Thinking about how dogs think

(Photo by Marvin Joseph/The Washington Post; Photo illustration by Betty Chavarria/The Washington Post)
Placeholder while article actions load

Back in 2002, when Alexandra Horowitz was working toward her PhD at the University of California at San Diego, she believed that dogs were a worthy thing to study. But her dissertation committee, which favored apes and monkeys, needed convincing.

“They were primate people,” she said. “They all studied nonhuman primates or human primates, and that’s where it was thought that the interesting cognitive work was going to happen. Trying to show them that there would be something interesting with dogs — that was a challenge.”

Oh, how things can change in just two decades, especially in a nation that includes about 90 million dogs among its residents — everything from beloved pets to working dogs doing all kinds of tasks, from sniffing out drugs in airports to assisting blind people with crossing a street. Today, Horowitz is a senior research fellow at Barnard College in New York City, where her specialty is dog cognition: understanding how dogs think, including the mental processes that go into tasks such as learning, problem-solving and communication. Dog cognition is now a widely respected field, a growing specialty branch of the more general animal-cognition research that has existed since the early 20th century.

“This field, and animal cognition, really, is all within our lifetimes,” Horowitz said. “It’s not as if nobody ever looked at dogs, but they weren’t looking at their minds.”

Looking at dogs’ minds, so far, has revealed quite a few insights. The Canine Cognition Center at Yale University, using a game where humans offer dogs pointing and looking cues to spot where treats are hidden, showed that dogs can follow our thinking even without verbal commands. The Max Planck Institute for Evolutionary Anthropology in Germany figured out that dogs are smart about getting what they want — they will eat forbidden food more frequently if humans can’t see them. Researchers from Austria, Israel and Britain determined that seeing a caregiver, versus a stranger, activated dogs’ brain regions of emotion and attachment much as it does in the human mother-child bond. Other European researchers showed that negative-reinforcement training (like jerking on a leash) causes lingering emotional changes and makes the dog less optimistic overall.

Read more stories of dogs, humans and the relationship they share

Some dog owners hear about this type of research and think: “They did a whole study to figure out that my dog looks where I point? I could have told you that.” But the studies aren’t just about what a dog is doing. They’re indicating areas to research so that we can better understand why and how the dog is doing it — in other words, what’s happening inside the dog’s mind.

“Maybe they’re not looking at your finger at all. Maybe they’re paying attention to your face and not to your hand,” said Federico Rossano, whose team at the University of California at San Diego is trying to determine whether dogs can translate their thoughts into words that humans can understand through a language device. “A lot of this becomes interesting in terms of how you can train them better.”

An evolving area of research

Right now, with no organizing body in the field, it’s hard to say exactly how many people are doing dog-cognition research. You can count on two hands the number of dedicated university spaces led by professors with graduate students and funding grants. When the leaders from those places get together once a year, it’s usually at someone’s home.

But researchers at universities doing studies on dogs? There are now many dozens of those, and there’s no lack of students wanting to at least dabble in the work.

“The thing that gets my students all abuzz is that people always want to know whether their dog loves them back,” said Ellen Furlong, associate professor of psychology at Illinois Wesleyan University and leader of its Dog Scientists Group.

Every semester, on the first day, she asks students if their dogs are happy. It’s her way of helping them understand why the study of dog cognition is important.

“They’re always kind of offended — ‘Of course my dog is happy. I love my dog,’ ” she said. “But then you dig a little bit and push them and say: ‘Your dog’s life is different from your life. You get to decide when your dog gets to eat and play and go outside. You decide everything about your dog’s life, but your dog isn’t human. They have different wants and needs than you do.’ They have a semester-long assignment where they have to consider how their work on cognition can help to design some enrichment activities to improve the dogs’ lives.”

With restrictions lifting and schedules changing, both pets and humans are experiencing separation anxiety. (Video: Monica Rodman, Sarah Hashemi/The Washington Post)

The topics that dog-cognition researchers focus on today often are chosen based on personal interests. While Furlong is most curious about ethics, welfare and how humans can meet dogs’ psychological needs, Horowitz is focusing her research on what dogs understand through smell. At the Duke Canine Cognition Center in North Carolina, Brian Hare is trying to determine — when a dog is still a puppy — whether the way a dog thinks might make her a good candidate for different jobs as an adult.

“We’re saying, ‘Here are some cognitive abilities that are critical for training for these jobs,’ ” Hare said. “It’s a little bit like talking about personality, but we’re talking about your cognitive personality, in a way. Maybe you have a really good memory for space, or maybe you’re good at understanding human gestures. The question is whether we can identify some of these dogs really early, in the first two to three months of life, who will do well in these programs.”

How the research is done

One example of dog cognition research with a potential training application is a study that Horowitz did on nose work — an activity that lets dogs use their natural abilities with scents to find everything from a treat hidden under a cone to marijuana in somebody’s suitcase.

Horowitz and her team showed the dogs three buckets and taught them that one of the buckets always had a treat under it, and one did not. Then she measured how quickly the dogs went to the “ambiguous bucket” in the middle.

The dogs then attended nose-work classes. These types of advanced classes are widely available at the same types of schools that teach basic obedience. In the nose-work classes, dogs are encouraged and trained to use their noses to search for and find treats or favorite toys that are hidden under boxes or cones, inside suitcases or in other places.

After a few weeks of nose-work classes, Horowitz repeated the bucket test.

“What we found was the dogs in the nose-work class got faster at approaching ambiguous stimulus,” she said, adding that the results suggest that for some dogs, taking nose-work classes could help them feel more optimistic. “The group that had nose work changed their behavior afterward, so I have to say it’s something about the nose work. I don’t know exactly what it was, but if the effect is profound and we keep seeing it, we would go in and try to see what it was that made it useful for the subjects.”

Hare is widely credited with having jump-started America’s dog-cognition research field. In the late 1990s as an undergraduate, he was doing research with chimpanzees when he realized they couldn’t do something that his dogs could do: follow a human’s pointing gesture to find food. Chimpanzees are the closest animal relatives humans have, and dogs could do something they couldn’t. Researchers suddenly wanted to know why dogs could understand something that chimpanzees could not.

In his most recent study, published in July, Hare and his team looked at the difference between wolf and dog pups. There had been some debate in the dog-cognition field about where dogs’ unusual abilities to cooperate with humans originate — whether those abilities are biological or taught. So the team gave a battery of temperament and cognition tests to dog and wolf puppies that were 5 weeks to 18 weeks old. The pups of both species were given the chance to approach familiar and unfamiliar humans to retrieve food; to follow a human’s pointing gesture to find food; to make eye contact with humans, and more. The team found that even at such a young age, the dog pups were more attracted to humans, read the human gestures more skillfully, and made more eye contact with humans than the wolf pups did.

The conclusion? The way that humans domesticated dogs actually altered the dogs’ developmental pathways, meaning their abilities to cooperate with us today are biological — a research result that is likely to have many practical implications.

“It’s highly inheritable, and it’s potentially manipulatable through breeding,” Hare said, adding that dogs might be bred to specialize in certain types of thinking. The finding opens up the idea of studying dogs in ways that could make deep-pocketed entities like the U.S. government want to fund more dog-cognition research, Hare said.

By way of example, he talked about dogs he has worked with for the U.S. Marine Corps, compared with dogs he has worked with for Canine Companions for Independence in California. The Marines needed dogs in places like Afghanistan to help sniff out incendiary devices, while the companions agency needed dogs that were good at helping people with disabilities.

Just looking at both types of purpose-bred dogs, most people would think they’re the same — to the naked eye, they all look like Labrador retrievers, and on paper, they would all be considered Labrador retrievers. But behaviorally and cognitively, because of their breeding for specific program purposes, Hare said, they were different in many ways.

Hare devised a test that could tell them apart in two or three minutes. It’s a test that’s intentionally impossible for the dog to solve — what Star Trek fans would recognize as the Kobayashi Maru. In Hare’s version, the dog was at first able to get a reward from inside a container whose lid was loosely secured and easy to dislodge; then, the reward was placed inside the same container with the lid locked and unable to be opened. Just as Starfleet was trying to figure out what a captain’s character would lead him to do in a no-win situation, Hare’s team was watching whether the dog kept trying to solve the test indefinitely, or looked to a human for help.

“What we found is that the dogs that ask for help are fantastic at the assistance-dog training, and the dogs that persevere and try to solve the problem no matter what are ideal for the detector training,” Hare said. “It’s not testing to see which dog is smart or dumb. What we’ve been able to show is that some of these measures tell you what jobs these dogs would be good at.”

What comes next in the field of dog-cognition research is probably a bit more of everything. Some researchers are following their interests, while others are following the research grants. Those grants can come from a wide array of sources, including the government trying to help soldiers with post-traumatic stress disorder, shelters trying to rehome animals and neuroscience institutes looking for insights across species.

“It’s a really exciting moment,” Hare said. “I think we can continue on with individual researchers pursuing fun, interesting things — the students and the universities love it — but most successful academic endeavors have two parts. Being intellectual is wonderful, but that kind of research tends to struggle with funding. Academic endeavors with practical application tend to be incredibly well funded, and then the field grows.

“If you can have both of those things, then it will grow, and it will grow phenomenally,” he added. “If it’s just, ‘We’re going to do this because people love dogs,’ that’ll be fun, but it will stay small like it is now.”