The Gap in Theistic Arguments

1 January 2007

In his paper The Gap in Theistic Arguments,1 Michael Martin attempts to show that “all the major proofs for a theistic God contain a gap.” Therefore, Martin asserts that, even if arguments like the Cosmological Argument were entirely persuasive, theism itself would still not be proven.

This is a very common tack taken by nontheists in response to theistic arguments, and so it is important to respond to Martin’s claim. In this brief critique I will show that Martin makes several mistakes which undercut his general strategy.

The Definition of Theism

One of the most common strategies taken by nonbelievers is to define theism as robustly as possible in order to make it more difficult to demonstrate that theism is true. However, this is an incorrect approach, because theism itself is a relatively modest position. Consider the American Heritage Dictionary’s definition:

“Belief in the existence of a god or gods, especially belief in a personal God as Creator and ruler of the world.”

According to this definition, polytheism (belief in many gods) and Deism (belief in a god who simply created the universe but is not involved in it or interested in it) are actually forms of theism. Thus, if the Cosmological Argument demonstrates the truth of Deism, then it also demonstrates the truth of theism, since Deism is just a type of theism. I think that this is actually the way that theism should be understood, especially when discussed in a philosophical context.

Some may complain that this definition of theism is much too broad. defines it as:

“1. The belief in one God as the creator and ruler of the universe, without rejection of revelation.”

According to this definition, theism is distinct from polytheism and Deism. Note, however, that the position is still a modest one. This definition does not even guarantee that the theistic God would provide revelation- it only claims that such revelation is a possibility.

These two definitions are appropriately conservative definitions of theism. However, in line with the common atheist strategy, Martin offers a much more robust definition. According to Martin, theism is “belief in one God who is personal and worthy of worship, who transcends the world but takes an active interest in it, and who reveals His goals for human beings through certain individuals, miraculous events or sacred writings. The theistic God is personal in that He can be understood on analogies drawn from a human person and human beings can enter into a personal relation with God, petitioning Him in prayer and referring to Him as “Thou.” He is worthy of worship since He is morally perfect and is infinitely knowledgeable and powerful.”

First, one should ask where on earth Martin got this definition. This type of definition certainly isn’t found in any dictionary or encyclopedia; it seems, rather, that Martin pulled it out of thin air. In any case, defining theism as such, Martin is able to contrast it with polytheism, deism, pantheism 2, and “belief in a finite God.” By stacking the deck with his imaginary definition, Martin is able to produce skepticism about the success of theistic arguments.

This aside, however, Martin must make yet another logical blunder in order to carry his argument. All throughout his article, Martin assumes that, if some nontheistic position is compatible with the evidence provided by a certain argument, then theism (as defined by Martin) is not supported. However, this type of reasoning, while common in atheistic writings, is clearly flawed. There are an infinite number of positions which are “compatible” with just about any set of evidential considerations.

For example, all evidence of an external world (with real physical objects and so on) is “compatible” with the position that we are actually just a brain in a vat being prodded by a team of mad scientists to imagine a physical world. Clearly, however, mere compatibility with evidence does not mean that a position is reasonable to hold.

Using this mistaken approach, Martin is able to reject every actual argument for the existence of God. Actually, Martin’s approach makes it literally impossible for any argument to provide evidence for theism (as Martin defines it). I will demonstrate this fact by considering one attribute Martin assigns to theism- omnipotence.

Omnipotence means that God is able to do anything which it is logically possible for Him to do. If we follow Martin’s line of thinking, then any argument that actually supports theism must be incompatible with the Finite God Theory (FGT). What possible set of evidence could support Martin-theism but not allow the possibility of FGT?

As it turns out, there is no set of evidence that meets such criteria. To see this, simply imagine how God could prove to Michael Martin that He is omnipotent. Perhaps God would create a universe out of nothing in front of Martin’s eyes. “Ah,” says Michael, “that’s quite impressive. But surely a finite God could do the same thing. Let’s see you make two universes at the same time.” If God were to create, instead, 5 million universes at the same time, all as big as our own universe, and all incredibly fine-tuned to allow the possibility of life, Michael could still respond, “Yes, but a finite God could do such a thing. Let’s see you make 5 million and 1 universes.”

The point is easy to see. God could never fully satisfy Michael because Michael could always want to see God do something slightly more powerful. Therefore, no possible state of affairs could actually be entirely incompatible with FGT. Thus, Martin is implicitly making the following argument:

1. For any set of evidence, E, if that set of evidence is compatible with Finite God Theory (FGT), then theism is not supported.
2. All E are compatible with FGT.
3. Therefore, all E do not support theism.

Premise 1 is Martin’s premise (whether explicitly stated or not, it is clear that this is his approach throughout the article.) Premise 2 was proven with the example of the never-satisfied Michael, but this premise could also be proven with regards to His moral perfection and His omniscience. It therefore follow that no conceivable set of evidence supports theism. But surely this is a ridiculous conclusion! We must therefore reject Martin’s first premise.

It is clear that premise 1 must go, but what can it be replaced with? If evidence is compatible with two views, then how do we pick one over the other? There are actually several ways we can discriminate between positions on the basis of evidence:

1. One position may be justified by properly basic beliefs.

This approach is only appropriate if some form of foundationalism is true, a defense of which is beyond the present scope of this article. 3 Suffice it to say that this type of approach is a common response to the “brain-in-a-vat” type scenario that I mentioned earlier. Foundationalists will deny that it is rational to believe that we are just a brain in a vat, since our belief in the external world is properly basic (i.e., it is a foundational belief that is justified in a basic way, not with reference to other arguments.)

2. Evidence may argue more strongly for one position than the other.

Sometimes evidence meshes much more nicely with one position than another competing position, even if both are compatible with the evidence. For example, the historical evidence for the resurrection of Jesus Christ, is admittedly compatible with some naturalistic, non-miracle hypothesis. However, in my view, the evidence meshes much more nicely with the resurrection hypothesis due to factors such as explanatory scope.

3. We may be inclined to believe the simpler, less ad hoc hypothesis.

This principle follows the well-known maxim, Ockham’s Razor. According to Ockham’s Razor, all things being equal, we should prefer the simpler hypothesis which requires less ad hoc rationalizations. This is a very important principle that undercuts a great deal of Martin’s case. For example, Martin (like many atheists) claims that all of the arguments for God’s existence are compatible with polytheism. But if we are to follow Ockham’s Razor, we will favor the simpler hypothesis of monotheism unless there is specific evidence for multiple gods.

4. We may consider the cumulative force of several arguments to see whether one position is favored over another.

When considering what we should believe, we often base our reasoning on several different arguments that work together in harmony. Each argument for God’s existence does not need to fully demonstrate theism if several persuasive arguments can work in harmony to demonstrate the view. Thus, the Cosmological Argument demonstrates a Creator of the universe, the Teleological Argument provides evidence that this Creator cares about the universe (proving theism over deism), the Moral Argument shows that this Creator is the locus of moral values, and the evidence for the resurrection demonstrates that this Creator has provided His creatures with revelation.

The Arguments

With all this foundational work done, it is easy to see the flaws in Martin’s specific examples. He never considers factors like (1)-(4) when critiquing theistic arguments. Thus, little more need be said about the specific examples; the reader can easily recognize where Martin errs. Notice however, that Martin is even able to deny the relevance of the resurrection of Christ to establish theism:

“Even Jesus’s resurrection and the virgin birth might result from the interventions of polytheistic gods or a finite God.”

This serves to demonstrate my claim earlier- according to Martin’s approach, no conceivable set of evidence could possibly support theism. But, as in the other cases, Martin makes easy-to-recognize errors. We should rationally believe that one God caused the resurrection, since this hypothesis is simpler and there is no justification whatsoever for several gods. Additionally, the hypothesis that one God raised Jesus meshes more nicely with other evidence. If Jesus was actually raised from the dead by God, then this would likely serve as a vindication of Jesus’s mission and words. But it is clear that Jesus believed in one God. Therefore, the evidence meshes most nicely with the hypothesis that one God raised Jesus from the dead, as Jesus himself proclaimed.

Against Martin’s claim that a finite God could have been responsible, similar responses could be urged. Richard Swinburne has argued that the hypothesis of an omnipotent God is simpler than the hypothesis of a God with a finite, arbitrary amount of power. 4 Additionally, Jesus Christ seemed to believe in an all-powerful God, so the hypothesis that an all-powerful God raised Jesus from the dead is a hypothesis that meshes with the evidence much better than the finite god view.


Martin takes a very familiar atheist strategy in his attempt to undermine theistic arguments. He provides an arbitrarily robust definition of theism, and fails to recognize that the mere compatibility of a view with a certain set of evidence shows almost nothing about the actual likelihood of such a view. By stacking the deck both ways, Martin is able to dismiss any conceivable evidence. However, when we consider our everyday strategies for evaluating compatible positions, we see that at least some of the arguments for theism are successful.


1. See his paper on the Secular Web: The Gap in Theistic Arguments

2. Pantheism is the belief that the world is God. In my view, pantheism is nothing but a form of naturalism that uses religious language.

3. Foundationalism is a theory of justification. According to this view, all beliefs are either basic or nonbasic. Basic beliefs are immediately justified, and nonbasic beliefs are mediately justified in some way by their relationship to the basic beliefs. A properly basic belief is a belief that is basic and also meets some other condition that makes the belief proper. For example, the belief that I hear a thud on the door might be considered properly basic. The belief that Jim has come over and is knocking on the door is a nonbasic belief that may or may not be justified on the basis of evidence. My belief that Jim is causing the thud on the door might be supported with evidence such as 1.) You invited Jim over, 2.) People often knock on the door when they are invited over.

4. See Swinburne, Richard. The Existence of God (New York: Oxford) 1991.


  1. Nice work Deming. I like your argument of the finite god best because as I read his paper I found myself thinking of the same thing as your argument, no matter what Martin would never be fully satisfied. This type of thinking enables him to continue to argue his point by simply bringing out the “what if” factor. At some point a belief in religion requires your faith and trust and without that Martin can continue to simply say “well a finite god can do that so what” which means he can continue to argue his point, even though it may not be very good. I was not impressed at all by his paper and the more I read the more disappointed I was by him. That’s just what I thought.

    — kyle taylor    Jan 11, 06:30 AM    #
  2. Reading this made me wonder if the following has been posited to an atheist in any debates: What sort of evidence would be credible evidence for God?

    I believe Atheists set up the following circular argument:

    1. To be credible evidence for a thing, the evidence must be “real.”

    2. To be “real” the evidence must be fully verifiable by scientific means – observable, measurable, repeatable, etc.

    3. Yet, anything with a full scientific explanation needs no God to explain it.

    In short, given their assumptions, there is no logical possibility for evidence of God to exist. Anything “miraculous” enough to demand a “God,” such as the experience of consciousness and appreciation for this universe would be thrown out as not real enough. Anything banal enough to be fully explained already (such as an atom of Hydrogen) would be thrown out as not suggestive of God.

    Vern    Jul 10, 06:03 PM    #
  3. Kyle, nice work. I keep getting the sense when I read atheists articles and listen to their talks that they are violating some laws of logic or argument, or are hitting below the belt, but I can’t always put my finger on it explicitely. Many of their flaws I can spot explicitly, though they often be blind to them, but I need and want to become more aware of their strategies of defence and offence, especially regarding those that are flawed logic. Anyone who ‘clings’ to belief, including theists, I want to be aware of their faults.

    Good job with this! And good job with this website, brother!

    People’s thoughts are muddled and confused and their thinking is naive and foolish, and this website is a good remedy for that. Note, I’m not saying anyone should “just believe” or even that God exists, which I think he does. I’m just saying we need to clean up our thinking, and atheists are so guilty of errors that they seem to not even be able to spot.

    gerald    Nov 22, 08:59 AM    #
  Textile Help