A few weeks ago I published at Rationally Speaking a guest post by my former (undergraduate) student Maaneli Derakhshani, who made a case for the scientific status of parapsychology. Some of my readers criticized the choice as an instance of allowing pseudoscience to be represented in what I hope is a reputable science and philosophy blog. That sentiment is, I think, misguided. If we really pride ourselves on our critical thinking we ought to be able to take other people’s best arguments on board and show if and why they are mistaken. And Maaneli did make a very good argument in defense of parapsychology.
I also think Maaneli, in response to some of the many comments posted, is correct in saying that it took courage for him to “come out” in this manner. As I understand it, he is hoping for a scientific career in theoretical physics, and he rightly argues that writing favorably on behalf of parapsychology is not going to help his chances. I know that I would not hire in my department someone with leanings toward what I consider to be a pseudoscience.
But is parapsychology a pseudoscience, as Maaneli chides me for having opined both in podcasts and in my Nonsense on Stilts book? He thinks not, based on his analysis of a small but persistent literature concerning experiments performed under so-called "psi-conducive" conditions, like the Maimonides dream telepathy and the Ganzfeld (“total field”) experiments.
I am sure I will disappoint both Maaneli and some of my readers who were actually supportive of his analysis, but I will not engage his claims one by one. This is, I assure you, not a cop out, but a reasonable decision based on three considerations. First, other critics of the paranormal have done a much better and more in-depth job at criticizing the experiments produced by Daryl Bem and others (for the most recent example, see this devastating critique of Bem's porn-facilitated clairvoyance experiments, published in Skeptical Inquirer. Additional critical sources are listed at the end of this post). Second, neither Maaneli nor I have access to the raw data or have been in a position to double check in person (or at least try to duplicate) the experimental protocols under discussion. Without that, we are both reduced to trusting the analyses (or in my case, the criticisms) done by others. Third, the question asked by Maaneli is whether and why current parapsychology qualifies as science or should be relegated to pseudoscience, and this issue is broader and more interesting than endless skirmishes about p-values and meta-analyses. I will therefore focus this post solely on Maaneli’s fundamental question.
One argument made by supporters of parapsychology is that there is by now sufficient evidence to accord the discipline scientific status (and, presumably, academic status and funding to its practitioners) because the quality of the marshaled evidence is at least as good as run of the mill evidence published in mainstream psychology journals. Besides the fact that there actually is much reasonable doubt that the latter assertion corresponds to the truth, the argument fails for two reasons. First, one could respond that at best this shows that a lot of psychology is sloppy science. As a formerly practicing scientist (in evolutionary biology) I can assure you that quite a bit of below par science is done (and, unfortunately, published) all the time. An embarrassing number of papers in mainstream science is based on bad experimental procedures, presents woefully inadequate and biased samples, and reports flawed statistical analyses. The reason this isn’t a bigger deal (although it probably got tenure for a number of people who should have dropped out of science in graduate school) is that scientific peer review is an endless process that eventually sifts the few nuggets of gold and simply ignores the sea of useless crap.
Second, as every skeptic knows very well, Hume's dictum reigns (though often in Carl Sagan's paraphrase): extraordinary claims require extraordinary evidence. This is true of normal science as well. Nobody bothers to replicate or even critically re-analyze boring results that simply confirm what we already know, only under slightly different circumstances. But try to claim cold fusion, or faster than light particles, and suddenly the standards of proof become much much higher. And so they should, as unfair as the individual scientist may feel about this epistemic heuristic.
Now, parapsychological claims of telepathy, clairvoyance and the like are just about as extraordinary as they come. One can reasonably argue that if confirmed, these claims would overturn physics and biology as we understand them, possibly violating several fundamental laws (vague nods to “quantum entanglement” effects notwithstanding). That being the case, Bem and colleagues simply have to do a hell of a lot better than they have done so far, and my bet is that they simply won’t be able to.
By the way, this isn't a question of effect size, for which several readers have erroneously hammered Maaneli. There are plenty of examples in science, from ecology to quantum physics, where the effect size is very small indeed. But the results are statistically clear, methodologically unimpeachable, and repeatable ad nauseam by a large community of scientists. None of the above is the case for anything that parapsychology has produced so far.
What about the lack of a sound theory? Again, Maaneli is correct in citing examples from the history of science when we didn’t have a theoretical explanation for certain observations, and yet the scientific community has eventually accepted the results and incorporated them into mainstream science. But a closer look at some of these examples is very instructive. Take Wegener’s idea of continental drift, which turned out to be correct, and which was based on initially already strong evidence (much better evidence, I submit, than anything produced in parapsychology). Still, the idea was not accepted immediately, and it took among other things the development of a sound theory to explain the phenomenon before geologists came on board en masse. And so it should be, since science isn’t just a collection of facts, odd or not, it is an attempt to understand those facts and how they fit into everything else we know about how the universe works.
Parapsychology has had more than a century to produce compelling facts and reasonable theories. It has fallen very short on the first count, and embarrassingly so on the second one. Nobody seems to have any idea of what “psi” is and why it works in the way in which it allegedly works. And nobody seems to have any clue at all concerning how “psi” might fit with everything else that psychology, physiology, neurobiology, evolutionary biology, chemistry and physics tell us about human cognitive abilities (again, vague references to quantum mechanics won’t do, despite how easily Deepak Chopra can make them). This is a really tall mountain for parapsychologists to climb, and they seem stuck on the first or second step.
Which brings me to why parapsychology is best thought of as pseudoscience. Karl Popper famously thought that the so-called demarcation problem, finding something that distinguishes science from pseudoscience, had been solved by his own criterion of falsifiability. Not so in modern philosophy of science. Maarten Boudry and I, as I have mentioned before, are finishing up the editing of a new book on the demarcation problem, to be published by Chicago Press. One of the things we learned from the many contributors to the volume is that nobody any longer thinks of science (or pseudoscience) as a simple concept that is amenable to a sharp definition based on a small set of necessary and jointly sufficient conditions. But we have also learned that many philosophers think that a hallmark of pseudoscience is the persistence of its practitioners to make grand and revolutionary claims in spite of the equally persisting dearth of compelling evidence (and theory) to back them up. This scenario, I think, fits the situation of parapsychology very well.
However, Maarten and I have also seen many of our colleagues argue — again, correctly, we think — that a pseudoscientific status is historically contingent. That is, something may be confined to pseudoscience for a long time and then emerge triumphantly, or vice versa, may be considered good science for a while, only to be eventually relegated to pseudoscience. Phrenology is an example of the latter, evolutionary biology one of the former (I know, surprising, but see the essay by Michael Ruse that I discussed at RS in the context of a recent book on biology and ideology).
So the verdict is pretty much never final, and good skeptics really ought to keep an open mind. But the more time that passes without significant and widely acknowledged progress, the more one’s skepticism is reasonable and warranted. So here is my suggestion to Maaneli: either shelve this whole thing and concentrate on your budding career as a theoretical physicist, or get your hands dirty and work to produce the kind of evidence (and theory) that really has the potential to shock the scientific world into paying attention. If you don’t mind the advise, however, my bet is that you’ll be far better off taking the first route.
Additional critical sources:
Blackmore, S. J. (1980) The extent of selective reporting of ESP ganzfeld studies. European Journal of Parapsychology 3:213–220.
Blackmore, Susan (1987) A Report of a Visit to Carl Sargent’s Laboratory. Journal of the Society for Psychical Research. 54:186-198.
Frazier, Kendrick (ed.) (1986) Science Confronts the Paranormal. Prometheus Books.
Hansen, G.P., Utts, J., and Markwick, B. (1992) Critique of the PEAR remote-viewing experiments. Journal of Parapsychology. 56:97-113.
Hansel, C.E.M. (1989) The Search for Psychic Power: ESP and Parapsychology Revisited. Prometheus Books.
Marks, David (2000) The Psychology of the Psychic. Prometheus Books.
Milton, J. and R.Wiseman (1999) Does psi exist? Lack of replication of an anomalous process of information transfer. Psychological Bulletin 125:387-391.