Does your dog know what you’re thinking? Cognitive biologists have shown our canine companions understand what we mean when we point at something, like a hidden snack. But are dogs really reading our minds? Or have they just lived with us so long that they’ve simply learned to make an association between, say, a hand and a tasty treat?
“I still think there’s a lot of controversy about this particular point—as it were,” says Christoph Völter, a comparative psychologist at the University of Veterinary Medicine in Vienna.
So, in a new study, Völter and his colleagues took a different tack, using an experiment similar to one designed to determine whether human babies can read adults’ intentions. The researchers attempted to offer treats to a number of dogs, then either “clumpsily” dropped the snack or teased the canines—snatching it away just before the pup could chow down on it. Even though the basic hand gestures were the same, the dogs appeared more frustrated with the teasing condition, suggesting they understand the difference between our good and ill intent.
“It seems really well done; I think it moves the ball forward,” says Evan MacLean, director of the Arizona Canine Cognition Center at the University of Arizona, who has shown dogs’ ability to follow human pointing may be genetically hard-wired. Still, he says, there could be other explanations besides canine mind reading.
To conduct the work, a member of Völter’s team put herself in jail. Or at least that’s what it might have looked like to the dogs. In a lab room at the university, Maud Steinmann, then an undergraduate biology student at the HAS University of Applied Sciences, sat in a rectangular box with mesh on the sides and with a clear plastic panel in front of her. In the center of the panel, the team had drilled a golf ball–size hole.
The researchers then led a succession of 48 pet dogs of various breeds into the room. Eight cameras at various locations recorded them, and 3D tracking software assisted by artificial intelligence captured every movement, from a twitch of the tail to a slight shift in muzzle direction.
Then the mind games began. In one set of experiments, Steinmann held a piece of sausage near the hole, but every time the dog approached, the snack “slipped” out of her fingers, falling back inside the cage. Next up, she again held the treat to the hole, but jerked it away as soon as the dog’s snout got close. In a final test, the team covered the hole; Steinmann tried to push the treat through, but she couldn’t. Each trial lasted about 30 seconds.
The dogs seemed to know when they were being messed with. They stuck around in front of the plastic barrier for 89% of the trial, on average, when the researcher kept dropping the treat, the team reports this month on the preprint server bioRxiv. “Dumb human is trying her best—but she’s clumsy,” they seemed to be thinking.
When Steinmann teased them, however, the dogs were quicker to turn away, sticking around for 78% of the trial. And when the hole was blocked, the pups spent a mere 64% of the trial time near the plastic barrier, quickly moving to the side of the cage, where it looked like they could get the treat from a different angle. “She wants to give me food, but she can’t,” Volter imagines the dogs thinking. “So I’m going to try to get it another way.”
The dogs’ tails also gave clues to their mindset. The 3D tracking software showed the animals tended to wag their tails more on the right side of their bodies when the researcher was fumbling. “That’s really intriguing,” MacLean says, as previous research has linked rightward tail movements to positive emotions. “That’s consistent with the idea that the dogs thought the experimenter’s intentions were good.”
Though the study is preliminary, the findings are important, MacLean says, especially given how hard it is to understand what’s going on in the mind of another species. The work, he says, supports the idea that dogs are tuned into our thoughts as well as our actions.
MacLean adds that the 3D tracking adds a new level of rigor to these types of cognitive experiments by picking up on subtle cues and reducing human bias. For example, a human might not expect a tail movement and therefore miss a brief wag. “I think it has the potential to lead to great insights.”
Still, he says the new study doesn’t conclusively show the dogs were probing human thoughts. They may have just learned from past experience: “When my owner fumbled, I eventually got the treat, so it was worth sticking around,” the dog may have been thinking. “When she teased me, it wasn’t.” No mind reading required.
Oddly, although chimpanzees, our closest relatives, fail the pointing test, other research has shown they perform similarly to dogs in the kind of experiment conducted in the new study. That could mean chimps can also read our intentions. The ability would be useful for any social animal, Volter says, as they need to intuit whether their group mates are about to harm or help them when they approach, say, with a stick in their hand.
If dogs truly pass both the pointing and clumsy tests, it could suggest they have a rudimentary “theory of mind”—a true ability to understand what others are thinking, MacLean says. And that might mean they are “considerably more complex” in their understanding of the motivations of others than are chimps, he says. They wouldn’t just understand what we’re thinking, but also that we want them to know what we’re thinking. Still, MacLean says, much more research is needed to show this.
In the meantime, studies like this one show how important it is for dogs to try to connect with us on at least some mental level—to be our friends, our helpers, and just to understand what the heck we’re up to.
“Most of what we do must seem really weird to dogs,” MacLean says. “Trying to get inside our minds would be really useful.”