The human brain is a belief-forming machine. It is what defines us as a species: we guide our actions by creating mental models of the world and using those models to hypothesize about what will happen under a given set of circumstances. Our ability to create such models far surpasses that of any other species, and accounts for our unprecedented degree of deliberate technological control over the world.
However, though we are intelligent, we are not infallible. Humans of every culture and level of civilization can show remarkable astuteness in drawing inferences from scarce evidence, but we do make mistakes, and the principles of scientific reasoning appear to be far too complex for evolution to endow us with an instinctual grasp of them. Instead, as with so many other things, evolution has given us a jury-rigged solution that works more often than not. When two events frequently occur together, we make the inference that they are causally connected. This is usually the correct judgment, but the problem comes when events in our environment are random, when there is no pattern to be found. Under those circumstances, our belief-forming brains struggle to find a causal association – and if by chance there seems to be an association, the brain will seize on it. Apparently, it is better to perceive a pattern that doesn’t actually exist than to miss a pattern that does exist.
The brain’s susceptibility to finding patterns in noise was demonstrated in animal experiments by B.F. Skinner and his famous pigeon superstition experiments. Skinner set up an experiment where pigeons would be fed on a predetermined schedule, regardless of what they were doing at the time. With no reliable correlations to be found, the pigeons’ brains found unreliable correlations, assuming that whatever they were doing at the time they were fed was what caused them to be fed. As a result, the pigeons eventually formed “superstitions” about what caused the release of food: some repeatedly turned in circles, others pecked at the cage’s floor, others made peculiar head-bobbing or jerking motions toward one corner of the cage.
But surely, only pigeons and other animals of low intelligence would commit such a transparently simple error of judgment! Human beings are far smarter than that. We are creatures of magnificent reason, in possession of a brilliant and incisive mind that can carefully observe the evidence and draw the most likely conclusions. We could never be taken in by such a trick!
Could we?
Koichi Ono, a Japanese psychologist, set up one of my favourite experiments in which subjects were asked to sit at a desk on which were placed three coloured levers. In front of the subjects on the partition wall behind the desk was a signal light and electronic counter, which was apparently to keep track of points. Each session lasted forty minutes, and the subject was told to earn as many points as possible, though it was not explained how she was expected to do that. In fact, the points display on the counter had nothing to do with the levers on the desk, and was designed just to display new points at various intervals independently of anything the subject might do.
…After a while the Japanese volunteers could be seen repeating elaborate or simple combinations of lever-pulls, banging the sides of the partition, or even jumping up to touch the ceiling until exhaustion set in.
I first read about this delightful experiment in Derren Brown’s Tricks of the Mind, from which I draw the above excerpt, and was happy to find that the original study is available online. Similar results have been observed in the past. For example, my post “Seeking the Hidden Switches” showed another study where children instinctively searched for hidden causes rather than assume that events might happen at random.
So strong is the human brain’s pattern-seeking bias that it can lead us to believe in causal associations that a rational evaluation would dismiss as far too unlikely to be plausible. Does it make any sense to believe that a simple partition around a desk, or the ceiling above that desk, with no obvious means of manipulation or connection to the digital points display, could nevertheless influence that display? Yet that is just what some of Ono’s volunteers were driven to conclude.
Granted, not all the participants in this experiment developed stable superstitious responses. Only a few developed the persistent responses described above, while many more developed superstitious responses that did not persist, but eventually died out. This can be explained by noting that, as in all things, humans vary. Some people will be more or less predisposed than others to infer patterns in noise, while others will be more sensitive to evidence. Not everyone needs to be predisposed to superstitious belief for that tendency to have an effect on society. It’s tempting to wonder if, under other circumstances, the volunteers that did act in this way could have been the founders of a new religion.