The Ontological Argument


1.         Introduction

One of the distinguishing features of Ontological Arguments is that they are deductive and a priori. They attempt to prove God's existence by reason alone, independent of experience. There are many different versions of the ontological argument, but the general spirit behind all of them is this: just by contemplating our idea of God—without having to look out in the world or find any sort of empirical proof—we can come to conclude that God must exist
 
We will be looking at Anselm’s ontological argument(s) in his Proslogium. I argued in class that in Chapter 2 of the Proslogium, Anselm presents a reductio ad absurdum. He assumes the opposite of what he aims to prove—i.e., he begins by assuming that God (the being than which nothing greater can be conceived) does not exist (in reality). If this assumption leads to a contradiction or absurdity, then we are justified in rejecting the original supposition and accepting its contrary—namely, that God does exist (in reality).

We will also be looking at Chapter 3 of the Proslogium. In this chapter, Anselm (arguably) suggests a modal version of the ontological argument, which we will discuss in tandem with David Lewis’ article “Anselm and Actuality.”   

In the pages that follow, however, I only concentrate on what I take to be Anselm’s argument in Chapter 2.


2.         Summary of Argument
    
    Defining God

Anselm defines God as the greatest conceivable being, or as something than which nothing greater can be conceived. To get a grasp on this, take every great-making attribute you like—wisdom, power, knowledge, moral goodness, etc.—and amplify these great-making features to their limit. There. Now you've got God. If you can think of anything greater than what you've just thought of, then you haven't succeeded in thinking about the thing than which nothing greater can be conceived.

Of course, just thinking of such a being does not (yet!) commit you to thinking that such a being exists (in reality). For consider: even if you think that the greatest conceivable being (GCB) does not exist, you at least have the ability to understand or have an idea of such a being. You can, at the very least, understand what such a being would be like. And this is all the argument needs to get it going. So long as you can at least understand the idea of a GCB, the argument aims to show that, on pain of contradiction, such a being must exist.

If you disagree with Anselm’s definition here, it won’t matter. For if you understand what Anselm (or any other theist) means when he talks about the GCB, and you think that that thing—God defined as such—simply does not exist, then Anselm’s argument is aimed at you. Anselm’s point (eventually, in order to get the argument going) is that if you disagree with Anselm, if you claim “God does not exist”, you must understand what the word ‘God’ means, or else your claim is meaningless.

In class, I asked whether the statement “Flugblubs do not exist” is true or false. Since you do not know what ‘flugblubs’ means—because I made it up specifically to be a nonsense word—you cannot say one way or the other whether you think such things exist. As another example, ask yourself whether you agree with any of the claims made in Lewis Carroll’s Jabberwocky (attached on the back). If you do not (even minimally!) understand the terms being used in a sentence, you cannot coherently claim whether such things exist or not. This is a crucial point, and one that Anselm capitalizes on in order to generate his reductio.   
    
    Existing in the Mind vs. Existing in Reality

Next, Anselm invokes the distinction between existing in the mind and existing in reality. Intuitively, many things can exist in the mind alone without existing in reality. Unicorns, wizards, the fountain of youth, and Harry Potter are all things that exist in the mind alone, but do not exist in reality. In contrast, there might be some things that exist in reality, but do not exist in the mind. There could be some totally crazy looking beetle that has not yet been thought of in the minds of men, but nonetheless exists in the world. So there are some things that exist in the mind (alone), which need not exist in reality; there are some things that exist in reality, which need not exist in the mind; there are some things that exist in the mind and reality (you and me, for example), and there are some things that do not exist either in the mind or in reality (but of course we cannot say what they are for as soon as we do *poof!* they exist in the mind).  

    Setting Up the Reductio

Let's suppose that the atheist is right. Let’s assume that the Greatest Conceivable Being (GCB) exists in the mind alone, but not in reality. This might seem a really backwards way to the conclusion—after all, this is an argument for the existence of God, to the conclusion that God exists in reality, as well as in the mind. But this is the nature of a reductio ad absurdum. One assumes the contrary of what one is trying to prove in order to show that a contradiction or absurdity follows. So this argument will start out assuming that the Greatest Conceivable Being exists in the mind alone, but not in reality.

    Existing in Reality Makes Things Greater!

A crucial premise in Anselm’s (first) ontological argument is the claim that existing in reality is a great-making property. Why should we think so? Suppose we have an imagined beer that exists in the mind alone. Make this pint of beer anything you like—make it as cold as you like, as bitter or hoppy as you like, as big as you like, etc.—but make sure that it exists in the mind alone, and not in reality. Nice, right? But not much good to us, if we are thirsty for a cold one. So now compare this merely imagined beer to one that is not only imagined, but exists in reality as well. Presumably, it is better that the pint of beer exists in reality, as well as in the mind, rather than the beer existing in the mind alone. At least, it's pretty clear which one you'd prefer given that you've got a considerable thirst.
It is important to realize that the claim is not: anything that exists in reality and in the mind is greater than anything else that exists only in the mind. Rather it’s: Take something that exists in the mind. That same thing would be better if it exists in reality.  

(If this seems too quick, don’t worry. We’ll get into this in more detail after we have formally laid out the argument.)

    Showing an Absurdity Follows

To recapitulate, we have assumed that the GCB exists in the mind alone, and not in reality. But we have also agreed that whatever exists in reality, as well as in the mind, is greater than that same thing that exists in the mind alone. But now we're in trouble! Because now we can think of a being greater than the GCB—namely, we can think of a GCB that exists in reality as well as in the mind. Thus, the GCB—the being than which nothing greater can be conceived—is a being than which something greater can be conceived. But this is a contradiction. Therefore, the original assumption—that the GCB exists in the mind alone—must be false, and it's contrary must be true. So, the GCB must exist in reality as well as in the mind. So, God exists (in reality)!
 

3.         Formalizing the Argument

1.    Suppose that the Greatest Conceivable Being (GCB) exists in the mind, but not in reality. [Let's call him Rod.]
2.    Existence in reality is greater than existence in the mind alone.
3.    We can conceive a GCB that exists in reality as well as in the mind. [Let's call him Todd]
4.    Therefore, there is a being [Todd] that is greater than the GCB [Rod].
5.    But this is a contradiction: there cannot be a being greater than the Greatest Conceivable Being.
6.    Therefore, it is false that the GCB exists in the mind alone and not in reality.
7.    The GCB at least exists in the mind.
8.    So, the GCB must exist in reality as well. So: GOD EXISTS (in reality)!

 
4.          Objections
    
    Guanilo's Perfect Island Objection

Guanilo suggests that something has seriously gone awry in the Ontological Argument. To show this, he devises a similar argument, resulting in a conclusion that we know to be false. He suggests that we run a parallel argument with one minor alteration--replace the Greatest Conceivable Being with the Greatest Conceivable Island. The replacement argument can be formalized in the following way:

1.    Suppose that the Greatest Conceivable Island (GCI) exists in the mind, but not in reality. [Let's call it Ned.]
2.    Existence in reality is greater than existence in the mind alone.
3.    We can conceive a GCI that exists in reality as well as in the mind. [Let's call it Maud.]
4.    Therefore, there is a island [Maud] that is greater than the GCI [Ned].
5.    But this is a contradiction: there cannot be an island greater than the Greatest Conceivable Island.
6.    Therefore, it is false that the GCI exists in the mind alone and not in reality.
7.    The GCI at least exists in the mind.
8.    So, the GCI must exist in reality as well. So: the GCI EXISTS (in reality)!

Guanilo's point is that if the above version of the ontological argument goes through, then we can reason similarly for the existence of the greatest conceivable island. What’s more, we can run this kind of argument for the existence of the greatest conceivable anything-you-like: the greatest conceivable lover, the greatest conceivable beer, the greatest conceivable evil being, etc. Yet since it is clear that there is no greatest conceivable island (lover, beer, evil being, etc.), something is amiss with the argument.

Note: there are two main ways to refute an argument: (i) attack the form of the argument or (ii) claim that (at least) one of the premises is false. Guanilo's Perfect Island Objection is doing the first. Can you see why?  

     Anselm's Response

You can read Anselm's response to Guanilo's objection on-line here: http://www.fordham.edu/halsall/basis/anselm-gaunilo.html

I admit that I have difficulty trying to understand his response clearly. Often, Anselm’s response seems to be nothing more than a mere restatement of his original argument (and thus not really addressing Guanilo's objection at all). But he does say the following, which (read charitably) has some promise:  

        "BUT, you say, it is as if one should suppose an island in the ocean, which surpasses all lands in its fertility, and which, because of the difficulty, or the impossibility, of discovering what does not exist, is called a lost island; and              should say that the be no doubt that this island truly exists in reality, for this reason, that one who hears it described easily understands what he hears.

        Now I promise confidently that if any man shall devise anything existing either in reality or in concept alone (except that than which a greater be conceived) to which he can adapt the sequence of my reasoning, I will discover that         thing, and will give him his lost island, not to be lost again.” Anselm’s Apologetic, Chapter III


From the quoted passage above, we can perhaps tease out the following response: in my original formulation of the argument, and indeed, given Anselm’s definition of God, we were considering the greatest conceivable being. But what was really meant (Anselm might argue) is the greasiest conceivable thing. When it is left open as to what the greatest conceivable thing is, there do not have to be limitations on the object under consideration. But there would be limitations if we needed to qualify the object as a particular kind of thing or other. If we must conceive of the greatest conceivable island, for example, then there are certain attributes that this island must have in order to qualify as an island. It must be surrounded by ocean on all sides, for instance. Or it may have to have been formed in a particular way—by a volcano, say—and it must be a land mass, with sand and rocks, perhaps some coconut trees, or certain vegetation, or a particular monkey to palm leaf ratio, etc. But the very qualities that make something an island may conflict with the very qualities that make something the greatest conceivable thing.

To see this, let’s play the which one is greater? game.

Which one is greater? An island that knows some things or an island that doesn't. Clearly, an island that knew what was going on would be greater than one that didn't, or couldn't. OK. So now we imagine our island as an island that knows some stuff.

Which one is greater? An island that knows everything, or an island that knows only some things. Clearly, an island that knows everything. OK. So now we imagine our island as one that knows everything.

Which one is greater? An island that is benevolent and can do morally good actions, or one that is morally neutral, or even worse—morally malevolent. Clearly, a benevolent island is better than a non-benevolent one. OK. So now our island is morally benevolent.

Which is greater? An island that is benevolent all of the time and in all ways, or an island the is benevolent only some of the time and in only some ways. Clearly, the first. OK. So now our island is omnibenevolent.

And which is greater: an island that has powers, or one that is inert? Clearly, one that has power, and lots (if not all) of it.  OK. So now our island is omnipotent. And so on.

After we have played the which one is greater? game, what we presumably end up with is something more like a God and less like an island. In fact, we could even ask: Which is greater? An island that is located in only one spot and not others? Or an island that can be everywhere all at once? What we'd eventually end up with is an omnipresent 'island' that is nothing like an island at all! This is because those qualities that make something great simpliciter—regardless of what kind of thing it is—are in conflict with those qualities that make something an island.

Anselm's original argument, in other words, is a demand for the greatest conceivable thing-whatsoever or thing-in-general. A greatest conceivable thing-in-particular is necessarily limited in certain ways—indeed, in all of the ways it has to be in order to qualify as a particular thing or other. And so a greatest conceivable thing-in-particular can never be the greatest conceivable thing-in-general.

And this might be what Anselm is getting at when he says:

"Now I promise confidently that if any man shall devise anything existing either in reality or in concept alone (except that than which a greater be conceived) to which he can adapt the sequence of my reasoning, I will discover that thing, and will give him his lost island, not to be lost again." (Emphasis mine.)

Anselm seems to be saying: if you can think of a greatest conceivable thing, and that thing ends up being an island, then so be it. The island must exist. But, given that the greatest conceivable thing is probably not an island (for reasons alluded to by the which one is greater? game), Anselm is confident that the argument won't deliver us a greatest conceivable island. But it will deliver the greatest conceivable thing, which is God.

Note: the above proposal is just my take on Anselm's response, and I am no Anselm scholar. Nonetheless, the response is available in logical space. I will leave as a response on Anselm's behalf, even if he did not endorse this move himself. Do you think the move works? Why or why not?


    Existence is Not a Predicate (Kant)

Immanuel Kant objects that ‘existence’ is not a predicate in the way that ‘red’, ‘six feet tall’, and ‘wacky’ are. If someone says ‘the castle is blue’, the speaker is attributing a property—blueness—to the castle. Yet when you say ‘the castle exists’, you are not, strictly speaking, attributing any property or characteristic to the castle. Rather, you are saying that the castle is instantiated, or exists in the world, or something to that effect. In other words, we aren't really adding anything to our idea of the castle, but to our idea of the world, and the way that it is—namely, that the world is such that the castle is in it.
 
To illustrate the above point, consider the following examples.

1.    Consider a table. Now abstract away all of its properties except existence. That is, imagine that you could leave it existing without all of its other properties: color, shape, size, function, and so on. But what is the difference between existence alone and nothing at all?

2.    Marge and Lisa are both compiling lists of the qualities of the perfect mate so that they can each put an ad in the personals. Since Marge and Lisa have similar tastes, they both end up with the following add: "Intelligent, hot chick in search of good-looking male, with an affinity for doling out compliments, playing pool til dawn, and talking philosophy until the cows come home." In fact, the only respect in which Lisa's ad differs from Marge's is that Lisa has supplemented the forgoing blurb with the extra qualification: "and must exist." Now, is Lisa's ad really any better than Marge's? No. Moreover, it seems that Lisa has misunderstood the purpose if the list, which is to set forth the qualities of a perfect mate. It is another matter whether these qualities are exemplified in an actual guy. Marge and Lisa can decide what they want to put into their description of the perfect guy, but the world decides (or determines) whether anybody meets such a description.

3.    We cannot simply build existence into our concept of things, yet it seems we have no problem doing this with regular predicates. For example, suppose we define a unicorn as: a one horned magical horse that exists. One might then claim that unicorns exist. But, of course, this is just a verbal trick. Nothing but the world can determine whether something fitting our concepts exists or not, and wishing it doesn't make it so. Contrast this, however, with regular predicates. If I define "smelloozer" as someone who drinks heavily and smells funny, then we can (uncontroversially!) claim that smelloozers drink heavily and smell funny. So we can build predicates into our definitions, but not existence. So, existence must not be a predicate.

Be careful! This last point is not the following objection: we cannot define anything into existence, yet Anselm does (with his definition of God), so his argument stinks. This objection is misguided, first, because Anselm does not (in the version I have proposed) define God as an existent thing. Second, this objection is aimed at ‘existence’ as a predicate. It is a recognition that ‘existence’ does not behave in the way that other predicates do, because it, unlike other (legitimate predicates) cannot be built into definitions.

The strategy of this Kantian objection is to indirectly attack the truth of premise 2. For an underlying assumption of the truth of premise 2 is that existence is a predicate like 'blue', 'round' and 'groovy' are. So, unlike Guanilo's objection, which attacks the form of the Ontological Argument, Kant's objection attacks the truth of one of the premises (viz., premise 2).


    Inconceivability Objection

Consider the series of positive integers: 1, 2, 3, etc. Now suppose I ask you to imagine the greatest conceivable integer. Rightly so, you should be confused, as it is quite obvious that the positive integers are not the sort of thing that admit of a greatest member. That is, for any integer you can think of, there will always be one greater. (In fact, there will always be infinitely many integers greater than the one you have in mind.) In this way, the Greatest Conceivable Integer is an impossible object. This objection claims that the idea of the Greatest Conceivable Being is like the idea of the Greatest Conceivable Integer—impossible!

Like Kant's objection, the strategy of this objection is to undermine the truth of one of the premises—viz., premise 1. The truth of premise one relies on our ability to at least conceive or imagine the greatest conceivable being. But if the Inconceivability Objection is correct, then we cannot conceive or imagine the Greatest Conceivable Being, and hence, we won't be able to make the supposition—premise 1—that gets the whole reductio going in the first place.

Does this objection work? What then becomes of the atheist’s negative existential claim: God does not exist? Can one be a coherent atheist of the inconceivability objection is correct? How?
 

    Existence in Reality vs. Existence in the Mind

One problem we might have with Anselm’s ontological argument has to do with the purported distinction between existing in reality and existing in the mind. This distinction suggests that there are two ways for something to exist—in the mind and in reality. But what sense is there in thinking that a thing exists at all when it ‘exists in the mind’? Certainly,  it isn’t at all clear how the identity conditions work out for things that exist in the mind. If I ask two people to imagine a pink elephant, for example, are the two people imagining the same elephant or not? If a pink elephant were to run around the corner, would the two people think, “Aha! I see that the pink elephant that exists in the understanding exists in reality as well”? Or could one think, “No. That’s not the pink elephant that exists in my understanding; mine doesn’t exist in reality at all.”?

Another example: Suppose that two artists imagine a painting in their head—such that Anselm would count the ideas of the painting as things that exist in the understanding. Is it enough that only one of the artists actually paints the painting in order for whatever it is that exists in the understanding to exist in reality as well, resulting in only one painting? Or would each painter have to paint separately and individually whatever it is that exists in the understanding before the painting(s) exists in reality, resulting in two numerically distinct paintings? Is there any way to tell the difference? If not, it seems we don’t have a criterion of identity for things that ‘exist in the understanding’, and so we should be suspicious that this is a legitimate notion of existence.

Put another way: we know what it takes for something(s) to be one thing or two when it (they) exist in reality (note: we will hopefully be spending some time in class discussing this point, so don't worry for now if it isn't clear to you). But this isn't the case with things that exist in the mind. For how are we to know whether the pink elephant that one person understands and the pink elephant that another understands are one and the same elephant? Put in terms of Qualitative and Quntitative (or Numerical) Identity:

Numerical identity is the relation that each thing holds to itself—e.g., I am numerically identical to myself, you are numerically identical to yourself, Jon Stewart is identical to himself, etc. If x is numerically identical to y, then x and y are one in number; "they" are one.

Qualitative identity, on the other hand, is the relation that many things can have to many others, provided that they have the same properties in common. For example, in class I talked about how two markers could have many of the same properties—e.g., they could both be white on the outside, with a black felt tip, a black plastic cap, cylindrical in shape, so many inches long, kept in a cardboard box, etc.—yet since they are two markers they are not numerically identical. Rather, they merely share some of their properties, or qualities—they are qualitatively identical—but they are not one and the same, numerically identical, marker.

How are we to distinguish between merely qualitatively identical (or exactly similar) pink elephants that merely exist in the mind, from truly qunatitatively (or numerically) identical pink elephants that exist in the mind? Unless entities are extended in the world—unless they exist in reality—there seems to be no other sense in which ‘they’ can be, or exist.

This objection would be another way to reject the truth of premise 2, since this premise presupposes that the distinction between existence in the mind and existence in reality is a legitimate one, and this objection directly questions this assumption.

    
     What Does "Greater Than'" Mean?

Anselm's ontological argument relies on the claim that something that exists in reality as well as in the mind is greater than something that exists in the mind alone (premise 2). We might wonder in what way existence in reality makes something that exists in the mind greater.

Certainly if one were to consider particular things—such as an ice-cold pitcher of beer or the perfect mate—we would much prefer that these things exist in reality than merely existing in the understanding alone. We might even say that the beer and mate that exist in the understanding are somehow ‘bettered’—and, moreover, that the world itself is ‘bettered’—if these presumably good things come to exist in reality (as well as in the understanding). However, particular other things—such as an overly vinegary bottle of wine or the most imperfect, annoying mate—would be things that that we would not think would be made greater if they actually existed.

In so far as we think that ‘greater’ means something like ‘better’ or ‘more improved’, it doesn’t seem that annoying and evil things would be made ‘better’ or ‘more improved’ if they existed in reality as well as in the understanding. On the contrary, these things, among countless others, would presumably be worse if they existed. What’s more, the world itself would surely be worse if these things existed in reality rather than just existing in the mind alone. So ‘greater’ cannot mean something like ‘makes the world a better place’ or makes the thing itself a ‘better’ thing, or else Anselm’s premise 2 would clearly be false. Certain things that exist merely in the understanding should stay that way; nothing ‘great’ or ‘good’ would come from them existing in reality. And so, in an attempt to find a plausible interpretation of premise 2 above—to find an interpretation of this premise that might at least have a shot at being true—Anselm must have had something else in mind when he uses that predicate ‘greater.’  

So what else could Anselm have meant by ‘greater’? Well, first, it is clear that there are a few things he couldn’t have meant. He couldn’t have meant ‘bigger’ or ‘heavier’ or ‘more powerful’. For even though we typically, in ordinary usage, use the predicate ‘greater’ to mean these things, it is not clear that we could even make sense of these predicates when applied to the things that ‘exist in reality and in the understanding’ or to those things that ‘exist in the understanding alone’.

For example, is a five-pound turkey that exists in reality as well as the understanding ‘bigger’ or ‘heavier’ or ‘more powerful’ than a five-pound turkey that exists in the understanding alone?

On one way of reading the question, the answer seems vacuously true, since turkeys that exist merely in the understanding don’t seem to be any particular size, weigh any particular amount of weight, or have any sort of power. Indeed, since things that exist in the understanding don’t exist in reality, it’s hard to see how they have any properties whatsoever; for things that don’t exist (in reality) can’t have any properties at all. But then, if this is what Anselm meant by ‘greater’ then it doesn’t seem to be of much help. For this would just amount to saying: things that exist merely in the understanding don’t have any properties, so of course things that exist in reality are ‘bigger’ and ‘heavier’ and ‘more powerful’ than things that exist merely in the understanding. Unfortunately, this sort of ‘greatness’ is way too easily had. For not only are things that exist in reality ‘heavier,’ ‘bigger’, and ‘more powerful’ than things that exist merely in the understanding; they are also ‘lighter’, ‘smaller’ ‘and ‘disappointingly less powerful’ than things that exist merely in the understanding. Comparative properties come cheap when one of the compared objects fails to have any properties whatsoever.

However, if you thought that things that exist in the understanding might have properties—if you thought that turkeys in the understanding have as much size and weight and power that is attributed to them, then you will have to say that the turkey that exists merely in the understanding is the same size, has the same weight, and has just as much power as it would if it existed in reality as well. To see this, ask yourself which weighs more?: a five-pound turkey in the understanding, or a five-pound turkey in reality? If, in light of the objection given in the previous paragraph, you are going to claim that objects that merely exist in the understanding have all of the properties that are attributed to them, then you will have to maintain that both the five pound turkey that exists in the understanding and the five pound turkey that exists both in the understanding and in reality each weigh the same. In which case, however, existence in reality doesn’t make something ‘greater’, if by ‘greater’ we mean ‘heavier.’ And similarly for ‘bigger’ and ‘more powerful.’  

So either the things that exist merely in the understanding have the properties associated with them or not. If they do, then it is not clear how a thing that exists in the understanding can be ‘greater’ if it exists in reality, if by ‘greater’ we mean ‘bigger’ or ‘heavier’ or ‘more powerful.’ For a thing that exists merely in the understanding will have just as much size, weight, and power whether it comes to exists in reality or not, since size, weight, and power are simply attributes that (we have assumed) will be true of a thing whether or not a thing that exists in the understanding exists in reality as well. However, if the things that exist in the understanding do not have the properties associated with them, then the claim that existence in reality is greater existence in the understanding alone is vacuously true. And it would be just as true to say, then, that existence in reality (as well as in the understanding) is ‘less great’ than existence in the understanding alone. So there must be something else that Anselm meant be premise 2.   

Perhaps he meant something like ‘more causally efficacious’? Or perhaps he meant by ‘greater than’ something like ‘more vivid’, ‘louder’ or ‘more amplified’? We will discuss these options in class.   


5.      Some Questions to Think About

What do you think that Anselm can say in response to some of the objections above? In class I suggested some of the moves he might make. Do find any of them plausible? Which ones and why? Do you think that the argument ultimately works? Or do you think that one of the above objections (or any others you might have thought of) are ultimately devastating to the argument? If the latter, which objection do you think is the most damaging? Can you think of any other ways to run this argument than in the way that Anselm did? If so, is this alternative better or worse and if so, how?

Discussion of these questions and more in class.
 



Sources and Suggested Further Reading

Saint Anselm, Proslogium.
Ontological Arguments, SEP entry.
William L. Rowe, "The Ontological Argument" in Reason and Responsibility, ed. Landau & Shafer-Landau.
Gideon Rosen, Anselm's Ontological Argument


Page Last Updated: Jan 07, 2011
Back to Meg's Main Page
Back to Meg's Teaching Page