Clever Hans, the Amazing Horse
Early in this century, Mr. von Osten, a German mathematics teacher, set out to prove that his horse, Hans, had great intellectual abilities, particularly in arithmetic. To teach Hans arithmetic, he first showed him a single object, said “One,” and lifted Hans’ foot once. Then he raised Hans’ foot twice for two objects, and so on. Eventually, when von Osten presented a group of objects, Hans tapped his foot by himself, and with practice he managed to tap the correct number of times. With more practice, it was no longer necessary for Hans to see the objects. Von Osten would just call out a number, and Hans would tap the appropriate number of times.
Von Osten moved on to addition and then to subtraction, multiplication, and division. Hans seemed to catch on amazingly quickly, soon responding with 90-95% accuracy. Von Osten began touring Germany to exhibit Hans’ abilities. He would give Hans a question, either orally or in writing, and Hans would tap out the answer. As time passed, Hans’ abilities grew, just from being around humans, without any special training. Soon he was able to add fractions, convert fractions to decimals or vice versa, do simple algebra, tell time to the minute, and give the values of all German coins. Using a number-to-letter code, he could spell out the names of objects and even identify musical notes, such as D or B-flat. (Hans, it seems, had perfect pitch.) He responded correctly even when questions were put to him be persons other than von Osten, in unfamiliar places with von Osten nowhere in sight.
Given this evidence, many people were ready to assume that Hans had great intellectual prowess. But others were not. Why not? Certainly the evidence was replicable. The problem was parsimony. No previous research had led us to assume that a nonhuman animal could perform complex mathematical calculations. Was there a simpler explanation?
Enter Oskar Pfungst. Pfungst (1911) discovered that Hans could not answer a question correctly if the questioner had not calculated the answer first. Evidently the horse was not actually doing the calculations but was somehow getting the answers from the questioner. Next Pfungst learned that Hans had to see the experimenter. When he experimenter stood in plain sight, Hans’ accuracy was 90% or better; when he could not see the experimenter, he either did not answer or made a wild guess.
Eventually Pfungst observed that any questioner who asked Hans a question would lean forward to watch Hans’ foot. Hans had simply learned to start tapping whenever someone stood next to his right forefoot and leaned forward. As soon as Hans had given the correct number of taps, the experimenter would give a slight upward jerk of the head and change facial expression in anticipation that this might be the last tap. (Even skeptical scientists who tested Hans did this involuntarily.) Hans simply continued tapping until he received that cue.
In short, Hans was indeed a clever horse. But what he did could be explained in simple terms that did not involve mathematical calculations of any other advanced cognitive process. We prefer the explanation in terms of facial expressions because it is more parsimonious.
Introduction to PSYCHOLOGY 6edamazon.com
James W. KALAT
ISBN 0-534-53988-2
Chapter 2 : Scientific Methods in Psychology
pp 36-37 (in 3ed)