The Last Gasp

When I was just starting high school, I got on something of a philosophical kick. This wasn't the product of nor impetus for any great amount of study, it was simply some quirky conclusion I had drawn from observations that had the effect of irritating my fellow students. At any rate, there was one day where an older student was brow-beating one of my classmates about some finer point of the workings of the world.  To defuse the situation, I issued the silly  challenge "Prove that God exists."

What I got was the ontological argument.

Now, at the time I was a christian, and so was just about everyone I knew.  If they weren't , they were at the least religious. I didn't know anyone who did not seriously believe in god. At the same time, though, I was pondering questions of solipsism, skepticism, and epistemology (without even knowing what they were). The ontological argument seemed to me like a godsend, banishing the need for troublesome demonstration.

Now, though, I positively loathe ontological arguments.  Some of that probably is due to my being fooled for a time by them.  More important though, is the way in which they conjure success.  Ontological arguments, and especially the modal version we'll be discussing, remind me of something my father used to say (and probably stole from someone else): If you can't dazzle them with brilliance, baffle them with bullshit.

William Lane Craig uses Plantinga's modal ontological argument as his final entry in Five Arguments for God. Aside from a much needed but far too short introduction to modal logic and the concept of necessary/possible, Craig spends little time explaining or defending the argument. I get the impression that he believes it speaks for itself. But modal logic can be confusing for anyone not well versed in the subject and a bit more effort was probably warranted (One claim against the MOA that I have seen repeated several times is that even the people who formulated it don't know what they're talking about).

I will try to keep things as simple as possible here, both because I know less about modal logic than I'd like and because I hope to leave you less baffled than before.  Essentially the MOA says this:

1. It is not possible that god does not exist
2. Therefore god exists

Everything else, all the tricky wording about necessary and possible, is in the argument to obscure the basic idea I just stated.  Proponents of the MOA will foam at the mouth over this, vehemently denying the truth of it, but after having spent several days going over the rejections and counter-arguments, I haven't seen one solid refutation of my assertion that doesn't leave the MOA toothless.

Modal logic deals with things in terms of necessary and possible.  You may notice that this bears some resemblance to necessity/contingency mentioned in The Cosmological Argument from Contingency.  While the concepts can be related, they are not exactly the same.

Modal logic hinges on what is best described as alternative world concepts.  Alternative world concepts are basically a collection of statements that describe a world different from the actual. Note that these are not actual alternate worlds, they are merely descriptions of conceptually possible worlds. The actual world is seen as the alternative world concept that has been actualized.

Using the idea of alternative world concepts (AWCs), we can now explain what necessary and possible mean here.  To say something is possible is to say that it is true in one of these alternative world concepts. To say that it is necessary says that it is true in all of them.

Already, an issue presents itself.  The old spectre of equivocation has once again raised its head, this time around the term "possible."

Generally, when we speak of something being possible, we are speaking of uncertainty.  If I say, for example, that it is possible that non-terrestrial intelligent life exists, I am saying that I am uncertain as to whether it is true. This uncertainty also logically entails the opposite possibility statement, it is possible that non-terrestrial life does not exist. Replacing the first statement with the second does not change the meaning. This type of possibility is usually referred to as empirical or epistemic possibility because it is based on (a lack of) knowledge and data.

Modal possibility, on the other hand, is more akin to conceivability. Craig's example of  George McGovern being President is a good example. Empirically, this is not possible, but there is nothing in the statement that appears to make it inconceivable. Thus, in some alternative world concept, George McGovern is president and the statement is modally possible.

Obviously, the MOA uses the second type of possibility, but it attempts a bit of trickery here that its proponents don't seem eager to correct. Because we generally use possibility in the first sense, when we hear a statement such as, possibly X exists, unless  we know that it is not the case, we tend to accept the statement.  Really, though, what we have agreed to is that in some alternative world concept, X exists. This difference may seem trivial and innocuous, but the entire MOA rests on it. I'll explain why once we begin looking at the actual argument.

Craig doesn't even bother to mention or explain the term necessary, though it is crucial to the MOA.  The problem with necessity is that we have no good way of determining if a statement or entity is necessary.  One suggestion is that a something is necessary if it is impossible for it not to be true. The laws of logic are usually put forth as an example. I'm not so sure this works. Consider this modal statement: Possibly, different laws of logic exist. Similar arguments have been put forth to propose that logical inconsistencies may exist between AWCs.  The other option for modal necessity is that the definition of the entity requires it to be necessary.  But this doesn't really work either, as we could make everything necessary simply by adding that clause to its definition.

As if things weren't complicated enough already, there are different modal logic systems, and the MOA doesn't work in all of them. Neither Plantinga nor Craig offer any reason for selecting this system (S5) over the others.

For the sake of relative brevity, we'll turn now to the argument itself. The first premise seems uncontroversial  until one unpacks it. Firstly, Craig gives no explanation of what maximally great means. Plantinga offers up a definition including the three omni-attributes usually associated with the christian god, not forgetting to throw in that it is necessary.  Putting aside the internal problems with this definition, the description of a maximally great being bears a curious resemblance to the perfect being from Anselm's ontological argument and is, as such, vulnerable to many of the same criticisms. To their credit, Plantinga and others have presented decent defenses of their position. (I am not alone in thinking these defenses don't hold, but that is a discussion in itself.) It also appears that the property of being necessary was simply tacked on to the definition, which leaves it open to the criticism of arbitrary assignment.

The second , more damning issue with the first premise is the trickery I mentioned earlier. The statement "It is possible that a maximally great being exists" also entails the opposite "It is possible that a maximally great being does not exist." With empirical possibility, the statements are essentially interchangeable. With modal possibility they are not; substituting the second statement as the premise produces a contradictory conclusion.
Breaking it down by Craig/Plantinga's numbers:

1.It is possible that a maximally great being does not exist.
2. If 1., then a maximally great being does not exist in some possible world
3. If 2., then a maximally great being does not exist in any possible world (because it is necessary, if it exists, it must exist in all AWC. Since there is one AWC where it does not exist, it does not exist)
4. If 3., then a maximally great being does not exist in the actual world
5. If 4., then a maximally great being does not exist.
6. Therefore, a maximally great being does not exist

Using the same formulation of the argument, we have reached the conclusion that a maximally great being (god) does not exist. Given that this contradicts with the conclusion reached in the MOA, Craig/Plantinga have something of a problem.  The only real way they have of rescuing the argument is by arguing that our first premise above to be untrue. Or to put things another way:

1. It is not possible that god does not exist.

Really, though, if they could do that, why bother with the MOA?

That is the last gasp of Craig's five arguments, and I'm afraid that they haven't fared particularly well. We have plenty of reason to reject these arguments as invalid and useless. Craig wishes to criticize the New Atheists for not addressing them, but we can see that it wasn't necessary. Though dressed up in semi-modern terms (George McGovern, really?) they hinge on the same fallacies they always have.

Next time I'll begin examinations of some of the arguments against the existence of god.

(Those of you interested in learning more about refutations of the MOA should start with Paul Almond, Debunking Christianity, and R. Kane. They are more thorough than I have been in raising additional objections.)