In The Existence of God (2nd edition, hereafter: EOG), Richard Swinburne lays out a systematic cumulative case for the claim that it is more likely than not that God exists.

I have a specific objection to the third argument in this case, but I believe this objection throws a monkey wrench into the works, and creates a serious problem for the case as a whole.

To understand my objection, it is important to understand the general logical structure of Swinburne’s case for the existence of God. It is always natural and tempting to immediately focus in on the question of the truth of the premises of an argument for God, so in order to get a clear grasp on the logical structure of Swinburne’s case, it may be best to FIRST consider that structure apart from the specific content of the premises of the arguments in his case. The content of the premises will be important to make my objection, so we will get to the specific content at a later point.

One key idea in Swinburne’s logic is that we begin from a state of ignorance in which we are to imagine that we know ZERO empirical claims (both facts and theories). Swinburne thus controls the flow of empirical data, introducing one fact at a time, and arguing that in each case (with the exception of the problem of evil)  that the added fact increases the probability of the hypothesis that God exists.

The basic strategy is to (1) put forward an empirical fact, (2) show that the empirical fact is more likely to be the case if God were to exist than if there were no God, (3) conclude that the fact increases the probability of the hypothesis that God exists above the a priori probability that God exists (i.e. the probability based on ZERO empirical facts), then (4) introduce a second fact, (5) show that the second fact is more likely to be the case if God exists (and the first fact is the case) than if God does not exist (and the first fact is the case), (6) conclude that the second fact in conjunction with the first fact increases the probability that God exists above the probability based on the first fact by itself, (7) put forward a third fact, (8) show that the third fact is more likely to be the case if God exists (and the first two facts are the case) than if God does not exist (and the first two facts are the case), (9) conclude that the third fact in conjunction with the previous two facts increases the probability of the hypothesis that God exists above the probability based on just the previous two facts, and so on…slowly increasing the probability of God’s existence with each new fact.

Swinburne changes the strategy a bit when he gets to the argument from religious experience (in Chapter 13 of EOG), but the above pattern of reasoning is supposed to hold up until that point, and the above pattern of reasoning, filled in with the empirical facts that Swinburne has selected, is supposed to get us to the point where the probability of the existence of God is about .5 (meaning there is about a 50/50 chance that God exists).

Swinburne uses Bayes’ theorem to justify key inferences in his reasoning, so I will reformulate the above description of the logical structure of Swinburne’s case in terms of conditional probability statements.  Let’s use the letter e for evidence, plus a number to indicate which empirical claim we are talking about in the sequence of empirical claims introduced by Swinburne.  Thus, e1 represents the first empirical  claim in Swinburne’s case, and e2 the second empirical claim, and so on.

g: God exists.

k: [tautological background knowledge – analytic truths, truths of logic, math, and conceptual truths]

The probability of e1 being the case given that God exists is written this way:

P(e1|g & k) 

Here is how we represent the idea that the first factual claim is more likely to be the case if God exists (and we have only tautological truths as background knowledge) than if God does not exist (and we have only tautological truths as background knowledge):

P(e1|g & k) > P(e1|~g & k)

From this Swinburne makes use of Bayes’ theorem and infers that e1 provides evidence that increases the probability that God exists, over the a priori probability that God exists (the probability based on ZERO empirical facts):

P(g| e1 & k) > P(g| k)

Then Swinburne introduces a second factual claim e2. Again Swinburne argues that this factual claim is more likely to be the case if God exists than if God does not exist (now assuming e1 as part of our background knowledge, for after consideration of the first argument we are no longer completely ignorant of all empirical facts):

P(e2|g & e1 & k) > P(e2|~g & e1 & k)

From this Swinburne makes use of Bayes’ theorem and infers that the addition of this second empirical fact to the first empirical fact has again increased the probability of the existence of God, over what it was based on just the first fact by itself:

P(g| e2 & e1 & k) > P(g| e1 & k)

Then Swinburne introduces a third factual claim: e3. Again Swinburne argues that this factual claim is more likely to be the case if God exists than if God does not exist (now assuming both e1 and e2 as part of our background knowledge):

P(e3|g & e2 & e1 & k) > P(e3|~g & e2 & e1 & k)

From this Swinburne makes use of Bayes’ theorem and infers that the addition of this third empirical fact to the first empirical fact has yet again increased the probability of the existence of God, over what it was based on just the first two facts:

P(g| e3 & e2 & e1 & k) > P(g| e2 & e1 & k)

There are problems and objections that can be raised against each of the particular arguments that Swinburne uses to get up to the point where the probability of the existence of God supposedly reaches the halfway mark, but this post will focus on the third argument in the systematic cumulative case that Swinburne presents: The Teleological Argument from Spatial Order (hereafter: TASO).

TASO can be stated fairly briefly:

(e3) There is a complex physical universe that is governed by simple natural laws and the values of the constants of the laws and of the variables of the universe’s initial conditions make it probable that human bodies will evolve in that universe.

Therefore:

(g) God exists.

Remember, this is NOT a deductive proof for the existence of God.  (e3) is put forward NOT as a conclusive reason for (g), but merely as evidence for (g); (e3) is an empirical claim that is supposed to increase the probability of (g) relative to the probability of (g) based on just the two previous empirical claims:

(e1) There is a complex physical universe.

(e2) There is a complex physical universe that is governed by simple natural laws.

One problem is that it is not clear to me that (e3) is in fact true.  The fact that human bodies evolved once in this universe does NOT imply (by itself) that it was probable that human bodies would evolve in this universe.  I think a good deal of argumentation and evidence would be required to establish the truth of (e3).

Another more important problem with (e3) is one that Swinburne himself mentions and briefly discusses: “What reason would God have for taking an evolutionary route?” (EOG, p.188).  Swinburne goes on to talk about the beauty of the long cosmological “evolution” of the universe, and the beauty of plants and animals that resulted from the long history of biological evolution.  But this is all beside the point. God, being omnipotent and omniscient, could have brought about all of the beautiful plants and animals on earth including human beings in the blink of an eye.

God had no need to use the natural biological process of evolution, and no need to build such a process into the fabric of the universe.  The story in Genesis makes much more sense than evolution as the way that God would create animals and humans.  If there really was an omnipotent and omniscient person, then that person could have brought about all life on earth in an instant.  Most importantly, doing so would have bypassed hundreds of millions of years of animals suffering and dying from disease and parasites and predation and injury.  A huge amount of animal suffering was involved in the natural process of evolution, so a perfectly morally good person clearly would NOT have used evolution to produce human bodies when there was a much better solution ready at hand: create plants, animals, and humans instantly, as in the book of Genesis. So, it seems clear to me that contrary to Swinburne’s view, (e3) does not provide evidence in support of the existence of God, even assuming (e3) to be true.

But there is a deeper problem here than just the inductive inference from (e3) to (g).  What do we need to know in order to determine that (e3) is true?  I think we have to know, or have good reasons to believe, that the theory of evolution is true, and I think we have to know, or have good reasons to believe that the Big Bang theory of the universe is true.  What do we need to know in order to determine that the theory of evolution is true and that the Big Bang theory is true?  I think we need to know at least a little about: chemistry, biology, physics, paleontology, geology, cosmology, and astronomy.  We might not need to be experts in any of these scientific fields, but we need to have some grasp of some key facts, concepts, and theories in these areas of knowledge.

Furthermore, since the theory of evolution has been generally opposed by many Christian and Muslim religious believers, we need to have given some consideration to the problem of the apparent conflict between science and religion.  For example, if the Pope were to declare that evolution is a false theory, would that be a sufficient reason to reject this theory, even given all of the scientific evidence we have supports the theory?  What if the Bible clearly teaches that God created the world 6,000 years ago, is that sufficient reason to reject the theories and findings of geology and astronomy that indicate the age of the earth to be billions of years?  Unless one has done some thinking about science vs. religion, I don’t see how one can be fully justified in believing the theory of evolution. In sum, to have a justified belief in the theory of evolution and the Big Bang theory, one must have a bit of knowledge about the history and philosophy of science, in addition to knowing a good deal of scientific facts, concepts, and theories from several scientific disciplines.

OK.  Here is the big problem.  In order to know that (e3) is true, one must have a good deal of knowledge about science and about a number of important scientific disciplines, including a good deal of basic facts, concepts, and theories from a variety of scientific disciplines.  This means that the background knowledge that is in play in evaluating this third argument has grown exponentially.  A large portion of human knowledge has been pulled back into the picture, and Swinbure has completely lost control of the flow of data.  Because of the significant amount of empirical facts, concepts, and theories that are required to determine whether (e3) is true,  it is difficult to distinguish between such a sizable collection of information and knowledge and our normal everyday background knowledge.

One very important implication of this is that the problem of evil has itself been pulled back into the picture.  Knowing that the theory of evolution is true involves knowing that there has been hundreds of millions of years of animal suffering from disease, injury, parasites, and predation.  Swinburne’s strategy was to put off the problem of evil until after several empirical facts that favor the existence of God had been put forward one at a time, and the probability of the existence of God had been bumped upward several times.  But since the problem of evil has come rushing back in with just the third argument, it is no longer clear whether his logical strategy can work.  At any rate, the problem of evil cannot be dealt with after three or four more factual claims have been put forward in support of God’s existence.  The problem of evil must be faced as part of the consideration of the significance of (e3).