Puppies Are Born Ready to Communicate With Humans

A new study finds very young dogs with little human contact can understand pointing gestures—and that the ability has a strong genetic basis

Black Labrador puppy
A young puppy responds to a human pointing to a treat during an experiment conducted by scientists at the University of Arizona. Canine Companions for Independence

Dog owners might not be too impressed when they’re able to point out a fallen piece of chicken or a thrown stick to their pooch, but dogs’ ability to follow that seemingly simple gesture places them in rare air in the animal kingdom. Some research suggests that even chimpanzees, our closest evolutionary relatives, don’t understand pointing as well as dogs.

For decades, researchers have debated whether dogs obtain their ability to understand pointing by spending time with humans and learning it or if our furry companions are born with a capacity to comprehend this deceptively complex feat of communication.

Now, a new study, published today in the journal Current Biology, finds that even 8-week-old puppies with little exposure to humans can understand pointing and show sophisticated levels of social cognition in other tests. On top of that, the study found that each fluffball’s genetic makeup was a strong predictor of its ability to follow a pointed finger to a hidden treat as well as the pup’s tendency to pay attention to human faces.

Emily E. Bray, a psychologist studying animal behavior at the University of Arizona and lead author of the study, says that these feats of canine cognition are about as genetically based, or heritable, as human intelligence. "This all suggests that dogs are biologically prepared for communication with humans," she says.

Finding a genetic basis for dogs’ social intelligence fills in a big unknown in the story of how they became domesticated and could one day help breed better service dogs—which need to be whizzes at reading human cues, says Evan MacLean, a comparative psychologist at the University of Arizona and co-author of the study.

The team behind the study had two main questions. The first was whether young puppies who had yet to spend any significant time with humans were sensitive to human attempts to communicate. The second was whether there was a genetic basis for the puppies’ social smarts.

In 2017, Bray and her co-authors started testing 375 eight to ten-week-old Labrador and golden retriever puppies from Canine Companions for Independence, an organization that breeds service dogs to support individuals with disabilities or those suffering from post-traumatic stress. Crucially, Canine Companions also keeps extensive information on the pedigree of each dog going back decades, which facilitated the genetic aspect of the study.

At this age, the puppies had spent nearly every minute of the day either with their mom or with their littermates. This made the pups perfect collaborators for trying to get at the question of whether dogs’ feats of social cognition were learned or innate. However, on a more practical level, puppies were not always a dream to work with for the scientists.

“They’re adorable and it’s fun to work with them,” says MacLean. “But they’re puppies, they have short attention spans and they pee and poop on everything. At the start of this project, it was like, ‘Puppies!’ And by the end it was, ‘Puppies.’”

For the study, the researchers ran the young dogs through a battery of four tasks designed to test their social cognition.

Puppies playing games for science

The first two tasks were aimed at assessing the puppies’ abilities to understand human gestures. Out of view, the researchers placed a treat under one of two upside-down cups. When the researcher revealed the cups they either pointed at the one hiding the food or, in the second version, showed the puppy a small yellow cube that was then placed in front of the cup concealing the reward.

In both tasks, the puppies made their way to the right cup an average of 67 percent of the time, much better than the 50 percent accuracy one would expect if chance alone was responsible for the correct answers.

The researchers also made sure the pups weren’t just following their noses by taping a bit of kibble inside both cups to ensure they both had the smell of a treat and by conducting what they called an “odor control” test that involved no pointing. The dogs only got around 49 percent of their guesses correct in this test, suggesting the human gestures were the deciding factor in their improved accuracy in the other trials.

MacLean says that despite being just eight weeks old on average, the puppies could follow human gestures about as well as adult dogs. Moreover, each furry test subject had to perform the task upwards of 12 times and their cup-picking accuracy stayed quite consistent from the first trial to the last, meaning they didn’t improve with experience. “However they’re solving this problem they’re doing it above chance from the first exposure and they’re not getting better across time,” says MacLean. “That says they’re ready to do this and don’t need to learn it.”

A third task gauged the puppies’ tendency to pay attention to human faces by having an experimenter recite a 30-second script in a high-pitched voice similar to baby talk while gazing down at the dog. Researchers found the test resulted in an average of six seconds of attention to the experimenter’s face. Adult dogs tend to make even more eye contact with humans, suggesting that human faces are something they learn to attend to even more over time.

Puppy
Puppies in the experiment were tested to see how long they responded to baby talk by researchers. Canine Companions for Independence

The fourth and final test was what the researchers called the “unsolvable task.” For this trial the researchers presented puppies with progressively more difficult to access treats inside a plastic container. Eventually the researchers made it impossible to get the food out to see if the puppy might look to the nearby human for help—a behavior that has been well-documented in adult dogs. In this trial the puppies mostly ignored the nearby person, only looking for an average of about one second, suggesting that puppies aren’t born with an instinct to look to humans for help but rather learn that behavior as they interact more with our species.

The researchers then combined the puppies’ results across these four tasks with the extensive pedigree information on each animal provided by Canine Companions. This allowed the team to assess whether each dog’s family tree, and thus their underlying genetics, provided a better statistical explanation for their performance on the tasks than other factors including the puppy’s breed, sex, age and rearing location.

Success in the pointing task as well as a puppy’s tendency to look at a human face during the 30-second script were highly heritable, according to the paper. More than 40 percent of the variation in performance was due to genetics.

“Around 40 percent heritability is an incredible number,” says Bridgett vonHoldt, a geneticist studying dogs and wolves at Princeton University who was not involved in the paper. “In well-studied animals like pigs and red squirrels the heritability of behaviors is in the 20s or 30s, and 40 is very high for a trait like reading social cues that is probably very complex.”

Performance on the other two tasks was much less heritable, which MacLean says tells us that not all of these traits have an equally strong genetic component.

This isn’t the first study to suggest that the ability to follow human social cues like pointing might have a genetic basis. But, according to Monique Udell, a psychologist studying human-animal interactions at Oregon State University who was not involved in this research, this study is the first to show, with a large group of dogs and with known levels of relatedness between them, that the animals don’t have to acquire these traits through learning and that some forms of social cognition do indeed have a substantial genetic component.

“This paper gives us very strong evidence that following human social cues was likely a key trait that was selected for during the canine domestication process,” says Zachary Silver, a comparative psychologist at Yale University who was not involved in the paper. “But as exciting as these results are, they don’t tell us what the mechanisms are either cognitively or genetically.”

Bray says she and her co-authors are already at work on a follow-up genomic study of dogs from Canine Companions that will search for genes that correlate with the same types of social cognition explored in the current paper.

Beyond potentially filling in a missing piece of the domestication story, identifying the genetic basis of this social skill set in dogs could one day help us breed even more successful service dogs, says MacLean. “About half the dogs that enter training programs to become service dogs don’t complete them,” he says, “so figuring out which dogs will excel in those roles has the potential to save resources and help people.”

Get the latest Science stories in your inbox.