The Believing Brain - Michael Shermer [33]
Figure 1. Patternicity in Pigeons
Inside a Skinner box in Douglas Navarick’s laboratory at California State University–Fullerton, where I conducted research on learning in the 1970s, one of our pigeons has learned to peck at the two keys above to receive grain through a food hopper below. Skinner discovered that if he randomly delivered the food reinforcement, whatever the pigeon happened to be doing just before the delivery of the food would be repeated the next time, such as spinning around once to the left before pecking at the key. This is pigeon patternicity, or the learning of a superstition. PHOTO BY THE AUTHOR.
* * *
Inspired by Skinner’s classic experiments, Koichi Ono of Komazawa University in Japan ran human subjects through the equivalent of a Skinner box by having them sit in a booth in which there were three levers.6 Independent of their pulling the levers (but unknown to them) the subjects were then exposed to a number counter that granted them one point at a time, which was followed by a flashing light and buzzer (a scaled-down slot machine, as it were). The points were delivered in a VI schedule of reinforcement (just like the pigeons) of, on average, either 30 seconds (with a range of 3 to 57 seconds) or 60 seconds (with a range of 25 to 95 seconds). Before the experiment began the subjects were instructed, “The experimenter does not require you to do anything specific. But if you do something, you may get points on the counter. Now try to get as many points as possible.”
Since the subjects could not predict when the points would be delivered (because the schedule of delivery was variable), and people just seem to have a natural propensity to pull levers, some of them inferred a connection between (A) pulling the handles and (B) getting points. Patternicity. And there were some doozies. Subject 1 happened to get a point after pulling the levers in the order of left, middle, right, right, middle, left, and so repeated that pattern three more times. Subject 5 began the session with short pulls of all the levers, with the points accumulating quite independently of his pulls, but then by chance he happened to be holding the middle lever when a point was delivered, so thereafter he performed the superstitious ritual of three short pulls followed by holding the middle lever. Of course, the longer he held the lever the greater the chance that he would get another point (because they were delivered on a variable time schedule). After minute nine of the thirty-minute session, Subject 5 had his ritual down pat. Subject 15 developed the strangest rite of all. Five minutes into her session a point was delivered the moment she happened to touch the point counter. Thereafter she started touching anything and everything within reach, and, of course, since the points continued to be delivered, this odd touching behavior was reinforced. At the ten-minute mark she got a point just as she happened to jump on the floor, whereby she promptly abandoned touching and took up jumping as her new strategy, climaxing in a point being scored when she touched the ceiling, leading her to end the session early from ceiling-touching exhaustion.
Technically speaking, in Ono’s words, “superstitious behavior is defined as behavior produced by response independent schedules of reinforcer delivery, in which only an accidental relation exists between responses and delivery of reinforcers.” That’s a fancy way of saying that superstitions are just an accidental form of learning. This is patternicity. Can such learned superstitious patternicities be unlearned? They can. In 1963, Skinner’s Harvard colleagues Charles Catania and David Cutts put humans through the pigeon paces by instructing each of twenty-six undergraduate subjects to press one of two different buttons on a box whenever a yellow light flashed