A handful of the golden retriever pups used in a new study of pointing behavior

Canine Companions for Independence

Few scientific mysteries can be solved with the help of nearly 400 adorably naughty puppies, but a new study is a pleasant exception. Researchers have used the furballs to show dogs’ ability to understand human pointing—a rarity in the animal kingdom and key to social intelligence—appears to be hardwired in doggy DNA.

“Using puppies to answer this question is a great approach,” says Heidi Parker, a geneticist at the U.S. National Institutes of Health’s Dog Genome Project who was not involved with the work. “Behavior is the holy grail of dog genetics,” she says. Before scientists go searching for genes that may have turned dogs into our faithful companions, they need to make sure they’re there in the first place, she says. “I feel like this study shows that.”

Scientists have known for more than 2 decades that dogs understand the logic behind a surprisingly complex gesture: When we point at something, we want them to look at it. That insight eludes even our closest relatives, chimpanzees, and helps our canine companions bond with us. But it’s been unclear whether pooches acquire this ability simply by hanging out with us, or it’s encoded in their genes. “It’s the one piece of the puzzle we don’t have evidence for,” says Evan MacLean, director of the Arizona Canine Cognition Center at the University of Arizona.

Enter puppies. If social intelligence is genetic, dogs should display it at a very young age. And there shouldn’t be any learning required.

That’s what MacLean and his colleagues found. The scientists partnered with Canine Companions for Independence, which breeds dogs to assist people in the United States with post-traumatic stress disorder and physical disabilities. The group loaned the researchers 375 8-week-old Labrador and golden retriever pups: They were just old enough to participate in the experiments, but young enough to have had very little interaction—and thus experience or learning—with people.

Things didn’t go quite as pleasantly as they might sound. “Working with puppies is a lot like having young kids,” MacLean says. “It’s a balance between extraordinarily cute and rewarding moments, and frustration that leaves you at the brink of insanity. There is nothing that will not be chewed or peed on, including all of your research equipment, your clothes, and your body.”

The researchers put the puppies through three tests. First, they performed a classic pointing experiment, placing the young dogs between two overturned cups—one containing a treat—and pointing to the one with the treat (see video, above). The animals understood the gesture more than two-thirds of the time, approaching the performance of adult dogs. But they didn’t get any better over a dozen rounds, suggesting they were not learning the behavior, MacLean says.

In a second experiment, a researcher stood outside a large playpen and, for 30 seconds, engaged in the kind of high-pitched “puppy talk” familiar to almost anyone who has owned a dog: “Hey puppy, look at you! You’re such a good puppy.” The animals spent an average of 6 seconds staring at the person. Such eye contact is rare among mammals—including the pups’ ancestors, gray wolves—and it’s an important foundation for social interaction with people.

In a final test, the researchers taught the puppies to find food in a plastic container, then sealed it with a lid. In contrast to adult dogs, which usually give up after a few seconds and look to humans for assistance, the pups rarely gazed at their scientist companions for help. “Puppies seem to be sensitive to receiving information from humans,” as the other experiments show, MacLean says, “but they may not yet know that that they can solicit help from us.”

A scientist times how long a puppy looks at her face, an important foundation for social interaction.

Canine Companions for Independence

To confirm that the puppies’ successful behaviors were genetic, the researchers analyzed their pedigrees to see how related each dog was to the others. Then they compared this relatedness with the dogs’ performance on the tests. Approximately 43% of the variation in performance was due to genetics, the team reports today on the preprint server bioRxiv.

That’s on par with the heritability of cognitive traits like IQ in people, MacLean says. “It’s about as hardwired as things in psychology come.”

“It’s a really high number for a complex trait like behavior—it’s a pretty big deal,” agrees Noah Snyder-Mackler, an evolutionary biologist at Arizona State University, Tempe, who has collaborated with MacLean in the past, but was not involved with the current study. He says the finding suggests people strongly selected for these abilities in the past, paving the way for dogs to become the human mind readers they are today.

Parker notes that Labrador and golden retriever service dogs are “pretty amiable,” and she’d like to see the study replicated with a wider variety of breeds. She also says doggy social intelligence is likely spread out over hundreds of genes, which could make the exact DNA sequences hard to nail down.

But MacLean says he’s up for the challenge. His team has already begun a genomewide association study, which will scan puppies’ DNA to look for genetic variants linked to these social skills. What he finds in dogs will likely hold true for complex behaviors in other animals, he says, including humans. At the very least, the work will have one upside: DNA can’t pee on your leg.