Why I don't believe in precognition
You may disagree now because of a later post.
Posted October 24, 2010
In assessing any particular piece of evidence we must consider what else the paranormal explanation of this evidence would lead us to expect ~ Humphrey, 1996, pp. 80-81, italics in the original
In a paper published by the Journal of Personality and Social Psychology (JPSP, 2011), Daryl Bem notes that 34% of psychologists in one studied sample believe that psi is impossible. I would be among those 34% if I were sampled. Yet, I remain intrigued by attempts to prove the existence of psi. Bem defines psi as "anomalous processes of information or energy transfer that are currently unexplained in terms of known physical or biological mechanisms." In his paper, he reports the results of 9 experiments focusing on precognition and premonition, which he defines as "the anomalous retroactive influence of some future event on an individual's current response."
Bem says he has enough evidence to suggest that psi might be true. His experiments are noteworthy because they do not use one of those arcane experimental set-ups peculiar to parapsychological labs (e.g., the ones used in studies of the "Ganzfeld" phenomenon). Instead, Bem selects experimental tasks that are de rigeur in mainstream research. One favorite is affective priming (see Bem's experiments 3 and 4).
Bem replicates regular or proactive priming. To do this, he first presents positive (beautiful) or negative (ugly) prime words briefly on a screen. Then, after a short delay, he presents a pleasant (puppy) or unpleasant (snake) picture. Participants then decide, as fast as they can, whether the picture is pleasant or unpleasant. There is evidence of priming when they respond faster on trials on which the valence of the prime is congruent with the valence of the picture than on trials on which the valences are incongruent. This is what Bem finds, as many others have before him.
His mind-bending innovation is a change of sequence. To test retroactive priming, Bem presents the picture, collects the response, and then presents the prime. He destroys any conceivable forward-working causal effect by letting state-of-the-art randomization devices on each trial select the valence of the prime. In other words, he attempts to show retroactive causation by ruling out another psi effect, namely, psychokinesis. Participants cannot make the prime a positive (negative) one by responding particularly fast to a pleasant (unpleasant) picture.
Bem finds standardized effect sizes of .4 and .25 respectively for proactive and retroactive causation. From a traditional point of view, the latter should be very close to zero. It isn't, and here the gauntlet is thrown. Yet, Bem cannot demand "Explain this!" because he can't explain it himself (as he concedes in the article and reiterated on the Colbert Report). The whole point of psi experiments is to demonstrate the existence of something that is inexplicable by ordinary lights. If it were explicable, it would not longer be anomalous; it would not be psi.
That's the paradox. Psi research seeks to establish the existence of weird phenomena while at the same time refusing to offer a positive theory of why and how these phenomena come into being. This must be so because once you have a positive (i.e., intelligible) theory about the process underlying the phenomenon, the mystique is gone.
I remain unconvinced. I am not only bothered by the lack of a positive theory, but also by the contradictions between psi and basic scientific assumptions. The conventional view assumes that events can causally affect other events that have not happened yet. When the arrow of time is depicted as pointing to the right, causes [C] lie to the left of effects [E]. This view also assumes that any number of intermediary causes can be inserted between C and E. A typical JPSP article-hence not Bem's-sports at least one such mediator, although their number is theoretically infinite. Each event is a cause with respect to later events and an effect with respect to earlier ones. What we have then is an infinitesimal chain of causally connected events that run from the past through the present to the future.
In an extraordinary editorial comment specifically written to explain why they accepted Bem's article for publication, Chick Judd and Bertram Gawronski endorse this view as their own. They write that Bem's findings "turn out traditional understanding of causality on its head. A central assumption in lay and scientific conceptions of causality is that a cause precedes its effects, not the other way round [and] we openly admit that the reported findings conflict with our beliefs about causality and that we find them extremelely puzzling" (p. 406).
By skirting the issue of how the future acts on the present, belief in psi boils down to belief in processless causation. We are not only asked to believe that the future can influence the present (and by extension that the present affects what we think of as the past), but also that it does so without intermediate steps. In other words, the future leaps back across time to affect the present; it does not flow back through a chain of retroactive causes along an inverted temporal arrow.
So on the one hand, we have the ordinary forward flow of causation from C to E via intermediate Cs. On the other hand, we also have, according to Bem, a retroactive but unmediated leap from E to C. Once you assume that retroactive causation does not flow but leaps, you must ask "Which leap? From which event to which event?" If there is no time-reversed flow, then any event in the future might affect any event in the past. If causal leaps are possible, we must also ask why they should be limited to retroactive causation. Ordinary proactive causation should also be able to advance by leaps, no? In short, acceptance of non-process retroactive causation does not just mean that something is accepted in addition to conventional forward causation; it amounts to a destruction of the conventional view.
I am not ready to trade my traditional view of how causal processes work themselves out through time for an "anything-goes" view. Bem knows that massive paradigm shifts require powerful evidence. The evidence he presents hardly clears that high threshold. On the plus side, there are 9 experiments conducted on different but related topics. On the minus side, the effect sizes are rather small; some of the null hypotheses are rejected only by way of the flat-footed one-tailed test, and finally, it took apparently 20 years and extensive pilot testing to put together this package of studies. Bem also knows that skeptics will clamor for independent replication studies. To his credit, he encourages such replications and provides the programs for running these studies.
So why believe in psi? Bem, like many proponents of extraordinary claims before him, appeals to metaphor. He must do so because he is already committed to the view that ordinary mechanical processes cannot be at play, for if they were, we would not be dealing with psi (see above). The metaphor of choice is quantum mechanics. According to quantum mechanics, the behavior of subatomic particles is beautiful, weird, and completely out of whack with the laws of Newtonian physics. This missing explanatory step is how we get retroactive causation at the macroscopic level from indeterminacy at the subatomic level. To be fair, the problem may be more general: How do we get determinacy (even in the ordinary sense) from indeterminacy?
The central goal of this post was to apply Humphrey's insight (see epigraph) to our evaluation of Bem's paranormal work. I recently came across a paper by LeBel & Peters (2011) that strikes a similar note. I quote: "The rejection of central beliefs to account for observed data thus entails a major restructuring of the overall knowledge system, whereas the rejection of peripheral beliefs entails little or no restructuring" (p. 372). Here, forward flowing causation is a central belief that Bem questions; the presumed validity of his measurement instruments involves peripheral beliefs. And "Quine and Ullian (1978) referred to the use of belief centrality as a criterion for theory choice as conservatism: choosing the theoretical explanation consistent with the data that requires the least amount of restructuring of the existing knowledge system" (same page). This is good advice.
JPSP published an attempt by Galak et al. to replicate Bem's last 2 experiments, in which subjects recall words before rehearsing them (or not). The replication studies used the exact experimental protocol Bem asks the replicators to use and involved data from more subjects (N = 3,289). The result: bubkes. It seems that in order to remember well, you need to rehearse first.
Bem, D. L. (2011). Feeling the future: Experimental evidence for anomalous retroactive influences on cognition and affect. Journal of Personality and Social Psychology, 100, 407-425.
Humphrey, N. (1996). Leaps of faith. New York, NY: Springer-Verlag.
Galak, J., LeBoeuf, R. A., Nelson, L. D., & Simmons, J. P. (2012). Correcting the past: Failures to replicate psi. Journal of Personality and Social Psychology, 103, 933-948.
Judd, C. M., & Gawronski, B. (2011). Editorial comment. Journal of Personality and Social Psychology, 100, 406.
LeBel, E. P., & Peters, K. R. (2011). Fearing the future of empirical psychology: Bem's (2011) evidence for psi as a case study of deficiencies in modal research practice. Review of General Psychology, 15, 371-379.
Quine, W. V. O., & Ullian, J. S. (1978). The web of belief (2nd ed.). New York, NY: Random House.
For a primer on null hypothesis signficance testing, see
Krueger, J. (2001). Null hypothesis significance testing: On the survival of a flawed method. American Psychologist, 56, 16-26.
This paper is linked on my profile page.