Whatever Happened to Kant’s
Ontological Argument?

ian logan

Blackfriars Hall, University of Oxford

Introduction

1

In the New Elucidation of 1755, Kant first put forward an argument
for the existence of God from the concept of possibility, seeking to dem-
onstrate that it is impossible to think of the non-existence of that which
is the ground of the possibility of things.

2

He developed this argument

further as the only basis for a rigorously viable proof of God in The One
Possible Basis for a Demonstration of the Existence of God

(OPB) of

1763. But when he came to write the Critique of Pure Reason (CPR) in
1781, Kant had apparently discarded what he had previously referred to
as the one possible proof, and did not include it in the critique of the
proofs of God in the ‘Transcendental Dialectic’. The proofs that he did
subject to critique in the CPR were those that he had excluded in the
OPB

from the realm of possible proofs. Kant’s own proof is of partic-

ular interest, because it represents an ‘ontological’ or ‘ontotheological’
argument which is distinct from that of Descartes, and which for Kant
does not succumb to his critique of the Cartesian argument. As has been
well attested, the subject of this proof evolved from the ens realissimum
of Kant’s so-called ‘pre-critical’ thought into the transcendental ideal of

1

My particular thanks are due to Dr M.J. Wood for commenting on more than one
version of this paper.

2

I. Kant, A new elucidation of the first principles of metaphysical cognition, in D. Wal-
ford (Ed.), Immanuel Kant: Theoretical Philosophy 1755-1770, Cambridge 1992, pp.
1-45: ‘‘if you deny the existence of God, you instantly abolish not only the entire
existence of things but even their inner possibility itself’’ (p. 16); ‘‘as soon as you
deny the existence of God every concept of possibility vanishes’’ (p. 17).

Philosophy and Phenomenological Research
Vol. LXXIV No. 2, March 2007
2007 International Phenomenological Society

346

IAN LOGAN

the CPR.

3

The aim of this paper is to uncover what precisely happened

to Kant’s argument, and to suggest reasons why it evolved the way it
did and why it continues to pose problems for Kant’s position. My the-
sis has four main contentions: (i) that the evolution of Kant’s argument
was occasioned by a fundamental empiricist shift in his thought, (ii) that
this shift was ‘pre-critical’ and not established or justified by the argu-
ment of the CPR, (iii) that as a result Kant could no longer sustain the
conclusion of his ontological argument, but (iv) that he was unable to
discard this argument, since his account of the transcendental ideal
depended on it. Kant was caught on the horns of a dilemma—

—without

this proof he could not establish the subjective necessity of the transcen-
dental ideal, but with it he left the way open for ‘pre-critical’ metaphys-
ics. For, in its evolved state, the proof depended on the assumption of
radical empiricist limitations on our knowledge, i.e., that knowledge
‘‘can have no other object than that supplied by experience’’

4

. If any

argument such as that put forward by Kant in the OPB is successful,
then this assumption is refuted. Such a refutation would have serious
consequences for the doctrine of the transcendental ideal and for the
whole architectonic of Kant’s critical philosophy. To say that any such
refutation is not possible, because we cannot have any knowledge out-
side of these empiricist limitations, is to beg the question, and it is this, I
suggest, that Kant has done.

The argument of this paper is developed in four main stages. Firstly,

I consider Kant’s own version of the ontological argument in the OPB
and the ‘pre-critical’ basis of his critique of the Cartesian ontological

3

See, for example, J. Schmucker, ‘On the development of Kant’s transcendental the-
ology’ in L. Beck (Ed.), Proceedings of the Third International Kant Congress, Dord-
recht 1972, pp. 495-500, p. 496: ‘‘there has been undoubtedly a real and significant
evolution, which proceeds from his famous onto-theological argument in the Nova
Dilucidatio

of 1755 and in the Beweisgrund of 1763 and terminates in the doctrine of

the transcendental ideal.’’ Similar views are expounded in J. Schmucker, Die Onto-
theologie des vorkritischen Kant, Berlin 1980;

J. Schmucker, Kants vorkritische Kritik

der Gottesbeweise: Ein Schlu¨ssel zur Interpretation des theologischen Hauptstu¨cks der
transzendentalen Dialektik der ‘Kritik der reinen Vernunft’

, Mainz 1983; R. Dell’Oro,

From Existence to the Ideal: Continuity and Development in Kant’s Theology

, New

York 1994; P. Laberge, La The´ologie kantienne pre´critique, Ottowa 1973; G. Sala,
Kant und die Frage nach Gott: Gottesbeweise und Gottesbeweiskritik in den Schriften
Kants

, Berlin

⁄ New York 1990.

4

I.Kant, Immanuel Kant’s Critique of Pure Reason, translated by N. Kemp Smith, 2

nd

impression with corrections, London 1933, A471; B499. (All translations of the
CPR

in this paper are from Kemp Smith’s edition.) Kant continues: ‘‘if we trans-

cend the limits thus imposed, the synthesis which seeks, independently of experience,
new species of knowledge, lacks the substratum of intuition upon which alone it can
be exercised.’’ To be more precise ‘‘upon which alone it can be exercised’’ is true iff
knowledge is limited to experience. What sounds like an argument justifying Kant’s
assumption is in fact a restatement of it.

WHATEVER HAPPENED TO KANT

’

S ONTOLOGICAL ARGUMENT

?

347

argument in the CPR. Secondly, I expose the nature of the empiricist
shift that led to Kant’s rejection of the objective reality of the conclu-
sion of his ontological argument in the OPB. Thirdly, I detail the evo-
lution of Kant’s ontological argument from his early writings to the
CPR

by way of this empiricist shift. And, fourthly, I examine Kant’s

implicit ‘refutation’ of his own ontological argument from possibility in
‘The Transcendental Ideal’, and why it is that the critique of the proofs
of God in the CPR does not include a critique of Kant’s own ontologi-
cal argument. In conclusion, I discuss the difficulties facing Kant’s posi-
tion in the CPR, arising from the role he is obliged to give his own
version of the ontological argument.

1. Kant’s ontological argument

It is frequently overlooked that Kant’s use of the term, ‘ontological’
and his application of this term to certain proofs for the existence of
God predates the CPR.

5

This might not be so important, if the issue

were ‘merely’ one of linguistic usage, but in the OPB Kant actually pre-
sents an ontological or ontotheological argument, which, he believes,
constitutes the singularly effective possibility of the title. Kant’s know-
ledge of the origins of ontological arguments was limited. It is clear,
for example, from the lack of any serious discussion of Anselm that
Kant was not directly acquainted with his work, and that his view of
Anselm’s argument as an ontological argument (in the sense of that of
Descartes or Leibniz) was based on the reports of others who were lit-
tle or no better acquainted with Anselm than Kant himself.

6

Kant

seems to consider his own argument from possibility as the only alter-
native to the ontological arguments of Descartes and Leibniz, but he
does not establish that this is the case.

7

He is assisted in this view by

his ignorance of the history of ontological arguments.

5

See, for instance, J. Barnes, The Ontological Argument, London 1972, p. 1.

6

There are, for example, no references to Anselm in Kant’s writings prior to the pub-
lication of the CPR. It seems likely that Kant would have noted the reference to
Anselm in Wolff’s Theologia Naturalis, II, §13, ‘Ens perfectissimum possibile est’, in
which Wolff puts a Cartesian spin on Anselm’s argument, stating that Anselm
affirmed that the existence of God is to be inferred from the notion of the most per-
fect being. In Reflexionen 6214 (dating from the early 1780s) we find a reference to
Anselm’s Ontotheology (Gesammelte Werke (GW), XVIII, p. 500). There is also a
reference to Anselm in the Lectures on the philosophical doctrine of religion (GW,
XXVIII, p. 1003). And in 1791 Kant writes of Leibniz as adding to or supplement-
ing Anselm’s argument (GW, XX, p. 349). But, perhaps, most revealing of Kant’s
knowledge of the history of ontological arguments is the reference dating from
1794, in which he calls Anselm a Parisian scholastic, who first put forward the argu-
ment developed by Descartes and Leibniz (GW, XVIII, p. 782).

7

That Anselm’s argument is such an alternative is a view I hold, but will not attempt
to justify in this paper.

348

IAN LOGAN

In the OPB, Kant argues that ‘‘the great truth, that there is a God,

must, if it is to be of the highest degree of mathematical certainty, have
this property: that it can be achieved in only one way.’’

8

Kant goes on

to argue for a particular demonstration, a non-Cartesian ontological
argument from possibility, with the consequence for him that all other
demonstrations, in this rigorous sense, must fail. Although Kant ‘dis-
cards’, or to put it more accurately does not explicitly treat of, this
proof in the CPR, it is clear that he already views all other competing
arguments for God, including the Cartesian ontological argument, as
doomed to failure. For Kant there could only be one apodictic proof
for God. If it is not apodictic it is no proof at all, and as early as the
conclusion of the OPB, he already appears unsure of the apodictic
force of his argument:

9

‘‘Seek the proof here. And if you are not pre-

pared to meet it on this ground, turn from this unfamiliar path to the
highway of human reason. It is thoroughly necessary that one be con-
vinced of God’s existence; but it is not nearly so necessary that it be
demonstrated.’’

But, what was the only possible proof of God in the OPB that Kant

dismisses without an explicit critique in the CPR?

10

At the end of the

OPB

, having sought to refute the Cartesian ontological as well as the

cosmological and physico-theological proofs, Kant states that ‘‘experi-
ence of contingent things cannot give an adequate argument by which
to comprehend the existence of something of which it is impossible that
it not be. It is solely in that the denial of divine existence is absolutely
nothing that the difference between this [divine] existence and that of
other things lies. Internal possibility, the essence of things, is precisely
that whose negation cancels all thought.’’

11

It is this characteristic of

the proof, that to deny God is not possible, since this would involve a
denial of all thought, that constitutes its unique character. For Kant
the term, ‘demonstration’, refers to a particular kind of proof,

12

a proof

8

I. Kant, The One Possible Basis for a Demonstration of the Existence of God, trans-
lation and introduction by G. Treash, Lincoln

⁄ London 1994, p. 223. ‘That it can

only be achieved in one way’ does not seem an obvious truth to me, but neverthe-
less, I do not intend to investigate the claim further here, other than to note that
this is a useful methodological move, which simplifies the issue of proofs of God,
since once you have found one that works, then there is no reason to consider or
search out any others.

9

OPB

, p. 239.

10

Unlike B. Reardon, Kant as Philosophical Theologian, London 1988, p. 32, I do not
take the view that we ‘‘need no longer dwell on Kant’s argumentation, therefore,
since he himself subsequently dropped it.’’ It seems to me that that is all the more
reason for dwelling on it.

11

OPB

, pp. 237f.

12

OPB

, p. 223.

WHATEVER HAPPENED TO KANT

’

S ONTOLOGICAL ARGUMENT

?

349

‘‘with mathematical clarity,’’ i.e., an apodictic proof, of the existence of
a supreme and unique being, God. However, as the title of the work
suggests, OPB does not present a fully developed version of such a
proof, but what he calls the ‘‘basis’’ (Beweisgrund) for it, i.e. ‘‘the first
strokes of a master plan.’’

13

What he is putting forward is the outline

of a proof of God’s existence based on the possibility of things. For
Kant, if the proof of God is possible, then this is how it must work.

In the OPB, Kant divides the so-called proofs of God into two

groups of two:

14

A. The ‘a priori’ proofs that proceed from the concepts of the

understanding (the ontological concepts).

1. The proof from the concept of the possible as cause to

the existence of God as effect—

—the Cartesian proof,

which treats existence as a predicate.

2. The proof from the concept of the possible as possible to

the existence of God as cause of the possible—

—Kant’s

own ontotheological proof.

B. The ‘a posteriori’

15

proofs that proceed from the concept of

experience of existing things.

1. From what exists (the world), in so far as it exists, to the

existence of a prime

⁄ independent cause, God—

—Wolff’s

proof from contingence—

—though this cause is not neces-

sarily to be equated with God. In the CPR Kant refers to
this as the cosmological proof.

16

2. From what exists, in so far as it possesses certain charac-

teristics, to God—

—the inductive or what Kant refers to in

the CPR as the physico-theological proof.

17

13

OPB

, p. 45.

14

See Sala, 1990, pp. 163f. See also Reardon, 1988, p. 31.

15

The term ‘a posteriori’ should be treated with caution, in so far as Kant regards
these arguments as dependent on the ‘ontological argument’. Cf. A638; B666: ‘‘all
merely speculative proofs in the end bring us always back to one and the same
proof, namely, the ontological.’’

16

A604; B632: ‘‘This proof, termed by Leibniz the proof a contingentia mundi,

…’’

17

Kant refers to this as the cosmological proof in the OPB, p. 233. But the same
argument is defined more precisely as the physico-theological proof in the CPR,
A620; B648ff.

350

IAN LOGAN

For Kant, A.2. alone counts as a logically rigorous proof, i.e., a
demonstration, whilst B.2., lacking ‘‘geometric rigour,’’

18

possesses only

a moral certainty, and A.1. and B.1. are false. It is the subject of A.2.,
i.e., Kant’s own ontotheological argument, that returns under the guise
of the transcendental ideal in the CPR.

19

As for the three other argu-

ments, in the CPR Kant identifies them as the only possible ways of
proving God’s existence ‘‘by means of speculative reason,’’

20

and sub-

jects them to a critique which is intended to culminate in their refuta-
tion. In this way, Kant seeks to establish that no proof of speculative
reason is, in fact, achievable. L. Kreimendahl notes Kant’s continuing
satisfaction with the systematic classification that he employed in his
critique of the proofs of God in Part Three of the OPB, and that Kant
consequently carried it over with terminological modifications into the
‘Ideal of Pure Reason’ in the CPR.

21

What is distinctive about the cri-

tique in the CPR is that the other arguments are made dependent on
the Cartesian ontological argument. The refutation of the ontological
argument involves the refutation of the other arguments. But the refu-
tation of the ontological argument is the same as the pre-critical refuta-
tion in the OPB. At the heart of the ‘Transcendental Dialectic’ we find
evidence of Kant employing arguments that do not simply arise out of
his critical philosophy as it is developed in the ‘Transcendental Ana-
lytic’, but derive from his so-called ‘pre-critical’ period. In other words,
the refutation of these proofs of God does not have its basis in, and
can be understood independently of, the working out of Kant’s critical
philosophy.

It is in OPB Part 1, Observation 2, that we find the central argument

of Kant’s proof. In §1, Kant makes what he calls a ‘‘necessary distinc-
tion’’ between the formal or logical and the material or real elements
of possibility. Something is formally or logically possible in so far as it
contains no contradiction. However, in §2 Kant makes it clear that for
something to be possible it requires not only formal or logical possibil-
ity, but also matter or data for thought, i.e., material or real possibility.
Thus, although there is no logical contradiction in the denial of all
existence, there is a contradiction in asserting that absolutely nothing
exists, but that something is possible. This contradiction arises,

18

But not entirely lacking cogency, pace Reardon, 1988, p. 32, since Kant makes it
clear that argument type B.2. ‘‘is not only possible, but also worthy of being
brought to its proper perfection through concerted efforts’’ (OPB, p. 231).

19

Sala, 1990, p. 7.

20

A590; B618.

21

L. Kreimendahl, Kant-Index, Band 38: Stellenindex und Konkordanz zu ‘Der einzig
mo¨gliche Beweisgrund zu einer Demonstration des Daseins Gottes’, Stuttgart

⁄ Bad

Canstatt 2003, p. xiii.

WHATEVER HAPPENED TO KANT

’

S ONTOLOGICAL ARGUMENT

?

351

because, according to Kant’s notion of material or real possibility there
has to be matter or data for thought, and clearly if there is absolutely
nothing, there is no such matter or data and accordingly there is no
possibility. Three points are worth making at this juncture. Firstly, for
Kant the possible is restricted to the thinkable.

22

So, only what can be

thought, i.e., that to which the rules of thought can be applied, is poss-
ible. Whatever restrictions are placed on the thinkable have a similar
restricting effect on the possible. Secondly, Kant’s notion of possibility
is bi-polar. It involves not simply the absence of contradiction, but also
the presence of some actuality. Thus, he cannot accept the Cartesian
ontological argument, since this only addresses one element of possibil-
ity, the logical or formal.

23

The third point follows on from the second.

For Kant the possible presupposes the actual, thus all possibility has to
be derived from something actual either directly (as a determination) or
indirectly (as a consequence).

24

It is this understanding of the relation-

ship of the possible and the thinkable and the possible and the actual
that opens the door to the all-pervading effects of Kant’s empiricist
shift. For, once the actual is confined to the empirical, then so are the
possible and the thinkable

.

Kant argues that the ‘‘internal possibility of all things presupposes

some existence.’’ Although there is no internal contradiction in denying
all existence, such a contradiction is found when one tries to assert that
there is ‘‘some possibility and yet nothing actual.’’ This is because ‘‘if
nothing exists, nothing conceivable is given and one would contradict
himself in nevertheless pretending something to be possible.’’

25

In §3,

Kant concludes that it is absolutely impossible that nothing should exist,
since ‘‘if all existence is denied, all possibility is also abolished.’’ Kant’s
argument appears to be based on the following steps, which are not
explicitly stated: Something exists, therefore something is possible. To
attempt to think of a state of affairs in which nothing exists, is to think
of a state of affairs in which nothing is possible, but in fact we know
that something is actual and consequently that something is possible.

22

For the connection between possibility and thinkability in Kant’s proof, see A.W.
Wood, Kant’s Rational Theology, Ithaca

⁄ London 1978, pp. 66ff., esp. p. 66: ‘‘Kant

begins his proof by distinguishing two kinds of impossibility. A supposition, he
says, is ‘formally’ impossible when its concept is unthinkable because it contains a
contradiction. It is ‘materially’ impossible, however, when it is unthinkable because
‘no material, no datum is there to think’.’’

23

Dell’Oro, 1994, p. 89.

24

See §4—

—OPB, p. 71.

25

OPB

, p. 69.

352

IAN LOGAN

In other words, it is impossible (unthinkable) that nothing is possible.
Therefore it is impossible (unthinkable) that nothing should exist.

26

In Part 1, Observation 3, §1, Kant moves on from the notion that it

is impossible that nothing should exist to the positive concept of abso-
lutely necessary existence. There exists something which is ‘‘absolutely
necessary,’’ since the undeniability of the possible demands this: ‘‘if I
annul all existence in general and if through that the final real ground
of all thought is abolished, then likewise all possibility disappears and
there is nothing left to be thought.’’

27

It is in §2, that Kant draws the

conclusion that there exists an absolutely necessary being:

28

‘‘All possi-

bility presupposes something actual in which and through which every-
thing conceivable is given.’’ Thus there is something whose annulment
would annul all (internal) possibility. ‘‘But that whose annulment or
negation eradicates all possibility is absolutely necessary.’’

29

Of course, establishing that an absolutely necessary something exists

is not yet to have demonstrated that God exists.

30

In the remainder of

Part 1, Observation 3, Kant seeks to show that this absolutely neces-
sary being is unitary, simple, immutable and eternal, containing the
highest reality. In the next Observation, he provides the outline of an
argument to show that this necessary being is a spirit and concludes
that this spirit is ‘‘a God,’’ though again he does not intend to go so
far as to give ‘‘a definite definition of the concept of God here.’’

31

He

provides a summary of the nature of his argument: ‘‘the ground of
proof we give for the existence of God is built simply upon [the fact]
that something is possible. Thus it is a proof that may be adduced
completely a priori. Neither my existence nor that of other minds nor
that of the corporeal world is presupposed.’’

32

But can Kant’s proof

26

See G. Oppy, Ontological Arguments and Belief in God, Cambridge 1995, p. 273,
for a brief criticism of Kant’s argument.

27

OPB

, p. 77.

28

OPB

, p. 79.

29

Kant appears to make an illicit move here from ‘it is necessary that something
exists’ to ‘a particular thing necessarily exists’. (Cf. Wood, 1978, pp. 70f.) However,
our concern is not to defend Kant’s argument, but to understand the grounds on
which Kant limited its effectiveness to that of a regulative principle.

30

Sala, 1990, p. 116 notes that the structure of Kant’s proof is thoroughly traditional,
and can be found already in the medieval scholastics. The existence of a necessary
being (Wesen) is established, then the attributes of this being are worked out, in
order to show that this necessary being is God. (Cf. A629; B657.) This thoroughly
traditional approach to the proofs points to the weakness Kant finds in the ‘a
posteriori

’ proofs. Without the ontological argument, they cannot even attain to

the ens necessarium, let alone God.

31

OPB

, p. 91.

32

OPB

, p. 95.

WHATEVER HAPPENED TO KANT

’

S ONTOLOGICAL ARGUMENT

?

353

from possibility be classified as a priori in this way, since real possibil-
ity is derived from the actual? Already we see a tension between Kant’s
intention of setting out an a priori argument for God in the OPB and
his developing (empiricist) understanding of the role of the actual.

For Kant the argument for God’s existence is ‘‘deduced from the

internal possibility of everything conceivable.’’

33

He writes: ‘‘if one is to

conclude from the concepts of possible things, no argument is possible
for the existence of God except the one which regards the internal pos-
sibility of things as something presupposing some existence.’’

34

Kant

terms this proof ‘‘ontological.’’

35

Ernst Cassirer has already identified in Kant’s own ‘ontological argu-

ment’ in the OPB a forerunner of the transcendental method that Kant
was to develop later.

36

It is this transcendental character of Kant’s argu-

ment that distinguishes his from the Cartesian. (Although it should be
noted that Kant at different times refers to both types of argument as
transcendental.

37

) In so far as it is transcendental, Kant’s argument pro-

ceeds, not by demonstrating ‘‘the necessity of God’s existence as an
immediate consequence of conceiving him as the perfect being,’’

38

but by

demonstrating that the non-existence of that which is absolutely neces-
sary, God, is unthinkable, that God is the ground of the possible (the
thinkable), and is not to be derived as a consequence from the possibil-
ity of things: ‘‘if you take away God, not only all existence of things but
even their internal possibility is absolutely abolished.’’

39

For Kant, the

significance of his own version of the ontological argument ‘‘is that we
must affirm the existence of God as the presupposition of all possibility
and of all determination.’’

40

In this argument he does not pass immedi-

ately from concept to existence, but from the inner possibility of things
to ‘‘the affirmation of the existence of that which makes possibility itself
possible.’’

41

The absolute necessity of ‘‘that which makes possibility

itself possible’’ is derived not from analysis of the concept and of the

33

OPB

, p. 227.

34

OPB

, p. 231.

35

OPB

, p. 233.

36

E. Cassirer, Kant’s Life and Thought, translated from the German by J. Haden,
New York

⁄ London 1981, p.65: ‘‘We have before us essentially a prelude to the

transcendental method to come, since the ultimate justification for positing exist-
ence in an absolute sense resides in the fact that without this assertion the possibil-
ity of knowledge is inconceivable.’’

37

See Wood, 1978, p. 66 n. 59.

38

OPB

, p. 11.

39

F. E. England, Kant’s Conception of God, London 1929, p. 50.

40

England, 1929, p. 55.

41

England, 1929, p. 54.

354

IAN LOGAN

identification of existence as a perfection or predicate contained in the
concept, but from the fact that its non-existence cannot be thought.

Kant soon showed signs of dissatisfaction with his own proof. Thus

in the Reflexionen from the mid- to late-1760s, he called into question
the conclusion of his own version of the ontological argument, for the
ens realissimum

might be a successive series (the World) rather than a

single being (God).

42

However, he does not develop such objections

explicitly in his published works. But not long after the publication of
the CPR, he makes clear the precise force that this proof has for him.
In the Lectures on the philosophical doctrine of religion (1783

⁄ 4) he is

reported as saying of the OPB argument:

43

‘‘Here it was shown that of

all possible proofs, the one which affords us the most satisfaction is the
argument that if we remove an original being, we at the same time
remove the substratum of the possibility of all things. But even this
proof is not apodictically certain; for it cannot establish the objective
necessity of an original being, but establishes only the subjective neces-
sity of assuming such a being.’’ However, Kant immediately goes on to
say that the proof is irrefutable, ‘‘because it has its ground in the nat-
ure of human reason.’’ As we shall see in the next section, it was
Kant’s unjustified empiricist assumption that led him to the view that
this argument did not establish God’s objective necessity. Of course,
there may have been other objections available to Kant here, but the
point is that Kant did not employ them, since they would put at risk
the coherence of his doctrine of the transcendental ideal.

2. Kant’s empiricist assumption

In the Prolegomena, Kant writes: ‘‘I freely admit: it was David Hume’s
remark that first, many years ago, interrupted my dogmatic slumber
and gave a completely different direction to my enquiries in the field of
speculative philosophy.’’

44

It was not Hume’s conclusions, but the ques-

tion he tackled, that was the cause of Kant’s ‘awakening’. This was the
question of a priori knowledge, and although Kant did not accept
Hume’s empiricist conclusion, ‘‘that all [reason’s] ostensibly a priori
knowledge is nothing but falsely stamped ordinary experiences,’’

45

it was

42

Reflexionen

, 3795: ‘‘nondum constat, utrum maxima realitas in ente compossibilis sit

simultaneo vel in serie successiva.

’’ (GW, XVII, 295.) See Sala, 1990, pp. 202f.

43

I. Kant, Lectures on the philosophical doctrine of religion in A. W. Wood &
G. Di Giovanni (Edd.), Immanuel Kant: Religion and Rational Theology, Cam-
bridge 1996, pp. 335-451, p. 357. (GW, XXVIII, 1034.)

44

I. Kant, Prolegomena to any future metaphysics that will be able to present itself as
a science

, translated from the German by P. Lucas, Manchester 1953, p. 9.

45

Prolegomena

, p. 6.

WHATEVER HAPPENED TO KANT

’

S ONTOLOGICAL ARGUMENT

?

355

the Humean insight concerning the problem with a priori knowledge
(in particular in relation to cause and effect) that awoke his interest
and admiration. Kant, however, took a different direction to Hume. For
him, concepts such as that of the connection between cause and effect
‘‘have their origin in pure understanding.’’

46

Nevertheless, in spite of

Kant’s rejection of Hume’s empiricism, somewhere along the line (per-
haps shortly after 1763

47

), he incorporates a fundamentally empiricist

principle into his philosophy. The effect of this is that Kant interprets
his own statement of this principle, ‘‘Our nature is so constituted that
our intuition can never be other than sensible

…,’’

48

in such a way that

all knowledge of objects is to be limited to as well as by sense experi-
ence.

49

It is this ‘limited to’ that needs to be established, rather than

assumed, if Kant is to succeed in the general task he sets himself in the
CPR

of establishing the illusory nature of speculative thought about

God, and the particular task of avoiding opinions and treating anything
‘‘which bears any manner of resemblance to an hypothesis’’ as ‘‘contra-
band.’’

50

As we have seen, the OPB had already opened the door to the

devastating effects of Kant’s empiricist principle, the roots of which are
already to be found in Kant’s understanding of the relationship between
possibility and thinkability, which meant that ‘‘for a reality to be an
object of thought it must be instantiated somewhere in an existing
thing.’’

51

However, in the OPB, such an existing thing was not yet

restricted to the realm of sensibility.

It is interesting to observe how in the CPR Kant seeks to manoeuvre

us into sharing his empiricist assumption: ‘‘We must therefore make
trial whether we may not have more success in the tasks of meta-
physics, if we suppose that objects must conform to our knowledge.’’

52

The effect of this Copernican turn is to ensure that all objects are
objects of knowledge and that objects of knowledge are limited to
objects of experience. By this simple move, objects that are not given
to us in experience are defined out of existence. As a consequence, in
the ‘Postulates of Empirical Thought’, Kant is able to conclude: ‘‘Our
knowledge of the existence of things reaches, then, only so far as per-
ception and its advance according to empirical laws can extend. If we

46

Prolegomena

, p. 10.

47

See for example Reflexion 3716 (1763

⁄ 66): ‘‘Die metaphysic ist nicht eine philosophie

u¨ber die obiecten, denn diese ko¨nnen nur durch die Sinnen gegeben werden, sondern
u¨ber das subiect, nemlich dessen Vernunftgesetze.’’ (GW, XVII, p. 259.)

48

B75.

49

See Reardon, 1988, p. 42.

50

Axv.

51

Wood, 1978, p. 67.

52

Bxvi.

356

IAN LOGAN

do not start from experience, or do not proceed in accordance with
laws of the empirical connection of appearances, our guessing or
enquiring into the existence of anything will only be an idle pretence.’’

53

But, as Allen Wood points out, even abstract and a priori arguments
for God ‘‘cannot be dismissed simply by appealing in some vague way
to an empiricist epistemology.’’

54

Wood goes on to say that Kant in

fact does not follow this route in trying to demonstrate the impossibil-
ity of any proof of God’s existence: ‘‘Instead [Kant] proceeds by divi-
ding all possible theistic proofs into kinds, and arguing in each case
that no successful proof of that kind can be given.’’ However, there is
at least one argument for God, which is excluded from Kant’s appar-
ently exhaustive list of all possible proofs in the CPR. Furthermore, it
is precisely this proof, which he had previously regarded as the only
possible proof in the OPB, that Kant dismisses by appealing to his
empiricist assumption, prior to his critique of the proofs of God in the
CPR

. In the next section we shall see in more detail how Kant’s empi-

ricist shift impacted on his ontological argument.

3. Kant’s treatment of his own ontological argument in the CPR and the

evolution of the ens realissimum into the transcendental ideal

In the CPR Kant reiterates his attack on the Cartesian form of the
ontological argument, but does not discuss his own 1763 version of the
argument, although he appears to allude to it in one place where
he says of the ontological argument that it ‘‘contains (in so far as a
speculative proof is possible at all) the one possible ground of proof
[den einzigmo¨glichen Beweisgrund] with which human reason can never
dispense.’’

55

However, this one possible proof refers to the argument of

Descartes, which he had been discussing a few pages previously. Thus,
although we might expect from the OPB that Kant would be referring
here to his own non-Cartesian ontological argument, he ignores that
argument and is only concerned with that of Descartes. There appears
to be a certain disingenuousness on Kant’s part here, since he had
already determined in the OPB that Descartes’ argument was ‘‘false
and completely impossible.’’

56

He now refers to that argument as the

only possible proof, although he has never previously considered it to
be such. In this way he is able to exclude his own version of the onto-
logical argument from the critique of the proofs of God, since it no
longer belongs to the realm of possible proofs. Furthermore, by stating

53

A226; B273f.

54

Wood, 1978, p. 96.

55

A625; B653.

56

OPB

, p. 237.

WHATEVER HAPPENED TO KANT

’

S ONTOLOGICAL ARGUMENT

?

357

that the argument of Descartes is the only possible proof Kant has
eliminated the need to explicitly refute his own or other versions of the
ontological argument, since only one proof is possible.

Kant’s treatment of the arguments for the existence of God in the

‘Transcendental Dialectic’ is well documented and it is not my inten-
tion to add yet another general discussion to this topic. Rather
I want to examine the ‘guise’ under which Kant’s own version of the
ontological argument is present in his ‘critical’ work, in order to see
whether Kant’s position has evolved in such a way as to undermine
the argument of the OPB. Certainly, between the production of the
OPB

and the CPR Kant had undergone a shift in his understanding

of the relationship between the necessity involved in judgements about
concepts and the necessity in things.

57

It was no longer possible from

his new perspective for judgements to attain to any existence uncon-
strained by sensibility. Existence can never be added to a concept by
a judgement in such a way that ‘renders’ the referent of the concept
actual.

The process by which Kant’s notion of the ‘transcendental ideal’

evolved has already been traced clearly in the work of Schmucker, Sala
and others.

58

This process was virtually complete by the mid-1760s and

is consequently ‘pre-critical’ in the sense the term is usually employed.

59

According to Schmucker,

60

the ‘‘decisive step’’ in this evolution is con-

nected with a problem left unresolved in the OPB: ‘‘how can the mater-
ial contents of possibilities as a consequence of the supreme infinite
Being be given to our thinking, if these possibilities must be conceived
as ontologically prior to the actual existence of the things themselves?’’
In Kant’s Reflexionen of the mid-1760s he resolves this problem: ‘‘that
the material contents of possibilities may be given to our thinking only
in and through the existing things themselves, so that our knowledge of
possibilities does not reach further than our knowledge of the objects
of experience. This means that practically speaking, possibility is noth-
ing other than a metaphysical dimension of the existing things.’’ As
Sala points out, for Kant our concept of what is really possible (vom
Realmo¨glichen) is dependent on experience: ‘‘With this new conception
of the possible, the basic presupposition of the OPB was abolished.’’

61

Or perhaps, it was not so much abolished as taken in a new, restrictive

57

Cf. A593; B621.

58

See note 2 above.

59

Schmucker, 1972, p. 496. See also Kreimendahl, 2003, p. xiii.

60

Schmucker, 1972, p. 497.

61

Sala, 1990, p. 205: ‘‘Mit dieser neuen Auffassung vom Mo¨glichen war die Grun-
dvoraussetzung des EmBg aufgehoben

.’’ (My translation.)

358

IAN LOGAN

direction, i.e., the positive, material element of Kant’s bi-polar notion
of possibility was restricted to the realm of sense experience. Thus, only
that which is given in sense experience is possible.

Kant’s empiricist shift is tied up with this ‘decisive step’ in the evo-

lution of the ‘transcendental ideal’. For objects of experience consti-
tute the material basis for all our knowledge, whether of principles or
of objects, and in the case of objects of knowledge constitute the set
of such knowledge. Thus knowledge can only be knowledge of princi-
ples or of objects of experience. It is the assumption that Kant adopts
in his empiricist shift that allows him to come to the final stage of
the development of the ‘transcendental ideal’, but it is only an
assumption.

In §9 of the 1770 Dissertation, Kant articulates this new perspective

in the doctrine of the theoretical ideal of pure reason, in which God is
not required to be an object, in order to function as a principle.

62

In

the ‘Ideal of Reason’ in the CPR, the evolution of the argument from
possibility is complete. The ens realissimum is no longer an object of
knowledge, ‘‘but rather the transcendental representation of an Ideal.
That is to say, the Prototypon Transcendentale is the condition for the
representation of all possibility, but it is nevertheless a merely subject-
ive representation.’’

63

What in 1763 was an argument that attained to

the ground of things from their possibility, now operates on a purely
conceptual level: ‘‘It was in this remodelled form, which was at the
same time a critique of the earlier ontotheology, that the ens realissi-
mum

was absorbed into the main theological section of the CPR.’’

64

Thus, implicit in the development of the transcendental ideal is a refu-
tation of the argument of the OPB, in so far as the latter claimed to
have established the objective existence of the ens realissimum. It is to
this refutation, which is fraught with dangers for Kant’s critical enter-
prise, that we shall turn in the next section, for, although he cannot
grasp the objective existence of the ens realissimum, he cannot simply
discard it.

65

62

I. Kant, On the form and principles of the sensible and the intelligible world [Inau-
gural Dissertation] (1770)

, in D. Walford (Ed.), Immanuel Kant: Theoretical philos-

ophy, 1755-1770

, Cambridge 1992, p. 388. (GW, II, p. 396.)

63

Dell’Oro, 1994, p. 167. See also W. Ro¨d, Der Gott der reinen Vernunft: Die
Auseinandersetzung um den ontologischen Gottesbeweis von Anselm bis Hegel

,

Munich 1992: ‘‘Ein Gottesbeweis mit Hilfe der Annahme objektiver ‘Mo¨glichkeiten’
la¨bt sich somit nicht mehr fu¨hren’’ (p. 133).

64

Sala, 1990, 209: ‘‘In dieser Form, die eine Umgestaltung und zugleich eine Kritik der
fru¨heren Ontotheologie ist, wurde dann die Ableitung des ens realissimum in das theo-
logische Hauptstu¨ck der KrV aufgenommen.’’ (My translation.)

65

See Sala, 1990, p. 213: ‘‘Wir brauchen ihn, ko¨nnen ihn aber nicht einsehen.’’

WHATEVER HAPPENED TO KANT

’

S ONTOLOGICAL ARGUMENT

?

359

4. The ‘refutation’ of Kant’s ontological argument

In the section on the ‘Transcendental Ideal’ Kant reiterates a version
of the proof from possibility. But on this occasion he does so to help
explain the illusion by which ‘‘we treat the empirical principle of our
concepts of the possibility of things, viewed as appearances, as being a
transcendental principle of the possibility of things in general.’’ This
illusion arises because we omit the fact that it is only objects of the sen-
ses that can be given to us ‘‘in the context of a possible experience.’’

66

However, it is not clear why we should accept that no objects ‘‘besides
those of the senses’’ could be given to us (even if we should accept that
they must be given in the context of possible experience). This is pre-
cisely the point to be established if Kant’s ‘refutation’ of his own onto-
logical argument is to work.

Kant had raised the question: ‘‘how does it happen that reason

regards all possibility of things as derived from one single fundamental
possibility, namely, that of the highest reality, and thereupon presuppo-
ses this to be contained in an individual primordial being?’’

67

In other

words how was he able to come to the conclusion of his own ontologi-
cal argument? He placed his answer firmly in the context of ‘‘the possi-
bility of the objects of the senses’’. In that context it is reasonable to
limit knowledge to objects of the senses, and thus to come to the con-
clusion that his argument ends in an illusion. But in order to establish
that it is an illusion he has to establish that the context of the possibil-
ity of the objects of the senses is the only possible context, and that the
possibility of objects, which are not limited to the senses, must be
excluded. But once again, we have to say that this is the point at issue.
If we are to capture this point, then we should render Kant’s argument
thus: ‘If no other objects, besides those of the senses, can, as a matter
of fact, be given to us, then nothing is an object for us, unless it presup-
poses the sum of all empirical reality as the condition of its possibility’.
The success of Kant’s ‘refutation’ of his own version of the ontological
argument is dependent on his empiricist assumption. But he fails to
establish this assumption, and consequently it is to be regarded as dog-
matic: ‘‘But when empiricism itself, as frequently happens, becomes
dogmatic in its attitude towards ideas, and confidently denies whatever
lies beyond the sphere of its intuitive knowledge, it betrays the same
lack of modesty

….’’

68

This assumption is an hypothetical, and as

Kant himself says in the Preface to the first edition of the CPR:

69

66

A582; B610.

67

A581; B609.

68

A471; B499.

69

Axv.

360

IAN LOGAN

‘‘Everything, therefore, which bears any manner of resemblance to an
hypothesis is to be treated as contraband.’’

The importance of the passage on the transcendental ideal, with its

implicit or indirect attempt at a refutation of the argument of the
OPB

, is easily missed. Thus Kemp Smith appears somewhat bemused

by it and ‘absolves’ the reader ‘‘from the thankless task of attempting
to render Kant’s argumentation in these paragraphs consistent with
itself,’’

70

apparently unaware of the Kantian basis of this argument,

believing rather that it derives from Wolff and Baumgarten. A failure
on the part of this refutation undermines the notion that the arguments
in the succeeding sections constitute the ‘‘only three possible ways of
proving the existence of God by means of speculative reason

’’

71

or that

the Cartesian ontological argument ‘‘contains (in so far as a speculative
proof is possible at all) the one possible ground of proof with which
human reason can never dispense.’’

72

The whole project of the Trans-

cendental Dialectic as Kant envisaged it falls apart if Kant is unable to
provide a non–hypothetical foundation for his empiricist principle,
since his rejection of the constitutive nature of the ens realissimum is
dependent on it. And without that rejection, ‘pre-critical’ metaphysics
remains potentially viable. Of course, it might be possible to refute
Kant’s ontological argument on other grounds. But any such refutation
would potentially leave the argument of the Transcendental Dialectic in
disarray, since it would undermine the coherence of the notion of the
ens realissimum

and hence the coherence of the notion of the transcen-

dental ideal.

Conclusion

In the New Elucidation and the OPB, Kant developed an ontological
proof of God as the ens realissimum and a refutation of other proofs,
including that of Descartes. Soon after completing the OPB, in the
mid-1760s, a decisive development took place in Kant’s thought, the
door to which was opened in the OPB, and which I have referred to as
his empiricist shift. Objects are objects of experience, and knowledge of
objects is limited by and to sense experience. The kind of reasoning
Kant had employed in the OPB could no longer be understood as able
to grasp objective reality, since such reality could only be grasped in
and by sense experience. As a consequence, by the time of the 1770
Dissertation

, Kant had developed the notion of the transcendental

70

N. Kemp Smith, A Commentary to Kant’s ‘Critique of Pure Reason’, 2

nd

ed., revised

and enlarged, Atlantic Heights, N.J., 1962, p. 525.

71

A590; B618.

72

A625; B653.

WHATEVER HAPPENED TO KANT

’

S ONTOLOGICAL ARGUMENT

?

361

ideal, which could account for the necessity in thought of what the old,
dogmatic metaphysics took to be God, in terms of a subjective neces-
sity governing the possibility of thought. The ens realissimum had
evolved from a constitutive object to a regulative idea. In the CPR, in
place of the proof of the objective existence of the ens realissimum, we
find an argument concerning the subjective necessity of the transcen-
dental ideal. In the CPR Kant incorporated his critique of the Cartes-
ian ontological argument, but did not provide an explicit refutation of
his own version of the argument. The requirement for such a refuta-
tion, Kant appears to have assumed, was vitiated by the introduction
of the transcendental ideal in the ‘Transcendental Dialectic’, the argu-
ment of which derives from his earlier thought and is independent of
the argument of the Transcendental Aesthetic and Analytic.

Kant’s refusal to give constitutive, rather than merely regulative, sub-

stance to what in the CPR had become the transcendental ideal was
based on his empiricist assumption. This assumption was the ‘dogmatic’
basis of his ‘refutation’ of his own non-Cartesian ontological argument
in ‘The Ideal of Pure Reason’. Unless he can successfully provide a refu-
tation of such ontological arguments, he is unable to justify (1) this
assumption and (2) the claim that the Cartesian ontological argument
constitutes ‘‘the only possible ground of proof’’ of God,

73

whose refuta-

tion undermines all other proofs of God. With regard to (1), it would
constitute no serious argument on behalf of Kant to say that an onto-
logical argument cannot be sound, because all metaphysical arguments
are limited to and by experience, since this is the point at issue. With
regard to (2), it would seem that Kant avoids attempting a direct refuta-
tion of his own ontological argument, precisely because he needs it to
establish the coherence of the notion of the transcendental ideal. In fact,
as we saw above, he continued to state that it was ‘‘irrefutable’’, even
after he had written the CPR. As a result, he fails to provide a compre-
hensive and successful refutation of the proofs of God in the CPR.

In sum, the answer to the question, ‘Whatever happened to Kant’s

ontological argument?’ is that, caught on the horns of a dilemma,

74

Kant both sought to discard this proof of the objective existence
of God, since knowledge cannot grasp that which is beyond sense

73

A625; B653.

74

The precise nature of the dilemma can perhaps be best expressed in terms of the
following argument: If Kant refutes his own ontological argument, then the neces-
sity of the transcendental ideal is not established, and the basis of his critique of
theology is undermined.
If Kant does not refute his own ontological argument, then his empiricist assump-
tion is not established, and the basis of his critique of theology is undermined.
Kant either does or does not refute his own ontological argument.
Therefore, the basis of his critique of theology is undermined.

362

IAN LOGAN

experience, and to retain it, in order to support the notion of the trans-
cendental ideal. However, he failed to provide a critical justification
for either move, and one must consequently conclude that the foun-
dations of his critique of theology were fundamentally flawed.

WHATEVER HAPPENED TO KANT

’

S ONTOLOGICAL ARGUMENT

?

363