Supplement to Hauptli’s Lecture on Critiques of Anselm’s Argument


    Copyright © 2015 Bruce W. Hauptli


There are so many critiques of Anselm’s argument that we could not begin to discuss them all without spending the rest of the semester on this topic.  In what follows I highlight some criticisms which point to central difficulties with the argument.  I will orient these critiques around my “eight step” analysis of the argument (though this does not necessarily lead to a list of the critiques in terms of their degree of importance).  For a more in-depth analysis of the argument I recommend the article on the argument in the Stanford Encyclopedia of Philosophy.  Of particular relevance here is the fourth section on objections to the argument. 


In what follows I will move through the ontological argument as I have outlined it in the lecture supplement, offering criticisms keyed to several of the major "steps" of the argument.  These criticisms move from a criticism of his move from the "language of praise" to the "language of definition," to a criticism of the use of definitions to prove existences; and then on to criticisms of his notions of "greater" or "degrees of reality."  Finally, I note some criticisms of his discussion of "exists in the understanding" vs. "exists in reality;" and whether or not his conception of an "absolutely greatest thing" needs to be established as, at least, a "possible thing." 


(A) In regard to the First Step:


If one doesn’t accept Anselm’s “definition,” the proof of course falls apart.  O.K. Bouwsma maintains that Anselm misinterprets the passage in the Psalms from which he derives the definition:[1] “the notion of “greatest” as used [in that context] as praise rather than as a description.  Thus, he claims, there is a linguistic error in confusing language of praise with language of description.”[2]  That is, Bouwsma contends that the Psalms Anselm refers to praise the transcendent being rather than describe it, and, he contends, it is a linguistic error to interpret praise as a definition. 


     In his “The Search for Certainty,” Hans Reichenbach offers a different sort of critique of the first step when he maintains that:


the premise, in fact, is analytic, because every definition is.[3]  Since the statement of God’s existence is synthetic,[4] the inference represents a trick by which a synthetic conclusion is derived from an analytic premise. 

  The fallacious nature of this inference is easily seen from its absurd consequences.  If it is permissible to derive existence from a definition, we could demonstrate the existence of a cat with three tails by defining such an animal as a cat which has three tails and which exists.  Logically speaking, the fallacy consists in a confusion of universals with particulars.  From the definition we can only infer that universal statement that if something is a cat with three tails it exists, which is a true statement.  But the particular statement that there is a cat with three tails cannot be derived.  Similarly, we can infer from Anselm’s definition only the statement that if something is an infinitely perfect being it exists, but not that there is such a being.[5] 


     Of course we need to recognize that Anselm himself would draw our attention to the fact that he begins “in faith,” and thus he has not doubt that his “definition” is solidly and objectively based, it is nonetheless true that a “necessary truth” may merely be a truth about definitions, concepts, essences, etc.; and thus it tells us nothing about existence (let alone telling us anything about “necessary existence”). 


(B) In regard to the Fifth and Sixth Steps:


Remember the chart I have offered which distinguishes “things” into four categories: “necessary things,” “contingent and actual things,” “contingent and possible things,” and “necessarily nonexistent things.”  When I introduced the argument and gave it my “eight step analysis,” I indicated that when “the fool” maintains that the deity exists only in the understanding, he or she is saying that it does not reside in either of the first two categories (which both are “in reality” in Anselm’s categorization of these things).  This leaves two possibilities: the fool could be saying that this deity is a “possible but not actual” thing, or that this deity is an “impossible” thing.  Now clearly Anselm’s argument presumes the former—if his concept of a deity belongs in the fourth category (that is, if it is an “impossible thing,” like round squares, or married bachelors), then it can not exist, and the proof is clearly flawed. 


     If you are still with me on this point, we can say that Anselm’s argument presumes (in my eight step version, the presumption is in the fifth and sixth steps) that his deity is a “possible thing.”  He doesn’t “prove” that this is the case however.  William Rowe questions whether this deity is, indeed, “a possible thing:” 


...the positive integer than which none larger is possible is an impossible object.  Perhaps this is also true of the being than which none greater is possible.  That is, perhaps no matter how great a being may be, it is possible for there to be a being greater than it.  If this were so, then, like the integer than which none larger is possible, Anselm’s God would not be a possible object.  The mere fact that there are degrees of greatness, however, does not entitle us to conclude that Anselm’s God is like the integer than which none larger is possible.  There are, for example, degrees in the size of angles—one angle is larger than another—but it is not true that no matter how large an angle is it is possible for there to be an angle greater than it.”  The question is, is God like the integer or like the angle.[6] 


     In his Think: A Compelling Introduction to Philosophy, Simon Blackburn maintains that:


in my own view, the crucial problem lies in an ambiguity lurking in the comparison of ‘reality’ and ‘conception’.  In the argument, things ‘in reality’ are compared with things ‘in conception’ (i.e. according to a definition, or in imagination or dreams), for such properties as greatness, or perfection.  This sounds simple, as if we are comparing things in two difference geographical regions, and we know that those in one region are greater or lesser than those in the other.  It would be like asking whether chickens in Germany are heavier than chickens in France.  But in fact it is not at all like that.  Consider this sentence:

Real turkeys are heavier than imagined turkeys. 

There seems to be a sense in which it is true.  In that sense, imagined turkeys weigh nothing (after all, you cannot make even a small meal from one).  But there is also a sense in which it is false, because you can imagine a turkey heavier than any real one—a five-hundred-pound turkey the size of a small barn, for example.  In the ontological argument, ‘God’ in imagination is compared with God in reality, like the imagined turkey compared to the real turkey, and found to weigh less.[7] 


This suggests that we must not think of ‘imagined turkeys’ or ‘turkeys in the understanding’ as kinds of turkey that can, in principle, be weighed against real ones but are always found to weigh less.  Yet the ontological argument requires just this kind of comparison.  It is here that it fails.  For even if God only exists in imagination…it does not follow that a greater being can be described or imagined.  After all, the description had the superlatives put into it.  But unhappily for Anselm’s proof, that does not settle the question whether anything answer s to it.[8] 


     William Alston offers an excellent treatment of the argument in his “The Ontological Argument Revisited.”  One of his points bears well on this step in the argument:


earlier we saw that an existential statement has the function of setting up a subject for predication.  Now that we have recognized different modes of existence we can add a further stipulation: the kind of existence which is being stated will place limits on the sort of predication that can be made with respect to the subject, that is, on the logical status of statements which can be made about it. [9] 


This principle might be defended by saying that a licensing bureau cannot authorize anyone to do anything it does not have the authority to do....[10] 


Existence in the understanding shares with other nonreal modes of existence the following features.  For each existent in some nonreal mode, we can specify two sorts of real existents.  First, there is some real existent of a given sort, which is always of the same sort for a given nonreal mode, the existence of which is entailed by the nonreal existence of the thing in question.  Whenever something exists in my dreams, there must be a real conscious dream state; whenever something exists in legend or myth, there are real activities of repeating, hearing, thinking about the legends and myths in question; whenever something exists in my understanding, there are real thoughts, ideas, images, and so forth, in my such a real existent the real correlate of a nonreal existent.  Secondly, we can specify something which really exists and has all the characteristics...of the nonreal existent.  Let us call this the real archetype....Now it seems to be a defining feature of all nonreal modes of existence that any statement about something which exists in such a mode will have no implications with respect to real things, except for its real correlate and any implications that might have.  In particular it has no implications concerning the real archetype.  This latter feature is an essential feature of the concept of different modes of existence.  If the existence of something in one mode should imply its existence in another mode, the distinction between these two modes would crumble.[11] 


(c) Regarding the Sixth Step:


In this step Anselm assumes that “existence in reality” is greater than “existence in the understanding.”  First, of course, he doesn’t prove that this is the case.  As I noted, Anselm needs the notion of “absolutely great-making” characteristics or properties—being golden, as we have seen, is only relatively better according to him (it depends upon the thing being discussed whether or not it would be “better” for it to be golden).  The question is whether “exists in reality” is an absolutely great-making characteristic or property of things. 


First, of course, one might maintain that it is not, that it is only a relatively great-making characteristic or property.  It is not better, one might argue that a debt exist in reality than that it exist in the understanding only (and similarly for any number of tragedies and hardships). 


Second, one might question whether ‘exists’[12] (as it is used in the contrast here) is a characteristic or property at all.  This takes us to the thorniest critique of Anselm’s argument.  Immanuel Kant, who named this argument “The Ontological Argument,” thinks there is a core difficulty because the argument wrongly assumes that ‘existence’ is a predicate.  This critique is very complex, but Ernest Nagel makes Kant’s point as follows: use Kant’s example, when we think of $100 we are thinking of the nature of this sum of money; but the nature of $100 remains the same whether we have $100 in our pockets or not.  Accordingly, we are confounding grammar with logic if we suppose that some characteristic is being attributed to the nature of $100 when we say that a hundred dollar bill exists in someone’s pocket.


....When we say that a lion has a tawny color, we are predicating a certain attribute of the animal, and similarly when we say that the lion is fierce or is hungry.  But when we say that the lion exists, all that we are saying is that something is (or has the nature of) a lion; we are not specifying an attribute which belongs to the nature of anything that is a lion.[13] 


     In his Contemporary Epistemology, Ralph Baergen clarifies this maintaining that:


it was argued (Anselm’s Ontological Argument) that God’s existence could be shown a priori by examining the concept of God; the idea is that God is a being with all perfect-making properties, existence is a perfect-making property, so God has the property of existence (that is, God exists).  The central problem with this is that existence doesn’t seem to be a property at all, let alone a perfect-making one.  (It seems, rather, to be a precondition for something’s having any properties.)[14] 


     In his “Nothing Ventured: A Bold Leap Into the Ontological Void,” Jim Holt maintains that:


theologians were chary of Anselm’s reasoning from the moment it was articulated.  Could a being whose existence is grounded in pure logic really be the God of faith, the God of Abraham, Isaac, and Jacob?  The argument fared better with philosophers.  Leibniz plumped for it; so did Descartes; so did Spinoza.  It was not until the eighteenth century, after hundreds of years of muddled controversy, that Immanuel Kant nosed out the fallacy.  Simply put, it is this: Existence is not a property of things, like size or color.  It adds nothing to a concept.  If it did, all kinds of new entities could be defined into existence.  Suppose, for instance, a unicorn were defined as the most perfect horse there could be; would it not follow, by the very reasoning Anselm employed, that unicorns exist?  No logical bridge can be built between a mere abstraction and concrete existence.[15] 


(D) Finally, in regard to the conclusion:


We must be certain to distinguish logical and ontological necessity.  The former applies only to propositions or statements (a proposition is logically necessary if it is true in virtue of the meanings of the terms composing it).  The latter notion is a uniquely philosophical one, and it is not immediately obvious that there must be any such “things.”  Indeed, one might contend, Anselm’s whole proof tries to move from thought to reality in a way which confuses these “categories.” 


Notes: [click on note number to return to text for that note]

[1] Anselm begins his proof in “Chapter 2” by referring to Psalm 14:1 which says “The fool says in his heart, ‘There is no God.’  They are corrupt, they do abominable deeds, there is none that does good”; and Psalm 53:1 which says the same thing.  The translation here is that of the Revised Standard Version  (N.Y.: Thomas Nelson and Sons, 1952). 

[2] Cf., O.K. Bouwsma, “Anselm’s Argument,” in his Without Proof or Evidence, eds. J.L. Craft and R.E. Hustuit (Lincoln: Univ. of Nebraska, 1984), pp. 48-49. 

[3] An analytic statement is one that is supposed to be true in terms of the mere meaning of the terms involved.  For example “squares are four-sided figures” is true because (a) ‘square’ means ‘four [equal]-sided figure’, and, thus, it reduces to an “identity statement.”  Such truths are conceptual truths. 

[4] A synthetic statement is one that is not true simply in virtue of the meanings of the terms involved.  Here the truth of the statement depends on the way the world is, rather than, simply, on the way words are used. 

[5] Hans Reichenbach, “The Search for Certainty,” in Certainty, ed. Jonathan Westphal (Indianapolis: Hackett, 1995),  pp. 104-120, pp. 112-113.  The selection originally appeared in Reichenbach’s The Rise of Scientific Philosophy (Berkeley: Univ. of California, 1951). 

[6] William Rowe, “The Ontological Argument,” in Reason and Responsibility (seventh edition), ed. Joel Feinberg (Belmont: Wadsworth, 1989), pp. 8-17, p. 13.  The essay originally appeared in the Third Edition of this work (1971).  Clearly Rowe is relying upon the idea that angles fall between 0° and 360°.  Of course many individuals talk about rotations larger than 360°, and if one allows this then his second example doesn’t do the work he wants it to do. 

[7] Simon Blackburn, Think: A Compelling Introduction to Philosophy (Oxford: Oxford U.P., 1999), pp. 157-158. 

[8] Ibid., p. 158. 

[9] William Alston, "The Ontological Argument Revisited," in Descartes: A Collection of Critical Essays, ed. Willis Doney (Garden City: Anchor, 1967), pp. 278-302, p. 290. 

[10] Ibid., p. 293. 

[11] Ibid., p. 295. 

[12] Philosophers use single quotes to surround a word when they are mentioning it rather than using it.  For example, in the sentence “‘Long’ is a short word,” the word ‘long’ is mentioned (discussed) while the word ‘short’ is used! 

[13] Ernest Nagel, “The Case for Atheism,” in Philosophy: An Introduction to the Labor of Reason, ed. Gary Percesepe (N.Y.: MacMillan, 1991), pp. 508-518, pp. 511-512.  The selection originally appeared in Nagel’s “Philosophical Concepts of Atheism,” in Basic Beliefs, ed. J.E. Farichild (N.Y.: Sheridan House, 1959). 

[14] Ralph Baergen, Contemporary Epistemology (Fort Worth: Harcourt Brace, 1995), pp. 211-212. 

[15] Jim Holt, "Nothing Ventured: A Bold Leap Into the Ontological Void," in Harper’s Magazine November, 1994, pp. 74 ff. 

Return to PHH 2063 Home page

File revised on 03/20/2015