5 / 06

#51 Ontological Argument for the Existence of God

April 07, 2008

Dear Dr Craig,

This is an extension to an e-mail I sent you about the ontological argument a while back. My first question concerns the soundness of the argument in general. The second, however, concerns its theological implications.

Firstly, then, the soundness of the argument.

I’m thinking particularly about the version of the OA you propose in To Everyone An Answer [Inter-Varsity Press, 2004] together with what you write about the argument in Philosophical Foundations [Inter-Varsity Press, 2003] (though my rendering of the argument below isn’t an exact quote):

(OA1) It’s possible that an all-surpassingly great being exists (i.e. a being greater than which nothing can be conceived). In other words, an all-surpassingly great being exists in some possible world.

(OA2) If an all-surpassingly great being exists in some possible world, then it exists in every possible world.

(OA3) If an all-surpassingly great being in every possible world, then it exists in the actual world (since the actual world is clearly a possible world).

(OA4) If an all-surpassingly great being exists in the actual world, then an all-surpassingly great being actually exists.

The problem I have with this argument it that we seem able to take the general form of the key premise (i.e. It’s possible that X exists) and then plug-in various different definitions of X in order to arrive at absurd answers.

Obviously, defining X as something like an all-surpassingly great island doesn’t work, since, as you and others rightly point out, islands are, amongst other things, inherently material (and are therefore contingent on the existence of space and time); moreover, it’s far from clear as to what properties make an island great (for some it might involve plenty of palm trees; for others, it might involve no palm trees at all). So defining X in such a way doesn’t seem to work.

But what if we define X as a being that necessarily exists but isn’t, say, necessarily all-knowing? The argument you give against this idea (if I understand you correctly) is that God must be such that nothing can exist independently of his power, meaning there must be a possible world where God alone exists. You then argue that the existence of God is therefore logically incompatible with the existence of X, meaning X is an incoherent concept.

But I have two problems with this argument.

Firstly, for God to be able to actualise a world where X doesn’t exist seems to be asking God to do the logically impossible (since X exists necessarily). So why should we think God must be capable of doing such a thing? Secondly, even if we grant that the existence of God is incompatible with the existence of X, why should it follow that GOD exists as opposed to X? That is, if the proponent of the OA in its original form can argue that, since God’s existence entails there being possible worlds where God alone exists, the concept of X is incoherent, why can’t the skeptic argue that, since X’s existence entails there being NO possible worlds where God alone exists, the concept of God is incoherent? This looks like a mexican stand-off to me.

I’ve considered appealing to factors like simplicity or likeliness at this point; but this seems misguided. Simplicity can be important in deciding on the most likely explanation for a given body of facts. But I’m not sure this is relevant here, since the OA isn’t an inference to the most likely explanation. Rather, the OA starts from the premise that God’s existence is either possible or it isn’t, and then tells us what logically follows as a result. If God’s existence is possible, then it’s necessary, meaning God exists; if it isn’t, God obviously doesn’t exist.

The OA is therefore different from something like the teleological argument, where we might want to argue that God is a simpler and more elegant explanation for the universe’s apparent fine-tuning than the extravagant ontologies posited by many-worlds-hypotheses. The OA is a question of metaphysical fact. From what we have said about X, it therefore seems to follow that either God exists or X exists (together with any number of other necessarily-existent beings; for if a modified OA works for X, why can’t it work for any other similarly-defined X? It might, I guess, be argued that to simply tack ‘necessarily-existing’ onto a being is somewhat gratuitous. But I don’t see why this would be any more gratuitous than tacking ‘all-surpassingly great’ onto a being, or why gratuity would even be important here. In any case, couldn’t X be a necessarily-existing number or a necessarily-existing set or something like that? Granted, the existence of such entities is problematic for the same reasons Platonism as a whole is problematic. But the existence of such entities doesn’t seem logically impossible; nor do the entities in question seem conceptually gratuitous or incoherent).

I therefore have concerns about the soundness of the OA.

But I also have concerns about its theological implications. For, in light of the OA, doesn’t it follow that God’s actions are a necessary consequence of his nature, since, given any set of conditions, an all-surpassingly great being will necessarily react in the ‘greatest’ possible way. Creating the universe, for instance, gives God opportunity to lavish his love and grace on other creatures, meaning only a lesser being would choose not to create the universe. It therefore follows that God’s creating the universe is necessary--which logic can be extended to God’s answering of prayers, and, worse still, Christ’s dying on the cross. In fact, on this view, the only thing that makes our world contingent is man’s free will; and, whilst this may be so, it seems wrong to me, since it leaves so little scope for the action of God’s grace and freedom and sovereignty. The OA therefore appears to eliminate God’s free will in the same way physicalism eliminates man’s free will.

As always, I’d be extremely grateful if you could spare the time to address these issues. (Though I don’t have a philosophical background, so I’d be grateful if you could assume a low level of technical knowledge as your starting point -- especially when it comes to the nature of necessity!). I get the impression you’ll be addressing the ontological argument in your forthcoming revision of Reasonable Faith. If so, I look forward to reading it. (That said, I look forward to reading it either way...).



United States

Dr. craig’s response


The ontological argument for the existence of God

Let’s deal with the soundness objection to the ontological argument for the existence of God first, James. What you’re really objecting to is not the soundness of the argument. An argument is sound if it is logically valid and has true premises. Since this argument is logically valid, your objecting to the argument’s soundness would require you to think that (OA 1) is false. But the parodies of the argument you mention don’t show that it’s impossible that there be a maximally great being, or, as you put it, an all-surpassingly great being. Rather the point of such parodies is that there is no good reason to think that (OA 1) is true. For any reason for thinking that premise to be true would also be a reason for accepting an obviously false premise in one of the parodies of the argument. So the argument, even if sound, is not a good argument because there’s no non-circular reason to think that (OA1) is true.

Now as you note, some of the parodies, like arguments for an unsurpassingly great island or a necessarily existent lion, are not well-thought out. We have good grounds for thinking that such concoctions are impossible, in contrast to the apparently coherent idea of a maximally great being. More difficult to assess is the notion of what I called a quasi-maximally great being: a being which is just like a maximally great being except that it lacks, for example, complete omniscience (like the God of so-called Open Theism, Who lacks knowledge of future free acts of men). My argument against such a parody is that any reason for thinking a quasi-maximally great being is possible also warrants belief in the possibility of a maximally great being, but if we think that a maximally great being is possible, then we must say that a quasi-maximally great being is impossible after all, since it’s impossible for the two to co-exist in the same world.

Now you object, why should we think that God must be capable of refraining from creating a quasi-maximally great being, since His refraining from creating it would be logically impossible? I think your question highlights the inadequacy of a definition of omnipotence simply in terms of what is logically possible for someone to do. To borrow a famous example, on this definition a person who is essentially capable of only scratching his ear could count as omnipotent since other actions are logically impossible for him to do! That is surely an inadequate concept of omnipotence! Similarly, if there exists another being outside God’s creative power, then this is plausibly not consistent with God’s being omnipotent. I’d say the same with respect to numbers and other allegedly necessarily existing abstract objects: given God’s necessary existence, they are broadly logically impossible. (For a brilliant, if technical, discussion of an adequate definition of omnipotence, see Thomas Flint and Alfred Freddoso, “Maximal Power,” in The Existence and Nature of God, ed. Alfred Freddoso [ Notre Dame, Ind.: University of Notre Dame Press, 1983], pp. 81-113).

The ontological argument for the existence of God – Is a maximally great being possible?

Your second objection to the ontological argument for the existence of God is the more difficult of the two: even if we grant that the existence of God is incompatible with the existence of a quasi-maximally great being, why should it follow that God’s existence is possible as opposed to the existence of a quasi-maximally great being? My answer here is that there’s an asymmetry between our intuitions about the possibility of such beings. Any intuition for thinking a quasi-maximally great being to be possible also warrants belief in the possibility of a maximally great being; indeed, the way we came to form the idea of the former was by diminution of the idea of the latter. But our intuition of the possibility of a maximally great being, once we understand its implications, tends to undermine our intuition of the possibility of a quasi-maximally great being. We begin to suspect that despite appearances, it’s not really possible after all.

Notice that all this is said solely on the basis of appeal to modal intuitions alone (i.e., our intuitions about what is possible or necessary). But one of the significant new wrinkles in discussions of the ontological argument is support for (OA1) which goes beyond mere modal intuitions. Here considerations of simplicity may indeed have a role to play. By appealing to such factors, one is not altering the form of the ontological argument for the existence of God but marshalling reasons in addition to modal intuitions for the truth of (OA1).

Now concerning your theological misgiving: I don’t see that this has anything to do with God’s being metaphysically necessary. Even if God exists contingently, so long as He is essentially morally perfect you can run your argument that He is morally obligated to do the best and therefore must create the best possible world. So this is a problem that faces any theist who thinks that God is morally perfect.

The misgiving is to be met, I think, by questioning the assumption that there is a best of all possible worlds. Worlds may just get better and better without limit. For any world God chooses to create there will always be a better one that He could have created. God must at most create a good world, not the best world (since there is no such thing). Moreover, there’s no reason to think that God must create anything at all. In a possible world in which God creates nothing, there is only He Himself, the paradigm and locus of goodness—the summum bonum. That’s a pretty good world, to say the least!

- William Lane Craig