On the Rationality of Believing Extraordinary Claims

It is surprisingly hard to define miracles with philosophical precision, but easy to cite significant examples associated with the Abrahamic religions. For example, many Christians are convinced that Jesus of Nazareth was able to come back from the dead, to walk on water, and to turn water into wine. Likewise, many Muslims believe that Muhammad literally flew to heaven on a winged horse, and that he split the moon. Is there any rational basis for such extraordinary claims?

In his Enquiry Concerning Human Understanding, David Hume noticed that religious belief in miracles is grounded primarily in the testimony of alleged witnesses. He also recognized a deep problem with beliefs formed in this way, which goes roughly as follows. Our experience with the physical world has taught us about various robust regularities of nature, of which we've never witnessed an exception. For example, we've all learned that the sun rises every day, and none of us has so far witnessed a day without a sunrise. Likewise, our experience with the social world has taught us various imperfect regularities about the behaviour of people we know, but also about human nature in general. For example, we know that people tend to be truth-seeking, trustworthy, and so on.

In many contexts, it is perfectly rational to assign a high degree of belief to the claims of our interlocutors. That being said, each of us knows that people don't always utter truths, and that sometimes our own beliefs that were confidently formed on the basis of testimony have turned out to be false. We’ve all had first-hand experiences of being confronted with faulty testimonial reports, but none of us has had the direct experience of witnessing an exception to any of the familiar uniformities of nature. Consequently, if someone informs us about an extraordinary event – which contradicts our firm everyday knowledge of the workings of nature – it is overwhelmingly likely that their testimony is mistaken, for one reason or another, on the balance of probability.

Hume's ‘Memetic’ Argument

Many empirically oriented scholars, including Pascal Boyer, Daniel Dennett, and Robin Dunbar, have examined religion through the lenses of cultural evolution. In this article, I will not deploy any sophisticated cultural-evolutionary framework, and I will informally avail myself of Richard Dawkins's rather popular notion of a meme. In the context of Dawkins’s work, it stands for the alleged unit of cultural information. In The Selfish Gene, Dawkins offers a handful of examples of his proposed cultural replicators:

We know that natural selection brings together well-cooperating genes. In a similar fashion, memes also inevitably get lumped together into groups as they evolve. A collective term for such gangs of cultural replicators is memeplexes, and their main purpose is to inhabit as many brains as possible. There is no constraint on memeplexes to track or preserve truth: they might, of course, but that is not crucial to their fitness quotient. As Susan Blackmore succinctly puts it in her ambitious book, The Meme Machine, ‘truth is not a necessary criterion for a successful meme. If a meme can spread, it will’ (p. 14).

Can miraculous religious stories be offered a memetic analysis or treatment? Are there any memetic defence-mechanisms in place to ensure their stability in the meme-pool? I think both questions should receive a positive answer and, indeed, religion is often given as a paradigmatic example of a memeplex whose cultural survival does not depend on truth. It is needless to say that Hume could not have explicitly worked in the memetic framework when he wrote the Enquiry. However, he provided one of the first attempts to explicate the memetic stability of stories with a miraculous flavour. His take appealed to some easy-to-confirm general observations about human nature that reveal ways in which the propagation of stories can ‘spread the intelligence’. How does Hume account for the fact that the meme-pool still contains many false supernatural stories? There are a few points that he makes on this issue. First, he takes such stories (henceforth ‘miracle-memes’) to evoke positive feelings in both the spreader and the consumer:

Second, Hume is explicit about his belief that miracle-memes can be spread without the communicator actually taking them as being true. Typically, that is because they are units in a bigger religious memeplex and the larger entity has memetic mechanisms in place which facilitate this effect. I will touch upon this again in the next section. Hume's take on the matter is:

Those quotes capture the essence of Hume's explanation of the cultural fitness of miracle-memes. To reiterate a previous point, just because one can account for the fitness of a meme, this does not reflect its truth-tracking properties. On this point, Hume also provides two reasons for not subscribing to miracle-memes. The first one is that the meme-pool contains too many miracle-memes and most of these stories have been debunked, so a rational inquirer should be doubtful about their truth on probabilistic grounds alone. The second reason concerns human nature's tendency to blindly spread interesting stories. This builds on the aforementioned quote whereby he touches on the surprising and seductive aspects of reporting (and consuming) certain news. To make his point, he deploys the example of marriages, whereby everyone in the community hears engagement news even when the couple barely made progress in that respect. In Hume's words:

Some philosophers think that the Humean parallel between marriages and miracles has weaknesses. In his recent book, Miracles, Yujin Nagasawa suggests that it would be a mistake to deploy psychological considerations of this sort with the aim of dismissing the contents of the reports in question. Importantly, Nagasawa agrees that it may be true that some innate human passion drives the force behind the rapid spread of engagement reports and he also grants that people are excited to consume and spread such information. However, he argues:

He then goes on to suggest that both miracles and marriages constitute what psychologists call a minimally counterintuitive concept, namely one that is simple enough to be remembered and transmitted. I think that Nagasawa's attempt to group miracles and marriages together this way rests on overlooking the difference between how marriage-stories and miracles-stories are reshaped and restructured in the course of repeated testimonial iterations. I maintain that these two kinds of stories do not undergo the same kind of memetic mutations. There is an unusual source of evidence that can illuminate how miracle-memes tend to evolve culturally, which I will now examine.

Insights from the History and Psychology of Magic

One intriguing art form which purports to invoke wonder into people – by making them witness ‘the impossible’ first-hand – is the art of magic. Conjurers mystified audiences from all around the world by performing feats which seem to violate nature's regularities. To see how word of mouth propagates when it comes to miraculous-sounding stories, it is worthwhile to analyse the reports of audiences which took part in magical acts. The history of magic is sprinkled with interesting events which reveal that Hume's intuitions were mostly on track. Even though his example of marriages was imperfect, it was not at all essential to his points, and the rest of this article is devoted to strengthening Hume's memetic argument by taking a detour through this unusual art.

There has been a recent increase in the number of neuropsychologists who believe that the study of sleight of hand can allow us to get a glimpse of the ‘magic show that goes on inside our heads’. Experimental researchers – such as Susana Martinez-Conde, Stephen Macknik, Jordi Camí, Luis Martínez, Gustav Kuhn, and others – all maintain that we can understand more about our own minds by studying the mental mechanisms that magicians attempt to hack during their performances.

At this stage, one may feel a worry that miracle-memes and magic-memes are not similar enough for this parallel to be productive. This is because of intuitions along the following lines: miracle-memes have a supernatural dimension of some sort, whereas magic-memes reside in brains of people who know that the art of magic and trickery are related. This reply makes the mistake of attributing modern, contemporary knowledge (i.e. that magic and trickery go hand in hand) to people of the past who did not have this knowledge. To make this point clearer by means of an example, the magic-memes which existed in the meme-pool 500 years ago were indistinguishable from miracle-memes. People were routinely witnessing conjuring acts that they believed to be real: their testimonies were constantly imbued with supernatural interpretations of what they have seen. Many deaths took place because of this supernatural interpretation of basic magic tricks. Reginald Scot wrote the earliest description of the work of street magicians, entitled Discoverie of Witchcraft. This book was essentially meant to be a protest against executing witches on grounds of consorting with the devil. Scot apologized to the community of conjurers for exposing the clever methods behind their tricks, but he felt the urge to show that those who were killed had nothing to do with the supernatural, despite the initial impressions and the popular beliefs. Witches were nothing but street magicians. Therefore, history clearly reveals that past audiences of magicians kept illusory beliefs in the supernatural alive, and they hurt conjurers whose performances seemed too real to be naturalistically explainable. Unfortunately, there are also contemporary instances of supernaturalistic interpretations of sleight-of-hand acts. Wayne Houchin, one of the foremost creators of magic tricks, was performing in 2012 in the Dominican Republic live on television when the host, Franklin Barazarte, actually set Houchin's head on fire after the performance, which resulted in hospitalization. This is a tragic situation which clearly showcases the psychological impact conjuring acts can have on audiences which are not operating on a default belief that trickery is involved.

Most educated people today do not attribute truth to magic-memes, but many do so in the case of miracle-memes. This is a puzzling state of affairs which raises the following dilemma. Many of the miracle-memes have a magic-meme analogue: magicians have performed demonstrations where they walk on water, change water into wine, and so on. Consider the ‘Jesus-Walking-On-Water’ meme: many people subscribe to the truth of the underlying story, despite it being grounded entirely in reports of people who lived thousands of years ago. On the other hand, the evidence we have that world-renowned magician Dynamo walked on the river Thames in 2011 exceeds it. On top of that, Dynamo clearly met Hume's criteria concerning the quality of the audience (number, education, and credibility) and, furthermore, the testimonial evidence is supplemented by recorded evidence – freely available on the internet in video and photographic format.

However, despite the amazement that these magical feats generate, even those who believe in religious miracles cannot help asking the following question when they are direct witnesses of an artful magical display performed by a conjurer: ‘How did he do that?’ This very question, which is common to audiences from all over the world, implies a search for a method. This is an act carried out automatically by the mind in response to noticing an event which goes against a physical regularity, whose place is perfectly secure in the web of beliefs. Uniformities of nature, such as the fact that people cannot walk on water, are here to stay, while ‘miraculous explanations’ have to go. All of this should make the rational believer wonder whether their acceptance of miraculous stories grounded in the testimonies of audiences from thousands of years ago is reconcilable with their naturalistic interpretational impulse that they have in the face of modern-day ‘miracles’ (for which they have additional evidence and/or first-hand experience).

Here we can see the power a memeplex can have. The reason for which truth is more likely to be attributed to miracle-memes as opposed to magic-memes is because magic-memes don't fit within a memeplex that is securely installed inside people's heads. People may have heard about Dynamo walking on water, but that magic-meme does not coherently fit into a memetic structure present in most brains belonging to people who make up a tight community. In her aforementioned work, Susan Blackmore went into the details of the memetic defences possessed by religious memeplexes (e.g. promises for the afterlife, rewards for fighting for your religion, etc.). Magic-memes do not have the luxury of being part of such a powerful memeplex, so they lack the protection offered by the larger entity. Interestingly, when the magic tricks in question are (unethically) disguised as effects that flirt with the supernatural (e.g. faith-healing acts), audiences are more likely to be deceived and believe in the reality of these tricks. Such magic-memes find it easier to infiltrate pre-existing memeplexes, unlike more regular tricks.

In the absence of powerful protecting memeplexes, the comparison between magic-memes and miracle-memes is particularly fruitful, because both of these type of memes should have the same fitness. If two twins were to witness an indistinguishable magical performance – one performed by a trickster and one performed by a person with actual supernatural powers – the two twins would have an identical subjective experience. Their testimonies, and the subsequent propagation of these testimonies, would resemble each other, despite the fact that one of the twins was just deceived by a conjurer.

A Case Study: The Indian Rope Trick

In the second section of this article, I said that Nagasawa underestimated the mutations that miracle-memes undergo in the course of repeated testimonial iterations. Indeed, the history of magic exhibits numerous cases which support my claim. I will now present one of the most well-documented ones, namely the story of The Indian Rope Trick, which illustrates various ways in which false legends can gain momentum and spread beyond control. The Indian Rope Trick received attention from many historians of magic, but also psychologists and neuroscientists. The memetic fitness of this story was so great that it reached many corners of the world. In our terminology, the Indian Rope Trick is a highly fit magic-meme whose fitness is independent of its truth-tracking properties. Understanding the particulars of this magic-meme's fitness can shed light on how stories about impossible events can circulate. Before I move on, I shall describe the trick. In their jointly published paper from 2001, ‘The Rise and Fall of the Indian Rope Trick’,Footnote ¹ Peter Lamont and Richard Wiseman provide a description of ‘probably the most famous secular miracle of all time’:

This description left plenty of magicians scratching their heads in search for an explanation, because the conditions of the trick (e.g. the trick was performed outside and in broad daylight) disqualify many of the standard methods which could be used to perform a magical effect with that description. What launched this story into the meme-pool? The answer is: a hoax-publication in The Daily Chicago Tribune newspaper from the 1890s by John Elbert Wilkie, which had publicity as its main aim. Wilkie certainly did not actually expect his story to reach so many minds. In December 1890, the newspaper had received a request from a man who wished to speak directly to the author of the story. Wilkie actually got back to him, saying: ‘I am led to believe that the little story attracted more attention than I dreamed it could, and that many accepted it as perfectly true.’ Even though he publicly acknowledged that the Indian Rope Trick was a mere fabrication, this barely affected the memetic fitness of the story. The news of a miraculous event reached all of Europe and the United States. The Humean explanation is quite obvious: such a magic-meme is great media to both consume and share. However, things get more interesting for our analysis.

Despite there never being any Indian Rope Trick and, therefore, no real witnesses, there are recorded cases of people who genuinely report having seen the trick performed live. In the article that we mentioned a moment ago, Lamont and Wiseman offer the example of S. T. Burchett, whose account was reported in the Journal of the Society for Psychical Research. Later investigations revealed that his memory of the event is not actually accurate, despite his honest belief in it. Burchett was not the only one: there were many people who had false memories of this miracle happening.

The previous remark is particularly interesting because it shows how some conditions enable some individuals to believe that they witnessed themselves the impossible act. It reveals how answers which explain the survival ability of miraculous stories that rely only on deception and gullibility can be overly simplistic. I modestly suggest that, in the religious context, it is not out of the question that some people of the past underwent similar experiences for similar reasons. The idea that people can have false memories may not be that surprising, given the emerging literature on the unreliability of eye-witness testimony. Also, the fact that audiences misreport what they have seen on stage is part of a handful of bits of knowledge that are very well known in the magic community. Psychologist Gustav Kuhn points out in his book, Experiencing the Impossible, that:

Given our remarks about the similarity of magic-memes and miracle-memes, looking at testimonials of magicians’ audiences may tell us something worthwhile about the relationship between our minds and miraculous stories. There are multiple factors which contribute to the trick-description's departure from what actually took place in the magical performance. I will only supply a summary of two of them.

The first factor is the exaggeration effect, which typically occurs when an audience member attempts to describe the impossible trick that they witnessed to someone who didn't witness the performance. Usually, the exaggeration effect takes place in this case because of a need to linguistically enrich what happened during the performance in order to (1) enhance the reactions of the people who did not witness it directly and (2) to secure the aspect of impossibility of the event in question through carefully chosen words. The first reason is usually applicable in the case of people who were astonished by the performance, while the second is usually applicable in the case of people who feel embarrassed that they cannot figure out the trick's method – hence, they exaggerate the trick's description in conversation with others in order to decrease their interlocutor's confidence that they could have figured out the trick if they were present at the performance.

The second factor is memory illusions, which can happen with or without the involvement of the magician. Thus, the exaggeration effect can take place even when the audience member doesn't try to enhance any details of the trick. In one of his popular documentaries on the art of magic, Wayne Houchin shares his experience of being routinely congratulated by people for performing a certain shocking spin-off of the classic Needle Swallowing trick. The interesting aspect is that he never performed the trick that the spectators describe – what he actually did was to perform two separate gory tricks at different points in the show, which spectators subsequently mentally merged into one single trick.

Sometimes, the faulty memory of the performance can be owed to the deliberate intent of the magician to make the audience members forget key parts of the trick. Neuroscientists and psychologists have started to show interest in the ways magicians can hack aspects of minds that pertain to memory and attention. Magicians artfully direct the attention of the audiences where it is needed through a process known as misdirection. Because during misdirection the attentional gaze of the spectators misses key elements of the trick, there is no surprise that important elements which contribute to the impossibility of the trick are absent from memory. However, there are other ways that magicians can ensure that key aspects of the trick are absent from memory. They don't need to hide those bits during the performance, they just need to make sure that they are forgotten afterwards. In the same book referenced above, Kuhn calls this memory misdirection and gives an example of himself being a ‘victim’ of magician Juan Tamariz's deployment of this technique, plus an explanation of how memory misdirection can work.

The reader might wonder whether our considerations have anything to say about chains of testimonies, not only isolated reports of a single audience member. Our interest in miracle-memes concerns what happens when an iteration of testimonies occurs, not just an initial report. Do memetic mutations occur even after the meme has passed through some generations of minds which were not first-hand witnesses? I suspect that the answer is yes, and I will bring back the Indian Rope Trick into discussion. In one of their other joint papers, Lamont and Wiseman carefully investigated how reports of this trick modified with time. In their own words, they went through ‘all of the English-language books, pamphlets, newspaper reports and magazine articles discussing the rope trick in order to collect eyewitness accounts’. It turns out that the degree of impressiveness of the trick's description is positively correlated with the passage of time: people mutated the story into something which does not resemble reality, and more mutations were added as time went by.