Hans, Too Clever by Half

In the late 1800s, a retired German school teacher named Wilhelm von Osten set in motion one of history’s most infamous demonstrations of animal communication. This experiment, involving Osten’s horse Hans, still casts a shadow over studies attempting to plumb the depths of human-nonhuman communication.

Von Osten thought Hans a clever horse, and set out to teach him to perform basic mathematics. The former teacher determined to use the same methods he had used with students. He began simply, but soon advanced the problems, surprised Hans could do so well. Hans solved double-digit subtraction problems, stamping one front hoof to count out the answers. He rarely erred.

Hans’ fame spread throughout Europe, drawing curious onlookers, including a number of eminent scientists. Few left doubting Hans’ skill. He could even answer questions posed by visitors when his trainer was out of sight.

Clever Hans in a photo taken before 1907

public domain photo from Wikipedia

Two doubters, however, determined to devise a test more clever than Hans. First they decided on a problem together. One whispered it to Hans, then moved out of Hans’ sight. Amazingly, Hans still solved the problem. But when the whisperer gave Hans a problem no one else knew and then left the room, Hans stumbled. His chance of success fell to the level of chance.

Hans was in fact clever beyond anyone’s imaginings—but not at math. Instead he was clever at reading human body language. Von Osten learned to his surprise that Hans simply started stamping and continued until von Osten’s body language indicated Hans had reached the right number. And Hans had learned to generalize this skill beyond von Osten to other questioners, and even bystanders who knew the answer. But when no one within his sight knew the answer, Hans had no clue when to stop tapping.

The Clever Hans incident chilled enthusiasm for human/nonhuman communication research for decades. No one today assumes horses can understand human language. But many assert great apes can. The 1960s and 1970s saw a resurgence of attempts to communicate with animals—chimpanzees, orangutans and gorillas this time—until the Clever Hans phenomenon struck again. One researcher analyzed videotapes of “conversations” between trainers and his subject, a chimpanzee named Nim, who had learned some word signs derived from American Sign Language. The frame-by-frame analysis revealed trainers unconsciously prompting and modeling each word, and Nim imitating. Unaware of their own prompting, the trainers had credited Nim with producing sentences. For the second time, enthusiasm for human-language training for animals waned. The few researchers currently pursuing these studies keep Clever Hans in mind, doing their best to design scientifically “clean” experiments.