Christian Max Andrews and Atheist Justin Schieber recently had a debate on the existence of the Christian god. Both audio and a transcript are available online. I think it’s well worth listening to or reading. In what follows, I want to offer my initial impressions of both debaters’ opening statements.

Max Andrews’ Case for Christian Theism

Andrews offers three arguments for Christian theism: (1) the Thomistic Cosmological Argument; (2) a fine-tuning argument; and (3) an explanatory argument for the historicity of Jesus’ resurrection. Let’s consider each argument in turn.

The Thomistic Cosmological Argument

(1) There are contingent constituents to the universe.

(2) Given the contingent constituents of the universe, the existence of the universe (U) is very, very unlikely under the hypothesis that these constituents are themselves uncaused or self-caused (~Cu): that is, P(U|~Cu & k) ≪ 1.

(3). Given the contingent constituents of the universe, the existence of the universe is not unlikely under the hypothesis of a first uncaused cause (Cu): that is, ~P(U|Cu & k) ≪ 1.

(4) Therefore, U strongly supports Cu over ~Cu.

Is this a good argument? Let us first turn to clarifying how Andrews defines his terms.

Constituents of the universe (CC): include galaxies, planets, stars, cars, humans, leptons, bosons, and other particles.

Metaphysically necessary: For something to be metaphysically necessary that means that it could not have failed to exist—it exists in every possible world.

Contingent: Something is contingent if and only if it is not necessarily false and not necessarily true.

These definitions immediately reveal a fatal flaw which lies at the heart of Andrews’ argument. Both (2) and (3) begin with the expression, “Given the contingent constituents of the universe,” and then proceed to make statements about the probability of U on the existence (or non-existence) of a first uncaused cause. As it stands, however, this way of formulating the argument is fatally flawed. Andrews does not seem to have noticed that, on his definitions, treating the “contingent constituents of the universe” as a given entails U. The upshot is that, if we already know that the constituents of the universe exist, we will know that U is true, regardless of whether there is an uncaused cause.

Andrews may reply that this objection misses the point, since the question is whether U favors an externally caused universe (CU) over a universe without an external cause (~CU). But this reply itself misses the point of objection: (2) and (3) do not ask whether U favors CU over ~CU. Instead, they ask whether U favors (CU & CC) over (~CU & CC). Those are not equivalent.

Fortunately for Andrews, however, his argument can be easily modified to avoid this defect.

(1) There are contingent constituents to the universe.

(2′) The existence of the contingent constituents of the universe is very, very unlikely on the assumption that the universe lacks an external cause, i.e., Pr(CC|~CU & k) ≪ 1.

(3′) The existence of the contingent constituents of the universe is not unlikely on the assumption that the universe has an external cause, i.e., ~(Pr(CC|CU & k) ≪ 1).

(4) Therefore, CC strongly supports CU over ~CU.

Even so modified, however, the argument still seems doubtful. Andrews–and the argument–asks us to choose between three mutually exclusive possibilities: (i) the universe is metaphysically necessary; (ii) the universe is self-caused; and (iii) the universe is contingent (and has an external cause). I agree with Andrews that both (i) and (ii) seem far-fetched. But these options do not exhaust the possibilities and so we cannot establish (iii) by eliminating (i) and (ii).

For example, it’s also possible that (iv) the universe is uncaused. Here is an argument for an uncaused universe.

(5) Cause and effect are always related to time, i.e., causation is necessarily temporal.

Consider the nature of causation. Cause and effect are always related to time. Causes always happen before their effects or, if you believe in such a thing, happen at the same time as their effects.

(6) Time itself must be uncaused.

Since it would be a contradiction in terms to speak of a cause “before” time, it follows that time itself must be uncaused.

(7) Time itself began with the Big Bang.

(8) Therefore, the Big Bang is uncaused.

(9) Therefore, the universe is uncaused.

A Fine-Tuning Argument

Next, Andrews defends the following version of the fine-tuning argument.

(10) Given the fine-tuning evidence, a life permitting universe (LPU) is very, very unlikely under the non-existence of a fine-tuner (~FT): that is, P(LPU|~FT & k) ≪ 1.

(11) Given the fine-tuning evidence, LPU is not unlikely under FT (Fine-Tuner): that is, ~P(LPU|FT & k) ≪ 1.

(12) Therefore, LPU strongly supports FT over ~FT.

Is this a good argument? Let us first turn to clarifying how Andrews defines his terms. First, what is the “fine-tuning evidence” (hereafter, E)? Andrews lists two items of evidence, which I will call E1 and E2.

E1. The special low entropy condition.

E2. Strong Nuclear Force (Strong nuclear force coupling constant, gs = 15)

Second, what does Andrews mean by a “life permitting universe” or (LPU)? As I read him, a LPU is a universe that “is finely tuned for the essential building blocks and environments that life requires.”

Again, these definitions immediately reveal a fatal flaw which lies at the heart of Andrews’ argument. Both (1) and (2) begin with the expression, “Given the fine-tuning evidence,” and then proceed to make statements about the probability of FPU on the existence (or non-existence) of a fine-tuner. As it stands, however, this way of formulating the argument is fatally flawed. Andrews does not seem to have noticed that, on his definitions, E entails LPU. The upshot is that, if we already know the facts and observations reported by E, we will know that LPU is true, regardless of whether there is a Fine-Tuner.

Andrews may reply that this objection misses the point, since the question is whether LPU favors FT over ~FT. But this reply itself misses the point of the objection: (1) and (2) do not ask whether LPU favors FT over ~FT. Instead, they ask whether E favors LPU and FT over LPU and ~FT. Those are not equivalent.

Fortunately for Andrews, however, his argument can be easily modified to avoid this defect.

(10′) LPU is very, very unlikely on the assumption there is no fine-tuner, i.e., P(LPU|~FT & k) << 1.

(11′) LPU is not very, very unlikely on the assumption that there is a fine-tuner, i.e., ~(Pr(LPU|FT & k) << 1).

(12′) Therefore, LPU strongly supports FT over ~FT, i.e., P(LPU|FT & k) >> P(LPU|~FT & K).

Even so modified, however, this argument doesn’t justify Andrews’ claim that it “gets us to an extremely intelligent mind.” This is for two reasons.

First, Andrews completely neglects the issue of the prior probabilities of the rival hypotheses. (3′) could be true and yet it could also be the case that ~FT has a prior probability that is very, very, very much greater than FT. In other words:

(13) Pr(FT | k) <<< Pr(~FT | k).

If (13) is true, then we could combine it with (12′) to show that there probably is no fine-tuner.

(14) Pr(~FT | LPU & k) > Pr(FT | LPU & k).

The upshot is that because the argument says nothing about the prior probabilities of FT and ~FT, the argument is, at best, incomplete. Unlike the previous objection, I do not how to repair the argument to overcome this defect. (Theism & FT) has the greatest intrinsic probability of all the variants of FT. Likewise (Metaphysical naturalism & ~FT) has the great intrinsic probability of all the variants of ~FT. Here’s the problem. While there seems to be no good reason to think that theism has a higher prior probability than metaphysical naturalism, there is good reason to think that metaphysical naturalism has a higher prior probability than theism. Thus, while it might be tempting to revise (13) to something like this,

(13′) Pr(FT | k) >= Pr(~FT | k),

the problem is that there is good reason to believe (13′) is false. But something like (13′) is what Andrews needs in order to justify his inference to FT.

Second, Andrews makes the evidence for FT over ~FT appear much more impressive than it actually is by understating the evidence. Suppose, but only for the sake of argument, that (3′) is true. Given LPU, the fact that so much of the universe is hostile to life is much more probable on the assumption that there is no fine-tuner than on the assumption that there is one. And it’s far from obvious that the fine-tuning evidence for FT outweighs the evidence of the universe’s hostility to life–what I will call the “coarse-tuning evidence”–against FT.

Andrews’ statement, “All the empty space in the universe, all the dead stars, all the non-life in the universe are necessary components of the fine-tuning of the universe,” completely misses the point of this objection. Indeed, it’s amazing how, when talking about the fine-tuning, we are asked to believe that the Fine Tuner is extremely powerful, so powerful that the Fine Tuner could choose whatever values he or she wanted. But then, when we consider the coarse-tuning evidence, we are supposed to believe that the Fine Tuner was somehow stuck with coarse-tuning as the inevitable outcome of fine-tuning. But this is purely ad hoc. If the Fine Tuner could choose literally any laws of physics and literally any values for the physical constants, then the Fine Tuner could have done a better job designing the universe to be consistently fine-tuned, not the strange combination of fine- and coarse-tuning our universe actually has.

Explanatory Argument for the Historicity of the Resurrection of Jesus

Finally, Andrews offers an explanatory argument for the historicity of the resurrection of Jesus. He argues there are five facts relevant to the Resurrection (R).

E1. Jesus died by crucifixion

E2. The apostles claimed to have seen resurrection appearances of Jesus

E3. The conversion of Paul

E4. The conversion of James

E5. The empty tomb.

Let E be the combination of these five facts (E1 & E2 & E3 & E4 & E5). Andrews’ argument seems to be as follows.

(15) The best explanation for the five historical facts in E is that God raised Jesus from the dead, i.e., Pr(E|R) > Pr(E|~R).

(16) Therefore, God raised Jesus from the dead, i.e., Pr(R) > 0.5.

This argument fails for many reasons, but here I will focus on just one. (15) is false. Even if we assume that all five facts are true–an assumption I am willing to grant–it still doesn’t follow that R is the best explanation. Why? Because R isn’t even an explanation. Something cannot be the best explanation of a fact if it is not even an explanation of that fact. Contrary to what Andrews suggests, R does not predict any of the statements in E. This is because R, by itself, tells us nothing about the death or postmortem activities of Jesus. In order to explain R, however, Andrews has to make dubious assumptions about the postmortem activities of Jesus. For example, Andrews has to assume that, after His resurrection, Jesus had the ability to pass through solid matter and to appear and disappear at will, the power to create ‘heavenly’ visions of glory, and so forth. The problem with these assumptions, however, is that they are not implied by either R or our existing (background) knowledge. Thus, R is not the best explanation for the five historical facts in E. Indeed, utterly lacking in explanatory scope and explanatory power, R is not even an explanation at all. Instead, R is merely an ad hoc hypothesis. But this entails that (15) is false.

(to be continued)