1. Don’t Criticize what you don’t understand.
I have been following this principle in my approach to Richard Swinburne. For more than a year now I have studied his case for God in The Coherence of Theism and The Existence of God. As an atheist the objective of finding significant problems in his case for theism is of interest to me, so that I can refute his case as part of a defense of my own viewpoint. But I’m not anxious to achieve that goal. So far, I have made no serious attempt to refute Swinburne’s arguments or to even raise objections to them, other than what just naturally pops into my mind as apparent potential problems or errors in his thinking.
Swinburne’s books defending theism have been available for over two decades, and lots of smart and well-informed philosophers and critical thinkers have read and commented on Swinburne’s case for God. If there are major problems or errors in his thinking, it is very likely that most of those problems have already been discovered and written about. It is thus unlikely that I will discover some major error in Swinburne’s thinking about God that no other philosopher or critical thinker has previously noticed and pointed out.
I’m in no hurry to evaluate Swinburne’s case for God. I can take my time to figure it out first, to get clear in my own mind how his arguments work and the logic of his case. If there are major flaws or errors in his reasoning, then slow and careful analysis of his case will eventually reveal most of those problems. If one first achieves a clear understanding of Swinburne’s case for God, then one will be in a good position to properly evaluate the quality and strength of that case.
2. Thinking out loud
In this particular post (and the next), I don’t plan to lay out an argument for a specific conclusion. I’m just going to do some thinking out loud about Swinburne’s use of Bayes’ theorem in his case for God, in the hope that some clarity or insight might be gained. So, I’m not promising any definite conclusions or even that I have important insights, just an honest effort to get a bit clearer about this topic.
3. GIGO: Garbage In, Garbage Out
GIGO is probably the most natural objection to make about Swinburne’s use of Bayes’ theorem to make his case for God. But I think this natural response is somewhat misguided. Bayes’ theorem serves as a conclusion-generating mechanism, in a way similar to the use of valid deductive argument forms by philosophers.
Often analytic philosophers will summarize an argument in a deductively valid argument form (e.g. 1. If P, then Q. 2. P. Therefore: 3. Q. – this is known as a modus ponens inference). This is done for a number of reasons.
First, by putting an argument into a valid deductive form, one simplifies the thinking by eliminating the issue of faulty logic. Anyone with a basic understanding or familiarity with deductive logic can verify that the logic of the argument is valid. The focus of thinking can thus be shifted to the question of the truth or justification of the premises of the argument. Furthermore, a deductive argument form ensures that all of the bases have been covered, that if one evaluates the truth (or justification) of each of the premises, then one has covered all of the issues necessary in order to arrive at a comprehensive evaluation of the argument. Also, in the case of arguments that have more than just one premise, the argument breaks down the thinking into pieces, so that one can focus attention on each of the premises, thinking about their truth (or justification) one at a time, thus helping one to achieve greater clarity about the argument, and greater confidence in one’s evaluation of the strength and quality of the argument.
It seems to me that Bayes’ theorem plays a similar role in Swinburne’s case for God, or at least It has the potential to play this kind of role. It is a bit of logic that takes bits of information as input, and then generates a conclusion based on that information. The inputs in this case are three conditional probabilities:
The posterior probability of the evidence: P(e I h & k)
The prior probability of the hypothesis: P(h I k)
The prior probability of the evidence: P(e I k)
The output is also a conditional probability:
The posterior probability of the hypothesis: P(h I e & k)
Like a valid deductive argument form, Bayes’ theorem helps us to set aside the issue of faulty reasoning (at least for the logic of the overall argument). Like a valid deductive argument form, Bayes’ theorem helps to ensure that our thinking covers all the bases, that our reasoning is comprehensive. Like a valid deductive argument form, Bayes’ theorem helps to break down the intellectual work into smaller bite-sized pieces, so we can focus on each piece one at a time, and achieve greater clarity about the argument, and greater confidence in our evaluation of the quality and strength of the argument.
The temptation is to object that the conditional probabilities given as input for use with Bayes’ theorem are speculation, guesses, subjective opinion, unfounded, or dubious for some reason or other. Such objections have some force, no doubt, but I think they are a bit misguided. So far as I can see, Swinburne is very reluctant to make estimates of conditional probabilities, and this means that he does not provide enough information to actually perform the necessary calculations that would make use of Bayes’ theorem. The big problem, it seems to me, is that Swinburne ought to have provided more estimates of conditional probabilities, or at least more claims that clearly constrain the range of acceptable estimates for the conditional probabilities that are needed to make use of Bayes’ theorem.
I would prefer “Garbage In, Garbage Out” over “Insufficient Data”, because if Swinburne had made some additional estimates of the relevant conditional probabilities, I would have a clearer understanding of his thinking, and a clearer target to aim at. Since he has not done so, I’m reduced to trying to fill in the missing data on my own; hence the next post on “Playing with the numbers”.
(To be continued…)