Chinese Whispers I can believe, but many of the miracle accounts I've heard were written by the eyewitnesses. Maybe my priors for human honesty and reliability are too high?

By some estimates, replicability in scientific cancer research is about 11%. We can reasonably assume that reports of miraculous events have at least as many flaws as laboratory experiments.

Only anecdotal, but it's hard to get good hard data on this because it would require collecting data in so many different languages.

You might be able to get better data by narrowing the field somewhat. For instance, by looking at the comparison in reported miracles between Mormons and conventional Christians (I recall from an earlier discussion on the topic here that Mormons reported a higher rate of answered prayers than any Christian denomination, except possibly devout Pentecostalists depending on how the measurement was taken.)

commands given by God that seem horrifyingly immoral

Just out of curiosity, what moral system are you using to make this judgement?

Aiyen:

I think the place I would start -- and did start -- is with the question, "Have I ever personally experienced a miracle?" I quickly discarded most possibilities and came down to just a few, all of which were fundamentally a strong of events, each of which was individually highly unlikely. But unlikely things happen every day, so that's not enough to say it is a miracle.

Next, I would ask if people I personally know and trust have experienced miracles. I'd ask them about their experiences. For me, I knew no one who had experienced anything that sounded like a miracle. Again, there were some unlikely coincidences, but that's not really enough. If one of them claimed to have experienced a true miracle, then I'd keep evaluating their credibility over a long period of time. Does their story change? Reasonable consistency would be more credible. Do they act like I would expect them to act if they had really experienced a miracle? A non-believer suddenly becoming a devout missionary in the absence of any personal crisis would be credible. Do they get anything by claiming to have experienced a miracle? Someone getting attention from conspicuous religiosity is less credible for me, though that doesn't necessarily make sense: if I had personally experienced a miracle, I'd think that I would sell off all my worldly possessions to talk about it as much as possible. Is this yet another in a series of flighty decisions? Going from new age to confused to born-again is not very persuasive for me. And so on.

Finally, given that people have cell phones with cameras now, and given that virtually everything else shows up on YouTube, I would think that there would be at least some credible video. I'd also think that I wouldn't be able to analyze whether it is credible, given the possibility of a good special effects studio. But I bet someone else would and that I would hear about a truly inexplicable video.

Max L.

So, a few observations on miracles.

1. There are miracle stories in every religious tradition and plenty of not-exactly-religious traditions. Unless there's some big difference in credibility -- which I'm not aware of any reason to think there is -- if you think "no smoke without fire" about one set then you should think the same about the others too. Which means you either have to believe in lots of different gods, or believe in one god and lots of evil spirits (or something) that just happen to do more or less the same sorts of miracle. (Or, I guess, believe that miraculous things happen but they're brought about by people's latent psychic powers or something, but that's pretty far from any religion's account of these things.)

2. When miraculous stories are investigated carefully, they consistently seem to evaporate. This happens even when the people doing the investigation belong to the religion that claims responsibility for the alleged miracle. For instance, consider something commonly cited as evidence for miracles: the shrine at Lourdes, to which pilgrims in their millions trek in the hope of miraculous healing. The Roman Catholic Church has a process -- to its credit, not a completely ridiculous one -- by which it certifies some healings there as miraculous. Although the process isn't completely ridiculous, it's far from obviously bulletproof; the main requirement is that a bunch of Roman Catholic doctors declare that the alleged cure is inexplicable according to current medical knowledge. As an example, the most recent case is of someone who had a tumour that went away after she bathed at Lourdes. (My understanding is that this is a thing that occasionally happens, miracle or no.) So, anyway, they appear to certify about one miracle per two million pilgrims, and I think pretty much all the pilgrims are there in hope of healing. One per two million! (If you think the alleged cures are so improbable that they couldn't happen naturally one time in two million, I have a bridge to sell you.)

3. In some situations (those in which a lot of these miraculous healings tend to occur) it really isn't difficult to get people to think more has happened than really has. Consider, for instance, the case of Peter Popoff. Lots of miraculous healings at his meetings -- but the whole thing was a fraud.

4. In general, unfortunately, people do lie. And make mistakes. And see what they hope or expect to see. And tales "grow in the telling", so that after a few steps of Chinese Whispers something sounds far more inexplicable and impressive than it ever really was.

You might try the following experiment: Talk to some of your Christian friends, and ask them for the most impressive examples they have personally experienced of miraculous interventions by God. If in fact there are no miracles, what you should expect is that (1) the things they cite won't, on the whole, be all that impressive; (2) the more careful and intelligent of them will have less impressive experiences; (3) the most impressive experiences will be the least verifiable.

Good answers have been given by commenters already, so I'll take a second to say: this reminds me strongly of C.S. Lewis' dissertation on miracles. Titled, can you believe it, Miracles.

I don't remember the entire discussion, but his opening expressed an idea of priors similar to what you've outlined here: those who review history looking for miracles will find evidence for them. Those who review history looking for evidence against miracles will find it. His idea was, simplified, "If a miracle occured, a person who does not believe in miracles would accept the evidence as being of an unexplained phenomena that more study will reveal. So, even if I were to raise the dead before your eyes, it would not constitute proof of a miracle to you unless you already accepted the idea that miracles can occur."

I... did not like this thesis. Now, I read this after my own falling out with Christianity, but I still had (and still have) a deep respect for Lewis. But this thesis reminded me very much of the argument made by creationists about how world views determine your interpretation of scientific fact. When laid over the field of paleontology, the idea starts to show its flaws. In geology, if you determine a rock is a million years old, you can't "interpret" that fact to mean it only looks a million but is actually a thousand years old. That's doublethink: the rock "is" a million years old (by looks) but actually "is" (by essence) a thousand. You're saying the evidence suggests two things: what the rock "looks" like (according to the evidence of your senses) and what the rock "is" (according to the evidence of your senses on the matter of religion).

I feel that Lewis' argument on miracles and priors relies on the same sort of difficulty. An event "is" (by looks) a miracle and either "is" (by essence) really or not really a miracle. That tells us nothing. The phenomena is or is not a miracle. Your priors don't matter. If Horus really did return from Hell, my disbelief can do nothing to change that fact. Horus' death and return, as real events, guarantee a world that physically resembles and acts like one where Horus' death and return did happen, not one where they did not happen. So the evidence can only mean one thing, not either depending on my priors.

So, all of this is to say, "miracles" are not different from other phenomena. They are not a special class. You judge a miracle (or report of a miracle) the exact same way you judge a rock. If I walked into my office tomorrow morning and found a large rock had appeared there overnight, I would use the same senses and tools to determine where it had come from as I would if it were the second incarnation of Christ. Neither event has a special set of circumstances regarding how I gather or consider its evidence. I still have to judge the second incarnation of Christ by sight, sound, smell, touch, etc and the evidence I gather is still subject to the same scrutiny and consideration as if it Christ were a rock.

Glad you decided to open up about this difficulty. As one raised Christian (you are obviously among fellows here), I can appreciate the power of Belief. This is not an easy issue to face, however you end up updating your beliefs. Good luck.

Yes, psychosis is a useful explanation for miracles only for people whose beliefs won't explain the whole category away. I doubt the op has such beliefs.

You're lucky! If it's only the argument from miracle keeping you from becoming a better thinker, then that's easily solved. It's really brittle.

First, "account of a miracle" is very, very different from "evidence of a miracle". The only datum here is someone saying "A miracle has happened to me!". From a Bayesian point of view, to assign a probability you must take into account all different models that could produce that statement, including:

• a real, honest miracle
• deception, inflicted on the witness
• deception, self-inflicted
• mistakenly reporting a perfectly natural phoenomenon for a miracle
• etc.

Having had some experience as an amateur mentalist, I'm always amazed at the degree at which people want to believe in the supernatural. Even when you tell them explicitly that it was all a trick, they still might argue you have power you don't know to have, and things like that.
Anyway, the simple fact that there are many ways to account for a witnessing of a miracle, automatically reduces the probability of such an event really happening.

Second, as others have pointed out, a miracle might point in the direction of the divine, but which divine? Islam, Induism, Buddhism... all have instances of miracles happening. Can you remember for example the case of the hindu milk drinking statues? It was famous all over the world. It even has its own Wikipedia page.

The third point I will explain with an analogy: suppose you travel back in time bringing with you a pair of walkie-talkie. Lo and behold, miracles can happen: suddenly people are capable of talking to one another even if they are very far apart!
Simply the fact that something very unusual has happened cannot confirm or deny any model of reality per se.

Bayesian statistics sum it all up pretty neatly: if you consider reporting a miracle as an outlier in an otherwise pretty standard distribution, you must decide the probability that the outlier is the indicator of some unusual model of the world by comparing it with all the other models. In math notation:

P( Christian god | miracle reported ) = P( miracle reported | Christian god ) * P( Christian god ) / P( miracle reported )

Assuming P( miracle reported | Christian god ) to be pretty much 1 and P( Christian god ) to be somewhat low but not fantastically so, the problem lies in the term P( miracle reported ), because it can be decomposed in this way:

P( miracle reported ) = P( miracle reported | Christian god ) * P( Christian god ) + P( miracle reported | no Christian god ) * P( no Christian god ).

Since we already estabilished that P( Christian god ) is low, then P( no Christian god ) must be high, but the real weight is in the term P( miracle reported | no Christian god ). As we already know, because of the tendency of the human mind to be deceived, this term is pretty close to unity, and so P( Christian god | miracle reported ) is pretty close to P( Christian god ).
This does not refute the existence of the Christian god, it simply says that, since witnessing can happen with equal probability whether a Christian god exists or not, then you cannot take that witnessing as evidence for a divinity.
But since you already hinge only on this for your belief, then you're pretty solved.

I have already read the Mormonism essay and mostly agreed with it.

However, I disagree that you would be using the same standard of evidence in this case. For example, all of the witnesses for Mormonism had readily understandable motives such as not breaking up the group or offending their leader. Something similar may be true about the boy and his parents, but it isn't true e.g. of the doctors who testified to amputating the boy's leg. They were from a different town, were not there when the supposed restoration happened, and had nothing to gain by agreeing with a made up story. Calanda could become famous by such a story, but the doctors would get nothing out of it.

That is only one out of a number of substantial differences.

First of all, congratulations! These kinds of questions are extremely challenging to even ask from within certain philosophical frameworks, and the fact that you're here at all means that you've accomplished something exceptional. Further, by using the question of miracles specifically, you've focused on empirical, testable claims with verifiable consequences. The epistemology that you're associating with atheism or agnosticism is fundamentally the ability to ask exactly these questions, the habit of doing so reflexively, and the willingness to follow those questions to real answers.

The basic Bayesian response to the question of miracles isn't just "are they lying, or is there a God?" Ask the question a different way: in a hypothetical universe in which Christianity is false, how many claims of miraculous events do we expect? In a hypothetical universe in which Christianity is true, how many true (and false!) claims of miraculous intervention do we expect? Do we expect a difference in the kind of miracles that are claimed to occur? For example, we experience people claiming that God cured infertility or cancer, but never people claiming that God cured their amputation. It's an interesting discrepancy, and which universe is that most consistent with? Etc. Don't think about it in terms of picking apart each individual claim. Just ask yourself about an interventionist God in terms of your honest expectations for such a God, and consider the world-as-it-is in comparison. Use the miraculous as a prediction that can succeed or fail, rather than simply as an explanation that is immune to correction.