In the Critique of Pure Reason, Immanuel Kant set out a framework intended to refute the ontological argument. It is said that the critique was directed at Descartes and Leibniz. And oddly, Pierre Gassendi expected such a criticism from Kant, even going so far as to write about it in his Objections to Descartes' Meditations. Kant's framework consisted of a number of interrelated but separate arguments.
Three main arguments are the backbone of Kant's refutation of the ontological argument. Primary is the argument that existence, or being, is not a real predicate. Second, Kant argues that it is possible to deny the subject and the predicate of a judgment in order to avoid contradiction. Kant theorizes that there are no claims about existence that can be termed analytic -- that is to say, tautological, as a result of the predicate expressing something already within a concept, therefore already known. Here, we note the ideas of predicates and concepts, and particularly that a predicate can be within a concept, or can link to something outside the concept.
The relation of concept and predicate. A concept is partially specified by what its properties make it. A concept, then, contains some predicates, but it also contains negations of predicates that do not make it up, and it may be silent about still other predicates[footnoteRef:1]. A concept is said to be determinable and can be extended by adding additional specification to it. Unlike a concept, an object can be said to be fully specified, such that, each object can have every predicate or every negation of that predicate belong to it. For example, the concept of a Persian cat is specified as fluffy, aloof, and quiet on its feet. However, the type of food preferred by the Persian cat or the color of the fancy collar it wears on any given day is unspecified. These unspecified properties will apply at the level of a specific individual Persian cat, each of whom will have a favorite brand of cat food and will wear a certain fancy collar. And all other possible predicates will have a determination, with regard to a particular Persian cat. Kant claims that existence, or "being is not a real predicate."[footnoteRef:2] Yet a claim that something exists is to assert that there is something that is within the concept that is of interest. To claim that a Persian cat exists is to assert that an object is within the concept of Persian cat. [1: Karl DeVries, Kant's Refutation of the Ontological Argument (April 2009). Concept from Zerby Prize Winning Essay 2009, Retrieved http://msuphilosophy.files.wordpress.com/2009/04/zerby_prize_2009_karldevries.pdf .] [2: From the translation, Norman Kemp Smith, The Critique of Pure Reason, (London, 1929): 504-5.]
"Being is obviously not a real predicate.[footnoteRef:3]" As Engel points out, Kant does not say that "Being is not a predicate." Nor does Kant say that "Being is obviously not a predicate." Kant says that "Being is obviously not a real predicate," thereby conveying that there are likely real predicates and predicates that are not real. The word real is problematic to the translation, in that, emphasis can be intended or reference can be made metaphorically to a relative condition. If Kant meant to emphasize the word predicate in his statement, then he might have chosen a different word. [3: Ibid.]
Kant argues that the function of a real predicate is to add to the domain of the subject term. According to Kant, the subject term would already have possessions over which it is dominant, and that the predicate term may serve to enlarge that domain, thereby adding to the possessions over which the subject dominates. All real predicates function in this way, except for "Being" which Kant considered to be a different type of predicate.[footnoteRef:4] [4: Morris Engel, "Kant's Refutation of the Ontological Argument," Philosophy and Phenomenological Research, 24 (1) (September 1963): 32.]
Further, Engel argues that the reader is mistaken to assume that Kant has created a dichotomy that restricts predicates to real or logical types. Further, Engel points out that Kant does imply more than two types of real when he writes "Im logischen[footnoteRef:5]" (Logically), and intends the reader to understand that he means from a logical point-of-view. From Kant's implied taxonomy, it becomes possible to consider three types of predicates. As described by Engel, these three types would be: "Real predicates," "nonreal predicates," and "logical predicates."[footnoteRef:6] [5: From Engel's translation in S. Morris Engel, "Kant's Refutation of the Ontological Argument," Philosophy and Phenomenological Research, 24 (1) (September 1963): 20-35.] [6: Ibid, p. 24.]
Engel proposes that Kant was constructing a "conceptualistic or two-stream theory of language in terms of which alone the doctrine can be defended, if defended at all."[footnoteRef:7] By establishing a linguistic stream and a conceptual stream, Engle argues, Kant could propose to do things at once: Refer to an "unit of meaning" in the conceptual stream that would relate to the expression "God" (referred to as the object of the concept by Kant), and could assert that unit of meaning or "object" is -- in the linguistic stream -- the objective correlative of the concept. In referring to this "God," Kant conceivably could argue that he references "is what I am thinking, is what I am entertaining."[footnoteRef:8] [7: Ibid., 27.] [8: Ibid.]
"For the real is an additional element which is not a product of my thought and therefore its correspondence with my concept can no longer be taken for granted as a priori certain…In positing, however, a concept of an existing thing what we posit is not some conceptual element of which we could say a priori that it is completely congruent with it, but the thing itself whose congruence with the concept is a matter to be proved or established and which cannot therefore be taken for granted.[footnoteRef:9] [9: Engel, Op. Cit., p. 30.]
If in saying that something exists, we set up a framework that the linguistic stream and the conceptual stream are equivalent then we can assert that we add nothing to the objective correlative by declaring that it exists. From this Kant could argue that "is" is not a real predicate. The function of the cupola (is) relates the unit of meaning that the word "omnipotent" conveys to the conceptual entity that the word "God" represents.
Avoiding the contradiction. Kant argued that a subject and its predicate can both be cancelled without committing a contradiction. Kant explained,
If I cancel the predicate in an identical judgment and keep the subject, then a contradiction arises; hence I say that the former necessarily pertains to the latter. But if I cancel the subject together with the predicate, then no contradiction arises; for there is no longer anything that could be contradicted. (A594/B622).
To clarify his point, Kant used a triangle as heuristic. The concept of a triangle invariably requires three sides as a property. One cannot at once claim that a concept is a triangle and that the concept does not have three sides. (A594/B622) There cannot be a two-sided or a one-sided triangle. If such an object existed, it would not fall under the concept of triangle. However, one could say that there are no objects -- no triangles of any configuration -- under the concept of triangle, thus the aspect of the concept of triangle does not hold for any object. Similarly, a contradiction is committed by claiming that God does not exist because existence belongs to the concept God.
An underpinning of Kant's framework was based on his precise and dogged differentiation of analytic and synthetic judgment. Consider that claims can be either synthetic or analytic, and that these categories also apply to existence claims Kant argued against the ontological argument that no existence claims can be analytic, and by the same reasoning, one cannot use a concept to prove the existence of God. Taking the position that an existence claim is analytic, you add nothing to your thought of the thing; but then either the thought that is in you must be the thing itself, or else you have presupposed the existence as belonging to the possibility, and then inferred that existence on this pretext from its inner possibility, which is nothing but a miserable tautology. (A597/B625)
An analytic existence claim generates a problem that pivots at the construction of the concept. Consider a situation in which a man in therapy recalls an incidence in which he was abused and, further imagines the person who abused him. The therapist brought the assumption of an abuser into the clinical dynamics, by asking the man if he had ever been abused. In essence, the therapist "presupposed the existence as belonging to the possibility."[footnoteRef:10] As a result, both the therapist and the patient have the conception of an existing abuser. Two possible scenarios exist with regard to this concept. The abuser exists only in the man's imagination, but still causes him to experience authentic distress. A non-existent abuser cannot…