This is an odd and deeply counterintuitive concept. To Santos, it became clearer after reading about a study with chimps, published last year by Katja Karg and colleagues. In that set-up, each chimp saw a box outside its enclosure, with some bananas. A human experimenter would either hand the bananas across or keep them for herself. The chimp, meanwhile, could control what the human saw by manipulating covers that either kept the bananas in view, or hid them. A smart chimp with a full theory of mind would open the covers (or leave them open) for co-operative humans, while closing them (or leaving them closed) for competitors.
Karg’s chimps did show the food to the cooperative humans, but they didn’t hide the food from competitors.
“They were good at opening things that were already closed. They’re good at showing; they want others to have the same facts,” says Santos. “But they don’t know to shut something because they can’t predict what it would be like for someone to not know the same facts as them. In a sense, their theory of mind is really easy. They import their own understanding to someone else. But as soon as that information changes for someone else, they have no representation. They can’t simulate a world that’s not the world they’re in right now.”
Josep Call, who was involved in Karg’s study, likes the idea. Still, he notes that there are studies in which chimps will hide themselves from competitors, or even steal food after first moving their arms out of sight. But Santos argues that they can do all this without understanding their rival’s ignorance. All they know is whether the competitor has the same knowledge as them or not. If not, they don't know what the competitor is thinking, and so all bets are off. “The crazy thing is that this ability gets them really far,” she says. “They can deceive each other. This dumb system lets you be an organism that’s pretty good at Machiavellian tasks.”
“I feel it is too early for such a proposal,” says Frans de Waal from Emory University. Just because primates fail at current tasks doesn’t mean they’ll always do so—history has taught us that much. “Humans are usually tested by their own species, and the apes are tested by us, another species. This situation hampers them from the outset. Absence of evidence is not evidence of absence.”
“He’s right in that tomorrow someone could come up with the right study and show we’re all wrong, but so far things have been pretty consistent,” says Santos. Sarah Brosnan from Georgia State University agrees. “It fits the data, seemingly better than other current theories,” she says, “although finding empirical evidence to support this interpretation over the others will be tricky, because of the challenges of testing these sorts of nuances.”
Martin and Santos already have ideas of possible experiments. Imagine that you show several containers to a chimp and then hide food in one of them. Now, another person comes in and goes straight to the hidden food. “That would be surprising to us—getting it right when you’re ignorant,” says Martin. “And the more containers there were, the more surprising it would be. But a monkey wouldn’t be surprised. Importantly, our theory can be falsified.”