Lewis Thomas, a distinguished American research pathologist, won bestsellerdom with The Lives of a Cell (1974), a collection of his scientific journalism. This extract is from Late Night Thoughts (1983).

Scientists who work on animal behavior are occupationally obliged to live chancier lives than most of their colleagues, always at risk of being fooled by the animals they are studying or, worse, fooling themselves. Whether their experiments involve domesticated laboratory animals or wild creatures in the field, there is no end to the surprises that an animal can think up in the presence of an investigator. Sometimes it seems as if animals are genetically programmed to puzzle human beings, especially psychologists.

The risks are especially high when the scientist is engaged in training the animal to do something or other and must bank his professional reputation on the integrity of his experimental subject. The most famous case in point is that of Clever Hans, the turn-of-the-century German horse now immortalized in the lexicon of behavioral science by the technical term, the ‘Clever Hans Error.’ The horse, owned and trained by Herr von Osten, could not only solve complex arithmetical problems, but even read the instructions on a blackboard and tap out infallibly, with one hoof, the right answer. What is more, he could perform the same computations when total strangers posed questions to him, with his trainer nowhere nearby. For several years Clever Hans was studied intensively by groups of puzzled scientists and taken seriously as a horse with something very like a human brain, quite possibly even better than human. But finally in 1911, it was discovered by Professor O. Pfungst that Hans was not really doing arithmetic at all; he was simply observing the behavior of the human experimenter. Subtle, unconscious gestures – nods of the head, the holding of breath, the cessation of nodding when the correct count was reached – were accurately read by the horse as cues to stop tapping.

Whenever I read about that phenomenon, usually recounted as the exposure of a sort of unconscious fraud on the part of either the experimenter or the horse or both, I wish Clever Hans would be given more credit than he generally gets. To be sure, the horse couldn’t really do arithmetic, but the record shows that he was considerably better at observing human beings and interpreting their behavior than humans are at comprehending horses or, for that matter, other humans.

Cats are a standing rebuke to behavioral scientists wanting to know how the minds of animals work. The mind of a cat is an inscrutable mystery, beyond human reach, the least human of all creatures and at the same time, as any cat owner will attest, the most intelligent. In 1979, a paper was published in Science by B. R. Moore and S. Stuttard entitled ‘Dr. Guthrie and Felis domesticus or: tripping over the cat,’ a wonderful account of the kind of scientific mischief native to this species. Thirty-five years ago, E. R. Guthrie and G. P. Horton described an experiment in which cats were placed in a glass-fronted puzzle box and trained to find their way out by jostling a slender vertical rod at the front of the box, thereby causing a door to open. What interested these investigators was not so much that the cats could learn to bump into the vertical rod, but that before doing so each animal performed a long ritual of highly stereotyped movements, rubbing their heads and backs against the front of the box, turning in circles, and finally touching the rod. The experiment has ranked as something of a classic in experimental psychology, even raising in some minds the notion of a ceremony of superstition on the part of cats: before the rod will open the door, it is necessary to go through a magical sequence of motions.

Moore and Stuttard repeated the Guthrie experiment, observed the same complex ‘learning’ behavior, but then discovered that it occurred only when a human being was visible to the cat. If no one was in the room with the box, the cat did nothing but take naps. The sight of a human being was all that was needed to launch the animal on the series of sinuous movements, rod or no rod, door or no door. It was not a learned pattern of behavior, it was a cat greeting a person.