Apes Might Know That You Don’t Know What They Know

The latest volley in a decades-long debate about apes’ theory of mind involved one scientist dressing up as King Kong and stealing from his colleague.

The face of an ape
How much do apes understand the minds of others? (Arterra / Universal Images Group / Getty)

In the pursuit of new knowledge, some scientists explore other worlds, discover new species, and develop cures for disease. Others film themselves being robbed by a colleague in a King Kong suit, to address a debate that’s been raging for more than 40 years.

Bedecked in ape cosplay, Satoshi Hirata from Kyoto University would grab a stone from his uncostumed colleague, Fumihiro Kano, and hide it under one of two boxes, all while Kano watched in mock indignation. Then, after Kano ducked behind a door, “Kong” would surreptitiously move the stolen stone to the second box. The duo filmed these shenanigans and then showed the videos to several chimpanzees, bonobos, and orangutans. They wanted to know how what the apes made of the scene. Specifically, when Kano returned and began looking for his stone, which box did the apes think he’d search first?

Humans would almost always say the first one. The items had of course been moved, but Kano doesn’t know that, and since we know he doesn’t know that, we can predict his actions accordingly. We have what psychologists call “theory of mind”—an understanding that others have their own mental states, their own beliefs and knowledge about the world. And crucially, we understand that those beliefs can be false and that knowledge can be wrong. Apes, Kano argues, share that ability. In his experiment, eye-tracking software revealed that they stared at the first box upon Kano’s return, presumably in anticipation of where he would search. This, he claimed when these results were published in 2016, was evidence of theory of mind. And like all such claims, it proved controversial.

Scientists and philosophers have spent decades arguing about whether theory of mind is a uniquely human trait. Early studies suggested that it’s absent in primates, but often used overly convoluted experiments. Later research, which used simpler and more naturalistic set-ups, showed that chimps do have a sense of their peers’ knowledge, goals, and intentions. Still, they kept on failing in one important way: They couldn’t seem to imagine that other individuals might hold false beliefs. Apes might have a theory of mind, but one that’s simpler than ours. As I described in 2016, “They can’t conceive of states of the world that are decoupled from their current reality … They can think about the minds of others, but only when those minds have the same contents as theirs.”

Kano’s Kong experiment, which was published mere months after I wrote that description, seemed to disprove that idea. Not everyone was convinced. Skeptics argued that the apes could have just been looking at the last place where Kano saw his pilfered stone—a simple associative rule that doesn’t require any theory of mind.

To address this critique, Kano and his colleagues ran a more complicated experiment. Once again, apes saw a video in which Kong stole an object from an actor, hid it under one box, and then moved it to another. This time, while Kong moved the object, the actor stood behind a seemingly opaque barrier. Crucially, the apes had a chance to inspect that same barrier before watching the video. Half of them were given a barrier that was as opaque as it seemed. The other half saw a barrier that, on closer inspection, was clearly made from see-through mesh.

If the apes were just looking at the last place where the actor saw the stone, both groups should react to the video in the same way. They did not. Those that inspected the opaque barrier, and knew that the actor couldn’t see the stone being moved, spent more time looking at the first box. Those that inspected the transparent barrier did not: They knew that the actor had actually seen everything. Even though all the apes saw the same video, they had different expectations of the actor’s beliefs and behavior, based on what they themselves had experienced. “Something important is going on in their mind,” Kano says.

This kind of experiment, known as a “goggles test,” was first proposed more than 20 years ago by the psychologist Cecilia Heyes. Can a participant use their own experience of novel goggles to deduce if others can see through the same goggles? Two studies, using opaque and transparent blindfolds, showed that 1-year-old human infants could do so. Kano’s experiment suggests that apes also pass, after just five minutes of inspection time. “I think that’s wild,” says Alia Martin, a psychologist from Victoria University of Wellington, who once argued that apes didn’t understand false beliefs. “It changes the way I think.”

But “there’s a lot to unpack about what this task is really telling us,” she adds. One hallmark of our theory of mind is our ability to hold and compare two competing views of the world: a false belief, and the reality. There’s still no evidence that apes can do this. Sure, they might infer that the actor thinks the stone is in the wrong place, but can they simultaneously remember where the stone actually is? Are they even thinking about the stone at all? “There’s an inferential leap to this being similar to humans,” she says.

Kano anticipates other critiques. Skeptics could still argue that the apes are using a more complicated rule: Look where the actor last saw the object with a clear line of sight. The type of tasks he used—anticipatory-looking exercises—are also controversial. They’ve been used to demonstrate theory of mind in human infants, but several researchers recently tried to replicate those studies with mixed results. “We did replicate it with great apes though, which is quite strange,” says Kano.

The problem, he thinks, is that many researchers use videos that are incredibly boring, where actors impassively and slowly move objects between bowls. When he initially showed such videos to his apes, they just didn’t care. That’s why he brought out the Kong suit, and added an element of theft. “The apes want to see more action, and maybe even violence,” he says. “And they’re slightly irritated by this weird ape-looking guy.”

Still, even Kano isn’t entirely convinced by his own studies, and admits there’s plenty for room for doubt. On a scale of one to 10, he rates his confidence that apes have a full theory of mind at a six. “We started these series of experiments believing that apes have theory of mind,” he says, “but there are so many alternatives that we need to test.”

For example, others have argued that tracking beliefs about an object’s location is a relatively simple feat that falls short of the full-fledged theory of mind that we possess. It would be more impressive if the apes could understand someone’s beliefs about an object’s identity: What happens, for example, if the pilfered stone is swapped for something else? That’s Kano’s next project. “Given the remarkable ingenuity and competence of this research team, I am sure it won’t be long before they succeed in testing this possibility, and I look forward to finding out their results,” says Renée Baillargeon, a psychologist from the University of Illinois who studies theory of mind in infants.

Determining what another species is thinking is always challenging. Apes are hard to work with, which means that studies are usually plagued by small sample sizes. They can’t speak or be questioned directly, so scientists must probe their thoughts through elaborate experiments, which can almost always be interpreted in different ways. The question is, as the primatologist Frans de Waal posed in his recent book, “Are we smart enough to know how smart animals are?”

“Alternative explanations will always be available for any one experimental setup,” says Kristin Andrews, a philosopher at York University who studies animal cognition. But “as clever researchers find that apes perform well in new situations, they generate more and more evidence that apes understand others’ minds, and the piecemeal skeptical explanations start to look like unwarranted conservatism.”