Leibniz and Kant on God

Holger Gutschmidt
Department for the Study of Modern Rationality, República Checa

Revista de Estudios Kantianos. Publicación internacional de la SEKLE

Universitat de València, España

ISSN-e: 2445-0669

Periodicity: Semestral

vol. 8, no. 2, 2024

Received: 16 November 2022

Accepted: 03 August 2023

Los autores conservan los derechos de autor y garantizan a la revista el derecho de ser la primera publicación del trabajo al igual que licenciado bajo una Creative Commons Attribution License que permite a otros compartir el trabajo con un reconocimiento de la autoría del trabajo y la publicación inicial en esta revista.

Abstract: Is Immanuel Kant’s critique of the proofs of God’s existence accurate? In order to answer this question, I analyse Leibniz’ proof in his Monadology and I determine the relation between the cosmological and the ontological version of this proof. Since Kant often refers (implicitly) to Leibniz’s argument, I examine Kant’s critique of rational theology in the Critique of pure reason in the light of Leibniz’ strategy. We will see that Kant’s critique of Leibniz’ argument is partly biased by his own concept of ‘transcendental idealism’ and fails to reject it. On the other hand, many questions in Leibniz’ attempt to solve the problem of the poof of God’s existence remain open.

Keywords: Leibniz, Kant, God, Ontological proof, Cosmological proof.

Resumen: ¿Es válida la crítica de Immanuel Kant a las pruebas de la existencia de Dios? Para responder a esta pregunta analizo la prueba de Leibniz en su Monadología y determino la relación entre la versión cosmológica y ontológica de esta prueba. Dado que Kant se refiere a menudo (implícitamente) al argumento de Leibniz, examino la crítica de Kant a la teología racional en la Crítica de la razón pura a la luz de la estrategia de Leibniz. Veremos que la crítica de Kant al argumento de Leibniz es parcialmente sesgada por su propio concepto de ‘idealismo trascendental’ y no logra rechazarlo. Por lo demás, muchas cuestiones en el intento de Leibniz de resolver el problema de la prueba de la existencia de Dios permanecen abiertas.

Palabras clave: Leibniz, Kant, Dios, Prueba ontológica, Prueba cosmológica.

1. Introduction

Famously, Theodor W. Adorno was still convinced that all philosophical thought ultimately circles around the demonstration of the existence of God (1975, p. 378).[2] And yet, the dominant view of the last two centuries has long been that the project of proving God’s existence was a clear failure. The support for this stance was provided in Immanuel Kant’s famous criticism of the proofs of God in his Critique of pure reason (1781, 1787.) and it was subsequently developed by Gottlob Frege in his 1884 Foundations of arithmetics where he defined the expression of existence in a novel way, leading to the general conviction that all the classical medieval and early modern demonstrations of God’s existence break down over semantic and logic errors. Only after World War II was there an increase of interest in these theorems when Kurt Gödel, Alvin Plantinga, Robert Spaemann and others offered attempts at a modern demonstration, sometimes based upon modal logic. The current discussion is characterized by a certain openness which has also led to a reassessment of the refutations attempted by Kant and Frege.

And this reassessment has also motivated a new reading of the classics of the project of proving God, including 17th-century authors such as René Descartes (who renewed the ontological proof of God, formulated in 1077 by Anselm of Canterbury), Baruch de Spinoza and Gottfried Wilhelm Leibniz. Leibniz’s main contribution was that he brought greater precision to the ontological proof of God, and he also presented an important version of the cosmological proof (for a survey, see Look, 2018). Moreover, it is in Leibniz that we find most distinctly an outline of a program of proving God, later attacked by Kant when he put forward the claim that the cosmological and teleological proof, far from being independent, presuppose the ontological proof, and thus all the types of proof must be analysed together. The failure of the ontological proof–so Kant’s summary–is a crucial reason of the failure of all the other proofs. Kant’s criticism was aimed at the formulations of these proofs presented by Alexander Gottlieb Baumgarten, Christian Wolff, and most importantly, Leibniz himself.

In order to understand and evaluate Kant’s criticism of this program, we thus have to reconstruct both Leibniz’s proof of God and Kant’s refutation and measure one against the other. It shall be seen that while Kant was correct in postulating a link between the various proofs, Leibniz’s own position is somewhat different from what Kant thought it to be. It will also be shown that even though Kant’s criticism of the cosmological proof does point out important problematic features of this type of proof, it stands and falls with his notion of transcendental idealism. And finally, we will demonstrate that Kant’s criticism of the ontological proof fails. This, to be sure, does not mean an automatic rehabilitation of Leibniz’s program, which includes some far-reaching presuppositions–such as, for instance, the so-called ‘principle of sufficient reason’–that require a separate discussion (cf., e.g., Pruss, 2009; Sobel, 2004, pp. 200-237). It does turn out, though, that the assumption–dear to Kantians up to this day–that the philosophical ‘ontotheology’ of the rationalists has been fundamentally destroyed by Kant (e.g., Höffe, 2011, pp. 261-265), is unpersuasive. In fact, Leibniz provides an excellent example of the potential contained in these theories. And, to take up and modify Kant’s well-known remark, it is yet another question–one we have to leave to the side at this moment–whether faith can be delimited so as to make room for thinking. Here we shall limit ourselves to the simple observation that Leibniz–as well as Descartes–was a convinced Christian and, in his own way, a pious man.[3]

2. Leibniz on the cosmological argument

The cosmological and the ontological proofs are presented together in Leibniz’s 1714 Monadology, §§ 36-45.[4] In the immediately preceding paragraphs (§§ 31-35) Leibniz explains some crucial presuppositions of his proof, especially the two ‘great principles’: the principle of the (excluded) contradiction and the principle of sufficient reason. According to § 32, the principle of sufficient reason applies both to events [faits] and to propositions [enontiations]. § 36 makes the consonant claim that sufficient reason is to be found both in ‘contingent truths or truths of fact’ and in the so-called ‘truths of reason’ which are necessarily true. That means that Leibniz does not limit this principle to causality (which, for him, includes ‘final causes’). The argument which Leibniz construes upon the basis of these distinctions in §§ 37-38 can be summed up as follows:[5]

1. The world consists of things and facts that are contingent, i.e. their opposite is possible. Anything whereof the opposite is possible does not exist necessarily.

2. Whatever does not exist necessarily cannot possess the reason (or the cause) of its being in itself. Rather, it must have its reason in something distinct from itself.

3. However, if this distinct thing were to have its own reason in something yet different, then it is by itself not a sufficient reason (even though it may be a part or an element of sufficient reasoning).

4. Therefore, a sufficient reason must be an ultimate reason, such that the reason of its own being does not reside in something distinct from itself. With regard of the world, this reason is called ‘God’.

5. Therefore, if the world consists of contingent things and facts, God must exist; and since–according to Proposition 1–the antecedent is true, God exists.

The key concepts of Leibniz’s argument are those of contingency and sufficiency. Let us look first at the concept of contingency. For Leibniz, to be contingent is to admit of not being or of being otherwise, i.e. to not be necessarily the case, or again, to not be necessarily the way it is. This allows various interpretations: a) The contingent could be without a reason. However, the principle of sufficient reason excludes this, and thus it co-determines Leibniz’s concept of contingency. Yet even so, it remains open as to what we are to say regarding the ultimate reason, the sufficient reason of everything. b) The contingent could be the epistemically contingent. That would mean that in merely observing it we could not determine whether it is necessary, just as from observing the fact that the sum of a particular triangle’s angles equals two right angles, we cannot tell whether triangles (and, therefore, this one triangle as well) necessarily exhibit such an equation. However, it is either the case that the sufficient reason of the epistemically contingent is also contingent itself, and then it possesses only a preliminary necessity (what Leibniz would call a case of ‘hypothetical necessity’) (cf. Balestra, 2003, p. 87f. and passim); or it is the case that this reason itself is necessary, and then its consequences are necessary just as well, their contingency being merely apparent. c) The contingent could be the intrinsically contingent. However, this is only possible if, either, its own sufficient reason is (intrinsically) contingent, or else if there is no necessary connection between the sufficient reason and what follows from it.

By sufficient reason of a fact or a thing Leibniz means a reason such that it provides a complete determination of this fact or thing. This means, first, that the fact or thing may have no properties undetermined by this reason, and it also means, second, that this reason itself does not receive its existence from something distinct from it, since otherwise its capacity to determine the fact or thing would not completely stem from itself; rather, at least in part, it would derive from whatever it is that determines the reason itself. Further, as can be seen from the text of the Monadology, in this context the term ‘reason’ designates not only the effective or the final cause but also a logical ground.[6] (Otherwise God as causa sui would have to be thought of as involved in a relationship of cause and effect, leading to paradoxes) (cf. Cramer, 2010, p. 20f.). Thus, if God exists necessarily, his existence must be necessary in a manner different from hypothetical necessity. Leibniz calls it ‘metaphysical necessity’ and its further clarification is provided only in the context of the ontological proof, following the cosmological one.

Finally, we have to point out a third, implicit premiss in Leibniz’s argument; namely, that everything there is, is conditioned either by hypothetical necessity (i.e. by something else), or by metaphysical necessity. For Leibniz, this constitutes a complete disjunction–tertium non datur. In this sense, the principle of sufficient reason is for Leibniz a positive formulation of the classical negative principle ex nihilo nihil fit.

3. The link between the cosmological and the ontological argument

Before we approach the ontological argument and the connection it has in Leibniz with the cosmological one, we must highlight a crucial observation. Regarding the ontological argument it is often overlooked that it relies upon a hermeneutical assumption, one that is clearly present already in its earliest version in Anselm’s Proslogion (see Anselm of Canterbury, Proslogion, cap. II): even the ‘fool’ who rejects God’s existence acknowledges that God is understood to be that than which nothing greater can be thought [quo maius nihil cogitari potest]. Anselm develops his argument based on this definition of ‘God’, shared by both sides. In renewing Anselm’s proof in his 1642 Meditations on First Philosophy, Descartes also has the subject of the meditation encounter in their own mind the idea of the ‘most perfect being’ [ens perfectissimum] (The Third Meditation [Adam and Tannery, 1996, VII, p. 40); and it is on the basis of this generally shared idea of God that Descartes then in the Fifth Meditation performs the ontological proof. However, Descartes notices that such an assumption is not free of its problems: the idea might be an arbitrary one, its general acceptation might be only due to convention. This is why Descartes demonstrates in the so-called ‘causal proof of God’ in the Third Meditation that a human could not possibly produce such an idea of their own, so that it must stem from God (cf. Gutschmidt, 2014, pp. 114-122). Descartes’ strategy of starting from a generally shared definition is tackled by his sparring partner Marin Mersenne from a different angle, in that Mersenne doubts whether one can correctly deduce the existence of God from a concept of God’s essence.[7] And the same problem is faced by Baruch de Spinoza in his own version of the ontological proof in Book I of the Ethics, when instead of construing the argument upon the previously current basis of a single definition of God’s essence (such as ens perfectissimum, ens infinitum, etc.) and analysing it further, he divides the concept of God into several separate notions (causa sui, substantia, Deus) and uses them, as (nominal) definitions, as the basis for his argument in Book I.[8] Finally, Leibniz follows to a degree the same procedure when, regarding each definition of essence upon which the ontological argument is founded, he demands a conceptual analysis which must show that the notions included in the definition are logically compatible–since, naturally, from mutually contradictory notional sets (even if they constitute a higher unity of meaning, as is the case with paradoxical concepts) one cannot demonstrate that God’s essence includes existence (cf. Gutschmidt, 2019). Leibniz highlights the necessity of such a conceptual analysis also in the part of Monadology we have summarized above (§§ 33-35, 44; cf. Oppy, 1995, pp. 24-26, 219-225).

This proves that even though the 17th-century authors mostly adopt Anselm’s strategy of starting the ontological proof from a generally shared definition of God’s essence, they also outline certain tests so as to secure this definition from arbitrariness or internal contradiction. Still, for one who does not share this presumptive definition of God, the whole procedure only shows that the proof–even in an ‘improved’ form–possesses no conclusive validity. Things stand differently with the cosmological proof, however, as it relies upon an assumption shared by everybody–that is, that the world (or at the very least, the person who is performing the proof) exists!

Let us look further at Leibniz’s argument in the Monadology. In § 37, Leibniz points out that the ultimate reason must lie ‘outside’ [hors] the sequence of all that is contingent, no matter how far this sequence may extend. By itself, this remark does not exclude the possibility that the reason itself be in some sense ‘contingent’.[9] It does state, however, that the ultimate reason’s existence has, by itself, no further reason that would be part of the thus grounded world. Nor should we overlook Leibniz’s reference to the potential infinity of the sequence of the grounded events and things of the world. Were it so that the world is determined by an infinite sequence of things and events, it might indicate that the sufficient reason itself, far from being an element (or even the beginning) of this sequence, stands towards it in a relationship fundamentally different from the relationship of any single element of this sequence to any other element. Kant will adopt this notion in his concept of a ‘cosmological idea’.

Once Leibniz deduces in § 38 a ‘necessary substance’ [substance necessaire] as the sufficient reason of the ‘intricate detail of changes’ and calls it ‘God’, there follows in §§ 39-41 a series of further and non-trivial characteristics of God: He is one (§ 39), there is nothing outside of him and he possesses an unlimited reality, i.e. objectivity (§ 40) and is infinitely perfect (§ 41). Via this sequence of predications attributed to the ‘necessary being’ (the sufficient reason of the world) Leibniz ultimately attains ‘the most perfect being’ (God), that is, the starting point of the ontological proof, performed in §§ 44-45.

However, as mentioned above, this transition is non-trivial, and the claim that it consists in a mere analysis of the result of the cosmological argument is implausible. It must be objected that: a) The affirmation of the ‘unity’ of God presupposes the unity of the world which Leibniz had not demonstrated. It further presupposes that there are no other ‘necessary beings’ that would ground either parts of the world or else their own existence. Leibniz’s claim also presupposes that the world consists of contingentia only and that these have one and only one ultimate reason. b) Nor does it follow from the result of the cosmological argument that God possesses an unlimited reality, i.e. that it is an ens realissimum. In Leibniz’s argument, one can speak of God’s reality only from the viewpoint of the world as grounded by God (i.e. from the viewpoint of the cosmological proof), not absolutely. Only if the world itself were to exhibit unlimited reality, could one state the same about its sufficient reason. Yet, up to now, a proof of this is just as lacking as a demonstration of the unity of the world. It does not even follow from Leibniz’s statement regarding God in § 40 (that He is ‘unique, universal, and necessary’), since Leibniz has not excluded that a reality greater than the reality of the universal God can at least be thought. c) Even though the statement of identity between the most real being with the most perfect being in § 41 relies initially only upon a simple act of assigning what these words mean (“since perfection is nothing other than magnitude positive reality, taken in the strict sense”), this step is far from being harmless. In the Fifth Meditation (Adam and Tannery, 1996, VII, p. 65n., and also see p. 115n. [“First Replies”]), Descartes deduced from the concept of the most perfect being that some such necessarily exists. As to the objection that, normally, one distinguishes between the concept and the being of a thing, Descartes replies that the concept of the most perfect being constitutes an exception from this rule. Students of Descartes have usually taken him to be saying that in the case of the most perfect being, ‘existence’ belongs as a notion to its concept, since the expression ‘the most perfect’ reaches beyond the realm of objective (‘real’) predicates.[10] Yet it is clear that regarding the determination of God as the ‘most real’ being, such an argumentation is impossible. Moreover, we have attained a more sharpened understanding of the principle of sufficient reason in the case of its application to ‘the necessary being’: the identification of ens necessarium with ens realissimum as well as that of ens realissimum with ens perfectissimum guarantees that the principle applies even to that which is the sufficient reason of everything, since that whose essence includes its own existence is nothing but the sufficient reason of itself!

Yet how are we to interpret Leibniz’s semantic expansion of the term ‘sufficient reason’? Are we to view it with Kant as ‘subreption’, i.e. argumentative deceit? The decisive transition to this expansion takes place already in § 38, when the sufficient reason of ‘things’ is denominated ‘God’. By itself, this identification seems spurious. If the cosmological argument is valid, then it always leads to something that explains the reality of the world. Thus, it leads to a highest or a last explanatory reason–whose characteristics, however, remain dependent upon that which it is supposed to explain. Thus, as we have seen, it is dependent upon how we understand the world. On the other hand, Leibniz–as we understand him now–has quite an ambitious notion of the sufficiency of the reason. For Leibniz, the concept excludes both that the reason of the world depends on something else (something transmundane) and that it simply ‘be there’ without our being capable to understand why. A first (or last) cause of the world, an ‘immovable mover’, might be contingent in this sense–whereas the ‘sufficient’ reason (in Leibniz’s expansive understanding) can never be contingent! Thus–interestingly–the very principle which, for Leibniz, makes the cosmological argument at all possible, also simultaneously points beyond it: the requirement of the sufficiency of the reason carries with it an interpretation of the result of the cosmological proof such that this proof is linked up to the ontological one. Therefore, the sufficient reason must be an ens necessarium in the strict sense, i.e. a being that necessarily exists (as against a being merely necessary for the existence of the reality which it grounds). Yet as long as it is not clear which property or properties make it possess this necessary existence, it seems to be a mere postulate. The requirement of the sufficiency of the reason is completely fulfilled only once we possess a description of the sufficient reason such that it states the reason why it necessarily exists–and then it is necessarily true that God necessarily exists.[11]

All this, however, can be understood only in connection with the corresponding definitions of God’s essence. Only by virtue of this connection does it become conceptually or ‘intrinsically’ true that God exists: to reject His existence, or even simply to put it in doubt, would make an impossibility out of the very definition of God as ‘the most perfect being’. This is the well-known link between the definition of essence and the statement of existence, as employed by the ontological proof and as explicitly referenced by Leibniz at the end of his argument in the Monadology (§§ 44-45). True, Leibniz also states in § 45 that God can be demonstrated in various ways, a priori (via the ontological argument) or a posteriori (via the cosmological argument). However, as we have seen, this is a half-truth at best. Rather, Leibniz’s version of the cosmological argument needs to be complemented by the ontological proof, since otherwise it is not completely done. Only the ontological argument provides knowledge as to how far the cosmological argument has really attained the ultimate reason of what is. Therefore, only the ontological argument guarantees that the reason of the real, as deduced by the cosmological argument, is sufficient in every sense. And finally, it is only the ontological argument that allows us to identify the ultimate reason with God.

Here, we can attach (though only in an outline) one further consideration. One can ask: is it not even the case that, in Leibniz, the ontological argument actually somehow presupposes the cosmological proof? As we have seen, even in the pre-Leibnizian versions of the ontological argument, there was some scepticism regarding the concepts employed in it. Can a clear meaning be attached to expressions such as ‘the infinite being’, ‘the most perfect being’, or ‘the being than which no greater can be thought’? Is it actually even possible to mark out such concepts consistently? They definitely acquire a meaning in the context of the cosmological argument–for only by means of them can one adequately interpret its result. This goes to show, however, that it is precisely the context of this argument which specifically legitimizes the use of such definitions. They cannot be simply accepted as mental or linguistic facts; rather, they constitute a presupposition for concluding the cosmological argument. It is true that in a sense this legitimizes them only functionally or ‘pragmatically’, yet at the very least this makes them more immune to attack than it is the case in an exclusively ontological proof. If this line of thought is correct, then Leibniz’s suggestion that there are various mutually independent ways of demonstrating God’s existence is not even ‘half-true’.

4. Kant’s interpretation and critique of the program of proving God’s existence

For Kant, propositions about reality possess a definite truth-value insofar as they can be experienced. However, all possible experience of the real is subject to transcendental conditions. These are, one, the pure forms of intuition, i.e. space and time, and two, the categories as the principles of synthetizing appearances. One among the categories is that of cause and effect. It constitutes a narrow version of Leibniz’s principle of reason: a narrow one because, on the one hand, ‘reason’ is limited here to efficient causes, and on the other, there is no criterion of ‘sufficiency’. Such a criterion establishes a weighing of reasons (distinguishing simple reasons from sufficient reasons) which is crucial for the cosmological argument. The principle of cause and effect, on the other hand, recognizes no such weighing. Yet it does possess constitutive significance for all objects of our experience, and thus also for our concept of reality. Kant discusses this distinction in the section on the so-called “Fourth Antinomy” (A452ff./B480ff.) and in the explanations attached to this section (“A critical resolution of the cosmological conflict of reason with itself”, A497ff./B525ff.). The antithesis between the stance that the world contains a ‘simply necessary being’ and the one that there is no such being in the world is, according to these texts, based upon the double application of the principle of reason: once to ‘things-in-themselves’, once to ‘appearances’. Whenever it is applied to things-in-themselves, the relationships between reason and its consequence consist in the specific relationships between these particular things. Thus, a full determination of reality must go back to the ultimate reason which reasons all the other relationships between reason and its consequence. As long as this is not the case, things and the entire reality are not fully determined, or conceptually grasped (A497f./B526f.). With regard to appearances, however, the ‘principle of reason’ functions quite differently, as it constitutes the objectivity of the objects of experience. This means that there can be no object of (possible) experience not grasped as cause and simultaneously as effect. Here, then, the principle has a generic significance. Moreover, the relationships between cause and effect are transitive, never reflective (so that nothing can be ‘cause of itself’). Therefore, an object of experience as a thing effected must always have its cause in a different object of experience, and as a causal thing it must always have its effect in a different object of experience.[12] This by itself makes it impossible that there be an ‘ultimate reason’ (or a ‘first cause’) in the world of appearances.

Still, Kant does not declare the concept of an ‘absolutely necessary being’ a conceptual nonsense. On the contrary, it is a concept stemming from reason itself. However, concepts such as ‘the unconditional’ or ‘totality’ (Kant calls them ‘ideas’), while possible and even required for experience (since together they make room for the concept of a ‘world’, see A407f./B434), have no experiences corresponding to them (on Kant’s doctrine of ideas, see, e.g., Klimmek, 2005). For Kant, ‘the world’ is not an object of ‘objective’ knowledge. That is why the claims of reason regarding the totality of the synthesis of appearances, and thus the series of their conditions, get fulfilled only in thoughts which are ‘problematic’ (i.e. possible but not binding), not in empirical or pure cognitions. With regard to the cosmological ‘idea’, this even leads to a rehabilitation of Leibniz’s principle of sufficient reason (A562/B590), insofar as it can be thought that “the entire [empirical] series could be grounded in some intelligible being (which is therefore free of every empirical condition, containing, rather, the reason of the possibility of all these appearances)” (Kant, 1998, p. 547f.). This intelligible being, however, is not understood to be the highest cause (since then it would have to be an object of experience) but rather the substance of the world. Insofar as the principle is applied here to an intelligible being, it lacks binding validity. Therefore, it cannot ever provide any basis for a proof, only for the formation of an ‘idea’.

Kant’s position that cognitions can be based only in ‘synthetic judgements a priori’ (pure judgements of experience) and that only in this context does it make sense to apply the ‘principle of reason’ (as causality), does provide the interesting piece of information that Kant’s transcendental idealism is incompatible with the cosmological proof, yet it might well be quite unconvincing for anyone who admires the proof. Kant cannot even admit the application of the principle of reason to facts or entities such as the ‘thing in itself’ (from whose existence Kant starts out) or the transcendental subject of knowledge (whose existence is indubitable for Kant). As one cannot be possibly reduced to the other (i.e. by themselves neither is an ens necessarium), the relationship between the epistemic subject and the ens (or entia) from which appearances are derived would provide a good starting point for a cosmological argument. However, since for Kant neither is an object of possible experience, no epistemically valid elucidation can be provided for them.

Since the notion of knowability of an ens necessarium is rejected by Kant a limine, it is rejected also in the form of those definitions from which one can deduce God’s existence, i.e. in the form adopted by the “necessary being” in the ontological argument (A606ff./B634ff.). And yet, significant light is thrown upon his understanding of the cosmological proof by his identification of both entia necessaria–the one from the cosmological proof as well as the one from the ontological proof. For Kant views the cosmological proof quite similarly to Leibniz. Leibniz’s criterion of sufficiency prohibits it that the reason of the world function as something contingent by itself. However, if we adopt this criterion not in the strict Leibnizian sense but rather only so as to secure the explicative force of the reason for what it grounds, then one can have a cosmological argument whose conclusion does not amount to an ens necessarium. This, however, is out for question both for Leibniz and for Kant, and thus, almost inevitably, there comes up for them a linkage between the cosmological and the ontological argument. Yet while for Leibniz the ontological proof provides only the definition which allows us to understand how far the sufficient reason of everything really ‘suffices’, Kant claims that the cosmological proof, arriving as it does at the concept of an ens necessarium, merely guides toward the ontological proof which then carries the actual burden of securing the ens necessarium. For Leibniz, on the other hand, the ontological argument is primarily a complement for understanding the result of the cosmological proof–which is why he could suggest in the Monadology that both proofs are mutually independent. For Leibniz, the cosmological argument by itself suffices for carrying the burden of proving the affirmation of the existence of an ens necessarium. Still: it does not clarify what such an ens is, or (in Kantian terms) how such an ens is ‘possible’. That is achieved only by definitions such as ens realissimum or ens perfectissimum, which also provide the semantic content for identifying the sufficient reason with God. The fact that one can go on to derive an ontological argument from them only serves to confirm their value as interpretations of the result of the cosmological proof.

All that is quite contrary to Kant. Since transcendental idealism admits of no cosmological argument other than in the form of the ‘cosmological idea’, it was handy for Kant to seek the argumentative burden of the proof of God in the concept of the ens necessarium, or again in those definitions of God’s essence which contain this concept. Thus, in Kant, only the ontological proof functions as a proof of God, whereas the other types of proofs (i.e. the cosmological and the teleological proof) have an introductory character at best.

To conclude, let us return to Kant’s rejection of the ontological argument (cf. esp. A592ff./B620ff.). Even though nowadays this finding is sometimes contested (Hindrichs, 2011, § 82, p. 105f.), Kant’s principal objection concerns the usage of the expression ‘to be’ (in the logical sense of ‘there is’). Kant counters that in the ontological proof, this expression is used as a denomination of a (first-degree) property, even though this expression contributes nothing towards characterizing the given object (cf. the well-known statement at A599/B627: “A hundred actual dollars do not contain the least bit more than a hundred possible ones” [Kant, 1998, p. 567]), and thus it cannot designate a ‘real’ predicate. Since, however, the burden of the proof hinges upon this, the proof–so Kant–must fail.

It would not be of any use to list here all the arguments for and against that have been presented in the history of the ontological proof. For our purposes of the comparison with Leibniz, a few references will suffice. a) Genetic or ‘causal’ definitions (the preferred type of definition in Leibniz) (see Discours de Métaphysique, § XXIV [Leibniz, 1880, p. 449f.]) contain what they say, insofar as they contain the prescriptions for producing it. Thus, there definitely are concepts such that the reality of the things described by them is part of their definition. Since they pertain exclusively to mathematics and logic, they belong for Kant among pure concepts of possible experience, and thus do not establish a counter-example to his criticism of the ontological argument. Still, Kant’s suggestion that only real experience can decide about the fulfilment of concepts is misleading. b) Propositions such as ‘God’s essence includes existence’ (see Monadology, § 44) do seem to designate the pertinent predicate expression in the relation of being contained, yet we do not find unequivocal statement regarding this point either in Leibniz or in other authors. On the contrary, there are good reasons for interpreting these formulations differently. c) Finally, the tools of modern logic allow us to demonstrate that the quantifier of existence can be made into a legitimate first-degree predicate which can then easily function as a property in the concept of an ens (Bromand, 2011, pp. 196-200). This fends off the logical objection formulated by Frege and often imputed to Kant as well. Therefore, the main task remains to understand what it means to say that something is of a ‘nature’ which allows it to necessarily be. The simple fact that ‘existence’ ought to be understood as a quantifier does not by itself demonstrate that such a thing is impossible or that it is impossible to form a concept of such a thing.

Due to Kant’s linking up of the cosmological and the ontological proof, his refutation of the ontological proof adopts central significance. This presupposes, however, that Kant’s reconstruction of this proof has general validity (i.e. that it encompasses all the variants of this proof). Kant did not even attempt to show this, and it is dubious whether he could succeed. If, however, his attempts at refuting individual variants does not persuade, what remains to reject the cosmological argument is exclusively Kant’s starting point: transcendental idealism. However, it may well be imagined that transcendental idealism will fail to convince an adherent of natural theology that we can merely believe in God’s existence but we cannot know it.

5. Conclusion

Kant’s thesis that the cosmological and the ontological argument belong together, forming a program of proving God’s existence, has been shown to be true for Leibniz. However, pace Kant, the reason is not that, by itself, the cosmological argument fails, but rather that Leibniz’s requirement of the ‘sufficiency’ of reason is especially ambitious. According to it, nothing that functions as sufficient ground can itself be ‘contingent’ in any manner or form, even if it were in a manner different from that of the reality which it grounds. This is also the reason for the broad definition of ‘reason’ in Leibniz’s ‘principle of sufficient reason’: it makes it possible to encompass logical relationships of reason and its consequence, thus (as Kant saw correctly) allowing the transition to the ontological argument. However, since Kant’s criticism of the cosmological argument clearly relies upon transcendental idealism as his philosophical stance, Leibniz is not bound to admit it. Yet even so, it remains a problem for him how are we to understand the concept of the ens necessarium, as resulting from the cosmological proof, and what kind of definition makes such a concept ‘possible’. The concepts employed in the ontological proof function here as ‘interpretamens’ for the concept of the ‘necessary being’ such that they also allow us to understand the ens necessarium as ‘God’. On the other hand, Kant’s attempt to declare such a concept void and to refute the ontological proof that is connected with it depends far too much upon the alleged use of the expression ‘to be’ in this type of argument. Therefore, a Leibnizian stance does not have to be seen as refuted by Kant’s attack. In any case, it remains a crucial task for everybody–both defenders and critics of the project of proving God’s existence–to determine the meaning and the inherent constitution of the ens necessarium.


Adam, C., and Tannery, P. (Eds.) (1996). Œuvresde Descartes. Vrin.

Adorno, T. W. (1975). Negative Dialektik. Suhrkamp.

Balestra, A. (2003). Kontingente Wahrheiten. Ein Beitrag zur Leibnizschen Metaphysik der Substanz. Königshausen & Neumann.

Bromand, J. (2011). Kant und Frege über Existenz. In J. Bromand and G. Kreis (Eds.), Gottesbeweise von Anselm bis Gödel (pp. 244-248). Suhrkamp.

Craig, W. L. (1980). The cosmological argument from Plato to Leibniz. Macmillan.

Cramer, W. (2010). Gottesbeweise und ihre Kritik. Prüfung ihrer Beweiskraft. Klostermann.

Gutschmidt, H. (2014). Objektive Ideen. Untersuchungen zum Verhältnis von Idee, Begriff und Begründung bei René Descartes und in der nachkartesischen Philosophie des 17. Jahrhunderts. Mohr Siebeck.

Gutschmidt, H. (2016). Notio Dei. Descartes’ “Antwort” auf Leibniz’ Kritik am ontologischen Beweis. In W. Li (Ed.), “Für unser Glück oder das Glück anderer”. Vorträge des X. Internationalen Leibniz-Kongresses, Hannover, 18. – 23. Juli 2016, Bd. II (pp. 609-617). Georg Olms.

Gutschmidt, H. (2019). Das Ens necessarium bei Leibniz und seinen Interpreten. Studia Leibnitiana, 51, 119-132.

Hindrichs, G. (2011). Das Absolute und das Subjekt. Untersuchungen zum Verhältnis von Metaphysik und Nachmetaphysik. Klostermann.

Höffe, O. (2011). Kants Kritik der reinen Vernunft. Die Grundlegung der modernen Philosophie. C. H. Beck.

Kant, I. (1998). Critique of pure reason. Cambridge University Press.

Klimmek, N. F. (2005). Kants System der transzendentalen Ideen. Walter de Gruyter.

Leibniz, G. W. (1880). Discours de Métaphysique. In C. I. Gerhardt (Ed.), Die philosophischen Schriften von Gottfried Wilhelm Leibniz, Bd. 4 (pp. 427-463). Weidmann.

Leibniz, G. W. (1890). De rerum originatione radicali. In C. I. Gerhardt (Ed.), Die philosophischen Schriften von Gottfried Wilhelm Leibniz, Bd. 7 (pp. 302-308). Weidmann.

Leibniz, G. W. (2014). Monadology. In: Ll. Strickland (Ed.), Leibniz’ Monadology. A New Translation and Guide. Edinburgh University Press.

Look, B. C. (2018). Arguments for the existence of God. In M. R. Antognazza (Ed.), The Oxford handbook of Leibniz (pp. 701-716). Oxford University Press.

Mojsisch, B. (Ed.) (1989). Kann Gottes Nicht-Sein gedacht werden? Die Kontroverse von Anselm von Canterbury und Gaunilo von Marmoutiers. Dieterich.

Oppy, G. (1995). Ontological arguments and belief in God. Cambridge University Press.

Pruss, A. R. (2009). The Leibnizian cosmological argument. In W. L. Craig and J. P. Moreland (Eds.), The Blackwell companion to natural theology (pp. 24-100). Blackwell.

Sobel, J. H. (2004). Logic and theism. Arguments for and against beliefs in God. Cambridge University Press.

Spinoza, B. (1984). Die Ethik. Reclam.


1 Academy of Sciences of the Czech Republic. Institute of Philosophy. Department for the Study of Modern Rationality, Czech Republic. Contact: ORCID:
2 More recently, the stance has been positively adopted by Hindrichs (2011, §§ 127-132, p. 152ff.).
3 As testified not only by the Theodicey, the only book he published himself (Essais de Théodicee sur la Bonté de Dieu, la Liberté de l’Homme et l’Origine du Mal, 1710), but also by the last paragraph (XXXVII) of his 1686 Discours de Métaphysique (Leibniz, 1880, p. 462f.), whereby metaphysical insights concerning God’s nature are revealed in Jesus Christ.
4 Here and below, we quote Monadology in the translation of Lloyd Strickland.
5 A more detailed version of the cosmological proof is to be found in Leibniz’s 1697 De rerum originatione radicali (see Leibniz, 1890, pp. 302-304).
6 ] On the distinct variants of the principle of sufficient reason in Leibniz see Craig (1980, pp. 259-268).
7 In the “Second Replies” to the Meditations (Adam and Tannery, 1996, VII, p. 127).
8 Ethica Ordine Geometrico demonstrata (1677 [1984]), Pars Prima, Definitiones.
9 It would not be implausible to think that the reason is permanent, i.e. uncreated and incorruptible, and itself needs no sufficient reason that would explain its existence. All that would be true for such a reason would be: it is simply because it is (corresponding to God’s well-known self-predication in Exodus 3, 14: ‘I am who I am’). Nor would it be implausible to think that the ‘outside’ in § 37 is to be understood as referring to a theory of plurality of worlds. Then, the sufficient reason of the world, while necessarily not being an element of the world it grounds, would be part of another world which again would have a sufficient reason for its own existence. In such a case, the cosmological argument would remain valid, yet the identification of its result with ‘God’ would be questionable. Also see below.
10 Not only Kant but Leibniz as well understood Descartes in this way. However, demonstrably there is no solid argument that shows that we must reconstruct the Cartesian argument so as to presuppose ‘existence’ as a predicate (on which see Gutschmidt, 2014, pp. 127-133, showing that the concept of an ens perfectissimum can be understood as a second-degree concept; see also Gutschmidt, 2016). Another piece of evidence for this view is Descartes’ outline of a proof from the concept of the ens summe potens (the “most powerful being”) in the “First Replies” (Adam and Tannery, 1996, VII, p. 119), which even excludes such an option explicitly, since ‘existence’ does not belong among the properties of ‘power’, thus a fortiori not of ‘the highest power’.
11 Thus, if not for Leibniz’s comprehensive understanding of ‘sufficiency’, it would be possible for the cosmological argument to end with the result: It is necessarily true that God exists. In such a case, the necessity would pertain only to the proposition (de dicto), not to God’s existence (de re). Thus, were someone to have a different definition of God, the thought would be possible for him that there is no God.
12 This is what Kant means by the “regress” and the task of a “continued empirical synthesis” of appearances (A499/B527).
Non-profit publishing model to preserve the academic and open nature of scientific communication
HTML generated from XML JATS4R