Why Science Can't Accept Miracles (Even if They Really Exist)

Steven Dutch, Natural and Applied Sciences, University of Wisconsin - Green Bay
First-time Visitors: Please visit Site Map and Disclaimer. Use "Back" to return here.

A Thought Experiment

Consider a banker, a devout Baptist and complete believer in miracles. During an audit he finds $100,000 missing. All the employees' books balance. Is he going to accept that the $100,000 just miraculously disappeared? Is he going to expect the police and banking regulators to accept it? Not very likely. Even if all attempts to find fraud fail, he's going to assume that somehow, somebody pulled off a theft.

Now let's assume there is a witness. A long-time, highly trusted employee who is a member of the same church as the banker, and whose character is above reproach. He tells the banker he was in the vault and saw the money simply vanish before his eyes. The banker is certainly not going to expect the police to believe this story or blame them because they suspect theft. If all attempts to crack the employee's story or find the money fail, we have a lot of options to consider. Maybe the employee blacked out or hallucinated momentarily, or had a small seizure. Maybe someone hypnotized, drugged, or distracted the employee momentarily and grabbed the money.

Suppose the vault has a video camera that shows the money sitting in plain view one frame and gone the next. Our hapless employee is in the clear. Or is he? Could someone have interrupted the video feed for a second or two and simultaneously have paused the recorder? Could someone have doctored the security tape? Could someone have fed a false signal to the camera system? Or, a la the old Mission Impossible TV series, used trick photography to fool the security system?

I can't imagine anyone in banking, no matter how devoutly religious, not exploring every one of these avenues before concluding a miracle had happened. Even after accepting a miracle as the only logical explanation, I think this banker would always be prepared for the possibility of a natural explanation. The methodology here is pretty close to that of David Hume 250 years ago, who held that no evidence would be sufficiently ironclad to demonstrate a miracle. The banker wouldn't go that far, but he'd explore every other avenue first.

So why do so many people have a problem when science rejects miracles? Why would people expect the police to dismiss claims that money miraculously vanished from a bank and angrily label scientists "skeptics" for drawing the same conclusion about a tumor gone from a cancer patient? Partly it's a prejudice that scientific theories, unlike $100,000 missing from a bank and possible prison terms for the bank employees, are really not of any practical importance, so what's the harm? Actually, scientific theories are a lot more important than $100,000 missing from a bank vault - in literal money terms, let alone the whole issue of truth. At $100 a foot, a 20,000 foot oil well will run $2,000,000. That's one well. It makes a difference whether the fossils that turn up in the cuttings from the well were deposited according to the conventional view of geologic time or as the result of a miraculous flood. It makes a big difference in money and lives whether we conclude someone's recovery in a $100 million clinical trial was due to the drug, the placebo effect, or a miracle.

Science rejects miracles for exactly the same reasons that accountants do when conducting audits, the police do when conducting forensics, and mechanics do when trouble-shooting cars.

The idea that we always seek natural explanations for phenomena is called methodological naturalism. It must be sharply distinguished from philosophical naturalism, which is the a priori assumption that only natural phenomena exist. It is perfectly possible to be a religious believer and still practice methodological naturalism.

From the discussion above we can draw two important conclusions about accepting miracles as explanations.

  1. All theologies that accept miracles admit they are exceptional events. That's what "miracle" means. So if there's a possible natural explanation of an phenomenon, we go with the natural explanation.
  2. If you stand to gain from explaining something away as a miracle, you don't get to play.
    • If you're from Enron, you don't get to claim your documents disappeared miraculously. It only happened if the FBI and the SEC said it did.
    • If you're a defendant, you don't get to claim your fingerprints miraculously appeared at a crime scene. Only the DA is allowed to say that.
    • If you're a bookkeeper, you don't get to say money miraculously disappeared from your company. If the auditors conclude that's what happened, all right, but not you.
    • If your religion needs to postulate a miracle to keep some doctrine from going south, guess what? You don't get to do that. Only someone with nothing to gain from claiming a miracle can say that.

Why Some Scientific Statements About Miracles are Wrong

Instead of a paltry sum missing from a bank, let's think big. Imagine that tomorrow, precisely at noon Central Standard Time, Chicago simply disappears. Everything inside a sharply-defined circle 20 kilometers in radius simply vanishes. Planes inbound to O'Hare airport see the airport simply vanish. Pavement is truncated as if by a razor. Inside the circle, there are no basements or foundations. Countless people witness the event, and it is caught on video. In many videos, the frame where the event happens shows the change in mid-frame. As nearly as anyone can measure, it is instantaneous.

Clearly, we would have to conclude that something extraordinary had happened, something outside the known laws of science. The abrupt disappearance of a city as important as Chicago would have global impact. The effects would be so intertwined that it would be impossible to counterfeit the totality of the evidence. Assuming the evidence were preserved in sufficient detail, even the most skeptical observer a thousand years from now would have to conclude that something singular had happened, even if the nature of the event was unknown.

In An Enquiry Concerning Human Understanding, David Hume formulated what has come to be called "the principle of minimum astonishment:" He basically argued that no amount of testimony could suffice as proof of a miracle, because the possibility of fakery or fraud would be so much more likely. Hume used the example of a claim, supported by many reliable witnesses, of the King of England dying and being restored to life. If the alleged interval were short, it would be more likely that the king had really not been dead but merely in a deep coma. If the interval were longer, say a month, he would conclude that fraud was at work. It would be far more likely that the king and his courtiers would fake the death and hide the king for a month than that he could really come to life.

A more commonplace example is claims of a perfect bridge hand; a player getting cards of only one suit. Lots of people have claimed to have seen or played such a hand, but there haven't been enough bridge hands dealt in history to make even one case remotely possible. It's far more sensible to conclude the witnesses are mistaken or the cards weren't shuffled well.

Hume's position is sometimes summarized by saying that to prove a miracle, it would have to be an even greater miracle for the evidence to be faulty. Since we can never achieve that level of certainty, goes the argument, therefore there can never be evidence capable of proving a miracle.

Or can there? As my thought experiment shows, Hume simply didn't think on a large enough scale. You could hide the King of England and fake a funeral. What if London itself simply vanished? A more or less mundane event that is claimed to be a miracle could be faked. But what if the event itself is so far beyond the scale of normal experience that no imaginable explanation of fakery would work, perhaps even no imaginable method could produce the effect?

Right off the bat, then, we see that one often-made statement about miracles is flatly wrong. It is possible, indeed easy, to imagine a singular event for which fraud is out of the question. But what about less extraordinary events? Although fraud or error is a possible explanation, being a possible explanation doesn't prove it is the explanation.

Also, Hume committed a fundamental fallacy of confusing the evidence for a phenomenon with the phenomenon itself. Suppose in 2100 someone examines the claim that two successive golfers in the 2004 Masters hit holes in one on the same hole. The odds of a hole in one in the Masters are about 4,000 to one. Figure 100 golfers per tournament and 4 rounds each, and an average of ten years between holes in one (there were 7 holes in one on that hole in 68 years.) The odds of two successive holes in one at that tournament and hole is one in 16,000,000, or one successive pair every 40,000 years.

Our future analyst watches the videotapes, but video can be faked. He reads eyewitness accounts, but eyewitnesses are notorious for describing famous events they never actually witnessed. There are the newspaper stories, but once a news hoax gets rolling, it has a life of its own. The Masters official statistics are maybe the most compelling evidence, but given a choice between the records being doctored and believing something happened that shouldn't happen before the next Ice Age, he concludes the story is a myth.

The reasoning is perfectly sound. It only has one tiny flaw. The event actually happened

Hume is probably right in saying that no alleged miracle could ever be documented well enough to rule out fakery for all time. However, as the example above shows, charges of fakery can be used to rule out real phenomena as well. Although fraud or error is a possible explanation, being a possible explanation doesn't prove it is the explanation.

A Historical Example

Writing off all alleged miracles as fraud is not only logically dubious (and lazy), but it has proven to be historically wrong as well. The classic example is the case of meteorites, which were once dismissed by almost all scientists as fakes or folk tales. When a meteorite fell in New England around 1800 and was described in the scientific literature, the normally scientifically astute Thomas Jefferson said (classic Hume reasoning here) that he could more easily believe that two Yankee professors could lie than that stones could fall from heaven.

Not long afterward, a large meteor broke up in the atmosphere and showered a village in France with fragments. The evidence was on a scale that eliminated fakery once and for all as an explanation. And meteorites went from being folk tales and frauds to real phenomena.

Some Conclusions

Is The Chicago Thought Experiment Definitely a Miracle?

No. One scientifically conceivable explanation that comes to mind is that there might exist physical processes that allow sudden translocations across space and time. Maybe pre-event Chicago reappeared in the Precambrian, or in outer space, or will reappear a billion years from now. There are theories, taken quite seriously, that space and time have extra dimensions, most of which are invisible to us because they close up on extremely short scales. Maybe space and time can close up on larger scales very rarely, and we just happened to witness one such event. Maybe it happens all the time, but if one patch of empty space-time gets translocated to a different point in empty space-time, how would we know? Maybe it is only noticeable when the event impinges on the earth.

Maybe we conclude, from the fact that the circle is so incredibly perfect and sharp edged, that the event was the work of some incredibly powerful intelligent entity. A miracle? Or is the entity natural but merely extremely powerful? One need only recall the cargo cults of the Pacific, where tribespeople thought that American World War II technology was supernatural.

Maybe this event is a real miracle, triggered by some conscious entity totally outside the laws of nature. But even if it is, even if a scientist personally believes it to be one, he will still be open to the possibility that he might be wrong, and that an explanation within the laws of nature might just be discovered.

Another Historical Example

Isaac Asimov once pinpointed the year 1752 as an overlooked turning point in history. Since earliest times, people had regarded lightning as supernatural. Benjamin Franklin showed that lightning was electricity and furthermore devised a way to control it. For the first time, a phenomenon went from supernatural not just to natural, but to something controllable by humans.

The device Franklin invented, of course, was the lightning rod. Contrary to widespread popular misconception, a lightning rod doesn't channel lightning harmlessly down a wire into the ground. If a lightning rod gets hit, it has failed, and there's a good chance the current down the ground cable will heat it red hot if not melt it, causing a fire. Also, a lightning strike anywhere within a hundred meters of a house will very likely fry electronic devices. I know from experience.

No, a lightning rod prevents lightning strikes. Franklin found that pointed wires could bleed off electrical charges into the atmosphere, a phenomenon called corona discharge. Since like charges repel, a wire from a charged object will collect charges because the charges will migrate down the wire away from the charge excess. If the wire has a sharp point, the charges at the end get crowded together so tightly that the repulsive force can actually push electrons off the wire (or attract them if the object is positively charged). The point of a lightning rod (pun intended) is to bleed off electrical charge before it gets strong enough to attract lightning.

There were, of course, people who denounced this interference with what had once been divine. But as Asimov noted, people soon noticed that the local church kept getting hit if it didn't have lightning rods, and the local brothel, if it had lightning rods, didn't. For Americans, utility trumps ideology.

Joseph Smith: An American Miracle?

One of my favorite test cases for beliefs in the supernatural is the tale of Joseph Smith, founder of Mormonism. Smith claimed to have been given gold plates by an angel, which he translated to become the Book of Mormon. The interesting thing is that he showed the plates to witnesses, who swore in notarized statements that they had seen them. Here we have an alleged miracle, not 2,000 years ago in Biblical times, but right here in the United States, recorded in American legal documents. The interesting thing here is that, apart from Mormons, most devout believers in miracles reject this account. And they do so for precisely the same reasons that most skeptics reject miracles in general. 

So it's not skepticism per se that annoys believers in miracles. It's skepticism directed against their miracle claims.

And here we get down to the real reason most scientists reject miracles. The vast majority of alleged miracle accounts are untrustworthy. If you're inclined to take offense at that remark, go look at any magazine put out by a religious denomination that accepts miracles, say Pentecostal Evangel. Almost invariably, the magazine will require that an alleged account of a miracle be certified by a minister. Out of thousands of alleged cures at Lourdes, the Catholic Church accepts only a few dozen as meeting its criteria for miracles. Why this skepticism on the part of religious bodies that believe in miracles? Because they themselves have found out the hard way that the vast majority of alleged miracle accounts are untrustworthy, even those claimed by their own adherents.

Certainly these groups don't dismiss all dubious claims of miracles as frauds, nor should anyone else. A large fraction are probably ambiguous. If someone in a storm cellar prays to be spared from a tornado, his survival may have been due to divine intervention, but may also have been due to the fact that tornadoes have narrow paths, and a specific house is a small target. Whether the tornado followed the path it did because of supernatural intervention is not something that can be tested by any known means. There are certain classes of miracles that never seem to happen. People have been alleged to be revived from the dead, but no decapitation victim ever has. Nor are there any reliable accounts of severed limbs regenerating. The fact that some types of miraculous cures are cited fairly frequently while others never seem to happen suggests strongly that other explanations are at work.

(C. S. Lewis, in Miracles, makes a rather similar point. He observes that Biblical miracles are generally amplifications or reversals of natural phenomena and that miracles of the sort found in classical mythology, such as people turning into animals or inanimate objects, are absent. In Lewis' theology, miracles are rational events that are perturbations of normal causality, not wholesale exceptions to it.)

Apart from alleged miracles that could also reasonably be interpreted as natural events, there are less innocent cases: attention seekers, people who want to advance their own splinter doctrines or attract a following, and finally outright fakes. Some of the hoaxes are pious frauds, some are cynical attempts to fleece the gullible, and yet others are designed to discredit belief in miracles.

What Can Miracles Tell Us?

Suppose, through some as-yet unknown line of reasoning, we come to accept that the disappearance of Chicago was a genuine miracle, performed by some extra-physical entity by means completely inexplicable in terms of the laws of nature. What have we proven?

Surprisingly little. So far we have proven only that there is a supernatural realm that occasionally impinges on our own. We still know nothing about the entity that caused it. Was it God? Satan? Or something totally outside of any existing belief system? Does this entity communicate with humans, or make demands of them? The Joseph Smith example shows clearly that an alleged miracle has to be interpreted within some religious framework to have meaning. Without that framework, a miracle remains an isolated anomaly.

In order for science to interpret something as a miracle, it would have to have some unambiguous criterion for distinguishing miracles from natural events, even those produced by yet unknown physical phenomena. This criterion would have to be objective and independent of any religious preconceptions. One way to achieve this goal would be to have a lot of data on miracles, but if miracles happened that often and showed that much regularity, would they really be miracles? Or would they be phenomena that obeyed laws of their own, and thus in some sense "natural?"

The "No Overlap" Theory

Some writers attempt to defuse the tension between science and religion by claiming that they apply to non-overlapping domains. As conciliatory as that approach sounds, it can at best offer a truce. Either there is, or there is not, some entity in the universe that can operate beyond the laws of nature. If there is, then there can be occasional interventions not explainable in terms of physical laws, possibly even in contradiction to them. If there is not, then there are no omnipotent entities in the universe. In addition, all religions view their dogmas as factually correct, not merely morally binding. Inevitably, religions will make claims that are susceptible to objective testing.

Defining mutually exclusive spheres of subject matter seems unworkable. Instead it seems far more sensible to define domains of methodology.

Science deals with phenomena that can be explored using the error-correcting methods of science. Replication and refutation are probably the most broadly applicable of these. Science deals with phenomena that can be repeated or otherwise verified, and which are potentially capable of being refuted.
Religion has no right to expect science to accept a religious doctrine unless it can be subjected to scientific testing, and unless religion is willing to accept that the doctrine can be refuted. Religion is perfectly free to believe that an event is a miracle, but it has no right to impose its assumptions on science.
Religion deals with phenomena that may be autonomous, beyond human control, intelligently directed, and volitional, that is, at the discretion of some conscious entity. Phenomena may or may not be manifested for reasons not connected to any physical constraints. Information about the phenomena and the entity producing it might be communicated in unique and non-replicable events.
Science has no authority to decree categorically that those phenomena accessible to scientific testing are the only ones that exist. Science cannot legitimately deal with religious claims unless it is prepared to admit the potential existence of autonomous and non-replicable phenomena.
So Where Does This Leave Paranormal Research? No place, I'm afraid. Paranormalists want to claim the paranormal is an objective part of physical reality but don't want to be held accountable to the normal rules of scientific evidence. In particular, they don't want to be held to replicability.

Why Science Can't Accept Miracles (Even If They Happen)

This page has been translated into Czech by Sandra Knudsen and is on her Web site at Czech version

Return to Pseudoscience Index
Return to Professor Dutch's Home Page

Created 8 December, 2001, Last Update 28 July, 2017

Not an official UW Green Bay site