What About the Real (or Surreal) Thing?

Acts of God and Man(Image: Columbia Business School Publishing)Excerpted from “Acts of God and Man: Ruminations on Risk and Insurance“, by Michael R. Powers.

Copyright © 2012 Michael R. Powers.

Used by arrangement with Columbia University Press.

The world of risk and insurance — replete with cagey underwriters, callous claim adjustors, and canny actuaries — is often viewed as staid and cheerless. In Acts of God and Man, Michael R. Powers challenges this perception, espousing a science of risk based upon both a “fundamentalist Bayesian” (i.e., subjective/judgmental) approach to modeling uncertainty and a “personalized scientific method” free of the restrictions of more orthodox paradigms. Readers of Truthout likely will be interested in the book’s focus on the boundary between “knowable” and “unknowable” aspects of risk, explored in the excerpt below by embracing the paranormal in the service of science. Other topics of interest — addressed with a sometimes offbeat sense of humor — include: an argument for eliminating individual underwriting in health insurance; a discussion of government as “insurer of last resort”; a critique of black-box methods in forecasting catastrophes; and a proposal for building a broader coalition to address the challenges of global climate change.

Having considered a variety of matters, both theoretical and empirical, related to the epistemology of randomness, I would like to complete this last chapter by addressing an issue at the margin of science that nevertheless is of importance in the overall understanding of risk: the possible connection between human consciousness and the physical events of our world. Unlike some science writers who have probed a possible role for human consciousness at the foundations of theoretical physics, I simply will present my thoughts by way of a single empirical anecdote.

As noted previously, I acknowledge a special relationship with trees — one that is both unusual and difficult to describe. I certainly am not overly fond of my woody neighbors and in no respect a “tree hugger” or other manner of dendrophile. In fact, I must admit to having as much animosity as kindliness toward trees (which are, after all, the autumnal enemies of the sedentary as well as a perennial peril to houses, automobiles, and other property). So suffice it to say that my relationship with trees is rather mixed and complex.

The remarkable event that first illuminated this relationship occurred during my senior year of high school. At that time, I lived in a quiet residential town about fifteen miles southwest of Boston. Although the incident in question took place in October, it was preceded by another remarkable occurrence five months earlier: a freak snow- and- ice storm in May that caused severe damage to the new growth of trees in the region. In my neighborhood several trees were toppled, and many others suffered broken and twisted limbs. There was one tree in particular that was left with a large upper branch dangling by a thin piece of wood, seemingly little more than a strip of bark.

Intrigued by the strange angle and tenuous connection formed by the broken branch, I found myself drawn to look at it from time to time. This occurred throughout the remainder of the spring and the ensuing summer and early fall. Then, one day in October, as I was unlocking the front door of my house, I happened to turn in the direction of the tree, and pausing for a moment to look at the broken branch, I was stunned to see it separate in the perfectly still noontime air, accompanied by only the gentlest of cracks. Of all the opportunities the branch could have taken — from rainy days when I stayed indoors to long dark nights when the winds howled — it chose this particular moment to consummate its secession. Was it just coincidence (i.e., chance) that I was there to witness this event, or was there an element of the paranormal to it?

To answer that question, of course, requires a working definition of the term paranormal. To that end, I would propose that an event be deemed paranormal if it involves a relation of cause and effect that cannot be explained by the current state of scientific knowledge. This is, of course, a very broad definition, simply requiring ignorance or incompetence on the part of science, rather than the presence of any particular type of phenomenon: clairvoyance, precognition, telepathy, psychokinesis, etc. But in the absence of any convincing theories outside current physics that could explain the relationship between my observation of the tree branch and its final breaking, such a definition will suffice.

Now let us do a little calculation. Suppose that the tree originally splintered on May 15, and then broke completely on October 15. That would give five months, or approximately 150 days, in which the branch could have come apart. Now suppose that I had looked at the tree five times per day during that period (a gross overestimate, since the actual frequency would have been no more than five times on the most active of days) and that each time I looked, my eyes beheld the tree for ten seconds (again, a gross overestimate, since the actual time elapsed would have been no more than ten seconds on the most obsessive of occasions). Then the total number of seconds during which I viewed the tree would have been at most 150 x 5 x 10, or 750 seconds, out of a total of 150 x 24 x 60 x 60, or 12,960,000 seconds. In other words, assuming that the final break was equally likely to have occurred at any instant during the five-month period, the chances of my having witnessed that particular event would have been no more than 750/12,960,000, or approximately 0.00005787.

Expressing this analysis in the language of hypothesis testing, one could say that the null hypothesis is given by H0: “The final tree break occurred in dependently of my observation,” and the p-value is 0.00005787, so for any value of a greater than 0.00005787 (including the commonly used standards of 0.05 and 0.01), the null hypothesis would be rejected. In short, it would seem that there is at least prima facie evidence of a paranormal event’s having taken place!

Before getting too excited about this discovery, however, let us consider what the skeptics would have to say about such anecdotal observations (i.e., observations not made under controlled scientific conditions).

First, skeptics would argue (quite reasonably) that anecdotal observations are not dependable for testing natural phenomena because of the likelihood of misspecifying the null hypothesis by defining the category of “unusual” outcomes too narrowly. In the case at hand, I used H0: “The final tree break occurred independently of my observation.” However, this null hypothesis presumably was identified after observing the tree’s final break, an event I considered unusual. Had some other unusual event occurred — for example, had I happened to be looking at the sky the moment a bright fireball appeared — then I simply would have specified the null hypothesis differently (i.e., H0: “The fireball’s appearance occurred in dependently of my observation”). Essentially, I seem to be guilty of the moving targets problem described in Chapter 14 — that is, fishing for statistical significance.

Skeptics also would argue (again, quite reasonably) that anecdotal observations are not dependable because of the likely selective adduction of evidence caused by remembering only statistically significant outcomes and ignoring the more mundane, negative outcomes. Specifically, in focusing only on witnessing the tree break, I presumably am neglecting all of the perfectly ordinary things that I saw at times when I actually expected to see something unusual. In other words, I seem to be guilty of the censorship problem (also described in Chapter 14).

So what do I have to say in my defense?

Interestingly, I believe that both concerns can be addressed satisfactorily. While acknowledging that anecdotal observations certainly are susceptible to the aforementioned shortcomings, I would argue that my observation was not as casual as might be supposed.

At the time of the events described, I had taken a limited interest in the paranormal — and in psychokinesis in particular. As a result of this interest, I occasionally tested my own psychokinetic abilities in either of two ways: by trying to bend metal keys — which I failed miserably to accomplish on at most ten occasions — and by trying to break the filaments in glowing incandescent light bulbs — which I thought would be easier than bending keys, but which yielded no better results on at most twenty occasions. The tenuously attached tree branch in my yard provided a fortuitous additional — and I should say, the only additional — context for such trials.

Thus, although I most assuredly would not claim that my efforts with keys, filaments, and the tree branch constituted scientifically controlled experiments, I do believe that they did not suffer that greatly from either the moving targets problem or the censorship problem. There was no moving target because the informal tests were carried out intentionally with the purpose of determining whether or not I possessed any psychokinetic ability, rather than simply as reactions to randomly appearing unusual events. And there was no censorship problem because I certainly did not neglect to remember my failures with the keys and filaments. Naturally, one might inquire why I did not continue my psychokinetic experiments with tree branches after this singular success. But the answer is rather obvious: tenuously dangling branches are simply too few and far between to permit a serious effort.

The point I wish to make here is not that there is sufficient evidence of a paranormal event to publish a research article according to the standards of the scientific method, or even to convince a fellow human being that such an event occurred. I would argue, however, that there is more than ample evidence for me, under my own personalized scientific method, to reject a null hypothesis that paranormal events do not occur. In fact, I would go somewhat further and say that to ignore this personal evidence by trying to pretend that it can be discredited by moving targets, censorship, or some other facile explanation would be intellectually dishonest. Thus, it is really not a question of choosing to believe in something viewed as dubious by practitioners of science and established religion alike. Rather, I am stuck with this particular outcome because of my a priori selection of an evaluation function, V(t), that implies my personalized scientific method, whether I like the outcome or not.

As a final, somewhat bemusing point, I would like to address one further possible criticism of the above analysis: If, as has been argued, observational studies constitute only a very crude and inferior type of science, then how could I possibly rely on an observational test of significance to support a conclusion of paranormal activity?

Strangely, the answer is quite simple: A conclusion of paranormal causality is the only one that is logically valid based upon an observational study! Recall that the null hypothesis was H0: “The final tree break occurred in dependently of my observation.” Thus, when I rejected the null hypothesis, it left me with the alternative that there was some connection between the tree break and my observation. However, since I could not identify that connection, I had to refer to it as paranormal. Now, recognizing that my observational test of significance may have failed to account for some unknown confounding variable that is either positively correlated with both the breaking branch and my observation or negatively correlated with both events, I must further acknowledge there could have been some unknown connection between the two events, mediated by this confounding variable. However, since such a confounding variable is unknown, it must be paranormal (according to the above definition) as well.


[A hospital room. Old man lies awake in bed; Grim Reaper approaches quietly.]

REAPER: Good evening, Mr. Wiley. It’s me again, the Grim Reaper.

MAN: A good evening to you, my friend.

REAPER: It’s been another thirty-three years. Time to call in my debt. But this time, no tricks. I’m quite aware that you’re a much better statistician than I.

MAN: Oh, Reaper, it’s kind of you to say so. But I can still sense a bit of regret in your voice. You’ve gotten in some trouble with your boss for this long delay, haven’t you?


MAN: And He instructed you not to accept any more probabilistic challenges from me, didn’t He?


MAN: But He didn’t say anything about other sorts of challenges, did He?

REAPER: [Opens eyes suspiciously.] No. But what do you have in mind?

MAN: Well, I suppose poker or other chance-based card games would be out of the question. So how about checkers . . . or racquetball?

REAPER: Somehow, those choices seem a bit too lowbrow for a mortality challenge.

MAN: Well, what about chess . . . or squash?

REAPER: Despite all the rumors to the contrary, I never liked chess much; and I’m afraid that you’re in no condition for squash right now.

MAN: Then how about Simon Says?

REAPER: [Eyes brighten.] Well, that certainly would be an interesting choice, although a bit juvenile for a man of your age. I suppose you’d want to be Simon?

MAN: Yes, of course.

REAPER: But how would we determine when the game is over? You see, I don’t anticipate being tricked again.

MAN: Oh, I’ll be able to beat you quickly enough. Just give me half an hour.

REAPER: Half an hour? With all the time I’ve already wasted on your case, that simply isn’t possible. Do you know how many wars and famines are behind schedule because of you? [Pauses.] Look, I can give you at most ten minutes.

MAN: OK, I guess that’ll have to do.

REAPER: Good. So let’s begin!

MAN: Yes, let’s! Simon says, “Touch your nose with your right index finger.”

REAPER: Easy. [Touches right index finger to nose.]

MAN: Simon says, “Put your right thumb in your mouth.”

REAPER: Rather infantile, but a simple matter. [Inserts right thumb into mouth.]

MAN: “Take your thumb out of your mouth.”

[Reaper makes incoherent sound with thumb remaining in mouth.]

MAN: OK. Simon says, “Twirl your scythe with your left hand.”

[Reaper makes incoherent sound, twirls scythe.]

MAN: Simon says, “Keep twirling your scythe, but put your right thumb in your right ear.”

REAPER: No problem. [Twirls scythe while inserting right thumb in ear.]

MAN: Very good. Now Simon says, “Go away and never come back.”

REAPER: [Stops twirling scythe and removes thumb from ear, face red with anger.] Damn, damn, triple damn! I’m going to catch hell for this. [Face suddenly relaxes into cunning expression.] You’re good, you know; very good. But you also must know that a delay is one thing, but giving you immortality is quite another. If I can’t take you, then I have to take somebody else. [Pauses.] There’s a young girl sleeping in the next room. She was going to make a full recovery from surgery. But I suppose your win is her loss. [Pauses.] Alas, she’s such a small and innocent child: only 5 years old! Are you sure you don’t want to reconsider?

MAN: “Just take the girl and go.”

REAPER: As you wish, then. [Moves toward door.]

MAN: [Smiles.] Where are you going, Reaper? I didn’t say “Simon says.”