You are on page 1of 24

British Journal for the History of Philosophy, 2015

Vol. 23, No. 5, 888–910, http://dx.doi.org/10.1080/09608788.2015.1047733

A RTICLE

KANT’S CRITICISMS OF HUME’S DIALOGUES


CONCERNING NATURAL RELIGION
Reed Winegar

According to recent commentators like Paul Guyer, Kant agrees with


Hume’s Dialogues Concerning Natural Religion (1) that physico-
theology can never provide knowledge of God and (2) that the
concept of God, nevertheless, provides a useful heuristic principle for
scientific enquiry. This paper argues that Kant, far from agreeing with
Hume, criticizes Hume’s Dialogues for failing to prove that physico-
theology can never yield knowledge of God and that Kant correctly
views Hume’s Dialogues as a threat to, rather than an anticipation of,
his own view that the concept of God provides a useful heuristic
principle for science. The paper concludes that Kant’s critique of
physico-theology reflects Kant’s deep dissatisfaction with Hume’s
manner of argumentation and suggests that Kant’s attempt to provide
a more successful critique of physico-theology merits continued
philosophical attention.

KEYWORDS: Kant; Hume; religion; God; science

1. INTRODUCTION

Hume’s 1779 Dialogues Concerning Natural Religion quickly piqued


Kant’s interest.1 Kant borrowed a private translation of the Dialogues
1
I provide author-title citations for all authors except Kant. Kant’s works are cited according to
the Akademie Ausgabe pagination with the exception of the Critique of Pure Reason for
which I provide the standard A/B pagination. I have utilized the translations in Kant (The
Cambridge Edition) when possible. All other translations are my own. I employ the following
abbreviations:

AA = Gesammelte Schriften [Akademie Ausgabe]


KpV = Critique of Practical Reason
KU = Critique of the Power of Judgement
Prol = Prolegomena to Any Future Metaphysics
Refl = Reflexionen
RGV = Religion within the Boundaries of Mere Reason
V-Met-L2/Pölitz = Metaphysik L2
V-Th/Baumbach = Danziger Rationaltheologie

© 2015 BSHP
KANT’S CRITICISMS 889

from J. G. Hamann on two separate occasions in 1780, and he acquired


Schreiter’s 1781 translation of the Dialogues almost immediately upon its
publication.2 Even though Kant refers to the Dialogues numerous times in
his writings, few commentators have attended to Kant’s reception of this
work.3 Those who have addressed this aspect of Kant’s thought have
largely emphasized similarities between the Dialogues and Kant’s own
views regarding theoretical philosophy. One prominent, recent example
of this trend is Paul Guyer, who states that Kant and Hume both deny
that physico-theology can yield knowledge of God but also both agree
that we should not banish God from theoretical philosophy entirely.4
More specifically, Kant claims that we cannot have knowledge of God
but also maintains that scientists should employ pure reason’s Idea of
God as a regulative, heuristic principle to guide empirical science. And
although Hume similarly denies knowledge of God, he also claims that
the imagination psychologically prompts belief in God and, like Kant, con-
strues this concept of God as a useful heuristic principle for empirical
science.5 According to Guyer, the most significant difference between
Kant and Hume regarding physico-theology actually concerns moral,
rather than theoretical, philosophy. Namely, Hume assigns no moral func-
tion to belief in God, but Kant links the concept of nature as a divinely

V-Th/Pölitz = Philosophische Religionslehre


2
For further historical details, see Hatfield (‘The Prolegomena and the Critiques of Pure
Reason’, 188n), Löwisch (‘Kants Kritik der reinen Vernunft und Humes Dialogues Concern-
ing Natural Religion’, 175ff.), Kemp Smith (‘Introduction’, 30), and Paulsen (‘Einleitung’,
26–7). The attribution of the 1781 translation to Schreiter remains conjectural; Kant still
owned a copy of this translation at his death (Warda, Immanuel Kants Bücher, 50).
3
The main discussions are Guyer (‘Natural Ends and Ends of Nature’; Knowledge, Reason,
and Taste), Hatfield (‘The Prolegomena and the Critiques of Pure Reason’), Kuehn
(‘Kant’s Critique of Hume’s Theory of Faith’, ‘The Reception of Hume in Germany’, ‘Von
der Grenzbestimmung der reinen Vernunft’), Logan (‘Hume and Kant on Knowing the
Deity’), Löwisch (‘Kants Kritik der reinen Vernunft und Humes Dialogues Concerning
Natural Religion’), and Santozki (Die Bedeutung antiker Theorien für die Genese und System-
atik von Kants Philosophie, 341ff.). For some passing references, see Brandt and Klemme
(David Hume in Deutschland, 7, 11), Byrne (Kant on God, 9–10, 26, 46), Gawlick and
Kreimendahl (Hume in der Deutschen Aufklärung, 183–5), Kemp Smith (‘Introduction’,
30), Maly (Kant über die symbolische Erkenntnis Gottes), McFarland (Kant’s Conception
of Teleology, 55–6, 79, 98, 105n, 107–8, 123, 126), Munzel (‘The Beautiful’, 304, 306),
Paulsen (‘Einleitung’, 26–7), Specht (Der Analogiebegriff bei Kant und Hegel, 34), and
Wolff (‘Kant’s Debt to Hume via Beattie’, 117).
4
I focus on Guyer as a prominent example of this longstanding trend because Guyer develops
this line of interpretation more fully than most commentators. Kuehn (‘Kant’s Critique of
Hume’s Theory of Faith’, ‘The Reception of Hume in Germany’, ‘Von der Grenzbestimmung
der reinen Vernunft’) also develops this line of interpretation in some detail, but Kuehn’s pos-
ition is close to that of Guyer. See notes 5, 6, 8, 13, and 32.
5
Guyer (Knowledge, Reason, and Taste, 208–9, 241); Kuehn (‘Kant’s Critique of Hume’s
Theory of Faith’, 252) also suggests that Hume might regard the concept of God as a heuristic
principle and would at least be open to this view.
890 REED WINEGAR

created, purposive system to his moral view of nature as directed towards


the highest good.6
Although I agree that Kant and Hume differ on the relationship between
physico-theology and morality, I will argue that there are two key respects in
which Guyer (along with similarly minded commentators) mistakenly exagger-
ates Kant and Hume’s agreement on the relationship between physico-theology
and theoretical philosophy.7 First, Guyer claims that Hume ‘thoroughly discre-
dited’ physico-theology and that Kant ‘agrees completely’ with Hume that
physico-theology ‘can never amount to theoretical cognition’ (Guyer, Knowl-
edge, Reason, and Taste, 241).8 But Kant does not completely agree with
Hume on this matter. Rather, Kant criticizes Hume’s Dialogues for failing to
prove that physico-theology can never yield knowledge of God.9 According
to Kant, Hume’s sceptical procedure in the Dialogues merely indicts the rational
credentials of particular physico-theological arguments and, thus, fails to prove
that no rational physico-theological arguments will ever be found.10 Moreover,

6
Guyer (Knowledge, Reason, and Taste, 252–3); Kuehn (‘Kant’s Critique of Hume’s Theory
of Faith’, 242, 248ff.; ‘The Reception of Hume in Germany’, 124ff.; ‘Von der Grenzbestim-
mung der reinen Vernunft’, 248) also emphasizes this difference.
7
I focus here on the role of God in Kant’s theoretical philosophy. Moral belief in God is
famously important to Kant, but space does not permit me to address this large topic.
8
For similar statements, see Kemp Smith (‘Introduction’, 30), Kuehn (‘Kant’s Critique of
Hume’s Theory of Faith’, 245–6), Logan (‘Hume and Kant on Knowing the Deity’, 144–
5), Löwisch (‘Kants Kritik der reinen Vernunft und Humes Dialogues Concerning Natural
Religion’), McFarland (Kant’s Conception of Teleology, 126), and Wolff (‘Kant’s Debt to
Hume via Beattie’, 117).
9
Although Guyer speaks of ‘theoretical cognition’, I will use the term ‘knowledge’. Kant
denies that we can have knowledge of God. However, Kant admits that we can think
(denken) about God from the theoretical perspective, and in the 1st Critique Kant even
claims that theoretical reason can ground doctrinal belief in God (A826/B854). There is a
longstanding question about how to reconcile Kant’s rejection of knowledge of the supersen-
sible (including knowledge of God’s existence) with Kant’s conviction that things-in-them-
selves are the grounds of appearances. Chignell (‘Belief in Kant’, 351) has recently
proposed that Kant’s commitment to things-in-themselves might be a case of ‘theoretical
belief’, rather than knowledge. Unfortunately, space does not permit me to evaluate this pro-
posal or address the many questions surrounding Kant’s commitment to things-in-themselves
as the grounds of appearances.
10
Chance (‘Kant and the Discipline of Reason’, 101) also notes Kant’s criticism that Hume
merely indicts the rationality of particular arguments. However, Chance does not acknowl-
edge Kant’s application of this point to Hume’s Dialogues. Instead, Chance takes Kant’s argu-
ment to focus solely on the Enquiry Concerning Human Understanding. Chance objects to
Kant’s criticism, because the Enquiry

attempts to show that reason is not the source of our conclusions from experience not
by examining individual actions of reason but by establishing the general claim that
there is no noncircular argument for the claim that the future will resemble the past.
(Chance, ‘Kant and the Discipline of Reason’, 102; emphasis in original)

But this misses Kant’s point. As I discuss in Section 2 below, Kant takes Hume’s theory of
causation (including its general claim that we cannot rationally demonstrate that the future
KANT’S CRITICISMS 891

Kant does not find Hume’s indictment of physico-theology’s rational creden-


tials very persuasive. Rather, Kant claims against Hume that physico-theologi-
cal argumentation rationally favours the view that nature is the product of
theistic intelligence (V-Met-L2/Pölitz, AA 28:604–5, V-Th/Pölitz, AA
28:1063–4; V-Th/Baumbach, AA 28:1277–8).11 However, this point does not
entail that Kant lapses into pre-critical metaphysics regarding physico-theology.
Instead, as I will illustrate below, Kant provocatively suggests that theoretical
inferences to the supersensible can qualify as rational yet fail to yield knowl-
edge. Thus, while Hume indicts the rationality of particular physico-theological
arguments, Kant argues that not even a rational physico-theological inference to
a supersensible, theistic intelligence would yield knowledge of God.12
Second, Guyer claims that Kant and Hume agree that the concept of God is
a useful heuristic for empirical science. As Guyer writes, ‘Kant will clearly
agree with Hume that the conception of God is the source of the useful strat-
egies for the investigation of nature … ’ (Guyer, Knowledge, Reason, and
Taste, 209). But a reexamination of Hume’s Dialogues reveals that Hume
adopts no such view. Rather, Hume aims to show that the concept of God
is indeterminate and, thus, useless for all human activities, including scien-
tific enquiry. Moreover, Kant explicitly identifies Hume’s Dialogues as a
threat to his own view that the Idea of God should serve as a heuristic
will resemble the past) to cast doubt on the cosmological argument, but the cosmological argu-
ment represents only one particular argument for the existence of God.
11
In the 3rd Critique, Kant limits the rationality of the physico-theological inference from
empirically observed order to intelligence to the inference from empirically observed
organic order to intelligence (KU, AA 5:437). The 3rd Critique’s published Introduction indi-
cates that we should represent a divine mind as the ground of a system of particular laws (KU,
AA 5:180). But this system of particular laws is a regulative principle of reflecting judgment,
rather than an empirically observed fact. In earlier writings, Kant seems to license the ration-
ality of the physico-theological inference from many aspects of natural order, such as nature’s
‘manifoldness, order, purposiveness, and beauty’ (A622/B650). In this essay, I will speak in
general terms about nature’s order, rather than specifically about organic order.
12
Wood (Kant’s Rational Theology, 76) notes that Kant refers to the ontological proof as sub-
jectively necessary, because it satisfies a demand of human reason but does not yield knowl-
edge. Similarly, Chignell (‘Belief in Kant’) argues that Kant takes the cosmological proof’s
inference to a necessary being to be rational but insufficient for knowledge. My view is
that Kant also takes the physico-theological proof’s inference to intelligence to be rational
(because it satisfies human reason’s demand for intelligibility) but insufficient for knowledge.
Chignell (‘Belief in Kant’, 346–7) emphasizes that the 1st Critique takes reason’s interest in
acquiring scientific cognition to ground doctrinal belief in God’s existence, because doctrinal
belief in God allows scientists to regard nature as a teleologically ordered system, which facili-
tates further scientific discoveries. Thus, doctrinal belief is rational. Strictly speaking, this
differs from (but is compatible with) my view, because doctrinal belief goes from the assertion
of an intelligent world-cause to the anticipation of natural order, while I hold that Kant also
endorses the rationality of the physico-theological inference from nature’s empirically
observed order to intelligence. But, like me, Chignell (‘Modal Motivations’, 596) and Paster-
nack (‘Regulative Principles and “the Wise Author of Nature”’, 415) also seem to take the 1st
Critique to maintain that the physico-theological inference to some intelligence is rational but
fails to yield knowledge. However, they do not explicitly relate this point to Kant’s evaluation
of Hume’s Dialogues.
892 REED WINEGAR

principle for empirical science, and Kant responds to this threat by appealing
to a theory of analogy that relies on his claim that the physico-theological
inference to theistic intelligence is rational but does not yield knowledge
(Prol, AA 4:356, 358–9; Refl 6136, AA 18:466).13
Therefore, Hume’s Dialogues are deeply at odds with Kant’s theoretical
philosophy. Of course, some aspects of Kant’s views regarding physico-
theology are likely to strike contemporary philosophers as antiquated. But,
as I will ultimately suggest, contemporary critics of the design argument
who share Kant’s conviction that knowledge of God is impossible might
benefit from supplementing familiar Humean attacks against the design argu-
ment’s rational credentials with Kant’s more radical view that not even a
rational physico-theological argument would yield knowledge of God.

2. KANT ON HUME’S RELIGIOUS SCEPTICISM

Allow me to begin by examining Kant’s reaction to the Dialogues’ scepti-


cism regarding knowledge of God.14 Kant maintains that scepticism is not
a permanent resting place for reason but, instead, a stimulus for critique
(Aix–xii). Kant presses this point against Hume’s Dialogues in the 1st Cri-
tique’s section on ‘The Discipline of Pure Reason’, where Kant introduces
Hume’s religious scepticism as follows:

If one were to ask the cool-headed David Hume, especially constituted for
equilibrium of judgment, ‘What moved you to undermine, by means of

13
Guyer does not discuss Kant’s theory of analogy in relation to Hume, but Kuehn (‘Kant’s
Critique of Hume’s Theory of Faith’, 248; ‘Von der Grenzbestimmung der reinen Vernunft’,
241–2), Logan (‘Hume and Kant on Knowing the Deity’, 137–41), Löwisch (‘Kants Kritik der
reinen Vernunft und Humes Dialogues Concerning Natural Religion’, 191ff.), and Maly (Kant
über die symbolische Erkenntnis Gottes) do. Yet, even these commentators do not seem to
recognize that Hume’s Dialogues threatens the Idea of God as a heuristic principle. For
example, Kuehn (‘Kant’s Critique of Hume’s Theory of Faith’, 252) finds ‘no reason to
think that Hume would have found major problems’ with Kant’s regulative employment of
the Idea of God for scientific purposes and, like Guyer, suggests that Hume might actually
hold such a view himself. More recently, Kuehn (‘Von der Grenzbestimmung der reinen Ver-
nunft’, 248) claims to doubt that Hume would accept Kant’s theory of analogy. But I do not
think that this revises Kuehn’s previous view. Kuehn does not explain why Hume would reject
Kant’s theory of analogy except by pointing to Kuehn (‘Kant’s Critique of Hume’s Theory of
Faith’, 240), which describes Hume’s view that the imagination, rather than reason, is the
source of belief in God. Presumably, then, Kuehn thinks that Hume disagrees with Kant’s
theory of analogy to the extent that Hume disagrees with Kant on the source of the concept
of God but simultaneously agrees with Kant that the concept of God is a useful heuristic prin-
ciple for science.
14
The straightforward identification of Philo as Hume’s spokesman is controversial. However,
eighteenth-century readers generally regarded the sceptic Philo as Hume’s spokesman in the
Dialogues (Gawlick and Kreimendahl, Hume in der Deutschen Aufklärung, 79). And this was
Kant’s interpretation too (Prol, AA 4:358, 360).
KANT’S CRITICISMS 893

reservations brooded on with so much effort, the persuasion, so comforting


and useful for humans, that the insight of their reason is adequate for the asser-
tion and determinate concept of a highest being?’, he would answer: ‘Nothing
but the intention of bringing reason further in its self-knowledge … ’
(A745/B773)

Although the 1st Critique does not explicitly mention Hume’s Dialogues,
Dieter-Jürgen Löwisch has shown that this passage constitutes a defence of
Hume’s personal character against Christoph Meiners’s negative 1779
review of the Dialogues.15 Meiners cites ‘vanity’ as one of Hume’s main
motivations for undermining religious faith, and he suggests that the Dialo-
gues represent a stain on Hume’s personal character (Meiners, ‘Review of
Dialogues Concerning Natural Religion’, 762).16 In opposition to
Meiners, Kant attributes Hume’s religious scepticism to the nobler goal,
which Kant shares, of bringing reason to ‘self-knowledge’ (A745/B773).
Yet, despite applauding Hume’s intentions, Kant deems Hume’s attempt
to bring reason to self-knowledge a failure.17 According to Kant, Hume
attempts to dispose of religious questions by placing them beyond the
horizon of human reason (A760/B788). For example, Hume’s theory that
the concept of causation is grounded in custom or habit, rather than
reason, denies any extension of the causal principle beyond common life
and, thus, threatens any attempt to use the causal principle to acquire knowl-
edge of God. Kant clearly identifies this implication of Hume’s position in
the 2nd Critique, where he writes:

Hume … asked nothing more than that a merely subjective meaning of neces-
sity, namely custom, be assumed in place of any objective meaning of neces-
sity in the concept of cause, so as to deny to reason any judgment about God,
freedom, and immortality …
(KpV, AA 5:13)18

15
Löwisch (‘Kants Kritik der reinen Vernunft und Humes Dialogues Concerning Natural Reli-
gion’, 177ff.) does not identify Meiners as the anonymously published review’s author, but
Gawlick and Kreimendahl (Hume in der Deutschen Aufklärung, 79) and Kuehn (‘The Recep-
tion of Hume in Germany’) do. As Löwisch (‘Kants Kritik der reinen Vernunft und Humes
Dialogues Concerning Natural Religion’, 177n) also notes, Kant’s reference to ‘the cool-
headed David Hume, especially constituted for equilibrium of judgment’ recalls a description
of Hume as ‘by nature calm and cool-headed’ in another anonymous review of the Dialogues
(Anonymous, ‘Review of Dialogues Concerning Natural Religion’, 71).
16
Cf. Anonymous (‘Review of Dialogues Concerning Natural Religion’, 70), which cites
Hume’s ‘inordinate striving after literary fame’.
17
Chance (‘Kant and the Discipline of Reason’), Kuehn (‘Kant’s Critique of Hume’s Theory
of Faith’, 245), Stern (‘Metaphysical Dogmatism, Humean Scepticism, Kantian Criticism’),
and Watkins (Kant and the Metaphysics of Causality, 374ff.) note this point.
18
Hatfield (‘The Prolegomena and the Critiques of Pure Reason’), Kuehn (‘Kant’s Critique of
Hume’s Theory of Faith’, ‘The Reception of Hume in Germany’), and Stern (‘Metaphysical
Dogmatism, Humean Scepticism, Kantian Criticism’, 109, 111) also emphasize Kant’s inter-
est in the religious implications of Hume’s theory.
894 REED WINEGAR

Although Kant ultimately agrees with Hume that the causal principle
cannot provide knowledge of God, Kant objects that Hume’s own arguments
do not successfully justify this conclusion. Kant explains this objection in
terms of a contrast between a censure of reason and a critique of reason.
According to Kant, Hume’s philosophy censures reason. Kant writes,
‘One can call a procedure of this sort, subjecting the facta of reason to exam-
ination and when necessary to blame, the censure [Zensur] of reason’
(A760/B788; translation modified; emphasis in original).19 Kant’s refer-
ences to the facta of reason and to the censure of reason are legal metaphors
that belong to Kant’s metaphorical description of the 1st Critique as a court
of law throughout ‘The Discipline of Pure Reason’. Here, a factum is a state-
ment of the alleged facts that serves as the ground for the judge’s enquiry,
and a censure is an official rebuke.20 Kant’s claim that Hume subjects ‘the
facta of reason to examination’ means that Hume acts as a judge who con-
siders and, then, passes a negative judgement on particular arguments
regarding God. For example, Hume concludes that the cosmological argu-
ment fails, because we cannot rationally infer that the world requires any
cause at all.
But, as Kant notes, the failure of a few particular arguments cannot
persuade the dogmatist that all future attempts will similarly fail. Rather,
the dogmatist will just hope to find better success in new arguments. Kant
writes:

All failed dogmatic attempts of reason are facta, which it is always useful to
subject to censure. But this cannot decide anything about reason’s expec-
tations of hoping for better success in its future efforts and making claims
to that …
(A763-4/B791-2)

In fact, the temptation to think that she might find better success in future
arguments leads the dogmatist, in turn, to doubt Hume’s success at arguing
for religious scepticism. Kant writes:

the same thing happens to him [i.e., Hume] that always brings down skepti-
cism, namely, he is himself doubted, for his objections rest only on facta
which are contingent, but not on principles that could effect a necessary
renunciation of the right to dogmatic assertions.
(A767-8/B795-6)

19
Guyer and Wood translate ‘Zensur’ here as ‘censorship’, but elsewhere they translate it as
‘censure’, for example, at A764/B792. I believe that the context of Kant’s discussion
favours ‘censure’ as a translation to capture the concept of a negative legal judgement.
20
Chance (‘Kant and the Discipline of Reason’, 102) interprets facta here as individual acts of
reason. Although the Latin ‘factum’ can mean act or deed, Chance overlooks how the term
‘facta’ relates here to Kant’s extended legal metaphor throughout ‘The Discipline of Pure
Reason’, which presents the Critique of Pure Reason as a court of law.
KANT’S CRITICISMS 895

In other words, the dogmatist recognizes that Hume has merely objected to
particular arguments and, thus, has not provided a sufficient reason to con-
clude that new arguments will inevitably fail. Thus, the dogmatist will
doubt Hume’s claim that no good arguments can ever be found and will
attempt to discover new arguments.21
To overcome these deficiencies with Hume’s position, Kant maintains that
we need a critique of reason, rather than a mere censure of reason, to prove
that reason lacks any right to knowledge of God. As Kant explains:

[Critical philosophy] subjects to evaluation not the facta of reason but reason
itself, as concerns its entire capacity and suitability for pure a priori cogni-
tions; this is not the censure but the critique of pure reason, whereby not
merely limits but rather the determinate boundaries of it – not merely ignor-
ance in one part or another but ignorance in regard to all possible questions of
a certain sort – are not merely suspected but are proved from principles.
(A761/B789; emphasis in original)

In other words, Kant’s critique of reason determines the boundaries of


human reason and proves that certain kinds of knowledge, including knowl-
edge of a supersensible God, are outside those boundaries. To continue
Kant’s legal metaphor, the critique of reason proves that theoretical reason
lacks any right to knowledge of the supersensible and that any theoretical
claim to knowledge of God is, thus, unlawful.
As we can see from the passages above, the 1st Critique’s discussion of
Hume’s religious scepticism focuses on Hume’s theory of causation. This
might not seem to sit well with Kant’s introductory allusion to the Dialo-
gues. Hume emphasizes his theory of causation in the Treatise of Human
Nature and Enquiry Concerning Human Understanding. But he does not
draw much attention to it in the Dialogues, and most of his arguments in
the Dialogues do not rely on it.22 Kant’s texts do not directly address this
apparent discrepancy. But I suggest that we can reconcile the matter as
21
Kant cites Sulzer, who edited the German translation of Hume’s Enquiry, as such a dogma-
tist (A741–2/B769–70). Stern (‘Metaphysical Dogmatism, Humean Scepticism, Kantian Cri-
ticism’, 109) interprets Kant to imply that Hume’s scepticism fails to convince, because the
causal principle strikes people as obvious, but I do not see how the last two quoted passages
in the main text could support this interpretation. Watkins interprets Kant to imply that
Hume’s scepticism is self-undermining, because a thoroughgoing scepticism requires that
one be sceptical about one’s scepticism (Watkins, Kant and the Metaphysics of Causality,
377). Again, I do not see how the last two quoted passages in the main text could support
such an interpretation. I cannot tell whether Kuehn’s (‘Kant’s Critique of Hume’s Theory
of Faith’, 245) remark that ‘Hume’s attempt to show our unavoidable ignorance by appeal
to contingent facts is insufficient’ anticipates my position or not. But I suspect not, because
Kuehn subsequently claims, ‘I doubt that Kant’s arguments would have been sufficient to
require Hume to modify his position on religious faith’ (Kuehn, ‘Kant’s Critique of
Hume’s Theory of Faith’, 252).
22
For this reason, Löwisch suggests the patchwork thesis that Kant originally composed the
section of ‘The Discipline of Pure Reason’ that refers to Hume’s theory of causation and
896 REED WINEGAR

follows. Hume’s contemporaries, including Kant, recognized that Hume’s


theory of causation threatened the cosmological argument for God’s exist-
ence. But the cosmological argument is just one particular argument.
Indeed, Hume’s characters in the Dialogues consider a long stream of par-
ticular arguments regarding God, including the cosmological argument,
and assess the merits of each particular argument in turn. Thus, Kant’s
general assessment of Hume’s religious scepticism applies to the Dialogues
quite well. After all, Kant’s assessment is that Hume evaluates particular
arguments regarding God but does not prove that knowledge of God is
impossible in principle. Although Hume’s characters cast doubt on many
particular arguments (including the cosmological argument), they arguably
fail to prove that any new arguments would inevitably fail.23 In fact, far
from agreeing with Kant that reason lacks any right to knowledge of God,
Hume’s apparent spokesman, Philo, argues (despite his sceptical tendencies
elsewhere in the Dialogues) that we should conclude from the world’s evil
that the source of the universe most likely lacks any human-like moral attri-
butes (Hume, Dialogues Concerning Natural Religion, 74).24
Of course, one could partially defend Hume against Kant’s criticism by
suggesting that the Dialogues refute many of the most powerful traditional
physico-theological inferences to theistic intelligence and, thus, strike a sub-
stantial (even if not totally devastating) blow against physico-theology. But,
interestingly, Kant will not grant Hume even this much. Instead, Kant argues
that physico-theology rationally favours the conclusion that nature is the
product of a theistic intelligence. A long remark in the Danziger Ratio-
naltheologie transcript of Kant’s lectures on religion makes this point
especially clear.25 We read:

David Hume in his Dialogues concerning Natural Religion has disputed the
[physico-theological] proof. His ground is this, that the mechanism of
nature produces purposiveness, order, even the human being’s understanding;
but understanding cannot make another understanding. He says he does not at
all comprehend how such a perfect understanding, as God is, could have come
into being, than he comprehends how the mechanism of nature produces so
much beauty. One is just as unbelievable as the other.–

later inserted the allusion to Hume’s Dialogues (Löwisch, ‘Kants Kritik der reinen Vernunft
und Humes Dialogues Concerning Natural Religion’, 179).
23
Perhaps one could argue that the Enquiry and Treatise provide Hume with further materials
for a more systematic refutation of the possibility of knowledge of God. But because our focus
is on Hume’s Dialogues, I will set this question aside.
24
See Gaskin (Hume’s Philosophy of Region, 68) for this interpretation. Holden (Spectres of
False Divinity) challenges this interpretation of Philo’s argument. But Holden still maintains
that Hume is happy to argue from experience for probabilistic claims about what any unknown
object, including God, is likely to be (Holden, Spectres of False Divinity, 32–6). Thus, Holden
still attributes to Hume the view that some physico-theological inferences are rational.
25
All of the transcripts of Kant’s lectures on religion cited in this essay stem from Kant’s 1783/
84 course (Kreimendahl, ‘Kants Kolleg über Rationaltheologie’).
KANT’S CRITICISMS 897

The question is whether we better comprehend how a most perfect under-


standing is possible than how a mechanism of nature could produce that.
The existence of a being in which all perfections are united, is not possible to
comprehend; but it is however not impossible.–
In contrast a nature without understanding is still less comprehensible; even
impossible; because the world is an aggregate of many substances, thus has
no understanding. One cannot at all imagine, that from the aggregate of
things without understanding an understanding arises. Nature can have no
principle of its perfection in itself.
Physico-theology thus has an entirely secure basis; we can namely infer from
the order of the world to a being with understanding.
(V-Th/Baumbach, AA 28:1277-8)26

Here Kant expresses his dissatisfaction with one of Philo’s most famous
objections to physico-theology. According to Philo, any attempt to explain
nature’s order in terms of a divine mind requires that we explain the order
of ideas in this divine mind. If we explain this order of ideas by claiming
that the divine mind in question is itself the product of some other divine
mind with orderly ideas, then we enter into a vicious infinite regress. But
if we claim that the principle of the divine mind’s order is within the
divine mind itself, then why can we not claim that the principle of
nature’s order is within nature itself and, thus, eliminate the inference
from nature’s order to a divine mind? (Hume, Dialogues Concerning
Natural Religion, 31).
Note that Kant responds to Philo’s objection by flatly denying Philo’s sug-
gestion that nature could contain the principle of its order within itself. In the
quoted passage, Kant makes this point by abruptly stating that we ‘cannot at
all imagine’ how human understanding could originate from a blind prin-
ciple within nature. And in the Philosophische Religionslehre transcript,
Kant also dismisses Philo’s objection as absurd. We read, ‘but what about
the whole of things, the totality of the world? Is it therefore generated by
some fertile cause? What a sophistry!’ (V-Th/Pölitz, AA 28:1064).
Obviously, Kant’s replies here are brief and would require significant

26
One might worry that the lecture transcripts are not reliable or reflect textbooks used in class.
However, we can compare this transcript with the following handwritten Reflexionen that
mention Hume. Reflexion 6045 reads, ‘The necessity of an author [Urheber] distinct from
the world is inferred from [aus … geschlossen] the contingency of order … ’ (Refl 6045,
AA 18:432). This echoes the final sentence in the lecture transcript passage. Reflexion 6136
states, ‘the things in nature cannot be represented as acting on each other purposively on
their own’ (Refl 6136, AA 18:466). This echoes the lecture transcript’s claim, ‘Nature can
have no principle of its perfection in itself’. Finally, Eberhard’s Vorbereitung zur natürlichen
Theologie, which Kant used in class, encourages the reader to refute Hume’s Dialogues, but
Eberhard presents no explicit reply to the Dialogues (Eberhard, Vorbereitung zur natürlichen
Theologie zum Gebrauch akademischer Vorlesungen, §27). Thus, I take the transcript’s criti-
cisms to be Kant’s own.
898 REED WINEGAR

elaboration in order to be evaluated. But the important point for our purposes
is simply to note that Kant dismissively rejects Philo’s alternative expla-
nation of nature’s order.27 According to Kant, Philo does not illustrate
how we can conceive the principle of nature’s order as existing in nature
itself; thus, Philo’s alleged alternative does not seriously challenge the
rationality of the physico-theological inference to intelligence.
Of course, Philo might reply to Kant by stressing that we also cannot com-
prehend the principle of the divine mind’s order. However, Kant deflects this
criticism by arguing that we do not actually need to explain the source of the
divine mind’s order.28 According to Kant, the Idea of God (unlike nature)
rests outside the boundary of human reason. Consequently, we can neither
objectively prove nor objectively refute God’s necessary existence.
Because we cannot prove God’s necessary existence, we lack the traditional
explanation of God’s perfections and their unity – namely, that they exist by
God’s absolute necessity. But because God’s necessary existence also
cannot be refuted, we cannot exclude the real possibility of God’s necessary
existence and, thus, cannot exclude the real possibility that God’s perfections
and their unity do exist by God’s absolute necessity. As the Philosophische
Religionslehre transcript relates:

And if we ask how this supreme being has sufficient perfections and whence it
gets them, the answer can be only that they follow from its absolute necessity
– into which, to be sure, on account of all the limitations of my reason I really
have no insight, but which for the same reason I also cannot deny.
(V-Th/Pölitz, AA 28:1064–5; emphasis in original)

Here Kant grants that we cannot comprehend how God could exist of
absolute necessity; thus, God’s existence as a being in which all perfections
are united might seem to require further explanation. However, this demand
for further explanation is misplaced. Our inability to comprehend God’s
absolute necessity merely reflects the subjective limitations of human under-
standing. God, along with the unity of God’s perfections, might exist of
absolute necessity, even if we cannot comprehend how such an existence
is possible.
As we have seen, Kant is antagonistic towards many aspects of Hume’s
Dialogues. But, obviously, Kant could have been impressed by some of
Hume’s arguments and less impressed by others. And although we have
seen Kant argue against Philo that physico-theology rationally favours an

27
Kant was hardly alone in his negative assessment of Philo’s arguments. Hume’s German
readers were mostly unimpressed by the Dialogues (Gawlick and Kreimendahl, Hume in
der Deutschen Aufklärung, 131ff.; Kuehn, ‘The Reception of Hume in Germany’).
28
Cf. KU, AA 5:420–1, where Kant’s response to Hume focuses primarily on God’s simplicity
as a substance. In the lectures, Kant is more sensitive to the fact that referring to God’s sim-
plicity does not completely answer the question of how God came to have the full set of per-
fections required to ground nature’s order.
KANT’S CRITICISMS 899

inference to intelligence, it is important to note that the 1st Critique’s section


‘On the Impossibility of a physico-theological Proof’ does at least concur
with Hume that physico-theological argumentation can prove neither that
nature is the product of a single intelligence (rather than multiple intelli-
gences) nor that this intelligence is omnipotent, omniscient, etc. (rather
than just pretty strong, rather knowledgeable, etc.) (A627-8/B655-6; cf.
KU, AA 5:437ff.). In other words, Kant concedes Hume’s suggestion that
physico-theology cannot rationally infer the existence of a single, supreme
intelligence, but Kant also maintains against Hume that the inference from
nature’s order to the existence of some kind of intelligence is rational.
Indeed, in the 1st Critique, Kant claims that he has ‘nothing to object
against the rationality and utility’ of the physico-theological inference to
some kind of intelligence (A624/B652; emphasis added).
Of course, Kant’s assertion that an inference from the order of nature to
some kind of intelligence is rational might seem like a surprising relapse
into pre-critical metaphysics.29 The long quotation from the Danziger Ratio-
naltheologie above might reinforce this impression when it concludes by
remarking, ‘Physico-theology thus has an entirely secure basis … ’ And
the 1st Critique’s claim that it has ‘nothing to object against the rationality’
of the physico-theological inference might seem to do so as well. But how
should we interpret such remarks? Do the 1st Critique and lectures on reli-
gion take dogmatic physico-theology to prove that some kind of intelligent
world-cause exists and to deny merely that physico-theology proves this
intelligence to be a single, supreme being? I do not think so. Kant does
not dispute the rationality of inferring from nature’s order to intelligence,
but a rational inference might still fail to yield knowledge if reason itself
were to lack any right to such knowledge. Such a suggestion would seem
to represent the logical culmination of Kant’s view in ‘The Discipline of
Pure Reason’. According to ‘The Discipline of Pure Reason’, the best way
to prevent dogmatists from advancing physico-theological arguments is
not to criticize the rational credentials of each particular argument but,
rather, to show that theoretical reason lacks any right to knowledge of the

29
Wood (Kant’s Rational Theology, 131) claims that Kant grants this inference’s rationality
merely for the sake of argument, because he mainly wants to prove that physico-theology
cannot yield knowledge of a single, supreme intelligence. Byrne (Kant on God, 38) and Pas-
ternack (‘Regulative Principles and “the Wise Author of Nature”’, 425n) argue that Kant’s
language is more approving of the physico-theological inference than Wood’s interpretation
permits. Byrne attributes this approval to the suggestion that physico-theology vivifies
Kant’s moral argument for belief in God. Like me, Chignell (‘Modal Motivations’, 596)
and Pasternack (‘Regulative Principles and “the Wise Author of Nature”’, 415) appear to
take the 1st Critique to maintain that the physico-theological inference to some intelligence
is rational but fails to yield knowledge. Although he briefly discusses Kant’s 3rd Critique, I
cannot tell whether Pasternack (who is mostly interested in arguing that the 3rd Critique
excludes doctrinal belief) takes the 3rd Critique to maintain, as I do, that the physico-theolo-
gical inference is rationally incumbent upon us (in order to satisfy reason’s demand for
intelligibility).
900 REED WINEGAR

supersensible. Even if physico-theology rationally leads to the Idea of an


intelligent ground of nature, Kant need not claim that this rational inference
yields knowledge of an intelligent ground of nature. For reason itself might
lack the right to such knowledge.
But does Kant hold such a view? Kant’s remarks in the 1st Critique and
lecture transcripts are brief and ambiguous. But such a position does seem
to represent Kant’s considered view in the 1783 Prolegomena, which
dates from roughly the same time as the lectures and only two years after
the 1st Critique. In a section of the Prolegomena that explicitly refers to
Hume’s Dialogues, Kant writes:

we think of the world AS IF it derived from a supreme reason as regards its


existence and inner determinations … we in part posit the basis of this consti-
tution (the rational form of the world) in relation to the highest cause of the
world, not finding the world itself sufficient thereto.
(Prol, AA 4:359-60; emphasis altered)

Here Kant claims that we cannot rationally regard the world itself as a suf-
ficient cause of nature’s order. Thus, we are rationally driven from nature’s
order to the Idea of an intelligent ground. But this rational inference does not
entitle us to claim that the world is, in fact, a product of intelligence. Rather,
we are merely ‘to think of the world AS IF it derived from a supreme reason’
(Prol, AA 4:359). This position becomes even more pronounced in the 3rd
Critique, where Kant claims that the discursive character of human under-
standing requires us to regard an intelligent world-cause as the ground of
the organic order found in nature. Kant writes:

It [i.e., physico-theology] can thus certainly justify the concept of an intelli-


gent world-cause, as a merely subjectively appropriate concept for the consti-
tution of our cognitive faculty of the possibility of the things that we make
intelligible to ourselves in accordance with ends …
(KU, AA 5:437)

Here Kant notes that, due to the discursive character of human understand-
ing, any attempt to render the existence of organic order intelligible to our-
selves will require appeal to an intelligent world-cause. Given human
reason’s demand for intelligibility, the inference from organic order to an
intelligent ground is, thus, rational for discursive creatures like us.30 But

30
I take this to be behind Kant’s point when he writes,

Now here this maxim is always valid, that even where the cognition of them outstrips
the understanding, we should conceive all objects in accordance with the subjective
conditions for the exercise of our faculties necessarily pertaining to our (i.e., human)
nature … .
(KU, AA 5:403)
KANT’S CRITICISMS 901

the 3rd Critique never claims that physico-theology yields knowledge of


God. Rather, Kant’s considered view in the 3rd Critique is that the
physico-theological inference from the existence of organic beings to an
intelligent ground of nature is ‘merely subjectively appropriate’ for discur-
sive creatures like us and does not yield knowledge.31 Thus, unlike Hume,
who simply criticizes the rationality of particular physico-theological argu-
ments, Kant maintains that knowledge of God is impossible in principle,
because even rational physico-theological inferences to an intelligent
ground of nature fail to yield knowledge.

3. KANT’S REGULATIVE IDEA OF GOD

Kant famously claims that we cannot know whether God exists but also
argues that we should employ the Idea of God as a heuristic principle that
regulates empirical science. More specifically, Kant claims that viewing
nature as if it were the creation of a theistic intelligence helps guide the dis-
covery of nature’s systematic order. As mentioned previously, Guyer also
takes Hume to claim that we should use the concept of God as a heuristic
principle for empirical science. Therefore, on Guyer’s view, Kant and
Hume agree that the concept of God is a useful heuristic principle for
science. Guyer bases this interpretation on a well-known passage from
Part 12 of the Dialogues, where Philo concedes that humans cannot help
but regard nature’s order as the product of intelligence.32 Philo remarks:

A purpose, an intention, a design strikes everywhere the most careless, the


most stupid thinker, and no man can be so hardened in absurd systems as at
all times to reject it. That nature does nothing in vain is a maxim established
in all the Schools, merely from the contemplation of the works of nature,
without any religious purpose; and, from a firm conviction of its truth, an anat-
omist who had observed a new organ or canal would never be satisfied till he
had also discovered its use and intention … The same thing is observable in
other parts of philosophy; and thus all the sciences must lead us insensibly
to acknowledge a first intelligent Author …
(Hume, Dialogues Concerning Natural Religion, 77–8; emphasis in original)

Given Philo’s sceptical posture earlier in the Dialogues, this embrace of


belief in God’s intelligence has puzzled many of Hume’s readers. Several
31
Cf. RGV, AA 6:65n for a similar statement in terms of a need for comprehensibility.
32
In defence of a similar point, Kuehn (‘Kant’s Critique of Hume’s Theory of Faith’, 252)
refers to the Natural History of Religion: ‘The whole frame of nature bespeaks an intelligent
author; and no rational enquirer can, after serious reflection, suspend his belief a moment with
regard to the primary principles of genuine Theism and Religion’ (Hume, A Dissertation on
the Passions, 33). But this passage refers only to belief in God, not to the employment of
that belief as a heuristic principle for science.
902 REED WINEGAR

recent commentators, including Guyer, have responded to this puzzle by


interpreting Philo’s belief in God’s intelligence as a Humean natural
belief, similar to a Humean natural belief in causation (Guyer, Knowledge,
Reason, and Taste, 208).33 According to such commentators, Philo’s
remark suggests that our experience of nature’s order psychologically
prompts belief in an intelligent designer. This belief is grounded in the
imagination, rather than in reason. But the imagination’s psychological
force is so strong that the belief is inevitable. According to Guyer, Philo
further implies that this natural belief is useful. More specifically, this
natural belief provides us with the heuristic principle of an intelligent
designer who has organized all things for a purpose, and this heuristic prin-
ciple guides empirical enquiry into nature. Guyer writes, ‘Hume’s spokes-
man Philo seems to allow that belief in the purposive design of the
universe and the intelligence of its author is not only natural but also
useful … ’ (Knowledge, Reason, and Taste, 208). And Guyer suggests that
Philo’s remark closely resembles the 3rd Critique’s claim that anatomists
should employ the reflecting power of judgement’s concept of purposiveness
to facilitate the discovery of new body parts or new mechanical operations in
living bodies (KU, AA 5:376).
But should we follow Guyer’s interpretation of Philo’s remark as an
anticipation of Kant’s own view? I think not. Carefully consider Philo’s
language: ‘from a firm conviction of its [i.e., the maxim that nature does
nothing in vain] truth, an anatomist who had observed a new organ or
canal would never be satisfied till he had also discovered its use and inten-
tion’ (Hume, Dialogues Concerning Natural Religion, 77). Unlike Kant,
Philo does not claim that anatomists use the concept of intelligent design
as a heuristic principle to seek out the discovery of new body parts or the dis-
covery of mechanical operations. Instead, Philo claims that anatomists who
have already discovered new body parts or mechanical operations then seek
to know their ‘use and intention’. In other words, Philo claims that scientists
ask what the purposes of already discovered elements of nature might be
simply for the specific reason of knowing their divinely bestowed purposes.
Moreover, Philo does not refer to a heuristic concept of God as the basis for
the maxim that nature does nothing in vain. Rather, Philo simply claims that

33
Kemp Smith (‘The Naturalism of Hume (I.)’, ‘The Naturalism of Hume (II.)’) introduced the
technical concept of a Humean natural belief as a belief that is not rationally justified but psy-
chologically universal, irresistible, and inevitable. Gaskin (Hume’s Philosophy of Region)
argues that Philo’s proclamation of belief in God does not meet these criteria for an official
Humean natural belief. Guyer does not address Gaskin’s criticisms. However, I will not
address that issue here, because my primary focus is on Guyer’s suggestion that Hume antici-
pates Kant’s view of God’s heuristic value, rather than on the topic of naturalness. Obviously,
a belief might be natural in another sense, even if it does not match Kemp Smith’s technical
concept of a Humean natural belief. For example, Hume’s Natural History of Religion exam-
ines the natural psychological sources of monotheism but does not claim that monotheism is
irresistible or universal.
KANT’S CRITICISMS 903

the maxim is empirically supported ‘merely from the contemplation of the


works of nature, without any religious purpose’ (Hume, Dialogues Concern-
ing Natural Religion, 77). Thus, Philo’s remark does not support the
interpretation that Hume, like Kant, regards the concept of God as a heuristic
principle that helps ground scientific enquiry into nature’s mechanical
operations.
Of course, one could retreat to the weaker claim that Kant interprets Hume
to claim that the concept of God is a useful heuristic principle for empirical
science, whether or not Hume actually holds such a view. As evidence for
this weaker claim, one could cite Reflexion 6136, which seems to imply a
connection between anatomy, the regulative Idea of God, and Hume. Kant
writes:

Anatomy, all things that are interconnected in a system. From this need there
springs the hypothesis of a rational cause – for the things in nature cannot be
represented as acting on each other purposively on their own. Hume.
(Refl 6136, AA 18:466; emphasis in original)

The final section of this Reflexion introduces the concept of anatomy and
the regulative Idea of a rational cause before mentioning Hume’s name. This
might suggest that Kant regards Philo’s discussion of anatomy as an inspi-
ration for his own views regarding the regulative Idea of God.
But despite initial appearances, this Reflexion cannot support such an
interpretation. Granted, the brief reference to Hume is ambiguous, but
Kant is probably connecting Hume’s name, once again, to Philo’s suggestion
that nature contains the principle of its order within itself. After all, the
clause immediately preceding Hume’s name states in apparent contrast to
Philo’s suggestion, ‘the things in nature cannot be represented as acting
on each other purposively on their own’ (Refl 6136, AA 18:466). Addition-
ally, the Reflexion comes from Kant’s copy of Baumgarten’s Metaphysica,
which Kant used for his lectures on metaphysics and religion. The extant
Nachschriften of these lectures contain multiple references to Philo’s objec-
tion but contain no passages where Kant claims that Hume advocates the
usefulness of the concept of God for promoting scientific discoveries in
anatomy. Instead, Kant actually singles Hume out as someone who denies
the need to employ the concept of God for the benefit of empirical
science. In §80 of the 3rd Critique, which concerns the need for scientists
to employ a regulative Idea of divine teleology, Kant writes, ‘Hume
makes the objection against those who find it necessary to assume for all
natural ends a teleological principle of judging, i.e., an architectonic under-
standing, that one could with equal right ask how such an understanding is
possible … ’ (KU, AA 5:420; emphasis in original).
In fact, rather than anticipating Kant’s own views, Hume’s Dialogues
threaten Kant’s regulative Idea of God. Towards the end of Part 12, Philo
claims that the concept of God is completely useless for all human activities.
904 REED WINEGAR

More specifically, Philo concedes an analogy between God and human intel-
ligence but emphasizes that this analogy is ‘remote’ and might not be any
closer than that between God and a rotting turnip (Hume, Dialogues Con-
cerning Natural Religion, 88). Given the remoteness of this analogy, the
concept of God is of no use to human beings. As Philo states, ‘it affords
no inference that affects human life, or can be the source of any action or for-
bearance’ (Hume, Dialogues Concerning Natural Religion, 88). If the
concept of God is not of use to any human activities, then it is not of use
to human beings’ scientific activities. Certainly, the remoteness of the
analogy raises the question of how exactly such a concept could be of
even heuristic use for science.
Kant directly addresses this issue in the Prolegomena. Kant couches his
discussion in terms of a contrast between deism and theism. Kant defines
‘deism’ as any theory that represents God merely by means of pure concepts.
He writes, ‘The deistic concept is a wholly pure concept of reason, which
however represents a thing that contains every reality, without being able
to determine a single one of them … ’ (Prol, AA 4:355). Here Kant notes
that the deistic conception of God as a being containing every reality fails
to determine God’s realities. Kant’s point is not that the abstract character-
ization of God as a being containing every reality fails to name those reali-
ties. Rather, Kant explicitly mentions eternity, omnipresence, substance, and
cause as pure ontological attributes of God (Prol, AA 4:357-8). Instead,
Kant’s point is that these pure ontological attributes are not further specified.
For example, according to the deistic conception of God, the kind of causal-
ity that God possesses remains indeterminate. Theism, however, attempts to
further specify God’s realities. For example, theism specifies how God’s
‘causality is constituted, e.g., by understanding and willing’ (Prol, AA
4:356).
Kant’s own opinion is that Hume’s Dialogues decisively refute dogmatic
theism but do not threaten deism. Kant writes:

[Hume’s] dangerous arguments relate wholly to anthropomorphism of which


he holds that it is inseparable from theism and makes theism self-contradic-
tory, but that if it is eliminated theism falls away and nothing but deism
remains …
(Prol, AA 4:356)

Although Kant does not provide any page references to Hume’s work, he
is almost certainly thinking of the exchange where Demea and Cleanthes
both observe that an anthropomorphic understanding conflicts with God’s
alleged status as a supreme being. More specifically, Demea and Cleanthes
acknowledge that features of human understanding, such as a succession of
diverse representations, contradict divine attributes like simplicity and
immutability (Hume, Dialogues Concerning Natural Religion, 27–9).
Given these contradictions, Demea concludes that we should abandon the
KANT’S CRITICISMS 905

claim that God’s intelligence is in any way similar to human understanding.


But Cleanthes insists that we must attribute a human-like understanding to
God in order to avoid falling into obscurantist mysticism. Thus, Cleanthes
urges us to attribute a human-like understanding to God and simply relin-
quish our conception of God as a supreme being (Hume, Dialogues Con-
cerning Natural Religion, 28–9).
Kant agrees with Demea and Cleanthes that a human-like understanding
contradicts God’s supremacy. But Kant focuses more on the worry that
having a faculty of understanding contradicts God’s independence than he
does on their worries about immutability and simplicity. Kant notes that
humans possess discursive understandings that depend on sensibility for
intuitions. However, a supreme being does not depend on anything; thus,
its cognition does not depend on sensibility. Consequently, a supreme
being must possess an entirely different kind of understanding – namely,
an intuitive understanding. As Kant notes, we can avoid attributing an
anthropomorphic understanding to God only by crediting God with ‘a differ-
ent understanding, which intuits objects’ (Prol, AA 4:355). However, we
have no insight into the character of an intuitive understanding. Instead,
we can conceive of an intuitive understanding only ‘negatively, namely as
not discursive’ (KU, AA 5:406).
Given Kant’s apprehensions about anthropomorphism, we might expect
Kant to restrict himself to a mysterious, deistic conception of God. But
Kant does not proceed in this manner. After all, we have already seen
Kant claim that the physico-theological inference rationally leads to a theis-
tic conception of God. Moreover, Kant worries that the deistic conception of
God, including the negative concept of God’s non-discursive understanding,
is too indeterminate to regulate either morality or scientific enquiry (Prol,
AA 4:356). According to Kant’s own interpretation of Hume, deism is a pos-
ition ‘from which nothing can be made, which can be of no use to us, and
which can in no way serve as a foundation for religion and morals’ (Prol,
AA 4:356). Here Kant refers to Hume’s suggestion, abhorred by Cleanthes
and embraced by Philo, that the deistic conception of God is too indetermi-
nate for morality. But Kant also maintains that the deistic conception of God
is too indeterminate for regulating scientific enquiry; thus, Kant insists that a
theistic conception of God should guide human reason’s ‘use within the sen-
sible world in accordance with principles of the greatest possible unity
(theoretical as well as practical) … ’ (Prol, AA 4:361; emphasis added).
And in the 1st Critique, Kant likewise claims that the regulative employment
of the Idea of God as a heuristic principle in scientific enquiry requires an
anthropomorphic conception of God (A692/B720).
Of course, one might question Kant’s claim that scientific enquiry requires
a theistic conception of God as a heuristic principle. But putting this worry to
the side, the main question for our purposes is how Kant can retain his con-
ception of God as a supreme being while also retaining a conception of God
in terms of human-like intelligence. Kant responds to this question by
906 REED WINEGAR

proposing an alternative to dogmatic theism and dogmatic deism. Kant refers


to this alternative as ‘symbolic anthropomorphism’ (Prol, AA 4:357; empha-
sis in original). If the physico-theological inference to intelligence yielded
knowledge, then physico-theology, which ascribes a human-like intelligence
to God, would contradict God’s status as a supreme being. But, as Kant notes
in the Prolegomena, his own theory regarding the boundary of reason entails
that we cannot presume to determine dogmatically the attributes of supersen-
sible things-in-themselves. Consequently, physico-theology does not prove
that nature’s ground actually possesses a human-like intelligence. Yet,
because we can make the existence of nature’s order intelligible to ourselves
only by ascribing a human-like intelligence to God, we can and should
indirectly specify God’s attributes by means of analogy. However, this rep-
resentation of God in terms of a human-like intelligence is merely a symbolic
representation of God, who would possess an intuitive understanding rather
than a human-like understanding.
At first, Kant’s appeal to analogy might not sound promising. After all,
Philo himself admits a remote analogy between God and human intelligence.
But Kant’s theory of analogy differs from Hume’s own.34 Hume’s theory of
analogy characterizes analogies as imperfect similarities. For example, Philo
indicates that human intelligence and God are similar but that the similarity
between them is remote. Kant, however, does not characterize analogies in
terms of imperfect similarities between objects. Instead, Kant endorses a
relational theory of analogy that concerns exact similarities between
relations (Prol, AA 4:357-8). As Kant observes, the relationship between
people in a case of perfect legal reciprocity would be identical to the relation-
ship between two bodies in a collision, because in both cases an action
immediately gives rise to an opposite and equal reaction (Prol, AA
4:358n). Thus, we can illustrate perfect legal reciprocity through the identi-
cal relationship between two billiard balls in a collision, even though a legal
person is entirely different in kind from a billiard ball. In the case of God,
Kant claims that (if God exists) the relationship between God and nature’s
order is identical to the relationship between human intelligence and
human artefacts. Therefore, we can indirectly specify the relationship
between God and world in terms of the identical relationship between
human intelligence and human artefacts, even though we regard God and
human intelligence as themselves entirely different in kind.
According to Kant, this analogy avoids Hume’s criticisms of theism and
deism. On the one hand, the analogy does not lapse into dogmatic anthropo-
morphism by claiming that God actually possesses human-like attributes.
Thus, the analogy does not contradict God’s supremacy (Prol, AA 4:358).
On the other hand, the analogy accommodates the alleged rationality of

34
Logan (‘Hume and Kant on Knowing the Deity’, 139) and Löwisch (‘Kants Kritik der reinen
Vernunft und Humes Dialogues Concerning Natural Religion’, 196, 200) notice this point.
KANT’S CRITICISMS 907

the physico-theological inference to theistic intelligence and avoids the


deistic problem of total indeterminacy (Prol, AA 4:358).
Thus, Kant and Hume do not agree that the concept of God should serve as
a useful heuristic principle in science. Instead, Kant recognizes Hume’s Dia-
logues as a threat to his own regulative Idea of God, and Kant attempts to
overcome this threat through his theory of symbolic anthropomorphism.
According to Kant, this theory of symbolic anthropomorphism renders
God’s relationship to the world comprehensible to human beings and,
thus, satisfies the rational demand for intelligibility. Additionally, while
the deistic notion of God is too indeterminate to guide empirical enquiry,
the heuristic representation of God as an anthropomorphic artisan provides
the example of human teleological artisanship as a comprehensible guideline
for directing empirical enquiry into nature’s systematic order. Note,
however, that Kant’s theory of symbolic anthropomorphism avoids dog-
matic anthropomorphism by arguing that the Idea of God is beyond the
boundaries of reason and, thus, not amenable to direct determination. In
other words, Kant avoids Hume’s worries about dogmatic anthropomorph-
ism by once again arguing that the physico-theological inference from
nature’s order to intelligence is rational but does not yield knowledge.
Thus, although the physico-theological inference yields no knowledge, it
satisfies the rational demand for intelligibility and provides humans with
the determinate, regulative Idea of God as an intelligent author of nature,
which facilitates empirical enquiry into nature’s systematic order.35

4. CONCLUSION

As we have seen, Kant takes issue with Hume’s Dialogues in two major
respects pertaining to theoretical philosophy. First, Kant argues that
Hume’s Dialogues do not successfully demonstrate the impossibility of
physico-theology. According to Kant, only a critique of reason can end all
physico-theological claims to knowledge by showing that even a rational
physico-theological inference would not yield knowledge. Second, Kant
defends his regulative theory of God against the threat of Hume’s Dialogues
by appealing to a theory of analogy, and this appeal similarly relies on Kant’s
view that the physico-theological inference to intelligence is rational but
35
One might ask what epistemic status (e.g. persuasion, opinion, conviction, belief) the
physico-theological inference provides for its conclusion, granted that it yields no knowledge
(see note 9). Kant does not answer this question. He only indicates that the inference is sub-
jectively necessary; that is, the inference satisfies the rational demand for intelligibility but
does not directly determine an object. Discussing the ontological and cosmological argu-
ments’ conclusions, Chignell (‘Belief in Kant’, 348f.) suggests that such cases might
amount to ‘theoretical belief’, although Kant does not state this explicitly. The 1st Critique
does endorse doctrinal belief in an intelligent designer, but Kant’s argument for doctrinal
belief differs from the physico-theological inference (see note 12).
908 REED WINEGAR

does not yield knowledge. Therefore, Kant’s own criticisms of physico-


theology are not mere footnotes to Hume’s Dialogues but, instead, reflect
Kant’s serious doubts about the adequacy of Hume’s position.
But what do we today stand to gain from Kant’s evaluation of Hume’s
Dialogues? Obviously, Kant’s regulative Idea of God does not play a
major role in contemporary science, and many contemporary philosophers
would reject Kant’s view that the physico-theological inference from
natural order to intelligence is rational. However, I think that contemporary
philosophers interested in criticizing current defences of the physico-theolo-
gical inference to divine intelligence (such as the fine-tuning argument)
might benefit from attending to Kant’s criticisms of Hume.36 As Kant
notes, Hume merely indicts the rational credentials of particular physico-
theological arguments and, thus, remains open to the worry that other philo-
sophers might one day discover rational, physico-theological arguments for
theism. Indeed, the physico-theologian would seem to require only one
rational inference to declare victory against Hume. Those who share
Kant’s conviction that physico-theology can never provide knowledge of
God might benefit from pairing Hume’s strategy of criticizing particular
physico-theological arguments with Kant’s attempt to argue that physico-
theology is impossible in principle, because even a rational physico-theolo-
gical inference (if such could ever be found) would not yield knowledge.37

Submitted 14 June 2014; revised 14 January and 26 April 2015; accepted 30


April 2015
Fordham University

BIBLIOGRAPHY

Anonymous. ‘Review of Dialogues Concerning Natural Religion’.


Brittisches Museum für die Deutschen 6 (1780): 42–72.
Brandt, Reinhard, and Heiner Klemme. David Hume in Deutschland.
Literatur zur Hume-Rezeption in Marburger Bibliotheken. Marburg:
Universitätsbibliothek Marburg, 1989.
Byrne, Peter. Kant on God. Aldershot: Ashgate, 2007.
Chance, Brian A. ‘Kant and the Discipline of Reason’. European Journal of
Philosophy 23, no. 1 (2015): 87–110.
Chignell, Andrew. ‘Belief in Kant’. Philosophical Review 116, no. 3 (2007):
323–60.
36
Collins (‘Hume, Fine-Tuning, and the “Who Designed God” Objection’) tries to defend the
fine-tuning argument against Hume.
37
I would like to thank Uygar Abaci, Paul Guyer, Michael Nance, Jeppe von Platz, the jour-
nal’s referees and audiences at the 2013 UK Kant Society, the University of Michigan, and the
2014 Leuven Kant Conference.
KANT’S CRITICISMS 909

Chignell, Andrew. ‘Modal Motivations for Noumenal Ignorance: Knowledge,


Cognition, and Coherence’. Kant-Studien 105, no. 4 (2014): 573–97.
Collins, Robin. ‘Hume, Fine-Tuning, and the “Who Designed God”
Objection’. In In Defense of Natural Theology: A Post-Humean
Assessment, edited by James F. Sennett and Douglas Groothius, 175–
99. Downers Grove: InterVarsity Press, 2005.
Eberhard, Johann August. Vorbereitung zur natürlichen Theologie zum
Gebrauch akademischer Vorlesungen. Halle: Waisenhaus, 1781.
Gaskin, J. C. A. Hume’s Philosophy of Religion. 2nd ed. Atlantic Highlands:
Humanities Press, 1988.
Gawlick, Günter, and Lothar Kreimendahl. Hume in der Deutschen
Aufklärung. Umrisse einer Rezeptionsgeschichte. Stuttgart-Bad
Cannstatt: Frommann-holzboog, 1987.
Guyer, Paul. ‘Natural Ends and Ends of Nature’. In Hans Christian Ørsted
and the Romantic Legacy in Science: Ideas, Disciplines, Practices,
edited by Robert M. Brain, Robert S. Cohen, and Ole Knudsen, 75–
96. Dordrecht: Springer, 2007.
Guyer, Paul. Knowledge, Reason, and Taste: Kant’s Response to Hume.
Princeton: Princeton University Press, 2008.
Hatfield, Gary. ‘The Prolegomena and the Critiques of Pure Reason’. In Kant
und die Berliner Aufklärung. Akten des IX. Internationaler Kant-
Kongress, Band 1, edited by Rolf-Peter Horstmann, Volker Gerhardt,
and Ralph Schumacher, 185–208. Berlin: de Gruyter, 2001.
Holden, Thomas. Spectres of False Divinity: Hume’s Moral Atheism.
Oxford: Oxford University Press, 2010.
Hume, David. Dialogues Concerning Natural Religion. Edited by Richard
H. Popkin. Indianapolis: Hackett, 1980.
Hume, David. A Dissertation on the Passions and the Natural History
of Religion. Edited by Tom L. Beauchamp. Oxford: Clarendon Press,
2007.
Kant, Immanuel. Gesammelte Schriften. Edited by Königlich Prueßische
Akademie der Wissenschaften, Akademie der Wissenschaften zu
Berlin, and Akademie der Wissenschaften zu Göttingen. Berlin: de
Gruyter, 1900–.
Kant, Immanuel. The Cambridge Edition of the Works of Immanuel Kant.
Edited by Paul Guyer and Allen W. Wood. Cambridge: Cambridge
University Press, 1992–.
Kemp Smith, Norman. ‘The Naturalism of Hume (I.)’. Mind 14, no. 2
(1905): 149–73.
Kemp Smith, Norman. ‘The Naturalism of Hume (II.)’. Mind 14, no. 3
(1905): 335–47.
Kemp Smith, Norman. ‘Introduction’. In Hume’s Dialogues Concerning
Natural Religion, 1–75. 2nd ed. New York: Social Sciences, 1948.
Kreimendahl, Lothar. ‘Kants Kolleg über Rationaltheologie. Fragmente
einer bislang unbekannten Vorlesungsnachschrift’. Kant-Studien 79,
no. 3 (1988): 318–28.
910 REED WINEGAR

Kuehn, Manfred. ‘Kant’s Critique of Hume’s Theory of Faith’. In Hume and


Hume’s Connexions, edited by M. A. Stewart and John P. Wright, 239–
55. University Park: Pennsylvania State University Press, 1994.
Kuehn, Manfred. ‘The Reception of Hume in Germany’. In The Reception of
David Hume in Europe, edited by Peter Jones, 98–138. London:
Continuum, 2005.
Kuehn, Manfred. ‘Von der Grenzbestimmung der reinen Vernunft’. In Kants
Prolegomena. Ein kooperativer Kommentar, edited by Holger Lyre and
Oliver Schliemann, 235–54. Frankfurt am Main: Klostermann, 2012.
Logan, Beryl. ‘Hume and Kant on Knowing the Deity’. International
Journal for Philosophy of Religion 43, no. 3 (1998): 133–48.
Löwisch, Dieter-Jürgen. ‘Kants Kritik der reinen Vernunft und Humes
Dialogues Concerning Natural Religion’. Kant-Studien 56, no. 2
(1965): 170–207.
Maly, Sebastian. Kant über die symbolische Erkenntnis Gottes. Berlin: de
Gruyter, 2011.
McFarland, John D. Kant’s Conception of Teleology. Edinburgh: University
of Edinburgh Press, 1970.
Meiners, Christoph. ‘Review of Dialogues Concerning Natural Religion’.
Zugabe zu den Göttingschen Anzeigen der gelehrten Sachen 48
(1779): 753–63.
Munzel, G. Felicitas. ‘“The Beautiful Is the Symbol of the Morally-Good”:
Kant’s Philosophical Basis of Proof for the Idea of the Morally-Good’.
Journal of the History of Philosophy 33, no. 2 (1995): 301–30.
Pasternack, Lawrence. ‘Regulative Principles and “the Wise Author of
Nature”’. Religious Studies 47, no. 4 (2011): 411–29.
Paulsen, Friedrich. ‘Einleitung’. In Dialoge über natürliche Religion: Über
Selbstmord und Unsterblichkeit der Seele. 3rd ed., 9–28. Leipzig:
Verlag der Dürr’schen Buchhandlung, 1905.
Santozki, Ulrike. Die Bedeutung antiker Theorien für die Genese und
Systematik von Kants Philosophie. Berlin: de Gruyter, 2006.
Specht, Ernst Konrad. Der Analogiebegriff bei Kant und Hegel. Köln:
Kölner Universitätsverlag, 1952.
Stern, Robert. ‘Metaphysical Dogmatism, Humean Scepticism, Kantian
Criticism’. Kantian Review 11, no. 1 (2006): 102–16.
Warda, Arthur. Immanuel Kants Bücher. Berlin: Martin Breslauer, 1922.
Watkins, Eric. Kant and the Metaphysics of Causality. Cambridge:
Cambridge University Press, 2005.
Wolff, Robert P. ‘Kant’s Debt to Hume via Beattie’. Journal of the History of
Ideas 21, no. 1 (1960): 117–23.
Wood, Allen W. Kant’s Rational Theology. Ithaca: Cornell University Press,
1978.
Copyright of British Journal for the History of Philosophy is the property of Routledge and its
content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email
articles for individual use.

You might also like