A horse that fooled all of Germany was never doing algebra

In the early 1900s, a horse named Clever Hans became a sensation across Germany. Ask him what four times three was, and he'd tap his hoof twelve times. Show him a written subtraction problem, and he'd tap out the right answer with what looked like genuine mathematical understanding. In 1904 the German board of education sent a thirteen-person commission, including a circus manager and a veterinarian, to investigate. Even separating Hans from his owner and using strangers to ask the questions didn't stop him from getting most answers right. It took until 1907, when the psychologist Oskar Pfungst dug further, to find the real explanation: Hans was reading the tiny, involuntary tension in his questioners' posture and expression as he approached the correct number of taps, and stopping when that tension peaked into relief.

Most retellings of this story treat it as proof that humans foolishly humanise animals, crediting them with abilities they don't have. Yuval Noah Harari draws the opposite lesson. Hans wasn't a mathematical genius; any eight-year-old could beat him at arithmetic. But in his ability to read human emotion and intention from body language alone, well enough to fool thirteen experts, he was genuinely remarkable, doing something no human could match without years of training. We underestimated animal cognition by looking for the wrong skill.

That correction opens onto a harder and unresolved question: how do we ever know, for certain, that any being other than ourselves is actually conscious rather than just behaving in a convincing way? Philosophers call this the Problem of Other Minds, and Harari points out it applies even to other humans; we simply take each other's word for it. Consider Santino, a chimpanzee at a Swedish zoo who quietly gathered stones each morning, hid them under straw, and waited calmly for visitors before launching a surprise bombardment hours later. Was he consciously planning ahead, or running a sophisticated but unconscious algorithm, the way a young squirrel who has never lived through winter still caches nuts? Nobody can prove it either way.

The same unresolved question, Harari notes, sits behind the Turing test, invented by Alan Turing, who was himself convicted for being gay in 1952 and forced into chemical castration, a man who knew from bitter experience that what mattered to society wasn't who you really were, only what others believed about you. As artificial intelligence grows more capable, and as it starts making decisions about and for us, this same unanswerable question, is anyone actually home in there, will decide how we treat both machines and the animals we've spent millennia assuming have nothing to say.

Start by catching yourself the next time you decide, almost automatically, whether an animal, a patient, or a chatbot has real feelings behind its behaviour, and name the specific unprovable assumption you just made. Where you can, look past the surface impression toward something closer to evidence, the kind of consistent brain signature scientists now use to detect awareness in patients who can't move or speak, rather than judging purely by how convincing the performance looks. And when you genuinely can't tell, lean toward taking the possibility of suffering seriously rather than dismissing it, the same shift that led leading scientists to sign a formal declaration in 2012 rather than keep assuming silence meant nothing was there.