Kant & Ontotheology

Here I discuss Kant’s theory of ground (a recap of last week’s discussion) and his (in)famous criticism of ontotheology – the position that God’s essence entails his/its existence. We’ll look at the basic elements of Kant’s view of existence and the general structure of his criticism of the ontological argument.

1 Kant on Ground

In his extensive pre-critical discussion of ground in the 1755 New Elucidation of the First Principles of Metaphysical Cognition, Kant defines a ground as “[t]hat which determines a subject in respect of any of its predicates” (1:391). The notion of “determine” here is supposed to be exclusive. A determining ground excludes all other predicates except the one so determined (1:393). Hence, for Kant, unlike Crusius, /all grounds are determining by their nature. Alternately put,

From the concept of a ground it is possible to understand which of the opposed predicates is to be ascribed to the subject and which is to be denied. (1:393)

Despite the appeal to the grammatical terms of “subject” and “predicate”, Kant’s conception of a determining ground here is at once both metaphysical and epistemic. The idea is that there is some basis in virtue of which a entity has a particular feature, and in virtue of which our concept of that entity attributes to it that feature. Hence, “subject” and “predicate” can be read here as designating both the subject position in a categorical judgment, as well as the object (the substance) so designated by the subject concept of that judgment.

To say any more than this quickly gets extremely messy. One important distinction, which we want to respect, is that between a logical ground and a real ground.

As Kant puts it in the ramp-up to the critical period,

Every ground is either logical, by means of which the consequence that is identical to it is posited as a predicate according to the rule of identity, or real, by means of which the consequence that is not identical to it is not posited according to the rule of identity. (Metaphysik Herder (1762-4), 28:11)

This characterization of ground is virtually identical with Crusius’s conception of the difference between ideal and real grounds. Both take “ground” to indicate a necessary relation between (possibly) existing objects or facts. The sense in which a ground “determines” is that it being posited excludes all but one consequence. So positing ground γ excludes all but the consequent δ.

What is meant by “exclusion” here? Consider Kant’s remark in a letter to Reinhold during the “critical” period.

A ground is (in general) that whereby something else (distinct from it) is made determinate {quo posito determinate^a ponitur aliud) [that which being posited determines something else]. A consequent (rationatum) is quod non ponitur nisi posito alio [that which is not posited unless something else is posited]. The ground must thus always be something distinct from the consequent, and he who can provide no ground but the given consequent itself shows that he does not know (or that the thing does not have) any ground! Now this distinction of ground and consequent is either merely logical (having to do with the manner of representation) or real, that is, in the object itself. (11:35 To Karl Leonhard Reinhold. May 12, 1789)

a: This expression must never be left out of the definition of “ground.” For a consequent too, is something that, if I posit it, I must at the same time think something else as posited, that is, a consequent always belongs to something or other that is its ground. But when I think something as consequent, I posit only some ground or other; -which ground is undetermined. (Thus the hypothetical judgment is based on the rule, “a positione consequentis ad positionem antecedentis non valet consequentia” [the movement from the consequent to the antecedent is not valid].) On the other hand, if the ground is posited, the consequent is determined.

This 1789 account of a ground corresponds closely with Kant’s account of ground in the pre-critical Negative Magnitudes as well as the Metaphysik Herder above.

2 Determination & Exclusion

Determination is a form of exclusion, which excludes according to a law. In the case of intellectual (i.e. logical) exclusion, the law is that of contradiction. However, in the case of real exclusion, the law of contradiction is not the basis of exclusion, but rather the structure of (for Kant) space and time (as the conditions of phenomena) itself—what we can literally think of as space or time’s form. Since the law of exclusion by which space or time determines something is fundamentally different from the law of contradiction (because Kant firmly distinguishes the intellect from sensibility), we cannot construe spatial or temporal determination as a function of the intellect’s activity, for the law of contradiction is a basic law of the intellect’s activity. Let me explain each of these points.

Determination is exclusion in the sense of “positing”, or setting, one of two opposing predicates as the predicate of a substance.¹ As Kant puts it in the New Elucidation, “to determine is to posit in such a way that every opposite is excluded” (1:393).² Here “opposing” can mean logically opposing, as in, e.g., the predicates <hot> and <not-hot>. But the opposition needn’t take this merely logical form—i.e. it need not be determined by the law of contradiction. Kant’s clearest examples of this are what he comes to call, starting in the 1760s but continuing through the critical period, “really opposing” grounds, as contrasted with merely “logically opposing” grounds.³ For example in his Negative Magnitudes he says,

Real opposition is that where two predicates of a thing are opposed to each other, but not through the law of contradiction. Here, too, one cancels what is posited by the other, but the consequence is something. The motive force of a body in one direction and an equal tendency of the same body in the opposite direction do not contradict each other; as predicates, they are simultaneously possible in one body. The consequence of such an opposition is rest, which is something. It is, nonetheless, a true opposition. For that which is posited by the one tendency is cancelled by the other tendency, and the two tendencies are true predicates of one and the self-same thing, and they belong to it simultaneously. (NM 2:171-2)

Hence, forces can exclude one another through opposition, but the opposition appealed to here is not that of contradiction. One piece of evidence Kant offers for this is that opposing (i.e. contradictory) predicates cannot be combined in a unity—i.e. in a judgment (A150/B189-90)—while opposing forces can be combined in a unity (i.e. in a substance). Kant’s example in the latter case is a stationary being that is so because of the outcome of being acted on by an opposing force.

So determination is the exclusion of a predicate or property in favor of its opposite. What “opposite” means here depends on the kind of law to which one appeals. In the case of logical determination, the exclusion is defined in terms of logically opposing predicates. In the case of real determination, the exclusion is defined merely negatively, in terms of predicates that oppose though not logically.

The critical period doctrine is thus one according to which a ground determines by virtue of (i) the predicate/property and (ii) the law governing the connection between positing that predicate and positing its consequent. In the case of a logical ground the law is that of contradiction, hence positing some predicate P has the consequence of excluding or opposing ~P. To use a simple example, predicating <bachelor> of John excludes John’s being <not-bachelor>. Since <bachelor> consists of “marks” (Merkmale) like (let’s assume) <male> and <unmarried>, predicating <bachelor> of John “determines” that he is male and unmarried, to the exclusion of his being non-male or married. Thus the logical ground <bachelor> is the basis of various other analytic truths concerning John, such as that he is male and unmarried.

Things work in an analogous manner for the case of real grounds/real determination. Let’s look at another simple example. To predicate of a figure that it has three sides is thereby the ground of its being three-angled, and thus excludes its being not-three-angled. What is the law which, when connected with the predicate <three-sided> determines the figure as three-angled? It is not the law of contradiction (B16, A220-1/B268). Compare what Kant says about two lines not constituting a figure.

in the concept of a figure that is enclosed between two straight lines there is no contradiction, for the concepts of two straight lines and their intersection contain no negation of a figure; rather the impossibility rests not on the concept in itself, but on its construction in space, i.e., on the conditions of space and its determinations (A220-1/B268)

Space, or more specifically its nature as cognizable a priori, determines the laws according to which a ground in space excludes its opposite. In the case of pure geometric figures, the laws of exclusion are not those of contradiction but rather those of space itself—i.e. its topological and metric properties. A three-sided figure must be three-angled; its not being three-angled is excluded, not because of any fact about a logical exclusion of predicates by the relevant concepts, but rather because of facts about space, primarily that it has exactly three dimensions.

These examples hopefully make it easier to see how analogous points might hold for time. What makes one temporal determination a ground of another requires, just as with conceptual judgment, or representation of spatial figures, that we appeal to an exclusion law. So, for example, what makes one temporal region T the ground of another R, e.g. as determining that it is later than T, is not the law of contradiction but rather a law deriving from time. In this case the law is that derived from time’s chronomorphic properties, and in particular that it is a continuous serial order.⁴ Thus predicating that region T is earlier than region R thereby determines the other as later. (FIXME: this seems a bit tautological as stated – the important issue is that the concept <earlier-than> refers to a relation that has the relvant features, even if the concepts <earlier> and <later> are analytically related.)

3 Ground & Intelligibility

In the pre-critical period Kant distinguishes two kinds of ground or reason – viz. ’antecedently’ (antecedenter) and ’consequently’ (consequenter) determining grounds.⁵

That which determines a subject in respect of any of its predicates, is called the ground. Grounds may be differentiated into those which are antecedently determining and those which are consequentially determining. An antecedently determining ground is one, the concept of which precedes that which is determined, i.e. that in the absence of which what is determined would not be intelligible. A consequentially determining ground is one which would not be posited [poneretur] unless the concept which is determined by it had not already been posited from some other source. You can also call the former the reason why, or the ground of being or becoming [rationem cur scilicet essendi vel fiendi], while the latter can be called the ground that, or the ground of knowing [rationem quod scilicet cognoscendi] (NE 1:391-2]

Kant articulates these notions of ground, at least in part, for epistemic purposes. His concern is with the basis upon which we assent or deny the truth of a particular proposition (i.e. judgment). Kant’s idea seems to be that, for any proposition, there must be some basis to which one can point that would allow one to decide between the truth of a proposition or its negation. If there were not, then we would have to suspend judgment concerning the proposition since each of its “two predicates [i.e. the predicate or its negation] remains possible, neither being established as true in respect of our cognition” (NE 1:392).

Hence, for Kant, two kinds of cognition are then possible. Cognition of the consequently determining ground is cognition that one of these predicates is excluded in favor of the other (NE 1:393, 396). Cognition of the antecedently determining ground provides not only cognition that one predicate is excluded in favor of another, but also why this is the case.⁶

One illustrative example Kant provides is that of the ground of evil in the world (NE 1:392). Our consequently determining ground for assenting to the proposition there is evil in the world is our own experience of such evil. So, with regard to the question ’is there evil?’ our experience provides us the basis for a positive answer. It is sufficient for us to decide between the propositions there is evil and there is no evil. But this only gives a reason for thinking that there is evil in the world, not why there is such evil. An antecedently determining ground would give us such a reason why by giving us not merely the fact of evil’s existence, but that in virtue of which it exists.

Despite its epistemic emphasis, Kant’s notion of ground is also ontological. In the example of the ground of evil, Kant’s notion of the antecedently determining ground is not simply one which gives an answer to a why question, but rather does so because it is the metaphysical basis of the evil. As Kant puts it, the “determining ground is not only the criterion of truth; it is also its source.” (NE 1:392). This could be the case for one of two reasons. On the one hand, the ground could be a necessary condition of there being evil. Thus, a full explanation of why evil exists would have to include this ground. But, on the other hand, since all the necessary conditions of evil’s existing could obtain and yet evil not exist, one must also ask after the sufficient conditions of this existence—i.e. after the cause of evil’s existence. Alternately put, an antecedently determining ground is a ground of “being or becoming” (essendi vel fiendi), and it is precisely due to the generative status of the ground that it has its explanatory status (NE 1:396-7, 398).

Consider again the proposition there is evil in the world. To have cognition or knowledge (we can leave niceties of this distinction aside) of a determining ground for this proposition is to have knowledge that excludes the proposition’s negation. In the case of a consequently determining ground, such as one’s experience of evil, the basis for excluding the opposing proposition – viz. it is not the case that there is evil in the world – is something that is a result or effect of the truth of the proposition in question—i.e of the fact that makes there is evil in the world true. So one’s ground for knowing the proposition there is evil in the world – e.g. one’s experience – is a consequence of there being a fact which obtains and makes that proposition true – viz. the fact that there is evil in the world.

In the case of cognition or knowledge of antecedently determining grounds, one’s knowledge is not a consequence of the fact which makes the proposition in question true, but rather knowledge of that which accounts for the existence of the fact in question. In this case, the ground for affirming the truth of there is evil in the world is not a consequence of the existence of the proposition’s truth-maker, but rather explains why that truth-maker exists. In the case of evil, this might be knowledge of god’s motives or intentions. To put things another way, if the proposition in question were the consequent of a material conditional, then one’s knowledge of an antecedently determining ground would be knowledge of the antecedent to that conditional. However, this is slightly misleading, since there is more than logical connection at work. Rather, the antecedent must also have explanatory priority.

4 Ontotheism & Ontological Arguments

“Ontotheism” is the position that God exists in virtue of his essence. In this sense, philosophers such as Aquinas, Descartes, Spinoza, Leibniz, Wolff, etc. are all ontotheists. All ontotheists thus endorse one version or other of what Kant terms the “ontological argument”. This is an argument that proceeds from the claim that, necessarily, <existence> is a predicate of the concept <God>, to the conclusion that, necessarily, God exists. Here are a few prominent statements of the position.

Descartes: whenever I do choose to think of the first and supreme being, and bring forth the idea of God from the treasure house of my mind as it were, it is necessary that I attribute all perfections to him, even if I do not at that time enumerate them or attend to them individually. And this necessity plainly guarantees that, when I later realize that existence is a perfection, I am correct in inferring that the first and supreme being exists. (7:67)

Spinoza: A substance cannot be produced by anything else (by P6c); therefore it will be the cause of itself, that is (by d1), its essence necessarily involves existence, or it pertains to its nature to exist (E1p7d)

Leibniz: For if there is reality in essences or possibles, or indeed, in eternal truths, this reality must be grounded in something existent and actual, and consequently, it must be grounded in the existence of the necessary being, in whom essence involves existence, that is, in whom possible being is sufficient for actual being. (Monadology §45)

Wolff: God exists through his essence, or his existence is essential. (TN §27. Cf. Dt.Met. §§929–30 and Ont. §308)

Baumgarten: From God’s possibility, it is valid to draw the conclusion that he exists, i.e. his existence is sufficiently determined by his very essence. (Meta. §820)

Kant introduced the term “ontotheology” to describe the part of theology that attempts to determine the predicates God has merely as a possible being (ens). Ontotheology, he says in his theology lectures, “considers God merely through concepts of possible beings in general [mögliche Dinge überhaupt]” (Volckmann Religion, 28:1142), and discusses the predicates of such a possible being, including substance, simplicity, immateriality, etc.

With this in mind we can say that an “ontotheological argument” for God’s existence is one that purports to prove the existence of God merely using modal concepts—i.e. from concepts concerning the possibility, or necessity of such a being. Kant refers to ontotheological arguments via his term “ontological argument” (MV 28:454 and ML_2 28:599). As Kant himself notes however, though he is the originator of these terms, the (generally agreed upon) originator of ontotheological argument itself (or “the” ontological argument) is Anselm (Volckmann Religion 28:1143).⁷

In lecture transcripts from 1783-4 Kant is reported as describing the ontological argument as follows:

The ontological proof for the existence of God is taken from the concept of an ens realissimum. One infers, namely: An ens realissimum is one which contains all realities in itself. But existence is also a reality; hence, the ens realissimum must necessarily exist. If one therefore asserts that God is not, then one thereby denies something in the predicate which lies already in the subject; consequently there is a contradiction here. (Pölitz Religion 28:1027)

Even though ontotheistic arguments are often lumped together under the definite descriptor of “the ontological argument”, there are in fact a variety of different ways one might pose an ontological argument. One might argue from a notion of greatest conceivable being (Anselm), or clear and distinct perception of essence (Descartes), or from the possibility of God’s existence to the necessity of God’s existence (Leibniz), etc. Again though, all these arguments share the view that <existence> is part of (or is “involved” in) both the concept <God> and the essence of God.

Here’s one way of putting the argument, reflecting Kant’s argument in the Pölitz Religion above:

1. God’s nature is such as to contain all realities
2. Existence is a reality
3. ∴ God exists

As we’ll see, Kant both rejects premise (2) of the argument and rejects the validity of the inference from what he construes as a set of conceptual claims in (1) and (2) to a claim about a extra-conceptual entity in (3).

5 Is <existence> a Determination?

Is <existence>, and its analogues, a determining predicate? Does its predication of a subject concept “exclude” its opposite? In what sense could there be a being to which the property of existence applies or not? Such questions are at the center of the debate as to whether ontotheology can mount a plausible argument for the existence of a being as included in its essence.

Kant’s criticism of ontotheism is rooted in his contention that <existence> is not a determining predicate. According to Kant, <existence> is not something predicated of a subject, but rather predicated of the concept, namely, that it is instantiated. Here are several contemporary statements of this view.⁸

Jonathan Bennett:

According to Kant, every existence-statement says about a concept that it is instantiated, rather than saying about an object that it exists. This is an important precursor of thc view of Frege that any legitimate existential statement must be built out of propositional atoms of the form `There is an F’, where F stands for a determining predicate. According to this Kant-Frege view, the real form of `Tigers exist’ is not like that of `Tigers growl’, but rather like that of `There are tigers’ or `The concept of tigerhood is instantiated’. Granted that Kant’s arguments fall far short of proving this hypothesis, they do at least illustrate and elucidate it; and the hypothesis itself is a philosophical contribution which deserves attention and which may even be true.⁹

James Van Cleve:

To say that God exists is to say something not directly about God, but only about the concept of God, namely, that there is an object corresponding to it. That does sound remarkably like Frege’s view.¹⁰

Nick Stang:

Kant claims that the judgment a narwhal is an existent animal does not assert that some predicate is contained in the concept <narwhal>, and does not attribute further predicates to the objects that fall under <narwhal>; it asserts that there is at least one object that falls under <narwhal>, i.e. that the concept is instantiated. I take this to be clear evidence that the fundamental (though not the only) role of the predicate <exists> is to apply to concepts: it applies to a concept if and only if that concept is instantiated by an object. While this is not yet the complete Fregean theory of the existential quantifier, it does anticipate it.¹¹

The position that existence is not a first-order predicate applied to objects, but rather a second-order predicate applied to concepts of objects is typically attributed to Kant on the basis of the following passage:

If the existence of the subject is not already presupposed, every predicate is always indeterminate in respect of whether it belongs to an existent or merely possible subject. Existence cannot, therefore, itself be a predicate. If I say: “God is an existent thing” it looks as if I am expressing the relation of a predicate to a subject. But there is an impropriety in this expression. Strictly speaking, the matter ought to be formulated like this: “Something existent is God.” In other words, there belongs to an existent thing those predicates which, taken together, we designate by means of the expression “God.” These predicates are posited relative to the subject, whereas the thing itself, together with all its predicates, is posited absolutely. (OPA 2:74; see also A598–99/B626–27)

Kant here distinguishes between what he calls “relative”, as opposed to “absolute”, positing. Relative positing concerns a relation between two concepts. In contrast, absolute positing concerns the attribution of a property to some existing object. As Kant puts it,

The concept of positing or setting [setzen] is perfectly simple: it is identical with the concept of being in general. Now, something can be thought as posited merely relatively, or, to express the matter better, it can be thought merely as the relation (respectus logicus) of something as a characteristic mark of a thing. In this case, being, that is to say, the positing of this relation, is nothing other than the copula in a judgement. If what is considered is not merely this relation but the thing posited in and for itself, then this being is the same as existence. (OPA 2:73)

Hence, one can assert a predicate of a subject term without (absolutely) positing that the subject term is instantiated. This is mere relative positing. Kant’s contention is that the ontotheist invalidly infers from premises that involve mere relative positing, to a conclusion that requires absolute positing. Hence the ontological argument confuses two distinct sorts of cognitive act, relative vs. absolute positing, and therefore invalidly concludes that God exists because the concept of the essence of God includes existence. There is, perhaps, a question here as to whether Kant’s distinction between relative and absolute positing commits him to a kind of “proto-Fregean” second-order view.¹² But even if we were to deny this connection Kant’s distinction between the two forms of positing and their relevance for his criticism of the ontological argument is clear.

However, what is somewhat less clear is the way in which Kant intends to present this view. Is it a premise in his argument against ontotheism? Or is it the conclusion of an argument (or arguments) against ontotheism?

Dai Heide (2021) argues that Kant presents arguments against ontotheism (or specific versions thereof) that do not rely on his positive view concerning the nature of the existence predicate, and moreover, that such arguments aim to show that ontotheism is internally inconsistent. Thus Kant hopes to show that the best way of resolving this inconsistency is via a rejection of the view that <existence> is a real determination or predicate.

References