1. Introduction
One interesting accomplishment of Kant's pre-Critical period is the a priori proof of God's existence that he initially presented in the New Elucidation (Reference Kant1755) (NE) and then reiterated, with some modification, in The Only Possible Argument (Reference Kant1763) (OPA).Footnote 1 The proof runs from the idea that every determinate possibility has a material or real content that must be grounded in actuality to the conclusion that there exists a necessary being which is the material ground of all possibility. What makes the proof historically significant is that Kant intended it to be an alternative to the traditional versions of the ontological proof that were propounded by Kant's rationalist predecessors. This is a clear indication that Kant was critical of traditional rationalist theology or, better, ontotheologie, even in his pre-Critical period, for the proof sprang out of Kant's criticisms of the modal notions utilized by the tradition, namely that it wrongly treated existence as a predicate and considered possibility and necessity only logically and not materially. There is also an air of mystery around the proof, adding to its interest. In his Critical period, Kant adopted a perplexing attitude toward his pre-Critical proof: even though he stated that it is formally irrefutable, he insisted that his proof at best establishes the subjective necessity of presupposing the idea of God and not the necessary existence of such a being.
It is therefore not surprising that this pre-Critical proof enjoyed a wide appreciation in twentieth-century Kant scholarshipFootnote 2 and is still attracting interest in more recent literature.Footnote 3 Among the recent works on the subject, Andrew Chignell's two articles, ‘Kant, Modality and the Most Real Being’ (Chignell Reference Chignell2009) and ‘Kant, Real Possibility, and the Threat of Spinoza’ (Chignell Reference Chignell2012), have a special place. For in these articles, Chignell presents a genuinely original reconstruction of the proof and locates it in the broader context of his account of Kant's modal theory. The reconstruction turns on a thick notion of ‘real possibility’. Namely, Chignell suggests that, in addition to the requirements of logical consistency and the givenness of positive real content, the satisfaction of which, he claims, suffices only for ‘logical possibility’, Kant introduces a third condition to account for the ‘real possibility’ of a thing: ‘The predicates of a thing must be really harmonious with one another.’Footnote 4 Chignell argues that this harmony requirement lies ‘at the heart’ of Kant's entire proof to the extent that it motivates both the ‘Grounding Premise’ of the proof, i.e. that every possibility must be materially grounded on some actuality, and the transition from the intermediary conclusion that it is necessary that some material ground of possibility should exist to the ultimate conclusion that this material ground is a necessary and unique being, the ens realissimum.Footnote 5 He further claims that the issue of real harmony is responsible for Kant's later ‘undoing’ of the proof in OPA. On Chignell's account, the shift in Kant's attitude towards his own proof was triggered by Kant's realization, sometime between 1763 and 1781, that the proof relied on the unmotivated assumption that there is a real harmony between the predicates of the ens realissimum.
I have four main objections. First, the pre-Critical Kant does not present the absence of real repugnance or the presence of ‘real harmony’ as a general condition of real possibility. Chignell's textual citations do not suffice to confirm his claim that Kant does so. Second, the real harmony requirement does not play a substantial role in Kant's proof itself. Consequently, third, the real harmony requirement is not what motivates the ‘Grounding Premise’ of the proof. Instead, this premise is sufficiently motivated by what Chignell labels the ‘content’ requirement. Finally, Kant's later ‘undoing’ or downgrading of the proof is not based on a concern regarding the real harmony of the predicates of God, but on Kant's Critical restrictions on cognition in general and modal cognition in particular. In the course of arguing for these four objections, I also hope to present what I take to be an adequate alternative reconstruction of Kant's proof.
2. On Real Possibility in the Pre-Critical Kant
Before taking up the question of the connection between real harmony and real possibility, I wish to start with a brief discussion regarding the notion of ‘real possibility’ itself in Kant's pre-Critical period. For it seems to me that, contrary to the impression that Chignell gives, it is not textually obvious that the Kant of the 1750s and early 1760s had a distinct notion of ‘real modality’ as opposed to the ‘logical modality’ of the Leibnizian-Wolffian school. Both in NE and OPA, the only two pre-Critical texts where Kant directly elucidates the concept of possibility, the distinction Kant offers is rather between two different elements or conditions, logical/formal and real/material, of possibility per se, and not between two different conceptions or kinds of possibility such as ‘logical’ and ‘real’ possibility (1: 395; 2: 77–8). In fact, Kant does not use the term ‘real possibility’ in any of his pre-Critical works.Footnote 6
Kant fully articulates real modality and sets it against logical modality in his Critical period. Accordingly, logical modality pertains to mere thinking and is properly predicated of our conceptual or propositional representations of things rather than things themselves. Logical possibility, for instance, is the possibility of thinking a thing through a concept or a proposition without logical contradiction.Footnote 7 Logical necessity expresses the necessity of the connection of concepts within a proposition and that the negation of that connection yields a logical contradiction.Footnote 8 Real modality, on the other hand, pertains to the being or existence of the thing in question, i.e. whether the thing or its state can exist, and if it actually does, whether it does so necessarily. Kant emphatically states – against the rationalists – that real modalities of things and logical modalities require different conditions to be satisfied. For while the former reflect the formal structure of human thinking, the latter reflect the real constitution of things.
Thus, on Kant's Critical exposition of real modality, the basic notion of real possibility expresses the possibility of being. Yet the Critical Kant introduces a further distinction into real possibility based on the very realm or context of being in question. If the question of real possibility is raised with respect to a certain realm of being, we are talking about relative real possibility of things restricted by the conditions of being in that specific realm. The very notion of real possibility with which Kant (Reference Kant1998) is mostly concerned in the Critique of Pure Reason is thus a relative (real) possibility of things, that is, relative to the realm of human experience and thereby conditioned by the formal conditions of it.Footnote 9 On the other hand, if the question of real possibility is raised against the background of the entire metaphysical space limited not by the conditions of any particular realm but only by the general conditions of being, then we are talking about the real possibility of things in general. Kant sometimes calls this kind of possibility ‘absolute possibility’, and states that this possibility, as ‘unconditioned’ or ‘unrestricted’, is valid in every ‘respect’, ‘relation’ or ‘hypothesis’, as opposed to the ‘relative’ or ‘conditioned’ possibility of a thing, which is valid only with respect to a specific set of conditions.Footnote 10
A parallel account of real necessity can be derived from this dual account of real possibility. The basic notion of real necessity is that of the necessity of being. Relative real necessity would be the necessity of the existence of a thing in a specific realm relative to the conditions of being in that realm. One such realm of being, for instance, is that of empirical reality. Kant's third postulate captures such a notion of relative real necessity as it defines necessity in relation to the conditions of being an object of human empirical cognition: ‘That whose connection with actuality is determined in accordance with the general conditions of experience is (exists) necessarily’ (B266). Absolute real necessity would be the necessary existence of a thing in the entire metaphysical space without regard to the conditions of any particular realm of being but as determined by the general conditions of being. In other words, something is absolutely really necessary if its existence is necessitated by the conditions of being in the metaphysical space of all really possible things. The non-existence of the absolutely really necessary would then cancel out the conditions of being in general and thus the entire metaphysical space and all real possibilities. As will be seen, Kant assigns a substantial role to this notion of absolute real necessity in OPA.
Now, even though the pre-Critical Kant does not articulate these distinctions, it can still be said with fairness that the proof in OPA is concerned with (absolute) real modalities in distinction from logical modalities. In the first place, Kant presents his proof as an alternative to the classical versions of the ontological proof purporting to demonstrate God's existence as a logical necessity. Kant points out that his previous rejection of the conception of existence as a predicate or determination that could be introduced into the concept of anything in the ‘First Reflection’ (2: 72) already makes it impossible for existential propositions to be logically necessary: ‘Existence is not a predicate at all, nor is the cancellation of existence the negation of a predicate, by means of which something in a thing is cancelled and through which an internal contradiction could arise’ (2: 81). Accordingly, the existence of everything, God included, can be denied without logical contradiction, or what is the same, nothing exists with logical necessity. Thus, Kant's proof in OPA replaces the idea of what could be called an ens logice necessarium, a logically necessary being, with that of an ens realiter necessarium, a really necessary being. The same point is also clear from Kant's repeated admittance that the cancellation of all existence, the state of complete nullity, is not logically impossible (2: 78; 2: 79). When Kant says that ‘it is absolutely impossible that nothing at all should exist’, he refers to a state of ‘(absolute) real impossibility’.
Therefore, Kant's proof talks about at least one real necessity (God's existence) that is not logically necessary, and one real impossibility (complete nullity) that is not logically impossible, or what is the same thing, one logical possibility that is not really possible. And without having to go further into the details of OPA, these considerations suffice to show that the Kant of OPA employs the notion of ‘real possibility’, even if he does not explicitly present it as a distinct notion of possibility.Footnote 11
3. Real Harmony as a Condition of Real Possibility?
While I agree with Chignell that Kant employs, though implicitly, a notion of (absolute) real possibility in OPA, I disagree as to what that notion consists in. Chignell argues that the pre-Critical Kant holds that, in addition to the logical consistency and the givenness of positive real content, which constitute the conditions of logical possibility, there is a third condition that must be satisfied for ‘real possibility’: ‘The predicates of a thing must be really harmonious with one another’.Footnote 12 I suggest that Chignell introduces the harmony requirement without sufficient textual ground and understates the content requirement.
First of all, in neither of the two pre-Critical texts in which Kant provides an analysis of the concept of possibility does he count real harmony among the conditions of possibility. In fact, in NE, Kant does not even mention real repugnance, and in OPA, the issue of real repugnance does not come up in the section that Kant devotes to the analysis of the concept of possibility, i.e. ‘The necessary distinction in the concept of possibility’ (2: 77–8). It arises only after Kant's presentation of the proof proper and for reasons that are not relevant to the question of how to conceive possibility.
Kant's pre-Critical interest in the concept of real repugnance does not lie in its purported relation to the concept of possibility. Kant provides his most extensive account of real repugnance in Negative Magnitudes, an essay published in Reference Kant1763, a few months after OPA. Kant's account relies on the discovery that the predicates of a subject may stand in a relation that is not logical and thus not captured by the law of contradiction or identity, namely a real relation.Footnote 13 Throughout Negative Magnitudes Kant focuses on the negative variant of this relation: real opposition (Realentgegensetzung) or real repugnance (Realrepugnanz).Footnote 14 Accordingly, two predicates that are logically consistent may still be in real repugnance, that is, they can simultaneously exist in one and the same subject but they nullify (vernichtigen) or cancel out (aufheben) each other's effects on the subject. Kant's typical example of real repugnance is that of two motive forces acting on the same body in opposite directions (2: 86; 2: 171). If, in such a case, the two forces have equal magnitudes, they nullify each other and the result for the subject is a state of equilibrium or rest.Footnote 15
Logical opposition or contradiction occurs when two logically contradictory predicates such as A and not-A are posited in the same subject. On the other hand, in real repugnance, both predicates represent positive realities and are logically independent from one another. They are opposed to each other only in the sense that their respective consequences cancel each other out. Therefore, whereas logical opposition is between a reality and its negation, in which case the quantitative degree of the reality in question is irrelevant, real opposition occurs between two quantifiable positive realities.Footnote 16 This leads to an important difference between the results of the two types of opposition. Logical opposition of two predicates in the same subject entails the logical impossibility of that subject in virtue of the law of contradiction. That is to say, in logical opposition, the subject is cancelled. ‘The consequence of the logical contradiction is nothing at all (nihil negativum irrepraesentabile)’ (2: 171). In real opposition, however, it is not the subject that is cancelled but only the opposing predicates or, more precisely, their effects. The result, therefore, is something that can be thought (cogitabile) or represented (repraesentabile),Footnote 17 which is the subject itself in a state of privation with respect to the effects of the opposing predicates. To follow Kant's example:
[A] body which is both in motion and also, in the very same sense, not in motion, is nothing at all … The motive force of a body in one direction and an equal tendency of the same body in the opposite direction … are simultaneously possible in one body. The consequence of such opposition is rest, which is something (repraesentible). (2: 171)
In other words, while logical contradiction leads to nothing in the absolute sense (nihil negativum), the result of a real opposition is nothing in the relative sense, i.e. relative to whatever magnitude the opposing predicates represent.Footnote 18
Therefore, real harmony or freedom from real repugnance does not seem fit to be a condition of real possibility in general, since the result of real repugnance is a really and thus also logically possible state of affairs: the cancellation not of the subject but only of the effects of the opposing positive predicates on the subject. Chignell admits that ‘many of the examples that Kant provides’ in OPA and Negative Magnitudes are this kind of ‘predicate-canceling real repugnance’.Footnote 19 He adds, however:
But in a few places [Kant] suggests that there is another variety of real repugnance that is subject-canceling. In these cases, the manner in which two or more predicates are opposed results in a ‘canceling’ not merely of their respective effects, but of the subject itself qua real possibility.Footnote 20
This is an extremely important claim for the overall tenability of Chignell's reconstruction of the proof in OPA. I will now argue that the pre-Critical Kant does not in fact hold that these cases constitute a subject-cancelling type of real repugnance, nor does he incorporate them into an account of ‘real possibility’.Footnote 21
Chignell draws his first example from Kant's discussion of ‘ceasing to be’ or ‘passing away’ (vergehen) in Negative Magnitudes (2: 190).Footnote 22 Kant's aim is to demonstrate that ‘ceasing to be’ could be the result of a real ground (i.e. the presence of a positive cause), just like ‘coming into being’. Yet Kant's own examples of such ‘ceasing to be’ strongly suggest that he does not have in mind ‘subject-cancelling’ real repugnance. Kant talks about the ceasing to be of certain states of subjects, e.g. ‘a movement’ (of a physical entity), ‘representations and desires’ or ‘inner accidents’ (of the soul) (2: 190–1), and argues that for ‘a movement’ to stop, there must be a positive motive force (acting on the opposite direction), just as there must be a positive force for a movement to start, and the elimination of a ‘sorrowful thought’ requires a positive effort, for instance, to entertain a happy thought (2: 190). These are all cases of ‘predicate-cancelling’ real repugnance, for they do not imply the cancellation of the subject that is the bearer of the states in question.
‘Predicate-cancelling’ real repugnance seems to be what Kant mainly has in mind in OPA as well. Since Kant introduces the notion of real repugnance in his exposition of the being that contains ‘supreme reality’, the case of the ens realissimum should be the most obvious example in OPA in favour of Chignell's interpretation. And it does really seem that in the case of the ens realissimum, construed as the subject containing as its predicates all positive realities to their maximal degrees, the real repugnance of predicates results in the cancellation not only of the predicates but also of the subject itself. However, the way Kant actually understands the cancellation of the subject here does not seem to align with Chignell's sweeping claim that for Kant freedom from real repugnance is a condition of possibility:
In the most real being of all there cannot be any real repugnance or positive conflict (Realrepugnanz oder positiver Widerstreit) among its own determinations, for the consequence would be a deprivation or a lack, and that would contradict (widerspricht) its supreme reality. (2: 86)Footnote 23
So, Kant thinks that in the case of a real repugnance among the predicates of the ens realissimum, the subject is cancelled not because of real repugnance, but because of the logical contradiction between its definition and the deprivation resulting from the cancellation of the predicates that are in repugnance to one another. It is logically impossible for the Supreme Being to have the maximal degrees of all positive realities and yet be deprived of some of them. In other words, in the special case of the ens realissimum, (predicate-cancelling) real repugnance leads to a logical contradiction. Thus, the pre-Critical Kant deems the ens realissimum, understood in its all-embracing Spinozistic sense, a ‘conceptual confusion’ (2: 85), not because the lack of real harmony makes the entity ‘really impossible’ but because it makes the concept logically incoherent. I will revisit this point in the next section.
In OPA, after establishing that God is not to be conceived as containing all positive realities as his own predicates or determinations, Kant discusses ‘whether the properties of understanding and will are to be found in the Supreme Being as determinations inhering in it, or whether they are to be regarded merely as consequences produced by it in other things’, and concludes that God must be a mind having both understanding and will as his own determinations (2: 89). Chignell regards this as another example of ‘subject-cancelling’ real repugnance, although Kant does not mention real repugnance in the relevant sub-sections of the text (‘Fourth Reflection’: 1, 2, 3).Footnote 24 Chignell takes Kant's conclusion to rely on the observation of a subject-cancelling real repugnance between the predicates of ‘being the Supreme Being’ and ‘being the natura naturans’ (with no will and understanding). In fact, however, Kant again observes a logical contradiction between the lack of understanding and will and the supremacy of the Supreme Being. More precisely, Kant suggests that ‘if the Supreme Being did not possess understanding and will, every other being … with these properties would … have to take precedence over the Supreme Being’ (2: 88), and such a god that has no understanding and will ‘would nonetheless be far inferior to what one must needs think when one thinks of a god’ (2: 89). Kant's reasoning is based on a conceptual analysis of the Supreme Being, with no need to appeal to a principle other than that of contradiction.
This crucial point also rules out a potential defence of Chignell. For instance, one might say that, even if Kant does not directly present cases of ‘subject-cancelling’ real repugnance, the ‘predicate-cancelling’ cases that he presents can be transformed into ‘subject-cancelling’ ones. For instance, if body A is acted on by forces F and F* of equal magnitude but opposite directions, then there is no logical contradiction among the predicates in the concept of ‘A acted on by F and F* and acted on by no other forces but moved in the direction of force F’ but it designates a (real) impossibility. Its logical possibility follows from the fact that the predicate-cancelling relation between F and F* is not a logical relation. Similarly, one might say that, since there could be no logical but only predicate cancelling conflict in the concept of ‘a being that possesses every positive determination’, that concept is not really possibly instantiated. What goes unnoticed here is that, even though there is not a logical contradiction between the predicates themselves but only a (predicate-cancelling) real repugnance, there is a logical contradiction between the result of the predicate cancellation and the concept itself. F and F* cancel out each other's effects on A and since there is no other force acting on it A will remain at rest. The concept of ‘A acted on by F and F* and acted on by no other forces but moved in the direction of force F’ is simply logically impossible, as being at rest and being in motion are logically contradictory. Similarly, in the case of ‘a being that possesses every positive determination’, the cancellation of some of those positive determinations would result in a deprivation that would stand in a logical contradiction with the very concept of that being as possessing all positive determinations. In both cases, the impossibility is not due to a metaphysical absurdity or incompatibility in the subject but to a logical incoherence in the concept of the subject.
One passage that Chignell cites does seem, however, to support his case for Kant's employment of the concept of ‘subject-cancelling’ real repugnance in OPA. Again in the context of the specification of what predicates God would have to possess, Kant writes ‘[All realities] can by no means co-exist together as determinations in a subject … The impenetrability of bodies, extension and such like, cannot be attributes of that which has understanding and will’ (2: 85). The ‘cannot’ here does not seem to express any logical contradiction but a metaphysical incompatibility between two natures, body and mind or the physical and the mental, based on a Cartesian sort of dualist assumption. And this incompatibility must be quite different than the standard or ‘predicate-cancelling’ real repugnance, in which the opposing predicates continue coexisting in the subject even though they cancel out their respective effects on the subject. Strangely enough, however, this reading is impugned when Kant goes on to explain the quoted statement. Kant first concedes that the incompatibility in the case is not a logical contradiction as ‘realities’ never constitute a logical contradiction, and against the rationalists who take logical contradiction to be the only kind of incompatibility that could occur between predicates, he brings up the notion of ‘real repugnance’. Yet, again, he explains ‘real repugnance’ in terms of opposing forces acting on a body and cancelling out each other's effects. He even goes as far as to say that these forces are ‘really possible in a body’ because the result of their opposition is rest, which is ‘indubitably, possible’ as opposed to the ‘absolute impossibility’ which results from a logical contradiction (2: 86).
This insensitivity to the nuance between ‘subject-cancelling’ and ‘predicate-cancelling’ kinds of real repugnance demonstrates either that Kant is not aware of the potential importance of this distinction for modality, and in particular for real possibility, or that he does not think that the examples of the real repugnance between the physical and the mental and the two opposing forces constitute genuinely distinct kinds of real repugnance. In either case, it is obvious that the Kant of OPA does not take the lack of ‘subject-cancelling’ repugnance or the presence of ‘real harmony’ between the predicates of a subject-concept to be a condition of real possibility in general. This observation alone, if correct, seriously jeopardizes Chignell's claim that Kant's proof in OPA operates on the notion of real harmony. However, in the next section I will take up Chignell's claim on its own terms, and investigate whether the proof really employs or even needs to employ the notion of real harmony.
4. The Role of Real Harmony in OPA
That the notion of real harmony does not have any substantial role in the proof in OPA is first of all strongly suggested by the fact that Kant introduces the notion of real repugnance only in his discussion of how to construe the idea of a Supreme Being that is the material ground of all possibilities (2: 85–6), and thus only after the proof proper for the necessary existence of such a being. The question is whether the Supreme Being is to be construed as containing all reality, i.e. all possible positive predicates. Kant answers in the negative for the reasons I explained above. Thus even if we grant that Kant has in mind subject-cancelling real repugnance at 2: 85–6, we can at best say that he intends the notion of real repugnance to refine the concept of the Supreme Being, whose existence is supposed to have been demonstrated by the proof proper (2: 77–83), in a non-Spinozistic direction.
It is also important to note that Kant does not talk about real repugnance or harmony in his first presentation of the proof in NE.Footnote 25 And it is actually the lack of consideration of real repugnance in NE which drives this version of the proof towards a concept of God that contains all realities.Footnote 26 Chignell's quote from NE as an example of Kant's early awareness of the importance of the notion of real harmony seems misleading.Footnote 27 For when Kant says ‘Possibility is only definable in terms of there not being a repugnance (non repugnantia) between certain combined predicates’ (1: 395), he is in fact referring to the logical consistency requirement on possibility. Similarly, in OPA, he calls the repugnance (Repugnanz) between the predicates of a concept the formal element of impossibility (2: 77). Following the common usage in the Wolffian texts, which itself is inherited from scholastic practice, Kant understands ‘repugnance’ broadly as contradiction.Footnote 28 He uses Repugnanz or repugnantia, interchangeably with ‘opposition’ (Entgegensetzung) and ‘conflict’ (Widerstreit), both of which usually signify logical contradiction (Widerspruch), unless Kant specifies the repugnance in question as ‘real repugnance’ (Realrepugnanz).Footnote 29
Why then suppose that real harmony is ‘central to the entire proof’ in OPA? Chignell's claim seems to be based on an overinterpretation of the ‘grounding premise’ of the argument. This is the principle that every real possibility presupposes some actuality for its material content, or more specifically, that if a concept is to represent a really possible object, then each of its predicates must be grounded such that either the predicate itself or a more fundamental predicate from which it can be logically derived must be instantiated in something actually existent. According to Chignell, what motivates the ‘grounding premise’ is the real harmony requirement. For, he claims, the other two requirements on possibility, consistency and content, could be satisfied without the presupposition of actuality, whereas the satisfaction of real harmony is warranted only when the predicates in question are actually co-instantiated.Footnote 30
I suggest that Kant does not need the real harmony requirement to motivate the ‘grounding premise’. He rather understands this premise as a natural extension or implication of the content requirement, which Chignell seriously understates. Chignell defines the content requirement in the following way: ‘The positive predicates of a thing must be given as data that possess real content’.Footnote 31 Yet it remains unclear what this condition ultimately means. In Chignell (Reference Chignell2009), he gives the impression that he diminishes this content requirement to a material condition of logical possibility: ‘the prior availability or “giveness” of such positive predicates is required if a concept is to have any real content at all, and if its object is to be logically possible’.Footnote 32 In Chignell (Reference Chignell2012), on the other hand, he seems to hold that the content requirement is more than just a condition of logical possibility and that it requires the predicates not only to have some content, but also to be really possible. In support of the latter reading, Chignell refers to Kant's notion of empty concepts or mere ‘thought-entities’, which obviously have some logically consistent but not really possible content.Footnote 33 The problem is that in both of his statements, Chignell remains vague about what he takes the ‘givenness’ or ‘availability’ of positive predicates as possessing ‘real’ content to consist in. In any case, if Kant has a distinct notion of real possibility in OPA and puts forth the content requirement as a condition of real possibility, then he cannot be understood as just talking about the necessity of the availability of some positive predicates to a mind for the possibility of concept construction. It is important to bear in mind that real possibility is the possibility of the existence of a thing and is distinct from the question of a thing's conceivability by a discursive mind. Furthermore, Kant sets out to lay the groundwork of an a priori demonstration of the existence of a being, which should not presuppose the existence of any epistemic subjects. On the other hand, that the positive predicates of a really possible thing must also be really possible, though certainly true, is not helpful in understanding the content requirement as a condition of real possibility. A condition should be explanatory of what it is a condition of, not vice versa.
So what exactly does it mean for a positive predicate to be ‘given’ or ‘available’ in the metaphysical space? Presumably, it means that it is not something that is merely thought by a mind but can be given to a mind as the material of thought by actual existence. The reference to actual existence is unmistakable in Kant's own introduction of the content requirement. For instance, in NE, he writes: ‘nothing can be conceived as possible unless whatever is real in every possible concept exists (existat)’ (1: 395). In OPA, he insists that the ‘givenness’ of the material element consists in actual existence: ‘if all existence (Dasein) is cancelled, then … nothing at all is given (gar nichts gegeben), there is no material element for anything which can be thought’ (2: 78). Kant also points out how this requirement distinguishes real possibility from merely logical possibility. In our logical possibility considerations, we only determine whether the predicates are logically compatible without regard to the question of their givenness. In real possibility, on the other hand, we are also required to determine whether these predicates are actually given, for after all they may all be ‘empty words’ without real content and signifying nothing at all. ‘[I]n the end, when you consider how this [material content] is given to you, the only thing to which you can appeal is an existence’ (2: 81).
Now, there are two ways in which the content requirement can be satisfied in any particular case. First, the predicates of the object in question themselves are actually instantiated (either by the object itself, in which case the object's real possibility directly follows from its actuality, or by some other object); second, not the particular predicates of the object themselves but more fundamental predicates, from which the former can be derived through negation, limitation or analysis, are actually instantiated. For instance, the predicate ‘not being green’ can be given through the actual instantiation of the predicate ‘green’ by a green tree; a limited degree of a quality such as ‘wisdom’ can be given through the actual instantiation of a higher degree of wisdom; the predicate ‘(being a) body’ can be given through the more fundamental predicate ‘being extended’ without there being actual bodies (2: 80–1).
The ‘grounding premise’ of Kant's proof that every (real) possibility presupposes something actual as its material ground is in fact an extension of the content requirement thus construed. It is because real possibility requires that its content be actually given that every real possibility requires an actuality as its material ground. This is clear from Kant's definition of material ground as ‘the actuality, by means of which, as by means of a ground, the internal possibility of other realities is given’ (2: 79). In other words, what ‘gives’ the content of a real possibility is nothing but what materially grounds it. Therefore, contrary to Chignell's claim, the harmony requirement is not requisite for motivating the ‘grounding premise’. And the first stage of the proof concluding that necessarily something exists can be reconstructed with the consistency and content requirements alone:
(1) For something to be really possible, (a) its content must conform to the law of contradiction, i.e. its predicates must be logically consistent with one another; (b) its predicates must be given through the actual instantiation either of themselves or of more fundamental predicates. (Premise)
(2) Each and every predicate F of any really possible thing must be grounded in some existent thing that actually instantiates either F or a more fundamental predicate F’ from which F can be derived. (1b, ‘grounding premise’)
(3) Real possibility, in any particular case, disappears (a) when there is a logical contradiction between the predicates; (b) when no predicate is grounded. (1, 2)
(4) If nothing at all exists, no predicate would be grounded and all real possibility would disappear. (3)
(5) ‘That, by means of which all [real] possibility is cancelled, is absolutely [really] impossible’.Footnote 34 (Definition)
(6) It is absolutely really impossible that nothing at all should exist = it is absolutely really necessary that something exists. (4, 5)
We can readily see that the conclusion of the first stage in Chignell's reconstruction, ‘necessarily, there is an actual y which materially grounds the possession of any really possible predicate F’ is entailed by (2) and (6). Hence, there is no need for the harmony requirement.
It might still be suggested, however, that the harmony requirement could be of use in the second stage of the proof, where Kant goes on to argue that there exists a unique absolutely necessary being which is the ultimate material ground of all real possibility (2: 82–5). This transition from the de dicto (real) necessity of the assertion that there exists something (6) to the de re (real) real necessity of a unique being is hard to swallow. Kant has to explain, first, why the material ground of all real possibility is necessary de re, and second, why this ground has to be a unique being or why a plurality of things cannot collectively do the grounding job. Whether he really provides a successful argument for the transition is a matter of controversy.
Chignell's real harmony reading offers an answer to the second question. It was said earlier that the predicates of really possible things must be grounded by the actual instantiation either of themselves or of more fundamental predicates. As long as the set of most fundamental predicates are actually instantiated, the matter of all real possibility, that is, all the predicates whose different combinations constitute all particular real possibilities would be grounded.Footnote 35 If we adopt Chignell's reading, we can make good sense of Kant's insistence that this set of most fundamental predicates must be instantiated in one being. For if real harmony of predicates is indeed a condition of real possibility in general, then all particular real possibilities constituted by different combinations of really possible predicates will collectively require that the most fundamental predicates be really harmonious, that is, really possibly co-instantiated. And only the actual co-instantiation of all fundamental predicates in one being can provide such a warranty. No other being that instantiates a subset of all fundamental predicates can warrant the real harmony of all the fundamental predicates. It follows that the ultimate material ground of all real possibility must be a unique being that actually instantiates the complete set of the most fundamental predicates. However, as Chignell himself admits, this is certainly not the path that Kant takes in OPA.Footnote 36
I suggest that the reasoning behind Kant's insistence on the uniqueness of the ultimate material ground of all real possibility should be understood in connection with his reasoning for the de re necessity of this ground. In the first stage of the proof Kant defines absolute real impossibility (5) and uses it for the conclusion of the de dicto necessity that something exists. In transitioning to the second stage, Kant directly presents a de re version of his definition of absolute real necessity:
(7) The being the non-existence of which cancels all real possibility (by cancelling all real content) is absolutely really necessary.Footnote 37 (Definition)
Following this definition, Kant writes, ‘All possibility presupposes something actual in and through which all that can be thought is given. Accordingly, there is a certain reality, the cancellation of which would itself cancel all internal possibility’ (2: 83). While at first sight this passage does not strike one as bringing anything novel to the ‘grounding premise’, it does in fact introduce a substantial premise. The implication of the ‘grounding premise’ was the idea that every particular real possibility presupposes the existence of a material ground, which could get us as far as (6). Here, however, the idea is that all real possibility, i.e. the totality of all particular real possibilities, presupposes a common material ground in which all real content, i.e. the totality of all really possible predicates, is actually given. And only a being that instantiates all of the most fundamental predicates can ground the totality of real possibilities on its own. This new premise can thus be formulated in the following way:
(8) All predicates (of all really possible things) must be grounded in one existing thing that actually instantiates the complete set of the most fundamental predicates. (Premise)
Once this premise is accepted, it becomes obvious that this being that is the common ultimate material ground of all real possibility fits the definition of the absolutely necessary being described in (7). And since it was established in the first stage that it is absolutely really impossible for there to be no real content, it follows that
(9) There exists an absolutely (really) necessary being. (6, 7, 8)
That this absolutely really necessary being must be unique is a trivial consequence of what it is to be an absolutely really necessary being. To further clarify this point, Kant presents the definition of real contingency as a correlate of (7):
(10) That the non-existence of which does not cancel all real possibility is really contingent.Footnote 38
Accordingly, if a being does not instantiate the complete set of the most fundamental real predicates, then it is a really contingent being. In the event that the instantiation of the most fundamental predicates and thus the job of materially grounding all real possibility is distributed over a multiplicity of being, we would not have the de re necessity of existence defined by (7).Footnote 39 Therefore,
(11) The absolutely really necessary being is unique.
Of course, the success of the entire second stage of the proof depends on the crucial new premise (8). The problem is that Kant does not seem to give any separate justification for holding this premise in OPA. It could be argued, however, that Kant brought some explanation for this premise retrospectively. There is a popular contention in the literature that The Ideal of Pure Reason in the Critique includes a restatement of Kant's own ontotheological proof.Footnote 40 Despite the lack of any explicit reference, the second section of The Ideal of Pure Reason does indeed seem to transform the ‘only possible proof’ into a syllogism of pure reason resulting in the generation of the idea of the most real being, the ens realissimum. Now if this reading is right, then Kant can be understood as introducing a new tool to directly connect particular real possibilities to the ultimate material ground of the totality of all real possibilities, namely the principle of thoroughgoing determination (durchgängigen Bestimmung). According to this principle, for any particular thing, ‘among all possible predicates of things, insofar as they are compared with their opposites, one must apply to it’ (A572/B600). This principle takes us beyond the principle of logical contradiction both in that it ‘deals with the content and not merely the logical form’ of possibility and that it ‘considers every thing further in relation to the whole of possibility, as the sum total all predicates of things in general’ (A572/B600). The idea is that we cannot thoroughly or completely conceive of a particular real possibility, e.g. Socrates, with reference to an incomplete set of positive predicates, e.g. being Greek, Athenian, mortal, philosopher, snub-nosed, etc., but only with reference to the totality of all positive predicates. That is to say, each complete particular real possibility presupposes the totality of all really possible predicates, and consequently, for the former to be materially grounded the latter must be materially grounded as well.
Yet the question still stands: why must the totality of all really possible predicates be grounded collectively in one actual being rather than distributively in multiple actual beings which would altogether constitute a material pool or substratum for all real possibility? Kant insists that, from the idea of such a material substratum, of an All of reality (omnitudo realitatis), the natural urge of our reason for simplicity drives us towards the ‘ideal of pure reason’, that is, the idea of an individual object that instantiates this substratum in a ‘refined’ way, ‘exclud[ing] multiplicity of predicates, which as derived through others, are already given, or cannot coexist with one another’ (A573–4/B601–2). This ‘ideal’, Kant states, is nothing but the idea of an ens realissimum, conceived as the being that instantiates the set of most fundamental predicates from which all other really possible predicates can be derived (A576/B604). Kant also suggests, as a further vindication of the ideal, that ‘among all the concepts of possible things the concept of a being having the highest reality would be the best suited to the concept of an unconditionally necessary being’ (A586/B614). Thus, when we reflect on the conditions of complete particular real possibilities, thanks to a natural urge of our reason for the absolute or unconditioned, we are led to hold the idea of an ens realissimum, which best suits the concept of an absolutely (really) necessary being.
Of course, this account of why we are rationally compelled to postulate a single material ground for all real possibility is hardly an objective justification for premise (8) or a basis for the demonstration of the existence of the ens realissimum. Nor is it intended to be so. Kant's intention is to show that a proof that moves in this way from particular real possibilities to the absolute (real) necessity of the ens realissimum can at best be taken as subjectively valid, that is, valid for the subjective conditions of our thinking about real possibilities. Such a proof falls short of objective validity. This insight underlines the Critical Kant's more explicit attitude toward the proof in OPA.
5. Real Harmony and the Critical Abandonment of the Only Possible Proof
Chignell's further suggestion that the notion of real harmony is what leads the Critical Kant to ‘undo’ his own proof is also problematic.Footnote 41 Chignell argues that, although Kant employed ‘real harmony’ as a crucial notion in his pre-Critical proof, only later, ‘at some point between 1763 and 1781’, did he notice ‘that his innovative theistic proof relies on the substantive and yet unmotivated Assumption that there is a real harmony between the predicates of the ens realissimum’.Footnote 42
First of all, as Chignell also points out, the Critical Kant never expressly refutes his pre-Critical proof. In fact, in his lectures on religion (from early to mid-1780s), he even states that ‘this proof can in no way be refuted, because it has its ground in the nature of human reason’ (28: 1034).Footnote 43 In those passages from the Lectures (28: 1015–16, 1024–5) that Chignell cites to support his account of the ‘undoing’ of the proof, Kant does not actually talk about his own proof or any other particular proof of the existence of God, but about the broader question of our cognition of the (absolute) real possibility of the ens realissimum. Along with a clear statement of the distinction between the logical possibility of a concept and the real possibility of the object of that concept, Kant suggests that we cannot cognize the (absolute) real possibility of God qua object on the grounds that we do not have the capacity to know a priori whether ‘the effects of the realities do not cancel one another’ (28: 1016, cf. 28: 1025).
Chignell takes Kant's suggestion to be a reference to the danger of ‘subject-cancelling’ repugnance in a concept of the ens realissimum and, consequently, an indication of Kant's realization that he had assumed in OPA, without proof, that there would not be such repugnance.Footnote 44 However, this seems to me a forced reading of the text. Now, it could be said that Kant just assumes in OPA that the synthesis of the most fundamental predicates can be instantiated in one subject. This is indeed right, and is clearly implied by premise (8) in my reconstruction. So it could further be said that Kant realized this deeper assumption and its shakiness later. Yet it is still not plausible to tie the shift in Kant's attitude toward the proof to a concern about ‘subject-cancelling’ real repugnance. For the passages cited above from the lectures suggest that Kant is concerned about a ‘predicate-cancelling’ repugnance as he talks about the cancelling of the effects of one another. And as I argued earlier, the predicate-cancelling type of real repugnance does not lead to the cancellation and thus to the real impossibility of the subject, except in one special case. This is the case of the ens realissimum in which predicate-cancellation leads to a logical impossibility. Thus we still would not have a genuine case of what Chignell means by ‘subject-cancelling’ repugnance. More importantly, Kant was sensitive to this danger already in OPA. In fact, it is exactly this sensitivity on the basis of which Kant refined the concept of God as containing all realities and reformulated it as one that grounds all realities in virtue of instantiating the complete set of most fundamental predicates. It seems to me both highly uncharitable and implausible to think that Kant displayed this sensitivity in the case of the former but noticed only later that the same danger would apply to the more refined concept as well.
The reasons that led the Critical Kant to the conclusion that we cannot know the real possibility of the ens realissimum are broader than a concern about real harmony. In a nutshell, the Critical Kant thinks that we cannot cognize the real modalities of supersensible or noumenal things because such things cannot be given to our intuition and our concepts of them are bound to remain without intuitive content and thus without objective reality. From the mid-1760s to 1780s, Kant's position on our modal cognition underwent a crucial shift. While the Kant of OPA, still in a rationalist mindset, operated with absolute real modalities without worrying too much about the epistemological conditions of our cognition of real modalities, the Critical Kant came to hold that absolute real modalities of things in general, without regard to the conditions of the givenness of things to our mode of experience, are beyond our cognition, and that our cognition of real modalities is restricted to that of relative real modalities, i.e. modalities relative to the conditions of our experience.Footnote 45 We should therefore remain agnostic about the absolute real possibility of things in general. All we can know is the relative real possibility of things insofar as they agree with the form of our experience, i.e. insofar as their predicates can be given in space and time and thought by the categories. This underscores another important point. In order to explain our ignorance of the absolute real possibility of noumenal things like God, we do not need the elaborate story about the real harmony requirement, which is, I have argued, difficult to textually support anyway. In fact, in the case of the absolute real possibility of noumena, the failure to satisfy the content requirement is sufficient to explain our ignorance. The predicates of noumena cannot be given to our mode of intuition, and consequently, we cannot know whether they are really possible predicates that are grounded in actual existence and not just empty words.Footnote 46
Having pointed out the problem of our cognition of absolute real possibilities, I would further diverge from Chignell regarding why exactly the Critical Kant ‘undid’ the pre-Critical proof. In the Lectures, Kant downgrades the conclusion of the proof some twenty pages after his announcement that we cannot cognize the real possibility of the ens realissimum and his formulation does not involve an appeal to this issue:
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 necessity of assuming such a being. (28: 1034)
Thus the Critical Kant, despite still sympathizing with its deductive form, strips the pre-Critical of its objective existential claim, i.e. the claim that the ens realissimum qua the object, necessarily exists, and downgrades it to a demonstration of the necessary presupposition of the idea of the ens realissimum, which is only a subjective necessity arising not from the constitution of things in general but from the specific constitution of the human cognitive subject. It is important to note that this approach to OPA is not peculiar to the Kant of 1780s alone. It is possible to observe a consistent tendency in Kant's thought, starting from the late 1760s, towards a downgrading of the proof as valid merely subjectively and consequently redefining the ideal of the ens realissimum as a subjectively necessary hypothesis of human reason. For instance, in a note from 1769–70 referring to his proof, he writes: ‘The conclusion here follows from the possibility of cognition through our reason, not from the given things, and therefore the argument is valid ad homines’ (Reflexion 4249, 17: 481). In another note dated 1772–5, he goes on to say that ‘the necessity of the divine existence is a necessary hypothesis of mere concepts of the possible’, but ‘the absolutely necessary cannot be proven’. He then adds, ‘The proof of God's existence is not apodictic, but hypothetical sub hypothesi logica’ (Reflexion 4580, 17: 600). In a reflection from 1773–5 he writes, ‘the concept of all reality as a substratum of reason is necessary for us; but we cannot on that account regard a highest reality as necessary in itself’ (Reflexion 4729, 17: 689).Footnote 47
This is a radical shift in Kant's perspective on what the only possible basis of proof actually demonstrates. Whether this shift is in essence part of Kant's developing Critical epistemology and goes hand in hand with the Critical direction his conception of modality started taking between OPA and the Critique is an interesting question in and of itself, but remains beyond the bounds of this article.Footnote 48 Yet it should at least be clear that Kant's devaluation of the conclusion of his proof is motivated by a concern about the validity conditions of existential claims which goes far deeper than a realization of our ignorance of the real possibility of the ens realissimum. Thus, even accepting Chignell's assertion that this ignorance is due to a failure on our part to verify the real harmony of the predicates of the ens realissimum, his overall explanation of Kant's ‘undoing’ of the proof would not work.
6. Conclusion
Chignell's ‘retroactive reconstruction’ concedes that ‘it is not always clear in Kant's early texts that he was aware of how crucial a role the Harmony condition plays’.Footnote 49 I have argued against this reconstruction on the ground that, while Kant does introduce the idea of real repugnance (and by implication, real harmony) as early as 1762, he does not incorporate this idea into his extra-logical or real notion of possibility as a condition. Moreover, I have argued, this idea of real harmony does not play the suggested role either in Kant's actual construction of the proof in NE and OPA or in his later downgrading of the proof.
