It was in the wake of Scottish philosopher David Hume (17111776) having waged war on Christianity with his work, Of Miracles (1748), that British minister Richard Price (1723-1791) was compelled to offer a Christian rebuttal to Hume’s critique of the probability of miracles. In his rebuke, Hume argued that it is much more probable that people inaccurately claim to have seen the risen Jesus than the probability that Jesus actually rose from the dead. More bluntly, Hume said of miracles, “It is nothing that men should lie in all ages.” Thomas Bayes (1701-1761), the creator of the theorem, never published his work concerning probability. Only after his death did his friend Richard Price dust off Bayes’ work and bring it to bear against Hume’s criticism in the form of a statistical tool known as Bayes’ probability theorem.
The Christian argument for the existence of God is best pursued much as this column series has progressed —through a cumulative case. The cumulative case approach to the existence of God is similar to a trial lawyer representing a client, presenting evidence that cumulatively points to a unified conclusion. That is what the rigor of Bayes’ theorem does. Simply stated, Bayes’ theorem is defined as “a mathematical formula used to calculate conditional probabilities.” More intricately, it is a successive and progressive mathematical approach to determining the relative probability of a particular outcome for each piece of evidence, and then the aggregate.
Today, Bayes’ theorem is used in various fields, including medicine, artificial intelligence, industrial risk assessment, and weather forecasting, among others, to assess the probability of different potential outcomes. Nevertheless, it was first used as a Christian apologetic to argue for the probability of the resurrection of Jesus, and ultimately for the existence of God. It is the latter purpose, of course, that is of interest here. Twenty years ago, quantum theory physicist Stephen Unwin took on the challenge to retrace, so to speak, the steps laid by Bayes and Price over two hundred years ago.
For those who speak the language of math, if we set about determining Bayes’ probability for the proposition, “God exists” (P(G)), each piece of evidence (E) is added, and the effect on probability is determined. Where P(G*) is the probability that God does not exist, and En is the number of pieces of evidence in the form of apologetic arguments, then: P(G)=P(GE1&. . .&En)/P(G*E1 & . . .& En) In his book, The Probability of God, Unwin considers only six items of evidence from a cumulative case in the form of the arguments from 1) The recognition of goodness; 2) The existence of moral evil; 3) The existence of natural evil; 4) inter-natural miracles; 5) Extra-natural miracles; and 6) Religious experience. Since Bayes’ theorem compares the calculated probability value for each piece of evidence against a prior probability, an initial prior probability of 50 percent was chosen. Then, cranking the Bayesian handle, the output of the theorem showed that the probability of the existence of God, with each result building upon the previous, went from the initial pre-probability of 50 percent to 91 percent. Subsequently, working through each of Urwin’s arguments for God’s existence, the final Bayesian probability that God exists, according to Urwin’s input, showed 67 percent—a considerable margin above 50 percent. That is to say, it is statistically more likely that God exists rather than not.
WhileIsaluteUrwin’swork,Iam not so sure 67 percent is an accurate assessment—it could be much higher (or lower). Any statistical model is only as good as the data that is input. The old saying “garbage in—garbage out” was never moreapt.Mycontentionstemsfrom Unwin’s limited choice in both the quantity and quality of arguments used as model input. Concerning statistical mathematics, more data points, all things considered, yield more accuracy. While there is nothing wrong with Unwin’s choice of evidence, there is, in my opinion, a vast array of evidential arguments that greatly strengthen the cumulative case and could further strengthen the quality of the Bayes theorem output if utilized. Most notably, the evidential weight from cosmology, morality, design, and the historical resurrection of Jesus could do nothing but improve the reliability of Bayes’ output.
There is much more to be said on this topic, and we will revisit this discussion later. But for now, know that a 67 percent probability that God exists is certainly not zero. Nor is it a mere 50-50 chance. Sixtyseven percent probability says “It is more likely than not that God does indeed exist.” So, for now, what do we do with 67 percent? Join us next time as we look to Blaise Pascal and his “Wager” proposition. Until then, 67 percent—think of it—Is God Dead?
Gloria in excelsis Deo!
Ty B. Kerley, DMin., is an ordained minister who teaches Christian apologetics, and relief preaches in Southern Oklahoma. Dr. Kerley and his wife Vicki are members of the Waurika church of Christ, and live in Ardmore. You can contact him at: dr.kerley@isGoddead.com.