHUR 0. LOVEJOY:

Laas h'as already raised the inquiry whether Kant regarded

the principle of contradiction in general as a valid criterion
of the nature of reality an sich; and he has justly insisted that
the point is one of capital importance.' "Upon a philo-
sopher," Laas observes, " who places between our world of
timne and space and the world of reality as such a chasm
so wide and deep that for the latter the most conflicting
possibilities remain open, it is decidedly incumbent to give
a definite answer to such questions as these: Is the formula
(of the impossibility of the self-contradictory) valid 'only in
relation to a sensuous perception '? Is its extension to
things in themselves inadmissible? Does it apply only to
phenomenal existence? For who would venture to declare
that there are other forms of being besides those known to
us, and then deduce from the mere concept of Being the
conclusion that all Being must be free from self-contra-
diction ? " Important these questions certainly are; but
there can be no doubt about Kant's answer to them. His
declaratioln on the subject is precisely the declaration of
which Laas speaks. In both the earlier and the later
peri.od of his thinking, Kant adhere, consistently to the
view that an object in which self-contradictory predicates
are said to be united is not truly an object of thought at all,
but a mere nihil negativum irreprcesentabile. In his earliest
philosophical writing, the Dilucidatio Nova, Kant, in criticising
the customary WVolffian way of formulating the principle of
contradiction expresses only the more clearly his recognition
of the importance of that principle in metaphysics. Under
the name of a single formula, he maintains, philosophers
have really been making use of two distinct principles,
nainely, impossibile est idem esse ac non esse and oujuscumque op-
positum est falsum, illud est verum. These two are not, Kant
thought, reducible to one another; but they are equally
and independently valid. As arn instrument of philosoph-
ical proof the latter is the more important. In the essay
oii the Idea of Negative Magtitude Kant declares that the self-
contradictory is gar nichts; and in the paper on "The Only
Possible Ground for a Demonstration of the Existence of
God " he writes: " All that is self-contradictory is inherently
impossible.... This logical repugnancy I call the formal
character of inconceivability ( Undenklichkeit) or im:possibility."
Similar expressions are abundant in the writings of the so-
called critical period; e.g., fromn the Reply to Eberhard: " what-
ever does not ac-ree with this principle (of contradiction) is

1 Kant's Analogien der Erfahrung, p. 33.

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 KANT' S ANTITHESIS OF DOGMATISM AND CRITICISM. 205

obviously nothing at all; for it is not even an idea (Gedanke);"

and from the Progress of Metaphysics: "anything of which
the idea is unthinkable, i.e., anything of which the concept
conitradicts itself, is an impossibility ". The passage in the
Kr. d. r. V. (first ed., p. 150) should be familiar: " Whatever
the object of our knowledge may be, and whatever the re-
lation between our knowledge and its object, it must always
be subject to that universal, though merely negative, con-
dition of all our judgments, that they do not contradict
themselves. . . . The Principle of Contradiction is a general
though negative criterion of all truth. No cognition can
run counter to that principle without destroying itself."
Kant's general recognition of the a priori validity and im-
portance of the principle of contradiction as an absolute
criterion of the nature of reality is thuis unmistakable.
Did he, however. accept the principle in one of those modi-
fied and restricted senses which I have indicated as open
to him ?
The answer must be that Kant's attempts, in the writings
of the critical period, to tell what he takes the principle of
contradiction to mean, are characterised, not only by a
curious vagueness and confusion of ideas, but also by a per-
sistent ignoring of just that distinction between the two
kinds of judgment a priori which Wolff and Baumgarten
had made so essential. One point, indeed, is clear: Kant
always speaks of judgments resulting from the principle of
contradiction as purely "analytical". This, however, does
not settle luch more than a question of nomenclature;
the usage of Wolff's own disciples was not well settled upon
this point. The important thing is not so much to deter-
mine whether the relation of predicate to subject in juidg-
ments per attributa can best be called analytical or synthetical
there are intelligible reasons for calling it either-but
whether such judgments are legitimate at all, and what their
logical nature is. And here Kant writes as if he had read
nothing in German philosophy since Leibniz-and only a
part of Leibniz.
For, in the largest class of passages on the subject, Kant
goes on as if nobody had ever suggested the possibility of
finding in judgments per impossibilitatem contrarii anything
more than the coccysmus inutilis which even Leibniz had
declared that they were not. In " analytical judgments,"
we are told, the connexion of predicate with subject durch
Identitdt gedacht wird; and hence " our knowledge is in no
way extended by them," their only service being " to put the
concepts that I already possess into better order, and to

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
206 ARTHUR 0. LOVEJOY:

make them more intelligible to myself " (Kr. d. r. V., first ed.,
pp. 7, 8). In the Prolegomena Kant goes so far as to say that
the notion which serves as the predicate of such a judgment-
is actiually always expressly present to the mind when the
subject is thought; the judgment " expresses nothing in the
predicate but what has already been wirklich gedacht (though
not ausdriicklich gesagt) in the subject ". Leibniz, in his
narrowest definitions, would have avoided such language; for
he at least placed judgments virttaliter identicae side by side
with the purely tautological ones. And occasionally Kant
himself reimembers that concepts sometimes implicitly in--
volve more than one happens, at any given moment, con-
sciously to think in them. He then adds that the predicate
of an analytical judgment may be contained in the subject
in a confused manner," or " without full consciousness '.
What this vague qualification precisely means Kant does not
further explain. It apparently refers, not to the implicit in-
terconnexion of " properties" in the connotation of an idea,
but merely to the fact that people sometimes forget just
what some of the essentialia are that they are accustomed to
signify by a term. The qualification cannot, therefore, be
construed as equivalent to the adinission of the distinctness
and the legitimacy of a priori judgments per attributa.
So far, then, and considering only Kant's principal writings,
we must say, not that Kant rejects the position of the ' dog-
matists,' but that he neglects to face or to oppose it at all.
With a degree of obtuseness rare in history, he entirely failed
to apprehend the distinction that had been the principal re-
sult of the previous half-century of refiexion upon the criteiia
of truth in mathematics and metaphysics-the distinction,
namely, between a priort judgments per essentialia and a
priori judgments per attributta-even though this distinction
contained an answer to just the question which he himself'
declared to be the fundamental one in all philosophy. His
attacks upon his predecessors implied that, having no cri-
terion of truth a priori save the principle of contradiction in
its narrowest analytical sense, they proceeded none the less
(in their ignorance of Hume) to construct an a priori system
of metaphysics. But since the criterion which they used
was quite other, and since their reasons for accepting it had.
been carefully explained, Kant cannot be said to have brought
any pertinent criticism to bear upon their position at all..
The longer treatises nowhere make it certain that, if Kant
had grasped the Wolffian distinction, he would not have
accepted the Wolffian method.
Fortunately, however, Kant was reminded of this dis-

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 KANT S ANTITHESIS OF DOGMATISM AND CRITICISM. 20T

tinction by one of his contemporary critics; and in his reply

to that criticism we find him at last brought to face the real
issue. Eberhard's Philosophisches Magazin, during the brief
period of its issue (1789-90), was devoted chiefly to proving
that Kant had not rendered Leibniz and Wolff obsolete, and
that he even did not depart so widely from their doctrines
as he appeared to suppose. It was in the execution of this
task that Eberhard pointed out those obvious facts to which
the present paper has again called attention. Long before
Kant's Kritik, Eberhard insisted, a distinction equivalent to
that between analytical and synthetical judgments had been
familiar in the Wolffian school. " Analytical judgments are
those whose predicates express the essence, or a part of the
essence, of the subject (das Wesen, oder einige von den wesent-
lichen Sti,cke, des Sublekts); those of which the predicates.
express qualities that do not belong to the essence are syn-
thetic." Now synthetic judgments are valid a priori when
(and only when) "their predicates are 'properties' (attributta)
of the subject, that is, determinations which do not belong
to the essence of the subject, yet have their sufficient ground
in that essence". When, now, this and much more of a
similar sort had been publicly pointed out, it was manifestly
incumbent upon Kant to define plainly his position with
respect to the central epistemological doctrine of the Wolffians,
which he had previously ignored.
Plain and unequivocal his answer is not; yet in the end
it amounts to an abandonment of his case against the 'dog-
matists'. Here was Kant's opportunity, if he really wished
to set his theory of a priori knowledge in sharp contrast with
that of the older school, to say, in so many words: 'I deny
that the principle of contradiction is susceptible of the ex-
tension which Wolff has given to it; I do not recognise the
existence of any such necessary relations of coinherence be-'
tween distinct concepts, going beyond the purely formal
necessity that a concept shall be identical with itself'. But
so far from doing anything of the sort, Kant readily accepts
Eberhard's (that is, Wolff's and Baumgarten's) distinction
between essentialia and attributa, and acknowledges that the
latter are no less necessarily, though they are less directly,
connected with their subjects, than are the former. In repeat-

I Ueber eine neue Entdeckmtng, nach der alle Kritik der reittee Vernmuift
entbehrlich. werden soll (1790). To any who wish to understand Kant's
relation to earlier logic and metaphysics, and so to determine his place
in the history of those sciences, this Reply to Eberhard is one of the most
important of his writings-for reasons that are made apparent in the
text. It is, however, full of flounderings and self-contradictions.

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
208 ARTHUR 0. LOVEJOY:

ing and making his own Eberhard's classification of the ways

in which predicates may be related to subjects a priori Kant
even proposes an einendation of the technical phraseology,
which has the effect of bringing out all the more clearly the
principle that, in attributa, we have predicates which do not
belong to the definition of a concept, and yet cannot be
negated of it without destroying its logical possibility. The
passage runs as follows: "A predicate which is ascribed to
any subject by an a priori judgment is thereby declared ne-
cessarily to belong to that subject, or to be inseparable from
it. Such predicates are said to belong to the essence, or
the inner possibility, of the concept (ad internam possibilitatem
pertinentia). Consequently all propositions that profess a
priori validity must have predicates of this character. All
other predicates (those, namely, which are separable from the
concept without destroying it) are called extra-essential marks
of the subject (ausserwesentliche Mlerkmale, extraessentialia).
Now predicates of the first sort may belong to the essence in
two ways, either as constitutive elements of it (BestandstAicke,
constitutiva), or as consequences of the essence that have in
it their sufficient ground (ut rationata). The former are
called essential elements (wcesentliche Sticke, essentialia), and, as
such, contain no predicate that could be dedutced from any other
predicate contained in the same concept; and in their totality
they make up the logical essence. The latter are called properties
(Eigenschaften, attributa)." Kant, then, here grants that
there may be predicates, technically to be named attributa
rationata, which form no part of the essentialia of a concept,
as it is defined, but yet are indispensable to the interna pos-
sibilitas of it. This is merely another way of saying-which
is the substance of the Wolffian theory-that not only tauto-
logical judgments but also judgments per attributa, are legiti-
mate sources of a priori knowledge. And the whole realm of
metaphysics, according to Wolff and (in his real intention)
according to Leibniz, lay within the limits of this last-
mentioned class of judgments, the possibility, validity and
potential fruitfulness of which Kant now adinits.1

I Paulsen has summarised this rather important part of the Reply to

Eberhard in a footnote in his volumrne on Kant in the Kltssik-r tier Philo-
.sophie (translation of Creighton and Lefevre, p. 143). Siince the book is
widely read, it is perhaps worth while to point out that here, as in
several other instances, Paulsen goes somewhat seriously astray in his
exposition. For he represents Kant as classifying predicables of the
sort called attributa or rationata, among the " extraessentialia, which can
be separated from the concept without affecting its nature ". But Kant
plainly and repeatedly classifies attributa as one of the two kinds of
predicates that are ad internacm possibil tatem pertinentia, inseparable front

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 KANT'S ANTITHESIS OF DOGMATISM AND CRITICISM. 209

Kant, h6wever, neither fully grasps nor long remembers

the significance and bearings of his own admissions; and he
attempts, furthermore, to save the originality of his doctrine
by insisting upon two additional points in regard to these
"judgments in which the predicates are inseparably connected
with the subject, though not a part of its logical essence
To these two points the whole distinction between the 'dog-
matic ' and the 'critical ' method now reduces itself.
(1) In the first place, Kant contends, Eberhard was wrong
in regarding all judgments per attributa as synthetical. They
may, he urges, be analytical; the distinction between ' ana-
lytic' and 'synthetic' runs, so to say, crosswise through
them. Eberhard had supposed that any a priori judgment
of which the predicate is an attrbibuttum-expressing, in Kant's
words, somiiething necessarily belonging to the subject, " yet
only as a rationatutm of the essence, not the essence itself nor
any part of it "-is ipso facto shown to be synthetical in
character. But in reality, Kant declares, " this shows no-
thina, more than that the predicate (mediately, it is true, yet
still as a consequence of the principle of contradiction) is
drawn out from (hergeleitet aus) the concept of the subject;
so that the proposition, notwithstanding the fact that it
expresses an attributum, mnay still be analytic, and therefore
lacks the distinguishing mark of a synthetical proposition".
In thus arguing that somne judgments of this sort are ana-
lytical, and are authorised by nothing less than the principle of
contradiction, Kanit commits himself to a still more thorough-
going adm-ission. He succeeds, indeed, in establishing, a point
of verbal opposition to Eberhard; but he at the same time
gives up all grounds of real opposition to the substance of
the Wolffian theory of a priorii knowledge. For it now turns
out that judgments in which, by virtue of some necessary
connexion between the concepts, predicates (as Kant else-
where puts it) are " developed out of " (entwickelt aus) the
subject miiay be analytical, and have the sanction of "the
highest principle of all analytical judgments ". But there is,
as we have seen, nothing that Kant asserts more emphat-
ically and more unequivocally thain that all analytical judg-
ments, involving the principle of contradiction, are valid a
the concept (vten Begr-iffe unabtrennlich), and capable of being deduced
from the essence of the concept as a necessary consequence thereof.
Paulseni's error, however, is doubtless chiefly due to a characteristic
awkwardness in Kant's own use of these technical terms. As Kant
employs it, the term essentialia is not, as one would expect, the antonym
to extra-essentialia, and the equivalent of ad essentiam pertinenti ; but is
mierely a species under the latter genus, the other species consisting of
attriblta.

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
2I0 ARTHUR 0. LOVEJOY:

priori of all possible reality, and in need of no further justifi-

cation or explanation. Kant, in short, for once abandons
,(though only to resume it on the next page) his usual language
about analytical propositions; and includes within their range
some judgments per attributa, which include distinctly more
,than need be wirklich gedacht in the original conception of
their subject. As for the mere question of phraseology, we
need not further discuss whether Kant or Eberhard is right.
Eberhard had sensibly remarked-what Kant could never be
brought to recognise-that " the controversy as to whether
a proposition is analytical or synthetical is a controversy of
no im-portance for the determination of the logical truth of
-the proposition ".' What is important is the determination
of the question whether or not there really subsists between
two given concepts, not defined as identical in essence, such
a relation that, if the first is negated of the second, the latter
becomes unthinkable within the termiis of its own essence.
When this is admitted-and Kant now admits that such a
relation is possible-the rest is merely a dispute as to the
sort of inetaphor which may most felicitously express this
logical fact.
(2) Thus far Kant has committed hinmself to the admissions
(a) that some propositions in which the predicates are at-
tributa rationata of the subject, are analytical; (b) that all
analytical propositions are true a priori of reality in general.
He further urges, however, that some propositions containing
attributa as predicates are synthetical; and he insists that
none of his precursors had shown-as he had done-how,
in this case, such propositions can be known to be valid.
Here Kant evidently has in mind the theorems of inatheina-
tics. These, he was sure, could only be called synthetical
propositions; and if synthetical, they required some justifica-
tion other than what Baumgarten would have called " the
internal marks of truth," their actual coinherence in the
mind. This justification Kant believed that he alone had
discovered, in his celebrated doctrine of the a priori percepts.
The reason why synthetical judgments are possible a priori is
that they are mediated through a reine Anschaunqny Here
is where Kant finally endeavours to come to a really sig-
nificant issue with his critic; and it is desirable to hear his
own words, which, in part, are not susceptible of adequate
translation: " Wenn gesagt wird, dass ich sie fiber die gege-
benen Begriffe hinaus, auch ohne Erfahrung, vermehren, d. i.
.a priori synthetisch urteilen konne, und man setzte hinzu,

I Philosophisches Magazin, 1789, p. 331.

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 KANT S ANTITHESIS OF DOGMATISM AND CRITICISM. 211

*dass hierzu notwendig etwas rnehr erfordert werde, als diese

Begriffe zu haben, es geh6re noch ein Grund dazu, um mehr,
als ich in jenen schon denke, mit Wahrheit hinzutun . . .
so will ich wissen, was denn das fuir Grund sei, der mich,
ausser dem, was meinem Begriffe wesentlich eigen ist und
was ich schon wusste, m-it mehrerem und zwar notwen-
dig als Attribut zu einem Dinge Gehorigen, aber doch
nicht im Begriffe desselben Enthaltenen bekannt macht."'
In so far as they are synthetic, then, we must look farther for
a " ground of the possibility" of these judgments containing
.cattributa as predicates. But, now, " any one may see that, in
what I have already set forth as the summary result of the
whole analytical part of the Kritik of the Understanding, I
have therein explained, with all needful explicitness, the
principle of synthetic judgments in general, namely: that
they are not possible except under the condition of a percep-
tion underlying their subject-one which, when they are a
posteriori judgments, is empirical, and when they are synthetic
judgments a priori is a pure a priori percept.... Compare
-now with this the pretended principle which is presented in
Eberhard's account of the nature of synthetic judgments a
priori: 'They are judgments which affirm of the subject one
of its attributa-that is to say, a predicate which necessarily
but only derivatively (nutr als Folge) belongs to the subject.'
. . . Now one is justified in asking whether the ground for
the affirmation of the predicate is to be sought for in the
subject, according to the principle of contradiction-in which
case the judgm-lent would still be merely analytical; or whether
the predicate is incapable of being deduced from the concept
of the subject by the principle of contradiction-in which case
the attributum is purely synthetic. Thus neither the name
4 attributum,' nor the principle of sufficient reason, serves to
-distinguish synthetical from analytical propositions; if the
former are to be affirmed a priori, one could say no more of
them (according to this phraseology) than that their predi-
cate is somehow or other grounded in the essence of the con-
,cept of the subject-and is therefore a 'property'-but not
simply in consequence of the principle of contradiction. But
how, as a synthetic property, this predicate comes to be bound
up with the concept of the subject-when it cannot be drawn
,out of the subject by analysis-cannot be explained by the
concept of a 'property,' so that Herr Eberhard's account
-of the matter is wholly barren. But the Kritik clearly sets
I Kant appears to have forgotten that, a few pages earlier, he had ex-
plained to us how predicates not contained in the essence may, by the
principle of contradiction, be ' developed out' of a concept analytically.

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
212 ARTHUR 0. LOVEJOY:

forth this ground of the possibility of such propositions, show--

ing that it can be nothing but the pure percept underlying
the concept of the subject. Only in such a pure percept is it-
possible to connect a synthetic predicate with a concept a
priori."
Thus the theory of the reine Anschautttng constitutes the last
trench into which Kant retires to defend his antithesis of
dogmatism and criticism. About this theory three things
may be said, of which only the two last are indispensable to
our present inquiry. First, the whole idea of a 'pure' per--
cept, which we are under no conditions to be allowed to
think as a concept, and which gets somne peculiar demon-
strative efficacy from its perceptual character-is a logical
chimera that belongs with the other hybrid monsters of
antiquity. But this is unimportant. Second, and what is
more to the point, the theory, though it were true, is, by-
Kant's own showing, not essential to the proof of the logical
validity of the a priori judgments in question; and it is
therefore irrelevant to the issue. Wolff and Baumgarten,
had, in effect, declared that, when you have analysed your
concepts to the uttermost and still find that two of themn,
not defined as possessing the same essence, coinhere indivulsa
nexu, you have reached the ultimate point of verificatiol;
beyond this de facto psychological necessity, this " internal
mark of truth," it is both needless and impossible to go.'
Now even if it be a fact that a reine Anschauukny 'underlies
our mathematical judgments, a knowledge of this fact is not
necessary in order to establish the truth of those propositions-
a priori (in so far as they are true a priori); nor does ignor-
ance of the fact subtract anything from the evidence of their
truth. The fact would be an interesting piece of psycho-
logical information; but the truth of the propositions, as
propositions a prtort, would still require to be tested by in-
quiring whether, when the meaning of the concepts involved,
as concepts, is fully understood, the contraries of the proposi-
tions prove to be inconceivable. All of which is admirably-
illustrated by Kant's own procedure when, in the Kritilk, he

I This must not be understood as implying that there is anything in.

common between the Wolffian theory of knowledae and the uncritical
affirmation of anything of which one has " clear and distinct ideas," or a
strong " emotion of conviction ". The prior analysis of concepts is pre-
suipposed, the careful fraining of definitions, the clear discrimination of
the several sorts of predicables, and the testing of attributa by the
attempt to think the given definition with the given attributumm negated.
But the Wolffian logicians had rightly maintained that the touchstone in
this final test can be nothing but the ultimate and inexplicable nmental
fact that the proposed negation is actually unthinkable.

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
 KANT'S ANTITHESIS OF DOGMATISM AND CRITICISM. 213

undertakes to prove that Space and Time are pure percepts

underlying the concepts of mathematics. He does it by
showing that in mathematical judgments the predicates are
necessarily and universally " connected with the subjects
a priori, and yet are not discoverable by a mere analysis of the
definitions of the subjects. Geometry, for example, " is a
science which determines the properties of Space synthetically
and yet a priori. What, then, must the idea of Space be, in
order that such a knowledge of it should be possible? It
must, primarily, be a perception; for out of a mere concept
no propositions can be drawn which go beyond the concept
itself '-as do the propositions of geometry. But this percep-
tion must be found in us a priori, and hence, must be a pure,
non-empirical perception. For geometrical propositions are
all apodictic, that is, they involve the consciousness of their
own necessity; and such propositions cannot be empirical"
(second ed., p. 41). Evidently, however, since the sup-
posed fact that Space is a reine Anschanung is inferred
the fact that geometrical theorems are necessary a priori, the
necessity and apriority of those theorems cannot in turn
depend upon a proof that Space is a reine Anschaunny. Kant
here, in a word, not only accepts the logical method of the
'doganatists,' but even uses it to establish that principle
which, so far as mathematical reasonings are concerned, is in
the same sentence supposed to show the illegitimacy of their
logical method. His own language exhibits better than any
commentary how irrelevant is the doctrine of the pure per-
cepts to the epistemological questions concerning the criteria
of a priori knowledge, which are the only questions at issue
between himself and the Wolffians. And now, in the third
place, the theory that judgments a priori presuppose a pure
percept, refers only to mathematical judgments; and it there-
fore does not affect Kant's admission, already noted, that
there may be valid a priori judgments of which the predicates
are (like those in mathematics) attributa rationata, but which
-being sufficiently explained by the principle of contradic-
tion-do not require the mediation of any such pure percept.
Only one point more needs to be touched upon. It may be
urged, in Kant's behalf, that he at least is distinguished from
his predecessors-and is more 'critical' than they-in that
he limits the scope even of our valid a priori judgments to
" objects of possible experience," and professes no knowledge
of the transcendent. Upon this much might be said, if

' Note again that this is inconsistent with the admissions which Kant
makes in the Revlii to Eberhard.
15

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms
214 A. 0. LOVEJOY: KANT. DOGMATISM V. CRITICISM.

space permitted; but I must limit myself to two remarks.

One is that, in Kant's own eyes, it is not upon this con-
sideration that his distinction of 'criticism' from 'dogma-
tism-i' turns, but upon the question how, even within the
narrower range, non-identical judgments may be justified a
priori. The other remark is that Kant himself, as we have
already seen, affirmed that no reality whatever " can possess
the formal character of Undenklichkeit," that is, can corre-
spond to a logically 'impossible' and self-contradictory con-
cept. And inasmuch as he at first fails to deny, and in the
end somewhat confusedly admits, that other predicates be-
sides those contained in the definition of a concept may be
involved in its interna possibilitas, and that other propositions
besides tautological ones may be tested by the criterion of
the inconceivability of their opposites-it must be said that,
even on this point, Kant did not successfully differentiate his
position from that of Leibniz and of Wolff.
And thus the whole distinction between ' criticism' and
'dogmatism '-in so far as it is intended to correspond to an
historical contrast between Kant and his German predeces-
sors-falls to the ground. The 'dogmatists' were not, in
Kant's sense, dogmatic; Kant, in his principal writings, did
not refute or even attack their real doctrine of the grounds
of the legitimacy of non-identical propositions a priori; when
eventually compelled by Eberhard to define his attitude to
that doctrine, he accepted-though evidently without quite
knowing what he was about-the essential principle of it;
and he was himself unable to formulate any different doctrine
of the grounds of the validity of such propositions-since the
theory of the reine Anschauung, even in the restricted field
where it applies, expressly presupposes the a priori legitimacy
of synthetic judgments in mathematics, and so cannot be
regarded as the logical ground of their legitimacy. Finally,
by his inability to deny the general applicability to all reality
of the principle of contradiction, in its wider as well as in its
more literal use, Kant undermined in advance his professedly
negative and agnostic position with respect to the possibility
of metaphysics as a theoretical science.

This content downloaded from 142.66.3.42 on Tue, 20 Sep 2016 15:32:33 UTC
All use subject to http://about.jstor.org/terms