Dunne Dreams – Science or Pseudoscience?

J W Dunne is remembered more for his precognitive dreams than for anything else. Indeed, such dreams are often referred to as "Dunne dreams." He recounted them in An Experiment with Time, along with his experiments on himself and his friends, to try and fathom what was happening. He also developed a pretty wild theory about it all, involving multiple levels of time and consciousness.

Despite his determination to take a rock-solid scientific approach and the persuasive evidence he offered, his ideas have never been widely accepted and the suggestion of dream precognition is written off by the mainstream view as pseudoscience, along with the study of psychic phenomena in general.

The more I study the subject, the more I find that the quality of the science presented by the sceptics is, in all honesty, no better than that of the advocates. After almost a century of continuing research, the best predictor for the outcome of any study is whether or not the lead researcher is inclined to acceptance or scepticism. Or, to put it the other way round, the only thing that any study will confirm with any reliability is the belief system of its author.

Science has to be done properly or it is not worth doing at all. The determined acceptance of bad science and its results as valid is a good definition of pseudoscience. So who are the real pseudoscientists here?

For a start, how scientific was Dunne? He certainly wanted to be and thought he was. He firmly believed that every thought and experience in life was accompanied by its associated brain signal. He accepted Einstein and Minkowski's four-dimensional spacetime as fundamentally sound, right down to the Lorentz contraction which presented physical reality relative to the observer and thus gave Einstein's theory its name. He thoroughly appreciated the human mind's ability to deceive itself and wrote off the contemporary fashions for spiritualism and fortune-telling with a disdain borne of personal experience. Moreover, he also thoroughly appreciated the technical challenge of drawing objective scientific results from what was an almost wholly subjective phenomenon: how do you identify a precognitive experience? Overall, he was a good deal more scientific in his methodology than anybody else at the time, and more so than most who have followed him since, in either camp.

His succession of deeply disturbing dreams ... but hold on, was he genuine or a charlatan, making them up in order to fool his audience? Well, he invoked various family members and friends in his accounts and not one of them ever raised a peep of objection to his tale. Similar reports have streamed in by the thousand whenever anybody has enquired ever since, and some of those have been from highly respected individuals with more to lose than to gain by the admission: Rudyard Kipling, Igor Sikorsky, Geoffrey de Havilland, Michael Bentine and MM Kaye among them. There is a vast body of literature devoted to collecting and recording such claims. If Dunne was a charlatan then he was blending skilfully with a wide body of earnest anecdotal evidence and at the same time silencing a loose and independent body of collaborators. Nor would he have left behind an extensive but haphazard pile of notes detailing his unpublished experiences, ruminations, fellow dreamers and so on. No, the idea of deliberate deception, right down to creating a false private record for posterity, simply does not stack up.

Well then, perhaps he was mistaken, looking back and subconsciously twisting his memories to fit the facts? He was acutely aware of such possibilities and documents his struggle with them, along with the experimental refinements he foisted on both himself and others to test for, identify and eliminate such effects. He kept a dream diary and referred back to it regularly to check for predictions of more recent events. In some cases he would discuss a dream with others before it came true, and none of those others ever spoke up against his claims. It becomes wholly untenable to suggest that he was not having dreams whose apparent predictive content deeply disturbed him.

But was that content really predictive, or were Dunne and his friends just fooling themselves, seeing prediction where none existed? This is surely the key question. Dunne developed statistical methods for analysing his dream records, adapted the format of his diary entries to aid that analysis. These methods became really quite sophisticated, probably more so than anybody had attempted before. But inevitably subjective judgement crept in. How probable was some particular coincidence? Indeed, how much was the coincidence more imagined than real? Building up a collection of such dreams, at least some would be down to coincidence, but might there be significantly more than could reasonably be accounted for by that? What was the probability of some of them being genuine as against wholly down to chance?

Dunne did his best to be objective, but the acid test must come with attempts to duplicate his results. The respected Society for Psychical Research decided to see for itself. But their research officer Theodore Besterman turned down Dunne's rigorous but complex procedures in favour of something simpler and easier to do in practice. It was therefore also a good deal less scientific and Dunne argued it was so badly designed it was no longer a test of his theory. When the results came through, Besterman and Dunne produced widely different assessments as to what was or was not a coincidence and what were the probabilities involved. Dunne went on to extract further information from the participants and used this to fill in some of the gaps that Besterman had opened up. The Society felt forced into the quite extraordinary position of publishing an editorial preface disclaiming their own research officer's report and he left not long afterwards. Yet every good sceptic ever since has seized on that report as proof that Dunne was wrong, conveniently ignoring the society's official disclaimer printed immediately and quite unmissably above. Dear sceptics, I call that blind eye of yours bad science. Your rejection of Dunne is more pseudoscientific than his claims ever were.

Matters were not helped by the psi hypothesis developed by JB and LE Rhine at that time. Psi is defined in a negative way, as "para"-normal, outside of accepted science. It is impossible to prove such a negative beyond all doubt, that any given claim can categorically not be explained by the known laws of science. Worse, although early experiments suggested positive results, the more rigorous and automated they became the more the results evaporated. From Dunne's perspective, all this automation was making fundamental mistakes. Dunne's theory, deduced from his own experiments, was that his dreams were previsions not of objective events but of his own experiences of those events. They were otherwise perfectly normal, they were just displaced in time by anything from a few minutes to many years. Had they occurred after the event in question, they would have been entirely normal. Returning to the Rhines' experiments, we can now appreciate their mistakes. Firstly, a long succession of similar tests yields a long succession of similar results. When the subject tries to predict the next result, there is in fact no knowing which experience of which result in the series might be being foreseen; the assumption that it is the next one is wholly unjustified. Then again, if the subject never knows the outcome of a given guess, there will be no way that they can have prevision of that knowledge. So automated test runs with only a statistical summary at the end leave the subject without anything to foresee in the first place. That wipes out the vast majority of psi-hypothesis experiments ever carried out. If Dunne was right then we can confidently predict that the psi programme would fail. The Dunne dream hypothesis is therefore actually (if weakly) supported by the outcome of the psi programme. But then of course, so too is the sceptical hypothesis. Ultimately, the psi hypothesis was not pseudoscience because it was falsifiable, it just happened to be proved wrong. But Dunne was not proved wrong, so once again any sceptics who write off Dunne dreams, just because psi failed, are the bad scientists, the pseudoscientists here.

One other argument is worth debunking here. Sceptics often claim that no genuine precognitive event has ever been verified. All can be explained, they claim, by conventional mechanisms such as fraud or mental slips. Well, technically such explanations can always be thought up to fit the known facts. Dream reports are particularly vulnerable to such criticisms because no individual event can be replicated in the laboratory; every dream and every apparent precognitive event is unique. Science has to find a more subtle approach, based on such less certain techniques as statistics and Occam's razor. As an unrelated example, the motions of the planets and moons across the sky can all be explained, to any desired degree of accuracy, by cycles and epicycles of gears using prime numbers of teeth. For the ancient Greeks, this was a stunning rationalist discovery and they used it to build practical supercomputers such as the Antikythera mechanism. Yet ultimately this whole Platonic Universe becomes so recursively complicated and convoluted that we found it more sensible to accept the Copernican model and Newton's theory of gravity (and eventually Einstein's). Back with precognition, just how tenuous and convoluted do the "Oh, well, he might have written the wrong number down and then misremembered, blah, blah..." inventions have to become – in the face of stringent precautions against that very thing or vast numbers of respectable claims – before it becomes more sensible to cut the knot and accept that a scientific basis for it might actually exist? The sceptic replies hotly that no evidence is no evidence and that is enough, without either realising or caring that their argument is circular: the judgement that no evidence exists is based on the blanket denial of the entire body of evidence presented, with that denial being itself based on the prejudicial belief that no such evidence can possibly exist. Totally circular, totally pseudoscientific. Even politicians such as Donald Rumsfeld famously understand that, in such a charged and uncertain atmosphere, absence of evidence is not evidence of absence.

Various other attempts have been made to test for dream precognition, some focused more closely on Dunne's methods than others. But few if any have applied as much methodological and statistical rigour as Dunne did, let alone bring in modern advances in methodology to tighten up on the looser areas such as rigorous assessments of probabilities. As I said at the beginning, the best predictor for the outcome of any study remains whether or not the lead researcher is inclined to acceptance or scepticism. That is not the mark of science, it is a red flag that the pseudoscientists have taken over the pitch. Dunne's theory is in principle falsifiable, that makes it good science. The real work that needs to be done is to clear the pseudoscientists out of the way, sceptic and advocate alike, so that Dunne's rational scientific heritage can be recovered in its full glory and taken forward to the next level. And that requires the goodwill of the scientifically genuinely open-minded, those who can put aside their predilections, be they to scepticism or acceptance, and collaborate on a clear and level playing-field to find out who is actually right. You see, I'd love to know, I really would, but I will not accept pseudoscientific rhetoric from anybody, not from either team. I want the referee to tell us the score.

As a postscript, all this says little about Dunne's theory of Serialism which he developed to explain the nature of time. Serialism could be barking mad and dream precognition still occur on some other basis. I am far from alone in making such a suggestion.

Updated 7 Jan 2022