Miracles and reasonable belief

March 2016

Strange Notions is a website devoted, according to its administrator, to “dialogue between Catholics and atheists.” One of the site’s regular contributors, Karlo Broussard, recently posted an essay ostensibly defending belief in miracles, titled Is It Reasonable to Believe in Miracles? Broussard does little to defend an affirmative response, though. His article rather presents rebuttals to a few arguments by Hume and some unnamed skeptics who say that we should not believe in miracles. I’m going to stipulate that the answer to the title question is a qualified yes. For Christians in general and Catholics in particular, belief in miracles can be reasonable. This is not to say it always is. For any particular person, a belief held for the wrong reasons is unreasonable, even if other people hold it for good reasons. What I’m stipulating, then, is that some Christians do, or at least could, have good reasons for believing in miracles.

But it does not follow that I’m being unreasonable if I don’t believe in miracles. I don’t need to prove that no reasonable person can believe in them. There are lots of things that reasonable people can disagree about. Just because some people reasonably believe X does not mean that everyone who disagrees with them is being unreasonable. I think it reasonable to doubt the existence of miracles. I will say why, and then address some of Broussard’s rebuttals.

Broussard offered no definition of “miracle,” but I’ll assume that the reference is to events that are (a) supernatural, meaning contrary to or inconsistent with the laws of nature and (b) caused by divine intervention. Because of (b), I’m not going to believe in miracles, so defined, unless I am antecedently convinced of God’s existence. However, it would be begging the question for me to defend my disbelief in miracles by appealing to my atheism. Furthermore, I do not agree that proof of supernatural occurrences would constitute proof of God’s existence. A denial of God’s existence does not contradict the affirmation the some law of nature has been violated.

So I prefer to address the question of whether my disbelief in miracles can be justified without regard to what might cause them. In other words, is it unreasonable for me to disbelieve in the occurrence of events that are contrary to or inconsistent with the laws of nature? If I were to admit that some supernatural event had occurred while maintaining my atheism, a Christian might accuse me at that point of being unreasonable, but we can postpone that debate for now. For the time being, my argument is only that I’m justified in believing that no supernatural event has occurred, no matter what might have caused it to occur if it did occur.

The history of science has demonstrated to my satisfaction that the laws of nature are inviolable. I don’t mean that science has proven their inviolability to an absolute certainty, and no responsible scientist would say it has. I regard the inviolability of natural law as a fact in the sense proposed by Steven Jay Gould: a proposition supported by observation to such a degree that it would be intellectually perverse to withhold provisional assent. The qualifier “provisional” is vital. Nothing in science is to be regarded as infallibly established. Science is a human enterprise, and some of its characteristic methodologies have evolved for the specific purpose of compensating for human fallibility, but the compensation cannot ever be perfectly effective. But I am convinced that no better method of discovering what happens or can happen in the real world has ever been discovered.

And so, suppose some person tells me that at a certain time in a certain place, something supernatural occurred. Should I believe it just because that person says so? No, not just because they say so. And this is not a response unique to atheists. For every one of us, there are certain things we will not believe happened just because someone says they happened. We will differ in what kinds of things we will put on that list, but we all have such a list.

A comment now about evidence. I am an evidentialist. I think we should not believe anything without sufficient evidence. But what’s the difference between sufficient and insufficient evidence? To answer that, I appeal to Bayes’ Theorem. If some fact or set of facts F is offered as evidence for a proposition P, then I should believe P if and only if F establishes a consequent probability for P of more than one-half. Without a digression into the relevant mathematics, we can note that whether F can do that depends among other things on the prior probability of P. If the prior probability of P is quite low, then for F to constitute sufficient evidence for P, it must be the case that both (a) F is very probable if P is true and (b) F is very improbable if P is false.

We may note at this point that one implication of Bayes’ Theorem is that no evidence of any kind can establish a proposition that is antecedently regarded as impossible. In such a case, the prior probability is zero by definition, and then the consequent probability is necessarily also zero. Anyone who appeals to Bayes in good faith, therefore, must admit a nonzero prior probability for the proposition at issue. What I am claiming, then, is that I regard the prior probability of supernatural events as very low and that I have seen no evidence of a kind that would justify changing my mind about them. I will now address some of Broussard’s rebuttals from a Bayesian perspective.

First: What about that prior probability? Broussard admits that it must be low:

A miracle, by definition, is an unusual event, something contrary to the ordinary course of things. So, according to Hume’s view, every miracle is disqualified from the start, because every miracle is a rare event.

I think he mischaracterizes Hume’s argument here, but that is beside the present point. If I am arguing in good faith, I can just ask my interlocutor how often he thinks miracles happen and, unless I have good reason to object to his estimate, I can use that for my own prior probability. To my knowledge, no Christian has ever tried to quantify the frequency of miracles, but for the time being, I don’t think any Christian supposes that, of all the events that have happened throughout human history, more than 1 out of 1,000 was a miracle. If that supposition is correct, then I can assign a charitable prior probability to supernatural events of 0.001. Notice that this estimate is based on the assumption that miracles actually do happen. My response is not based on assuming that they can’t happen. Bayes won’t let me do that.

Broussard also says, “Many skeptics consider only how improbable a miracle is but hardly ever consider the improbability of a miracle not occurring despite the testimony.” I’m not speaking for any skeptic but myself, and “a miracle not occurring despite the testimony” is exactly what I’m considering. Testimony is evidence. In some situations testimony alone is sufficient evidence. In other situations it is not. As noted earlier, this principle is one that we all accept. We only disagree about its application in some particular situations. Bayes’ Theorem instructs us to ask, quite particularly: For this testimony that this event occurred, how likely is it that we would have the testimony if the event had not occurred?

So what is our particular situation? Broussard again: “Take for example the Resurrection of Jesus, to which the early Christians testified.” Now, “early Christians” is not nearly particular enough. I cannot evaluate testimony if I don’t know anything about the witness giving the testimony, and I think it reasonable to doubt that we have any testimony to the resurrection from anyone who was in a position to know whether or not it actually happened. Maybe Catholic dogma says otherwise, but I’m not a Catholic. I have no epistemological obligation to take the church’s word for anything.

This renders some of Broussard’s other rebuttals irrelevant. I don’t need to address the likelihood of the disciples lying or reporting hallucinations about the resurrection if, as I believe, we have no testimony from any disciple. The evidence that I have for the resurrection is not anybody’s testimony. It is a set of documents by authors whom I cannot identify and who say nothing about the sources on which they based their narratives, and they seem to have been written at least a generation and possibly several generations after their subject’s lifetime. The probability that such documents might report an event that didn’t actually occur is a long way from zero.

Please indulge me while I clarify the significance of that. Let’s be extremely generous and assign a prior probability of 0.01 to a man’s returning to life three days after dying. This is like saying that we could reasonably expect 1 percent of all dead people to be resurrected within three days. This is obviously unrealistic, but let’s see where it gets us. Our evidence in the case of Jesus is all the documents comprising the New Testament. Bayes asks us to assess the likelihood that we would have such documents if Jesus had indeed risen from the dead. The higher that figure, the more credible the resurrection becomes, and so to be charitable, let’s say it’s 0.99. Bayes also asks us to assess the likelihood that we would have such documents if the resurrection had not really occurred. For this factor, a lower figure favors the hypothesis, and the lower the better. We can now ask, considering the values already stipulated, how low this last probability has to be in order to get a consequent probability of at least 0.5 for the resurrection. And the answer is: 0.01 or less. I don’t think I’m being unreasonable if I think there is more than one chance in a hundred that a new religion would produce documents claiming that its founder, who the religion claimed was God incarnate and had been killed by evil men, had been raised from the dead.

But what about Paul, who according to Broussard “records Jesus appearing to many different people on several different occasions as well as appearing to more than 500 disciples at the same time”? To start with, Paul didn’t use the word “disciples.” He never referred to anyone as a disciple, of Jesus or anyone else. Paul does mention people who, according to the gospels, were among Jesus’ disciples, and he does tell us that Jesus “appeared” to them after his resurrection. He tells us nothing specific about those appearances, though, and the context clearly suggests that they were equivalent to the appearance Paul himself had experienced. In his own writings, Paul himself says nothing further about that appearance: He says Jesus appeared to him, and that is all he says. Yes, the author of Acts provides considerably more detail, but I have no reason to trust Acts as a historical document. In Paul’s own words, “he appeared to me also,” the “also” implying “the same way he appeared to all the others I have mentioned.” And from everything else in the undisputed Pauline corpus, especially Galatians, I infer that this appearance was some kind of personal revelation, perhaps but not necessarily a vision. There are all kinds of subjective experiences that a religious believer can construe as a revelation from God. Maybe it was a vision, but we don’t have Paul’s word for that, and even if it was a vision, that would not constitute eyewitness testimony of Jesus’ returning to life after his crucifixion.

As for the others mentioned by Paul, including the unnamed 500, his assertion that Jesus appeared to them is not equivalent to their testimony. For believers in scriptural inerrancy, there may be no epistemological distinction, but as for the rest of us, we don’t have 500 witnesses. We have one witness saying that 500 people had a certain experience, without telling us how he knew they had that experience. Did he talk to all of them? He doesn’t say. Did he personally know any of them? He doesn’t say. Was this just a tradition he’d been told about? He doesn’t say. For a Bayesian analysis, Paul’s testimony, as we have it, is nowhere near sufficient to overcome the very low prior probability of a man’s returning to life after being dead three days.

Broussard opens his essay thus:

Should I believe in miracles? This question doesn’t pertain to whether I should believe in this miracle or that miracle. It has to do with whether I’m rationally justified in believing in miracles as such.

As I noted earlier, he presents no direct argument for the proposition It is reasonable to believe, as a generality, that miracles have happened. He addresses some objections to the proposition, but you don’t prove a proposition by rebutting arguments against it. Furthermore, it appears disingenuous of him to suggest that the question has nothing to do with “whether I should believe in this or that miracle.” I strongly suspect that he couldn’t care less how skeptical anyone is about the miracles alleged by any religion but his own. But that’s OK. If Jesus’ resurrection is the only miracle he really wants me to believe in, then I think I have justified my belief that it probably didn’t really happen.

(This page last updated on March 12, 2016.)

Site home.

Religion index.