Feedback archiveFeedback 2017

Reverse ontological argument?

A risky ploy for an atheist

Published: 30 September 2017 (GMT+10)
Anselm of Canterbury was the first to formulate an ontological argument for God

Ontological arguments are fascinating and convoluted (so this article might be a mental stretch for some readers). Essentially, they try to show that the very definition of ‘God’ entails that He exists. They have generated many responses and reformulations since Anselm of Canterbury first proposed one in the 11th century. One such response is a ‘reverse’ ontological argument for atheism, which parallels the form of an ontological argument, but is an argument against God rather than for God. It tries to show that the possibility of God not existing entails that God cannot exist. What are we to make of this argument? Nicholas C. from the United States writes:

Recently, I was presented by an atheist an argument against the existence of God known as the Reverse Ontological Argument. The argument is as follows:

  1. It is possible that God does not exist.
  2. If it is possible that God doesn’t exist, then God doesn’t exist in some possible worlds.
  3. If God doesn’t exist in some possible worlds, then God doesn’t exist in all possible worlds.
  4. If God doesn’t exist in all possible worlds then God doesn’t exist in the actual world.
  5. If God doesn’t exist in the actual world the God does not exist.

I questioned the validity of the fourth premise and the atheist responded that if the Christian God exists (who is omnipotent), then he should be able to exist in all possible worlds.

How do I refute this argument?

CMI’s Shaun Doyle responds:

God possibly can’t exist?

The first premise is crucial to understanding and refuting this argument. Unfortunately, the first premise allows for several misunderstandings. First, ontological arguments depend crucially on defining the term ‘God’, and this argument doesn’t offer any definition for ‘God’. Nonetheless, the form clearly mirrors the modal ontological argument set forth by Alvin Plantinga. As such, we should use Plantinga’s definition for the purposes of this argument—‘God’ is a maximally great being; i.e. a being that is omnipotent, omniscient, morally perfect, and metaphysically necessary (i.e. cannot fail to exist).

Second, many people when they hear “It’s possible that God doesn’t exist” don’t hear what the first premise is actually positing. They hear things like “As far as I know, God might not exist” or “God could’ve existed, but doesn’t actually exist”. Neither of these are right; the first is an issue of what we know, not an issue of what’s really possible, and the second makes God out to be a contingent being, which is nonsense. Rather, it’s asserting that ‘God’ possibly can’t exist, like how we could assert that ‘married bachelors’ possibly can’t exist. Remember that part of the definition of ‘God’ these arguments work with is that ‘God’ cannot fail to exist. In other words, the first premise isn’t simply asserting the idea that God might exist, but doesn’t actually exist; it’s asserting that the concept of God is possibly incoherent.

This is explained more in the second premise: “If it is possible that God doesn’t exist, then God doesn’t exist in some possible worlds.” Now, it’s important to understand what is meant by ‘possible worlds’. A ‘possible world’ doesn’t refer to another planet, or another universe, or anything like that. If we say ‘unicorns exist in some possible worlds’, we’re not saying that there are some planets or universes that actually exist where unicorns can be found. Rather, a ‘possible world’ is a total description of a hypothetical reality. For instance, saying ‘unicorns exist in a possible world’ is equivalent to saying, “unicorns possibly exist”. It’s saying, ‘There is a total description of a possible reality in which the statement ‘unicorns exist’ is true’. In other words, ‘possible worlds’ language is a way of talking about possibility and necessity. To say ‘God is necessary’ in this way of describing things, we would say ‘God exists in all possible worlds’. Likewise, to say ‘God cannot exist’, we would say ‘God doesn’t exist in any possible worlds’.

What are the implications of this understanding of premise 1, then? Since ‘God’ as defined above cannot fail to exist, if we grant this premise, we are forced to admit on pain of irrationality that ‘God’ so defined is an incoherent concept. Why? To say that a necessarily existing being, as God is, possibly doesn’t exist, is self-contradictory.

God possibly exists

How do we refute this argument, then? The simplest solution is to reject the first premise—‘God’ so defined must exist, and there is nothing incoherent about that idea. After all, consider what the atheist must prove to establish premise 1—that it could be that God couldn’t exist. The definition of God the atheist assumes in this argument forces them to demonstrate this. But how does the atheist prove that? They can only do so in classical ‘married bachelor’ style—show that there is something fundamentally incoherent about the concept of God. Frankly, that’s a burden of proof the atheist is welcome to—I wouldn’t want to be burdened with something so difficult to prove!

Moreover, if an atheist is seriously willing to posit this argument, then on pain of irrationality they cannot reject the validity of Plantinga’s modal ontological argument:

  1. It is possible that a maximally great being (aka God) exists.
  2. If it is possible that God exists, then God exists in some possible world.
  3. If God exists in some possible world, then God exists in all possible worlds.
  4. If God exists in all possible worlds, then God exists in the actual world.
  5. If God exists in the actual world, then God exists.
  6. Therefore, God exists.

It comes down to this: which premise do you think is more plausible: “It is possible that God does not exist”, or “It is possible that God exists”? And the way to answer that question is to answer this question: is ‘God’ (i.e. an omnipotent, omniscient, morally perfect, metaphysically necessary being) a coherent concept or not? On the face of it, it frankly seems far more plausible to think ‘God’ is a coherent concept than to reject it as incoherent. And we can add all sorts of other considerations to this, both intuitive and based on other arguments for God (See Does God exist? and Philosophical arguments for God), that can bolster our warrant for thinking ‘God’ is a coherent concept and thus possibly real.

‘God’ is incoherent?

Atheists who wish to defend the idea that ‘God’ is incoherent must demonstrate it. To do so, atheists generally try three tactics—single attribute objections (i.e. one attribute is incoherent—e.g. omnipotence), conflicts between multiple attributes (e.g. omnipotence and moral perfection), or God’s logical inconsistency with some fact about the world (e.g. God and evil, or God’s omniscience and human freedom). The theist has three options in response—refute the objection, modify the definition of the attribute/s, or discard the attribute as a divine attribute. Obviously, the four attributes crucial to Plantinga’s definition of ‘maximal greatness’ are indicative of intuitions about God the biblical theist cannot abandon. Therefore, we need to refute atheistic objections to them, or refine our definitions of those attributes to render the objections moot (or perhaps a bit of both!).

But even a brief look at the three atheistic options for argument shows that there are problems. Single-attribute incoherence arguments are essentially doomed to fail from the start. Why? Even if our current definition is shown to be inadequate, we won’t abandon the intuitive concept; we’ll reformulate our definition to account for those weaknesses. The theist isn’t bound to inconsistent definitions atheists insist on. Indeed, an inconsistent definition may seem like an argument against God to an atheist, but to a theist, it’s just an encouragement to define the attribute better. See If God can do anything, then can He make a being more powerful than Himself? and Questioning God’s many attributes.

Multiple-attribute incoherence arguments suffer much the same fate. Consider e.g. the idea that omnipotence, omniscience, and moral perfection are incompatible in a single being. But if sin is succumbing to a temptation to deviate from right moral order in the will, would an omniscient and omnipotent God ever be weak and foolish enough to sin? As such, omnipotence, moral perfection, and omniscience are plausibly mutually defining; God is too smart and strong to sin. My point is this: how does the atheist know that the four attributes of the maximally great being can’t be coherently co-defined? It’s hard to see how any of us could have such knowledge. Definitions are slippery things, as are intuitions (even ‘reasonable intuitions’) for what these ideas are about. How could the atheist craft an argument to anticipate all the possible reasonable options for defining those attributes? And why should the atheist’s assessment of what counts as a reasonable definition of these divine attributes worry the theist? Again, the atheist’s incoherence argument is just impetus for the theist to define his terms better. As such, it seems practically impossible to show that these attributes can’t be coherently co-defined.

Logical inconsistencies between God and some facet of the world (e.g. evil and God’s goodness and omnipotence, or human freedom and God’s omniscience) are probably the most common arguments one finds. But they too are weak. Why think God’s omni-attributes can’t be defined coherently so as to allow for the existence of evil or human freedom? In fact, there are several options on both counts that plausibly deal with these issues. The atheist has to refute them all, and then argue that no possible answer is likely, to hold that God is logically inconsistent with facets of the actual world.

Like I said above, the atheist has his work cut out for him if he wants to demonstrate that God possibly can’t exist. Many philosophers have realized this, and while incoherence arguments for God were once popular in the philosophical literature, they are not any more. Why? Theists have done the hard work showing just how difficult it is to sustain the idea that ‘God’ is an incoherent concept. The atheist’s incoherence argument is just the theist’s research opportunity. On the other hand, it’s positively easy to think that ‘God’ is indeed a coherent concept, and thus that God possibly exists. But if God possibly exists, then He must exist according to the modal ontological argument, because God cannot fail to exist by definition.

Reverse ontological arguments: risky business for atheists

The goal here has not been to defend Plantinga’s argument. None of what has been said implies that Plantinga’s modal ontological argument is a good argument (nor does it imply it is a bad argument). Rather, the point is that affirming the logical validity of any ontological argument for God is really risky business for the atheist who wants to stay an atheist. And proponents of reverse ontological arguments, to be rational, can’t help but affirm the logical validity of an ontological argument for God.

Related Media

Helpful Resources

Christianity for Skeptics
by Drs Steve Kumar, Jonathan D Sarfati
US $12.00
Soft Cover
Christianity for Skeptics
by Drs Steve Kumar, Jonathan D Sarfati
US $10.00
Kindle (.mobi)
Christianity for Skeptics
by Drs Steve Kumar, Jonathan D Sarfati
US $10.00
eReader (.epub)
The God Reality
by Rob Slane
US $8.00
Soft Cover
Does God exist?
by Drs Don Batten, Jonathan Sarfati
US $0.75
Soft Cover

Readers’ comments

Comments are automatically closed 14 days after publication.