science will also account for metaphysics properly conceived.
1 Cf. above, p. Ivi ; below, p. 571 ff. 2 Dissertation, 7.
20 PREFACE TO THE SECOND EDITION
Upon this important point Kant dwells at some length.
Even the negative results of the Critique are, he emphasises,
truly positive in their ultimate consequences. The dogmatic
extension of speculative reason really leads to the narrowing
of its employment, for the principles of which it then makes
use involve the subjecting of things in themselves to the
limiting conditions of sensibility. All attempts to construe
the unconditioned in terms that will satisfy reason are by
such procedure ruled out from the very start. To demonstrate
this is the fundamental purpose and chief aim of the Critique.
Space and time are merely forms of sensuous intuition ; the
concepts of understanding can yield knowledge only in their
connection with them. Though the concepts in their purity
possess a quite general meaning, this is not sufficient to con-
stitute knowledge. The conception of causality, for instance,
necessarily involves the notion of time-sequence ; apart from
time it is the bare, empty, and entirely unspecified con-
ception of a sufficient ground. Similarly, the category of sub-
stance signifies the permanent in time and space ; as a form
of pure reason it has a quite indefinite meaning signifying
merely that which is always a subject and never a predicate.
In the absence of further specification, it remains entirely
problematic in its reference. The fact, however, that the
categories of the understanding possess, in independence of
sensibility, even this quite general significance is all-important.
Originating in pure reason they have a wider scope than the
forms of sense, and enable us to conceive, though not to gain
knowledge of, things in themselves. 1 Our dual nature, as
being at once sensuous and supersensuous, opens out to us
the apprehension of both.
Kant illustrates his position by reference to the problem
of the freedom of the will. As thought is wider than sense,
and reveals to us the existence of a noumenal realm, we
are enabled to reconcile belief in the freedom of the will
with the mechanism of nature. We can recognise that within
the phenomenal sphere everything without exception is
causally determined, and yet at the same time maintain that
the whole order of nature is grounded in noumenal conditions.
We can assert of one and the same being that its will is
subject to the necessity of nature and that it is free-
mechanically determined in its visible actions, free in its real
supersensible existence. We have, indeed, no knowledge of
the soul, and therefore cannot assert on theoretical grounds
that it possesses any such freedom. The very possibility of
freedom transcends our powers of comprehension.
1 All these assertions call for later modification and restatement.
COMMENT ON PREFACE 21
proof that it can at least be conceived without contradiction
is, however, all-important. For otherwise no arguments from
the nature of the moral consciousness could be of the least
avail ; before a palpable contradiction every argument is
bound to give way. Now, for the first time, the doctrine of
morals and the doctrine of nature can be independently
developed, without conflict, each in accordance with its own
laws. The same is true in regard to the existence of God
and the immortality of the soul. By means of the Critical
distinction between the empirical and the supersensible worlds,
these conceptions are now for the first time rendered possible of
belief. " I had to remove knowledge, in order to make room
for belief"'*- "This loss affects only the monopoly of the schools,
in no respect the interests of humanity ."*
Lastly, Kant emphasises the fact that the method of the
Critique must be akin to that of dogmatism. It must be
rational a priori. To adopt any other method of procedure
is "to shake off the fetters of science altogether, and thus
to change work into play, certainty into opinion, philosophy
into philodoxy." 3 And Kant repeats the claims of the preface
of the first edition as to the completeness and finality of his
system. " This system will, as I hope, maintain through the
future this same unchangeableness." 4
Logic. 5 For Kant's view of the logic of Aristotle as com-
plete and perfect, cf. below, pp. 184-5. Kant compares meta-
physics to mathematics and physics on the one hand, and to
formal logic on the other. The former show the possibility of
attaining to the secure path of science by a sudden and single
revolution ; the latter demonstrates the possibility of creating
a science complete and entire at a stroke. Thanks to the
new Critical method, metaphysics may be enabled, Kant
claims, to parallel both achievements at once.
Theoretical and practical reason. 6 Such comment as is
necessary upon this distinction is given below. Cf. p. 569 ff.
Hitherto it has been supposed that all knowledge must conform
to the objects. 7 This statement is historically correct. That
assumption did actually underlie one and all of the pre-
Kantian philosophies. At the same time, it is true that Kant's
phenomenalist standpoint is partially anticipated by Hume,
by Malebranche and by Leibniz, especially by the first named.
Hume argues that to condemn knowledge on the ground that
it can never copy or truly reveal any external reality is to
misunderstand its true function. Our sense perceptions
and our general principles are so determined by nature
1 B xxx. z B xxxii. * B xxxvii. 4 B xxxviii.
6 B vii. 6 B viii. 7 B xvi.
22 PREFACE TO THE SECOND EDITION
as to render feasible only a practical organisation of life.
When we attempt to derive from them a consistent body of
knowledge, failure is the inevitable result. 1 Malebranche,
while retaining the absolutist view of conceptual knowledge,
propounds a similar theory of sense-perception. 2 Our per-
ceptions are, as he shows, permeated through and through,
from end to end, with illusion. Such illusions justify them-
selves by their practical usefulness, but they likewise prove
that theoretical insight is not the purpose of our sense-
experience. Kant's Copernican hypothesis consists in great
part of an extension of this view to our conceptual, scientific
knowledge. But he differs both from Malebranche and from
Hume in that he develops his phenomenalism on rationalist
lines. He professes to show that though our knowledge is
only of the phenomenal, it is conditioned by a priori
principles. The resulting view of the distinction between
appearance and reality has kinship with that of Leibniz. 3
The phenomena of science, though only appearances, are
none the less bene fundata. Our scientific knowledge,
though not equivalent to metaphysical apprehension of the
ultimately real, can be progressively developed by scientific
The two "parts " of metaphysics. 4 Kant is here drawing the
important distinction, which is one result of his new standpoint,
between immanent and transcendent metaphysics. It is un-
fortunate that he does not do so in a more explicit manner,
with full recognition of its novelty and of its far-reaching
significance. Many ambiguities in his exposition here and
elsewhere would then have been obviated. 5
The unconditioned which Reason postulates in all things by
themselves, by necessity and by right. 6 Points are here raised
the discussion of which must be deferred. Cf. below,
pp. 429-31, 433-4, 558-6i.
The Critique is a treatise on method, not a system of the
science itself. 7 Cf. A xv. ; B xxxvi. ; and especially A 1 1 =
B 24, below pp. 71-2.
The Copernican hypothesis. 8 Kant's comparison of his new
hypothesis to that of Copernicus has generally been mis-
understood. The reader very naturally conceives the
Copernican revolution in terms of its main ultimate conse-
quence, the reduction of the earth from its proud position of
central pre-eminence. But that does not bear the least
1 Cf. above, pp. xxvi-vii ; below, pp. 594-5.
2 Cf. " Malebranche's Theory of the Perception of Distance and Magnitude,"
in British Journal of Psychology (1905), i. pp. 191-204.
3 Cf. below, pp. 143 ff., 604. 4 B xviii.-xix. 5 Cf. below, pp. 33, 56, 66 ff.
6 B xx. 7 B xxii. 8 B xvi. ; B xxii. .
THE COPERNICAN HYPOTHESIS 23
analogy to the intended consequences of the Critical philo-
sophy. The direct opposite is indeed true. Kant's hypo-
thesis is inspired by the avowed purpose of neutralising the
naturalistic implications of the Copernican astronomy. His
aim is nothing less than the firm establishment of what
may perhaps be described as a Ptolemaic, anthropocentric
metaphysics. Such naturalistic philosophy as that of Hume
may perhaps be described as Copernican, but the Critical
philosophy, as humanistic, has genuine kinship with the Greek
Even some of Kant's best commentators have interpreted
the analogy in the above manner. 1 It is so interpreted by
T. H. Green 2 and by J. Hutchison Stirling. 3 Caird in his
Critical Philosophy of Kant makes not the least mention of the
analogy, probably for the reason that while reading it in the
same fashion as Green, he recognised the inappropriateness
of the comparison as thus taken. The analogy is stated in
typically ambiguous fashion by Lange 4 and by HofTding. 5
S. Alexander, while very forcibly insisting upon the Ptolemaic
character of the Kantian philosophy, also endorses this
interpretation in the following terms :
" It is very ironical that Kant himself signalised the revolution
which he believed himself to be effecting as a Copernican revolution.
But there is nothing Copernican in it except that he believed it to be
a revolution. If every change is Copernican which reverses the
order of the terms with which it deals, which declares A to depend
on B when B had before been declared to depend on A, then Kant
who believed that he had reversed the order of dependence of mind
and things was right in saying that he effected a Copernican revolu-
tion. But he was not right in any other sense. For his revolution,
so far as it was one, was accurately anti-Copernican." 6
As the second edition preface is not covered by the
published volumes of Vaihinger's Commentary^ the point has
not been taken up by him.
Now Kant's own statements are entirely unambiguous and
do not justify any such interpretation as that of Green and
Alexander. As it seems to me, they have missed the real
point of the analogy. The misunderstanding would never
have been possible save for our neglect of the scientific
1 Watson's The Philosophy of Kant Explained (p. . 37) is the only work in
which I have found correct and unambiguous indication of the true interpretation
of Kant's analogy.
2 Prolegomena to Ethics, bk i. ch. i. n.
3 Text-Book to Kant (1881), p. 29.
4 History of Materialism, Eng. transl., ii. pp. 156, 158, 237.
5 Geschichte der neueren Philosophic (1896), ii. p. 64.
6 Hibbert Journal, October 1910, p. 49.
24 PREFACE TO THE SECOND EDITION
classics. Kant must have had first-hand acquaintance with
Copernicus' De Revolutionibus, and the comparison which he
draws assumes similar knowledge on the part of his readers.
Copernicus by his proof of the " hypothesis " (his own term)
of the earth's motion sought only to achieve a more harmonious
ordering of the Ptolemaic universe. And as thus merely a
simplification of the traditional cosmology, his treatise could
fittingly be dedicated to the reigning Pope. The sun upon
which our terrestrial life depends was still regarded as uniquely
distinct from the fixed stars ; and our earth was still located
in the central region of a universe that was conceived in the
traditional manner as being single and spherical. Giordano
Bruno was the first, a generation later, to realise the revolu-
tionary consequences to which the new teaching, consistently
developed, must inevitably lead. It was he who first taught
what we have now come to regard as an integral part of
Copernicus' revolution, the doctrine of innumerable planetary
systems side by side with one another in infinite space.
Copernicus' argument starts from the Aristotelian principle
of relative motion. To quote Copernicus' exact words :
" All apprehended change of place is due to movement either of
the observed object or of the observer, or to differences in movements
that are occurring simultaneously in both. For if the observed
object and the observer are moving in the same direction with equal
velocity, no motion can be detected. Now it is from the earth that
we visually apprehend the revolution of the heavens. If, then, any
movement is ascribed to the earth, that motion will generate the
appearance of itself in all things which are external to it, though as
occurring in the opposite direction, as if everything were passing
across the earth. This will be especially true of the daily revolution.
For it seems to seize upon the whole world, and indeed upon every-
thing that is around the earth, though not upon the earth itself. . . .
As the heavens, which contain and cover everything, are the common
locus of things, it is not at all evident why it should be to the contain-
ing rather than to the contained, to the located rather than to the
locating, that a motion is to be ascribed." 1
The apparently objective movements of the fixed stars and of
the sun are mere appearances, due to the projection of our
own motion into the heavens.
" The first and highest of all the spheres is that of the fixed stars,
self -containing and all -containing, and consequently immobile, in
short the locus of the universe, by relation to which the motion and
position of all the other heavenly bodies have to be reckoned." 2
1 De Rcvolutionibus, I. v. 2 Ibid. I. x.
THE COPERNICAN HYPOTHESIS 25
Now it is this doctrine, and this doctrine alone, to which
Kant is referring in the passages before us, namely, Copernicus'
hypothesis of a subjective explanation of apparently objective
motions. And further, in thus comparing his Critical pro-
cedure to that of Copernicus, he is concerned more with the
positive than with the negative consequences of their common
hypothesis. For it is chiefly from the point of view of the
constructive parts of the Aesthetic, Analytic, and Dialectic that
the comparison is formulated. By means of the Critical
hypothesis Kant professes on the one hand to account for our
scientific knowledge, and on the other to safeguard our
legitimate metaphysical aspirations. The spectator projects
his own motion into the heavens ; human reason legislates for
the domain of natural science. The sphere of the fixed stars
is proved to be motionless ; things in themselves are freed
from the limitations of space and time. " Copernicus dared,
in a manner contradictory of the senses but yet true, to seek
the observed movements, not in the heavenly bodies, but in
the spectator." x
In view of Kant's explicit elimination of all hypotheses
from the Critique 2 the employment of that term would seem
to be illegitimate. He accordingly here states that though in
the Preface his Critical theory is formulated as an hypothesis
only, in the Critique itself its truth is demonstrated a priori.
Distinction between knowing and thinking. 3 Since according
to Critical teaching the limits of sense-experience are the
limits of knowledge, the term knowledge has for Kant a very
limited denotation, and leaves open a proportionately wide
field for what he entitles thought. Though things in them-
selves are unknowable, their existence may still be recognised
1 B xxii. n. z Cf. below, p. 543 ff.
3 B xxvi. Cf. above, pp. Iv-vi, 20; below, pp. 290-1, 331, 342, 404 ff.
I SHALL first 1 give a restatement, partly historical and partly
explanatory, of Kant's main argument as contained in the
enlarged Introduction of the second edition.
There were two stages in the process by which Kant came
to full realisation of the Critical problem. There is first the
problem as formulated in his letter of 1772 to Herz : how
the a priori can yield knowledge of the independently real. 2
This, as he there states it, is an essentially metaphysical
problem. It is the problem of the possibility of transcendent
metaphysics. He became aware of it when reflecting upon
the function which he had ascribed to intellect in the
Dissertation. Then, secondly, this problem was immeasur-
ably deepened, and at the same time the proper line for its
treatment was discovered, through the renewed influence
which Hume at some date subsequent to February 1772
exercised upon Kant's thought. 3 Hume awakened Kant to
what may be called the immanent problem involved in the
very conception of a priori knowledge as such. The primary
problem to be solved is not how we advance by means of a
priori ideas to the independently real, but how we are able to
advance beyond a subject term to a predicate which it does not
appear to contain. The problem is indeed capable of solution,
just because it takes this logical form. Here as elsewhere,
ontological questions are viewed by Kant as soluble only
to the extent to which they can be restated in logical terms.
Now also the enquiry becomes twofold : how and in what
degree are a priori synthetic judgments possible, first in their
employment within the empirical sphere (the problem of
immanent metaphysics) and secondly in their application to
things in themselves (the problem of transcendent meta-
physics). The outcome of the Critical enquiry is to establish
1 This restatement will continue up to p. 33. In pp. 33-43 I shall then give
general comment on the Introduction as a whole. In p. 43 ff. I add the necessary
detailed treatment of special points. 2 Cf. below, p. 219 ff.
3 Cf. above, p. xxv ff. ; below, pp. 6 1 ff., 593 ff.
EMPIRICAL AND A PRIORI JUDGMENT 27
the legitimacy of immanent metaphysics and the impossibility
of all transcendent speculation.
The argument of Kant's Introduction follows the above
sequence. It starts by defining the problem of metaphysical
knowledge a priori, and through it leads up to the logical
problem of the a priori synthetic judgment. In respect of
time all knowledge begins with experience. But it does not
therefore follow that it all arises from experience. Our ex-
perience may be a compound of that which we receive through
impressions, and of that which pure reason supplies from
itself. 1 The question as to whether or not any such a priori
actually exists, is one that can be answered only after further
enquiry. The two inseparable criteria of the a priori are
necessity and universality. That neither can be imparted to
a proposition by experience was Kant's confirmed and un-
questioned belief. He inherited this view both from Leibniz
and from Hume. It is one of the presuppositions of his
argument. Experience can reveal only co- existence or
sequence. It enables us only to assert that so far as we have
hitherto observed, there is no exception to this or that rule.
A generalisation, based on observation, can never possess a
wider universality than the limited experience for which it
stands. If, therefore, necessary and universal judgments can
anywhere be found in our knowledge, the existence of an
a priori that originates independently of experience is ipso
facto demonstrated. 2
The contrast between empirical and a priori judgments,
as formulated from the dogmatic standpoint, is the most
significant and striking fact in the whole range of human
knowledge. A priori judgments claim absolute necessity.
They allow of no possible exception. They are valid not
only for us, but also for all conceivable beings, however
different the specific conditions of their existence, whether
they live on the planet Mars or in some infinitely remote
region of stellar space, and no matter how diversely their
bodily senses may be organised. Through these judgments
a creature five feet high, and correspondingly limited by
temporal conditions, legislates for all existence and for all
time. Empirical judgments, on the other hand, possess only
a hypothetical certainty. We recognise that they may be
1 This statement is first made in the Introduction to the second edition. It is
really out of keeping with the argument of the Introduction in either edition. Cf.
below, pp. 39-40, 57, 85, 1 68, 222, 245 ff. (especially pp. 278, 288).
2 This is the argument of the Introduction to the second edition. In the first
edition Kant assumes without question the existence of the a priori. He
enquires only whether it is also valid in its metaphysical employment beyond the
field of possible experience.
overturned through some addition to our present experience,
and that they may not hold for beings on other planets or for
beings with senses differently constituted. Whereas the
opposite of a rational judgment is not even conceivable, the
opposite of an empirical judgment is always possible. The
one depends upon the inherent and inalienable nature of
our thinking ; the other is bound up with the contingent
material of sense. The one claims absolute or meta-
physical truth : the other is a merely tentative rsum of a
The possibility of such a priori judgments had hitherto
been questioned only by those who sought to deny to them
all possible objective validity. Kant, as a rationalist, has no
doubt as to their actual existence. In the Introduction to the
second edition he bluntly asserts their de facto existence,
citing as instances the propositions of mathematics and the
fundamental principles of physical science. Their possibility
can be accounted for through the assumption of a priori forms
and principles. 1 But with equal emphasis he questions the
validity of their metaphysical employment. For that is an
entirely different matter. We then completely transcend the
world of the senses and pass into a sphere where experience
can neither guide nor correct us. In this sphere the a priori
is illegitimately taken as being at once the source of our
professed knowledge and also the sole criterion of its own
This is the problem, semi-Critical, semi-dogmatic, which is
formulated in the letter of 1772 to Herz. 2 What right have we
to regard ideas, which as a priori originate from within, as
being valid of things in themselves? In so doing we are
assuming a pre-established harmony between our human
faculties and the ultimately real ; and that is an assumption
which by its very nature is incapable of demonstration. The
proofs offered by Malebranche and by Leibniz are themselves
speculative, and consequently presuppose the conclusion which
they profess to establish. 3 As above stated, Kant obtained
his answer to this problem by way of the logical enquiry into
the nature and conditions of a priori judgment.
One of the chief causes, Kant declares, why hitherto meta-
physical speculation has passed unchallenged among those
who practise it, is the confusion of two very different kinds
of judgment, the analytic and the synthetic. Much the
greater portion of what reason finds to do consists in the
analysis of our concepts of objects.
1 The argument of the first edition, though briefer, is substantially the same.
2 Quoted below, pp. 219-20. ? Cf. below, pp. 114, 290, 590.
NATURE OF A PRIORI JUDGMENT 29
"As this procedure yields real knowledge a priori, which progresses
in secure and useful fashion, reason is so far misled as surreptitiously
to introduce, without itself being aware of so doing, assertions of an
entirely different order, in which reason attaches to given concepts
others completely foreign to them and moreover attaches them
a priori. And yet one does not know how reason comes to do this.
This is a question which is never as much as thought of." 1
The concepts which are analytically treated may be either
empirical or a priori. When they are empirical, the judg-
ments which they involve can have no wider application
than the experience to which they give expression ; and in
any case can only reveal what has all along been thought,
though confusedly, in the term which serves as subject of the
proposition. They can never reveal anything different in
kind from the contents actually experienced. This limitation,
to which the analysis of empirical concepts is subject, was
admitted by both empiricists and rationalists. The latter
sought, however, to escape its consequences by basing their
metaphysics upon concepts which are purely a priori, and
which by their a priori content may carry us beyond the
experienced. But here also Kant asserts a non possibile. A