Thursday, September 30, 2010

Ontological Arguments for God's Existence

St. Anselm, the 11th Century theologian and philosopher (who also served as archbishop of Canterbury), is best known for two things. Among theologians, he is principally associated with his deeply influential understanding of the Atonement--that is, his account of how Christ's crucifixion is to be understood as securing salvation for sinners. But among philosophers Anselm is better known as the author of the so-called "ontological argument" for God's existence (this name for the argument originated with Kant, who is also credited historically with formulating the most telling objection to the argument).

Until fairly recently it was generally assumed that, in his Proslogion,  Anselm offered a single "ontological argument" for the existence of God based on his distinctive formal definition of God as "that than which a greater cannot be thought." Certainly, there is only one "ontological argument" that comes under Kant's scrutiny and becomes the target of his famed (well, famous among philosophers) objection.

That argument runs roughly as follows: Taking "God" to mean "that than which a greater cannot be conceived," let us assume that God in this sense is only an idea in our heads and doesn't actually exist. On this assumption, we can conceived of something that is greater than God--because we can conceive of this God as actually existing, rather than as being nothing more than an idea in our heads (and a God that actually existed would be greater than one that did not). So, on this assumption, it follows that we can conceive of a being greater than that than which a greater cannot be conceived. But, of course, that is impossible. Hence, our assumption has led to a logical impossibility and so has to be rejected. It is not the case that God is only an idea in our head and doesn't exist in reality. Rather, God really exists.

This is the argument that Richard Dawkins, in The God Delusion, calls "infantile" and frames in terms of a school-yard conversation in which one smarmy kid "proves" God's existence by playing with words in a way that makes the argument sound truly silly. Now it is true that most philosophers look at this argument and think there is something fishy going on. But I think it's also true that most philosophers are grateful that the history of reflection on the argument was shaped by serious engagement. In wrestling seriously with this argument (which is formally valid), critics and defenders alike were forced to refine their understanding of what it means to say that something exists, as well as how this kind of statement differs from saying that something is blue or round. One might even wonder about whether the emergence of formal predicate logic as we know it owes something to the lessons of engaging with Anselm (in the symbolic language of predicate logic, an entity's existence is expressed through the use of what is called the existential quantifier, as opposed to being attributed to the entity as a predicate).

In any event, this ontological argument was criticized by Kant on the grounds that, in his terms, "existence is not a real predicate." Put another way, when we say of something that it exists, we don't add to our idea of a thing. Rather, we say that this idea (described in terms of the "real" predicates), has an instance in the world. Thus, to say of God that He exists is not to add to our concept of God. Since positing existence adds nothing to our concept, an existing God is not conceptually different from a nonexistent one. And so there can be no conceptual incoherence with respect to the latter that doesn't also attach to the former.

(Interestingly, I suspect that Anselm would have some sympathy with Kant's objection, because Kant is describing what existence means in its ordinary usage as applied to ordinary things--and Anselm insisted that "God exists" has to mean something very different from "Eric Reitan exists," insofar as God is taken to be the source of existence--or, in Anselm's language, "that through which" things exist, or from which things derive their being. Very roughly, to say that I exist is to say that I participate in existence itself, which is that through which I derive my being; whereas to say that God exists is to say that God is existence itself, that through which other things derive their being. Anselm furthermore attempted to show that existence itself must also be goodness itself, and by implication must be Godlike in the ways that theists have traditionally thought. Whether this element of Anselm's theology can be developed into an answer to Kant, however, is something I won't explore here.)

In any event, many if not most philosophers after Kant found his objection compelling, and the objection has since been expressed in numerous ways--including by Richard Dawkins in The God Delusion. While Dawkins makes no serious effort to carefully articulate this line of objection, he does at least quote part of one philosopher's articulation of it. Unfortunately for Dawkins (at least if he cares about his credentials as a scholar), the philosopher he appeals to for this quote is Norman Malcolm. Even more unfortunate for Dawkins, the quote comes from an article, "Anselm's Ontological Arguments," that is justly famous in the history of discussion about Anselm and his case for God. Since (according to his endnotes) Dawkins extracted the quote from an online encyclopedia article, Dawkins apparently doesn't know that this quote comes from an article in defense of Anselm. As such, Dawkins doesn't know that his dismissal of the ontological argument is, we might say, premature. To put the point more bluntly, Dawkins succeeds in completely ignoring the version of the ontological argument that is still alive today (in the sense of having active defenders) largely by virtue of the efforts of the philosopher Dawkins invokes to justify his dismissal of the argument.

It is true enough that Malcolm accepts the Kantian objection to Anselm. But what Malcolm then does (which is what makes his essay historically important) is to point something out about Anselm's original text that philosophers had historically overlooked. If one looks at that text, one sees Anselm wording his argument in a couple of ways, and it reads as if his second version is intended to be just a different way of saying the same thing. In fact, on a cursory reading it looks like the same argument restated slightly differently.

What Malcolm does in "Ontological Arguments" is show that this cursory reading is a mistake. The two arguments aren't different ways of saying the same thing, but different arguments. While there are important similarities, there is a crucial difference as well. And the difference makes a huge difference: By virtue of that difference, Malcolm argues, Kant's famous objection to the ontological argument doesn't apply. Whereas philosophers had widely assumed that Kant had dealt the death blow to the ontological argument in the 18th century, Malcolm not only showed that there was a version of the argument that avoided Kant's challenge, but he then proceeded to develop the argument in an effort to show that it was, in fact, a powerful and compelling argument for the existence of God--far more compelling, in fact, than the version for which Anselm became famous. Others have followed Malcolm's lead (including Alvin Plantinga) in developing so-called "modal ontological arguments" that trace their lineage to Anselm's second argument.

In addition to these arguments that trace back to Anselm (by way of Malcolm's discovery of the second argument), there are various ontological arguments sketched out by Gödel in his notebooks, especially in terms of the notion of "positive properties," that have come under discussion. Interestingly, however, Malcolm's development of Anselm's second argument develops themes that are also found in Gödel's argument. For these reasons I think Malcolm's version of the argument is especially deserving of closer attention. In my next post, then, I will outline Malcolm's argument.


  1. Dear Eric-

    One would think that real philosophers would be positively mortified that such rubbish has tainted their field for nearly a millenium. That they seem to have failed to put a firewall between themselves and incatatory wish-ism. And calling it modal S5 logic just indicts the so-called philosophers further for sophistry and obfuscation.

    Yet again, we come up against hoary old magical thinking.. whatever I can think of, that must be so. And if it supports the reigning myth ... that is not just real, but terribly important. And they say positivism is passe! Good grief.

  2. I guess that one thing that philosophers like far less than arguments that ultimately don't work is the practice of dismissing arguments on the basis of ridicule and name-calling--especially when the name-calling is delivered in a tone of self-righteous indignation intended to imply that anyone who fails to join in the ridicule is a fool.

    We don't like this because it leads people to jump on a band-wagon of dismissiveness without ever investigating whether the dismissal is justified (but simply out of fear of being the object of ridicule themselves). Such a technique can succeed as well with sound arguments as it can with unsound arguments--and philosophers in general have internalized Plato's warnings (in his dialogues) against such techniques.

  3. At the same time, the practice of extending such philosophical forebearance to selected arguments constitutes attention and legitimation that is appalling if they are not worthy, and not promptly disposed of as historical curiosities. Do you also extend such care to Douglas Adam's Hitchhiker philosophy, equally worthy of deep and respectful analysis? Or to the philosophy of Vedanta, Gnosticism, etc? It looks very much like an exercise of carrying water for the traditions of (your) religion, whether philosophically sound or not.

    Sorry to be a little irritated here .. will cease immediately.

  4. I haven't been reading the blog for a few weeks. Glad to see Burk still thinks calling his opponents stupid and brutely asserting that they're wrong without further argument is going to... do... whatever Burk thinks it's going to do...

    Honestly, dude, if you don't think this stuff is worth engaging with, why the *hell* do you think it's worth spending months insulting people without offering substantive criticisms? Are you just hoping that Eric will get tired and quit or something?

  5. Philosophical jokes--such as those offered by Douglas Adams, should be appreciated as what they are. Were someone to say, "You know, I think this argument, intended as a joke, actually has merit," I'd want to know why the person thought so. If I disagreed, I' say why.

    As for arguments advanced within the Vedanta or Gnostic tradition, my basic thought is this: the tradition out of which an argument emerges is not a basis for dismissing it--but understanding that tradition may be necessary for properly interpreting and thus fairly assessing it. Furthermore, in some cases arguments rely on premises that are core assumptions of the tradition and thus show what else you should believe if you are committed to this tradition--and that can be very useful information for understanding what it means to live out the worldview embodied by the tradition, and hence for refining it.

    And yes, I believe that the path to truth lies in progressively refining holistic worldviews in the testing ground of lived experience--which requires that Vedantists live out their worldview critically, that Gnostics do the same with theirs, that Christians do the same with theirs, and that naturalists do the same with theirs.

  6. Eric and Burk

    For me, one of the reasons for seriously considering something like the ontological argument (apart from the sheer fun of it) is that it often helps me better understand the thing that is being discussed, which is to say what people might have in mind when they use a word like God. Not clearing this up can, it seems to me, lead to a great deal of confusion.

    For example, the first time I saw the ontological argument my reaction as 'well, wouldn't a God that can be great without actually existing be far greater than one that has to stoop to the messy business of being?'

    Once we're dealing with something as mind stretching as God, I'm not sure how our normal rules of logic can be expected to either constrain or uncover it. If the concept of God is to human what calculus is to sea squirt then how deeply strange and unknowable might it be? I think for instance of the logic bending implications of Relativity or the Heisenberg uncertainty Principle and think, well how very much weirder is God likely to be if it exists? Way too weird to be captured in an argument about whether or not existence is a predicate, I'd guess.

    Seeking to establish proofs for either the existence or non-existence of God seems to imply that we have in advance some way of defining this thing being argued about. To a theist that makes sense, as part of the end package is a God that wants us to know it. But if you take a non-theistic starting point, how do you know enough about God to say it can't happily encompass the dual state of both being great and not existing?

    I don't think Dawkins' real objection is to the notion of God per se, I think it's an objection to a God we can know anything about, and our habit of imagining qualities into that gap. So, the big question I think is not, does God exist, but rather, if God did exist, how on earth would we find anything out about it?


  7. Bernard,

    You raise several important ideas here. First off, one reason I engage with some arguments is the same as the reason I occasionally engage with logic puzzles. It's just fun.

    But in the case of the ontological argument, that's not the only reason. Another has to do with precisely the point you gesture towards in your own way: wrestling with arguments for God's existence helps to sharpen our understanding of WHAT people (ourselves included) might have in mind (both historically and today) when they speak of "God." That Aquinas devotes time to trying to prove the existence of an unchanging source of change, an uncaused cause, a necessary being that has its necessity from itself, etc.--along with the fact that theists generally have no trouble treating arguments for such conclusions AS arguments for God--tells us something about the conceptual space gestured towards by the term "God."

    But in terms of the ontological argument(s), which wrestles with the idea of an infinite or unlimited being, the persistence of our struggles with arguments that pertain to such a being may help to drive home the fact that if "God" occupies this conceptual space, then God is going to defy our understanding. If there is something in this conceptual space, then our limited concepts and categories won't be able to characterize it in anything like an adequate way--and the pretentions of various kinds of fundamentalists, who make confident pronouncements about God in terms of very terrestrial categories, are exposed (which isn't quite the same as saying that NOTHING can be said).

    For example, in our ordinary understanding of things, abstract thought can identify impossible states of affairs and possible ones, but it cannot tell us what states of affairs are actual. That is, what actually exists requires empirical investigation. Furthermore, in our ordinary understanding of things, existential claims are contingent, never necessary (it doesn't seem that denying the existence of something is incoherent). Even if the ontological argument doesn't succeed in REFUTING these ordinary assumptions, I think it raises questions about whether these ways of thinking can properly be extended to an infinite being. These points may become clearer in connection with my next post.

  8. Oh, and your thought that a nonexistent God would be even greater than an existent one has actually been developed by a couple of philosophers (Douglas Gasking and William Grey). They argue it would be an infinitely greater achievement for a nonexistent being to create the universe than for an existent one to do so. As you can imagine, there are objections to this line of thought.

  9. Thanks Eric

    I look forward to seeing how this argument can be reformulated.

    I have been tying to view the ontological argument through a novelist's lens. Often writers play a sort of a confidence trick on our audience. We use language deliberately designed to stop the reader asking themselves the obvious question, so that we can then spring a surprise on them later that, although they didn't see it coming, seems sort of obvious in hindsight. The sort of trick the film The Sixth Sense is famous for, for instance. Is it possible the ontological argument contains just such a ruse?

    Starting with the definition, 'that than which a greater thing can not be conceived', in my case at least I realise upon examination I have no idea what this means. I have an intuitive sense that I do, which encourages me to skip lightly over it in the way a confidence trickster would like, but by what criteria is greatness to be measured? It collapses into either subjective measurements like goodness, or love or power, or into terms I don't ultimately understand, like time, cause or infinity. (My greatest thing would certainly taste of chocolate, and probably last a long time but not forever. And it would enjoy arguments).

    The other trick is when we jump to: now let us assume this God is only an idea in our heads and doesn't actually exist. Compare this to a God that does actually exist.

    Here, I think language is playing its games with me again. The calculation seems to be that I'm not going to object with 'no, hold on, I definitely imagined my God existing' which would make the two Gods exactly equal and leave no contradiction. So a sleight of language is used with the words 'rather than as being nothing more than an idea' which masks the just referenced 'we can conceive of' quality of this actually existing God. It too is explicitly just an idea.

    I don't know how the above holds up to philosophical scrutiny, but my background gives me a prejudice towards believing the problem might be with the words themselves.


  10. I appreciate Anselm’s argument, and agree that God does indeed exist as defined. By this I mean that there truly is that which nothing greater can be thought. However, I do not believe that Anselm’s logic has any merit, at least not how it is outlined. All Anselm proves is that the concept of God, as defined, to exist in reality is a greater concept than the concept God exists only in the mind. This does not, however, imply that God actually exist, but only that the concept of (God=that than which nothing greater can be thought) must refer to something real in order to be a valid concept.

    So the concept {(God=that than which nothing greater can be thought) = God must exist} is a valid concept to have for God, as defined, but does not imply that God must exist 'actually apart from concept' only 'actually as a valid concept.'

    One could run this concept {(God=that than which nothing greater can be thought) = God must exist} thru the same argument and end up with a new concept, namely this: ({(God=that than which nothing greater can be thought) = God must exist} = actually exists) and so on ad infinitum. However, all one has done is proved one concept greater than another concept. It does not logically follow that because a concept is conceptually valid that it is actually valid. Therefore, one cannot ‘think’ God into existence as Anselm’s argument seems to do.