Does your dog know what you are thinking?

Groundbreaking research offers surprising new clues

Photo illustration by Levi Nicholson

James Scharien, a truck driver, first met his three dogs while making deliveries to the Waywayseecappo First Nation reserve, a 320-km drive west of Winnipeg. All were strays, and he took them to his home outside Portage la Prairie. Scharien had never owned dogs before, but found himself quickly bonding with them. There was Buster, a shepherd-Lab cross, Byron, who’s “maybe a little Rottweiler, maybe something else,” and Roger, who “looks like someone crossed a German shepherd with a fox.” Like many pet owners, Scharien, 39, got to wondering what his dogs were thinking.

He signed up with Dognition, a company launched in February that purports to help owners “find the genius” in their dogs. Led by Brian Hare, associate professor of evolutionary anthropology at Duke University and founder of the Duke Canine Cognition Center, Dognition offers web-based games that owners conduct at home with their dogs, to reveal various traits—a way to help them better understand their animals.

In one, Scharien would “yawn and see if they yawned along,” a sign of empathy. In another, he tested whether the dogs could follow his pointing, an indication of how much they rely on gestural communication. Other games tested memory, cunning and reasoning. In Dognition terms, Buster, the shepherd-Lab mutt, turned out to be an “Einstein,” with strong reasoning and inferential skills. (Scharien was surprised: “He’s always bumping his head on the coffee table.”) Byron, who died in July, was a “Protodog,” a strong communicator similar to the first dogs to befriend humankind. Roger, the youngest, is a “Socialite” who gets by on charm. Gaining a better understanding of his dogs has helped Scharien build a stronger relationship with them, he believes. “I don’t want them digging up the yard and chewing the couch while I’m at work,” he says. “The more I fulfill their needs, the more obedient they are to me.”

It’s also enabled Scharien to participate in a massive citizen science project—one that’s building the “largest data set on dogs that’s ever existed,” Hare says. His game results, along with other participating pet owners’, are stored by Dognition to be used by scientists, Hare included, in ongoing studies about how dogs think, feel, understand and interact with the world. In the few short months since its launch, thousands of dogs from dozens of countries have participated in Dognition. (It’s free to try the Dognition program, but a full assessment costs $39.) Although this may sound like the latest craze in a pet-obsessed culture—like doggie pedicures, massages, or jewellery—there’s real science behind it. “At Duke, we can only test a few hundred dogs a year,” says Hare, co-author of The Genius of Dogs, published earlier this year, who adds that every test is based on an experiment that’s been used in the lab. “With Dognition, we can get data on a few hundred dogs in a weekend. It’s a game changer.”

Scientists have long known that a range of animals exhibit qualities like intelligence, empathy, problem-solving, even self-awareness. A recent study found that dolphins can remember each other’s signature whistles (which function like names) for decades. Chimpanzees improve their use of tools by watching what others do. In experiments, elephants have been seen to co-operate on tasks, showing evidence of deliberate thought and teamwork. But, for all we’ve learned about these and other creatures, dogs—a species that’s lived and worked alongside us for thousands of years—have barely been studied, until recently. “We still know more about the honey bee than we do about the dog,” says Ádám Miklósi, head of the department of ethology at Eötvös University in Budapest, and an adviser to Dognition.

A growing number of researchers are peering into the mind of the dog and discovering that man’s best friend is a complex creature. Remarkably, some scientists believe that dogs might have developed unique cognitive skills to understand what we humans are thinking. It’s early days in this booming area, and though debates over dog cognition flare up within the field, one thing virtually all can agree upon: There’s more going on inside the dog’s brain than many of us suspected.

In the mid-nineties, as an undergraduate student at Emory University in Atlanta, Hare was studying animal psychology with Michael Tomasello. “He was interested in social skills that [appear] at around nine months in human children,” Hare says: the ability to interpret the gestures of parents and other family members. After discovering that chimpanzees (who are among our closest relatives) weren’t very successful at following human gestures, Tomasello suggested that this ability might be unique to humans. Hare, then 19, disagreed and said: “I think my dog can do that.”

Hare’s childhood pet, a Labrador named Oreo, was enlisted. Oreo could retrieve tennis balls by following where Hare pointed. In another experiment, a group of dogs, including Oreo, were able to find food hidden under upside-down cups based on gestures. Hare and Tomasello’s now-famous study, published in 1998, showed that dogs follow human social cues more readily than the great apes. That same year, by coincidence, Miklósi’s group published a similar paper, and the modern field of dog cognition was born.

Dogs have been studied for centuries, most famously by Ivan Pavlov, who rang a bell when his were fed; as they began to associate the sound with food, mere ringing was enough to make them drool. Still, for a long time, they weren’t considered a valid subject for psychological research. “There’s a tradition of thinking of domesticated animals as not real animals,” says Monique Udell, an assistant professor and dog researcher at Oregon State University. “Especially those living with us as pets.” That’s changing. At a 2010 international conference on comparative cognition (the study of how different species think and behave) held in Melbourne, Fla., 10 per cent of presentations focused on dogs and their relatives, compared to none in 1994. (The study of cats still lags, partly because they make trickier research subjects. “Cat research is probably where dog research was 20 years ago,” Udell says. “I think that’s going to be the next up-and-coming thing.”)

Dogs have been known to recognize hundreds of words; but to Hare, the number is less relevant than the way in which they learn. One of the most remarkable findings of the last decade, he notes, is that some dogs seem to learn words using the same kind of “inferential reasoning” used by human babies—the only species, other than us, known to do this. Dogs aren’t only tuned in to our words and gestures. They seem even to pick up on the intention to communicate. In a 2012 study, Hungarian researchers showed dogs video of a person either saying, “Hi dog!” in a high-pitched voice, looking straight at the animal, or uttering a low-pitched “Hi dog,” while avoiding eye contact. In both cases, the person then turned and looked at a plastic pot. Using eye-tracking software, the team saw that dogs were much more likely to look to the pot when the person made eye contact with the dog.

Many pet owners might suspect their dogs are mischievous, manipulative or cunning. Research from Juliane Kaminski of the University of Portsmouth, another Dognition adviser, suggests they may in fact understand the human point of view enough to try to pull one over on us. In one study, when a person forbade dogs to take food, she found they were four times more likely to disobey in a dark room than a lit room, implying that they understood the human couldn’t see them.

Udell, the Oregon dog researcher, did a set of experiments on pet dogs, shelter dogs and wolves raised by humans, a proxy for dogs’ ancestors. All three had a choice to beg from a person who was looking at them, or from one with his vision blocked somehow. Shelter dogs, pet dogs and wolves were all more likely to beg from an attentive person, compared to one with his back turned. Only pet dogs knew not to beg from a person reading a book. “It makes sense, because only pet dogs have that experience,” Udell says. Surprisingly, all three groups begged from the person with a bucket over his head. “It brings into question whether [dogs are] taking our perspective,” she says. “If they can see our eyes are covered, it should tell them we can’t see them.”

It’s one of many questions raised by dog cognition researchers. Whether these animals evolved to take the human perspective into account or are just good at picking up on cues is one of the liveliest debates in the field. Dogs were domesticated 15,000 years ago, possibly longer, according to Kaminski. “This new phase of dog psychology argues that the dog has acquired new cognitive skills to understand us,” says Clive Wynne of Arizona State University, who specializes in behaviour and cognition in canines. “I do not agree with that point of view.” Of the one billion dogs on the planet, maybe five per cent are pets, he says. Others might gather around settlements, and scavenge from our garbage, but they don’t share our beds: the term “man’s best friend” wasn’t commonly used until the 19th century, he notes. Udell, Wynne’s former student, agrees. “The idea that dogs have an evolutionary predisposition to bond with humans, specifically, may not be quite right,” she says, noting that herding dogs, for example, also form tight bonds with their flock. Yet domesticated dogs do share a special relationship with humans, especially their owners. In a recent study, 25 dogs were much more likely to yawn contagiously after their owner did, as opposed to a stranger.

Still, other animals can follow human pointing, including chimpanzees and hand-raised wolves; in at least one study, some adult cats were shown to follow pointing. But, according to Hare, none of the other species are nearly as successful at it as dogs. “Puppies as young as six to nine weeks can use these gestures to find things,” he says. It takes human infants about nine months to do the same.

Dog cognition research often says as much about dog owners as it does about their pets. John Bradshaw of the University of Bristol’s Veterinary School, author of Dog Sense and the forthcoming book Cat Sense, has studied how dogs react when the owner leaves for several hours. “Fifteen per cent of people report that they know the dog is upset,” he says: The animal is howling, behaving destructively, or urinating or defecating indoors. “Many will interpret this as the dog being spiteful, or trying to make you feel guilty. Well, those are complicated emotions that don’t even develop in human children until they’re five or six.” Instead, the dog’s response may be more like an infant’s: It’s distressed at the loss of the person to whom it’s attached, he explains. Because dogs don’t easily understand the concept of a future, they can’t rationalize the absence as temporary. (Another five per cent are “silent sufferers,” he says.) Dogs seem to form these attachments remarkably quickly. Working with shelter dogs who’d been abandoned, Bradshaw found that about half formed a strong attachment to “a person who’s being nice to them” after just 15 minutes.

Erica Feuerbacher of the University of Florida, who’s been completing her Ph.D. under Wynne, has studied what happens when dog owners return after several hours out of the house, during which time the dogs haven’t eaten. “We rigged it so the door opens remotely, and they can choose between [the owner] and a bowl of food,” she says. “We’ve run about four dogs with four-to-eight trials per dog, and in all but one, the dog chooses to greet the owner first.” (Feuerbacher also found that dogs prefer petting to vocal praise, suggesting petting is an important interaction between our species.) Domestic dogs “like other dogs,” Bradshaw says, “but they love people.” They’re bred that way.

And yet, they aren’t little people, no matter how human they might seem. Alexandra Horowitz, who runs the Horowitz Dog Cognition Lab at Columbia University’s Barnard College, has studied whether a dog’s guilty expression really suggests it’s feeling badly that it misbehaved. In an experiment, owners ordered their dogs not to eat a treat, and left the room. Horowitz then fed some of these dogs the treat, unbeknownst to the owners. When they returned, some were told their dog had eaten the treat; others were assured the dog had behaved, although what the owners were told wasn’t always the reality. Dogs looked most “guilty” if they got in trouble for eating the treat, whether they’d really eaten it or not—implying their look was a response to the owner’s behaviour.

One of the most telling studies was done by Krista Macpherson, a Ph.D. candidate at Western University’s Dog Cognition Lab, and her supervisor Bill Roberts. They tested our assumption that, Lassie-like, dogs will rush to a person’s aid. In the first experiment, dogs were walked by their owners through a field while a bystander sat nearby reading a magazine. The owner faked a heart attack and collapsed to the ground. The dogs (including two rough collies, one Brittany spaniel and a Portuguese water dog) were filmed for six minutes to gauge their reactions. Most of them hung around the owner or roamed through the park, but didn’t approach the bystander for help—in fact, none of the dogs actually touched a bystander, except a toy poodle, who leapt into the bystander’s lap. In the second experiment, the owners pretended to be crushed by a falling bookshelf, and cried out to the dog to get help. None of the dogs obliged.

Of course, dogs can be trained as service animals, and are used in search and rescue missions; there are high-profile reports of dogs saving people. Recently, a foster animal named Chloe rescued a Pennsylvania man who’d suffered a heart attack while mowing the lawn, by barking to alert the neighbours. In their study, Macpherson and Roberts acknowledge that maybe these dogs knew their owners were faking it; maybe the dogs assumed they’d simply fallen asleep. “We come to expect that the dog loves us so much, they would do anything for us,” says Roberts, a professor emeritus at Western. “But if the dog doesn’t understand the nature of the situation, it might not know to go get help.” Even though researchers must separate themselves from their subjects, there’s some benefit to seeing dogs as “human,” just as the science emphasizes our differences. “I’m on the record as damning people for anthropomorphizing them,” says Wynne. “But if you couldn’t anthropomorphize your dog, I’m not sure what the point would be of sharing your home. If its behaviour wasn’t consistent with the idea that it’s a member of your family—then what would be the point?”

Macpherson sometimes uses her own rough collies, Cash and Sedona, in her experiments. In her 2012 investigation into whether dogs can count, using a method adapted from chimpanzee studies, she had 27 pet dogs watch as different amounts of food were dropped into two bowls, one kibble at a time. The dogs were allowed to pick a bowl to eat from. Only when the ratio of food in the two bowls was one to zero did the dogs show a significant preference for the bowl with more food: a “food or no food” scenario, as the paper describes it.

But in a second experiment, Macpherson showed Sedona various geometric shapes attached to two magnet boards. When she chose the board with more items, she got a treat: a piece of food hidden under the board. Sedona performed much better than the dogs in the first trial. This could be because the items were shown to her at once, notes Macpherson, who says that monkeys have an easier time with simultaneous presentation, too. And when Sedona chose wrong, she got no reward at all, which might have provided her with extra motivation. Reasons for this will be the subject of future study.

“In some tasks, dogs are worse than rats and pigeons. Rats are a lot smarter than people realize,” Macpherson says. Rats, for example, outperform dogs when it comes to remembering where food is located inside a radial maze. “We’ve selectively bred out of dogs the ability to find food for themselves,” she explains. “We find it for them.”

Future research will focus on differences between breeds, although that can be surprisingly daunting to study. There are 177 recognized by the American Kennel Club; to perform cognitive experiments to compare them “you’d need a huge sample size,” says Macpherson, who handles rough collies in dog shows. (Hare thinks Dognition’s growing database will help.) Dogs’ perception, Macpherson believes, could make a difference in how they perform on these tests. “Look at a pug versus my breed,” she says. “They have to be experiencing the world differently.” Rough collies, which were bred to herd sheep and are descended from sight hounds, track visually; in experiments with Sedona, Macpherson can observe the dog’s eyes moving back and forth between displays. A beagle, on the other hand, lives nose to the ground. With so many different breeds, diets, home environments and other factors, dogs make a much more varied study subject than plain old lab rats. But “at the end of the day, they’re all the same species. You shouldn’t see whopping differences in cognitive ability,” Macpherson says.

As Dognition continues amassing information about canines and their owners, surprising findings are already starting to pop up. So far, data seems to suggest that mixed breed dogs are more reliant on their memory, Hare says, whereas purebred dogs depend on human gestures, although it’s impossible to say why right now. Individual owners like Scharien are discovering aspects of their pets’ personalities that they never suspected. After learning that Buster is more intelligent than he seemed, Scharien now gives him treats locked inside a toy; the dog has to puzzle them out. “I never would have thought he’d enjoy that, but he does.”

Dogs and humans evolved side by side, and it’s clear that—even among domesticated species—these animals are different, and somehow closer to us, than virtually any other. “They’ve been bred to pull our sleds, guard our homes, herd our livestock,” Macpherson says. “They have a unique relationship with humans, and it’s important to understand that.” Ultimately, understanding how they experience the world, she says, “is about understanding us.”

Does your dog know what you are thinking?

  1. The heart attack experiment could well be an example of how people have failed to design experiments properly, rather than showing something about dogs. For instance, the dogs might have known full well that nothing was wrong because they might be able to detect that nothing is wrong with their sense of smell (e.g. some dogs can tell someone’s going to have a seizure well before they have one).

    • I agree, the dogs know it’s a ruse.
      Most dogs can tell if you’re really asleep or just faking it. I’ve played this game with quite a number of dogs and they never fall for it. Fake snoring is likely to get you a tongue in your mouth whether you like it or not… and a lot of tail wagging.

    • my thoughts exactly!

  2. Not only do Dogs NOT know what YOU are thinking, They do not know what THEY are thinking because Dogs do not have the power of thought, they only know how to eat sleep, sh_t, and make other little Dogs. End of story!

    • It takes one to know one

    • You clearly know absolutely nothing about dogs mark.
      And if you hate dogs that much to be so arrogant in your statement
      Why are you reading an article about dogs????
      I’m sure there are plenty of other things for you to read… It’s the Internet!!!

    • well you just made yourself look rather stupid, clearly a dog is far more intelligent than yourself!

  3. You need to truly be in a state of panic for most dogs to think something’s wrong,
    Eg; being afraid of spiders, jumping around like a lunatic because a spider fell on you,
    My boy Toby very quickly came and got me to aid the person, he even growled at thin air and he didn’t know why the persons was panicking, nevertheless he still rushed to get me to come outside and help.

  4. Also on the heart attack experiment … why would a dog rush to rescue to a fake one? If they’re really so well in tune with their humans, wouldn’t they know it’s fake? I feel they would.

Sign in to comment.