Humans Can’t Quit a Basic Myth About Dog Breeds

Breed doesn’t have that big an effect on a dog’s personality.

a row of puppies
Bettmann / Getty

After four decades of training and studying dogs, Marjie Alonso has lost track of the number of pets she’s seen because their humans felt they weren’t acting as they “should.” There were the golden retrievers who weren’t “friendly” or “good enough with kids,” and the German shepherds who were more timid scaredy-cats than vigilant guard dogs. There was the Newfoundland (who later turned out not to be a Newfoundland) who had been adopted to fulfill a Peter Pan–esque fantasy of a devoted dog nanny, but acted so aloof that his owners put him on meds. And then there was the horde of Shih Tzus, acquired by a woman who was “super pissed,” Alonso told me, to find the little dogs regularly escaping her home and terrorizing her neighbors’ yards—nothing, she complained, like the regal pooches whose “idea of fun is sitting in your lap acting adorable as you try to watch TV,” as advertised by the American Kennel Club.

Alonso, who’s now the executive director of the International Association of Animal Behavior Consultants Foundation, gets it; she really does. Stereotypes about breed “personalities” are hardwired into almost every interaction people have with dogs: They influence which canines are adopted first, which are routed into service jobs, which are allowed to inhabit apartment buildings. Breed is one of the first things people ask about a dog, and the answer has a way of guiding how they’ll treat that animal next. Which is exactly the problem. “Any good dog trainer will tell you those stereotypes are a disaster,” says Marc Bekoff, a dog-behavior expert at the University of Colorado at Boulder. “Breeds don’t have personalities. Individuals do.”

Maybe that logic seems obvious. Of course dogs’ behavioral quirks, much like those of humans, aren’t mere products of genetics or pedigree; of course experiences factor in. Even Brandi Hunter Munden, of the AKC, which details breed personalities on its website, acknowledged to me in an email that “every dog is different.” And yet, breed—a concept predicated on purity, sameness, predictability passed from parent to pup—is an undeniably powerful force in dogs. “I don’t think you’re ever starting from a totally clean slate,” says Gita Gnanadesikan, an evolutionary biologist at the University of Arizona.

Dogs are, in part, “a human creation—they are not something that existed before us,” says Isain Zapata, of Rocky Vista University, in Colorado, who’s studied dog genetics and behavior. And over millennia, we have sculpted them to fit a multitude of functions and forms. Purebred canines are a product of human preferences and prejudices; they should be expected to have certain proclivities written into their genes, exactly because humans decided they should. “Does breed matter? Does breed not matter?” says Kathleen Morrill, a dog geneticist at the Broad Institute and UMass Chan Medical School. “Really, it’s both.”

Experts agree that dog behavior is the product of a multitude of factors—among them genes, development, socialization, and environment; they disagree on the ratios, the measurements, the ways in which they swirl together. The key ingredient in every recipe, though, is always us: the people who fancy themselves the arbiters of what makes a dog a dog. The sway of breed, even over personality, is not fiction—our species has made sure of that. But its influence isn’t just over dogs, and it’s not as simple as we might like to think.

Many different versions of the dog origin story exist (and more than one may be true), but the gist tends to go like this. Some tens of thousands of years ago, wolves and humans started spending a lot more time together, and began to coevolve. It’s not clear who made the first move—maybe it was the canines, lured into encampments by their noses; maybe the two species just found themselves thrown together, and bonded over a mutual love for meat. In any case, the chillest and chummiest canines of the bunch kept coming back. At first perhaps lopsided, the relationship soon became more mutually beneficial: People realized that dogs could enhance humans’ ability to feed and protect their families, and eventually, corral their sheep and cattle; the animals would do this in exchange for calories, shelter, and maybe some well-earned belly rubs.

The first chapter of the dog-human relationship, then, was about function. People noticed behaviors they liked in the animals, and started to favor them, “maybe giving them extra food, giving them a chance to breed,” says Kathryn Lord, a dog-behavior-and-evolution expert at UMass Chan and the Broad Institute, where she’s working with Morrill. Slowly, a wolfish lineage shed some of its fear of people, and some of its grump; it lost the sharpness of its lupine features, and its apex-predator edge. Even the animals’ tightly tuned predatory sequence—search, stalk, chase, grab, kill—fractured, yielding groups of dogs that specialized in, for instance, stalking and sprinting (herders), pursuing and catching (retrievers), all of the above (terriers), or none of the above (livestock guardians). Under the pressures of employment, dogs diversified.

Then, in the 1800s, dog-rearing underwent a massive pivot. “Victorians changed the way we think about dogs,” says Michael Worboys, a historian of science at the University of Manchester, in England, and an author of The Invention of the Modern Dog: Breed and Blood in Victorian Britain. It was in this period—the era of “fancy everything,” as Lord puts it—that the modern concept of breed was born. Suddenly, people prized dogs more for their looks than the tasks they could perform. Puppeteering the sex lives of dogs became ultra-purposeful, ultra-fashionable; the idea of breed became so valuable that it needed to be policed by stringent criteria and formal clubs. And as the goalposts shifted to achieving purity of blood and physical ideals, canine evolution bent fast. “Once you start selecting on form—your coat color, your shape,” Lord told me, “it’s so much more powerful than selecting on what it behaves like.” The number of distinct breeds ballooned, and the dogs within them grew more and more alike.

Nowadays, that uniformity seems a scientific dream: Purebred genomes have been stripped of much of diversity’s noisiness, making patterns within groups easier to spot; with the dog genome sequenced, it should be easy to go in and figure out how human meddling has, at various points, cemented both physical and behavioral propensities into DNA. But behavior is extremely complicated—sometimes involving many, many genes that may each have only a small influence—and it’s been repeatedly wrung through humans’ changing ideas about what makes for a very, very good boy.

To untangle the gene-behavior snarl, researchers first need to find lots and lots of dogs—thousands, tens of thousands, the more the better—to represent enough diversity in both behavior and genetics, and draw connective tissue between them. Morrill, Lord, and their colleagues recently wrapped up one such gargantuan study, one of the most sprawling and in-depth to date. They distributed behavioral surveys to the human companions of roughly 20,000 dogs, asking the same sorts of queries that psychologists use to suss out personality in people, with a canine-focused kick: Does your dog behave fearfully toward unfamiliar people? Cower during storms? Ignore commands? Get pushy with other dogs? They then whole-genome-sequenced the saliva of about 2,000 of those dogs, and searched the DNA for signatures that might help explain the owners’ answers. Unlike other studies of its ilk, Morrill’s also took care to enroll a lot of mutts—dogs whose appearance and personality are “naturally shuffled up,” she said.

The team’s findings confirmed that some aspects of canine behavior do seem quite heritable—and sometimes even echo kennel-club dogma. Work, it turns out, is a pretty good motivator, and several of the traits with genetic ties were probably the ones that kept a lot of early dogs employed. Many herder dogs, for instance—border collies and the like—remain very herder-y. They’re still, on average, more likely than other pups to comply with human commands, be curious about their surroundings, and make an enthusiastic bid for toys. Retrievers, too, seem to have fetching written in their genes; it was “the most heritable behavior we could find in our study,” Morrill said. A few other patterns might be similarly rationalizable: Great Pyrenees, which originated as livestock guardians, tended to be tougher to rattle than other dogs. Beagles, historically tasked with zooming after prey, generally trended toward being headstrong. Centuries of career focus have clearly left a legacy in canine genes. And in at least a few ways, some breeds are still what humans bred them to be.

The further behaviors drift outside the professional arena, though, the trickier they can be to assess, and the trickier their genetic roots are to nail down. Owners might reliably describe how their dogs go bounding after balls, but they might have a less objective sense of whether their canine is especially apt to be calm or skittish, aloof or clingy, assertive or easygoing—all traits that can fall victim to the vagaries of human perception. Scientists seeking a bit more objectivity will sometimes try out laboratory experiments: Someone interested in gauging timidness, for example, can put a dog in a pen with an unfamiliar object, such as an unnervingly plush robotic cat, and watch how much it wigs. But not all behavioral quirks, or dogs, lend themselves to tests in a weird building staffed by strangers, where it’s easy for animals to completely lose their cool. And few researchers are willing to do that thousands of times over for a giant genetic study, already weighed down by time and cost. None of that makes behavior data useless—just tougher to interpret, and then straightforwardly explain.

But, other experts told me, with so many dogs banked in their study, Morrill and her colleagues seem to have hit upon some solid connections. And at least as important as any of the genetic trends they found, Morrill said, were the ones they very much did not. In the end, they couldn’t find a single behavioral trait that was either absent from any of the breeds they surveyed, or present in every dog within a given breed. Sure, greyhounds, on average, were more blasé about toys than other dogs, while German shepherds seem genetically poised to think they’re swell. Lots of Chihuahuas are trembly little nubbins, many Brittanys eat their own poop, and Shiba Inus, as a group, will not be the first to jump in your cuddle puddle. All of these, however, are predilections, not prescriptions. There exist border collies who won’t herd, and pugs who will; there exist high-strung Great Pyrenees, and beagles who will obey every command. Breeds may draw loose borders around dog behaviors. But the boundaries are far-flung and poorly patrolled—easy for an individual to stray outside of, if the conditions are right.

According to Morrill’s team, breed explains just a small fraction of the mind-boggling variation in behavior seen in the species that is dog—less than 10 percent. Which is to say, most of the mishmash can be attributed to something else. That might seem like too small a proportion to some; other, older studies, which analyzed their own data somewhat differently, have made higher estimates. The AKC, unsurprisingly, isn’t totally on board with what the new study found. In a statement, the organization reiterated that “breed and type of dog does inform about general and instinctual behavior,” and said it thinks owners should let those penchants guide decision making. Elinor Karlsson, the computational biologist who led the study, doesn’t quite agree. “You might be able to take a random dog off the street and make a prediction about it based on its breed, and you might be right more often than if it was a totally random guess,” she told me. “But it’s not going to be particularly effective.”

None of this means that intel on a dog’s breed is worthless. Encoded into breed is still a rich lineage, a history, a wealth of information about how a dog’s body will be built and how it will be set up to navigate its surroundings. Purebred dogs will still, generally, look a certain way. They may even be more likely to behave a certain way. Those propensities just have to contend with the real world, not just once, but over and over and over again. These collisions fascinate Flavio Ayrosa, of the University of São Paulo, who has studied how a dog’s height, weight, and nose size, alongside factors such as genes and socialization, might affect its temperament. Small dogs have a different experience of the world than big ones; long snouts bisect a dog’s vision in a way that short schnozzes don’t. All of that matters. “These morphological factors affect how an animal will interact with its environment,” he told me.

Breed information can also help set expectations for what certain dogs might need to stay happy and healthy, and what might be physically feasible for them. Dogs are flexible, but not infinitely so. “You cannot make a Chihuahua race with greyhounds, or make a Chihuahua a sled dog,” says Carlos Alvarez, a geneticist at Nationwide Children’s Hospital, in Ohio. It helps, too, to consider the professional legacies that humans have seeded across the dog family tree. Even allowing for individuality, someone who wants a dog content to stay cooped up in a studio apartment all day might be taking a gamble by adopting one whose ancestors were sprinters and chasers.

The point, then, isn’t to discount breed’s influence over dogs, but to rethink its sway over us. People who go after particular dog breeds may do so under the pretense that their new pets will act a certain way. And then they treat them as such, emphasizing and exaggerating the very behaviors they wanted out of their dogs in the first place, while suppressing others. They teach a “clever” dog more tricks because they assume that the animal will learn them; they give an “aloof” dog more space because they figure their pet needs the time alone. Stereotypes become “self-fulfilling prophecies,” Bekoff, of CU Boulder, told me. Dog behaviors are what we breed them to be, but also what we expect them to be. “How much of how breeds behave is how we behave toward breeds?” Alonso, of the International Association of Animal Behavior Consultants, said. “That is the million-dollar question.”

The answer, in fact, may be: a decent bit, at least for some dogs. In the UMass-Broad study, purebred golden and Labrador retrievers “tended to score exceptionally,” Karlsson told me, on people-friendly metrics—exactly as the AKC website says they should. But those effects evaporated when her team turned their lens to mutts with retriever ancestry, who are harder to typecast by appearance alone. (Most people, by the way, aren’t actually that good at correctly guessing a dog’s lineage.) Even after the researchers accounted for the mutts’ mixed heritage, they found that the part-retrievers weren’t any more eager to mingle than the average pooch.

While friendly dogs are considered desirable, the stereotype pendulum swings the opposite way with aggression, a murky label that some behaviorists dislike and yet is often inappropriately pasted onto dogs who then end up banned from housing complexes, abandoned in shelters, even euthanized based on breed alone. Dogs who fall into the pit-bull category are a famous, and particularly controversial, example of this: Bred to fight other animals, they’ve acquired a reputation for violence and unpredictability, a stigma worsened, scholars have argued, by racism against America’s urban Black and Latino communities, to which the dogs were culturally linked in the mid-20th century. Some experts argue that caution around pit bulls is warranted, given their history; people who look at pictures of the dogs tend to rate them unfavorably. And yet, studies done by Alvarez, Zapata, and others have found that pit bulls don’t seem to be more aggressive or volatile than other dogs. If any dogs are a bit more apt to react when provoked, Alvarez told me, it might be the shrimpy ones—Chihuahuas, dachshunds, and so on—perhaps because their teenier brains have a harder time reining in impulsive behaviors … or because they’re just smaller dogs, constantly being loomed over, picked up, or accidentally kicked.

Plenty of personality descriptors leashed to the dog world might seem far more benign. They’re almost horoscopic in nature—snazzy yet vague and catch-all: Shelties are bright; Boston terriers are amusing; Yorkshire terriers are tomboyish; toy poodles are self-confident; Clumber spaniels are gentlemanly; Chow Chows are serious-minded. Hunter Munden, of the AKC, defended her organization’s descriptors, explaining that they’re usually drawn directly from breed standards—detailed criteria that lay out the traits for a breed’s “ideal” specimen and include temperament and behaviors that are, she said, “innate to a breed.” But according to experts such as Ádám Miklósi, a dog-cognition researcher at Eötvös Loránd University, in Hungary, terms like these are ridiculous anthropomorphizations, so squishy that they’re meaningless, meant more to market dogs than accurately describe them. And they can verge on harmful when they set expectations that can’t be met.

Assumptions aren’t easy to shake. Even Alonso, who started working with dogs in the late ’70s, still hasn’t quite broken her “decades-long habit yet of ascribing certain behavioral traits to breeds,” she told me. She never faults individual dogs for wandering beyond a stereotype. She’s just not that surprised when she encounters one who fits its own breed bill: an overprotective Akita, a cattle dog who just won’t stop chasing the kids. “I know it’s wrong,” she told me. “I’m still wrestling with myself.”

Maybe that’s simply human nature. We buy into the idea that personality is ultra-heritable, because it supports the idea that we made dogs the way they are, that we warped wild wolves into workers, guides, companions, and teammates, with personalities as transparent as Tinder profiles. There is a comfort in the notion that dogs are predictable, categorizable, easy to bin into the boxes we have created for them; that in their behaviors are the motivations and emotions, the University of Manchester’s Worboys told me, that we humans feel. Many of the descriptive traits we cling most to in dogs—loyal, friendly, loving—mirror the traits that “we want to associate with ourselves,” Zapata told me. But dogs are their own animals, as individual as we humans are. Dogs, like humans, can buck the trends of heritage. Dogs, like humans, can shift over a lifetime, trading bad habits for good ones. Dogs, like humans, are “continually changing systems, always in development,” Ayrosa, of the University of São Paulo, told me, from conception to death. And dogs, like humans, can alter the trajectory of other species, as they have with us.

Alonso’s beagle, Nellie, scored about as stereotypically beagle as they come on the surveys used in the UMass-Broad study. But she refuses to be captured by a single data point. She and Alonso met six years ago, when Nellie was a “separation-anxious, resource-guarding, bitey, anxious dog,” Alonso told me—so riddled with tough-to-take behaviors that she’d been adopted, then unadopted, by multiple owners before. Now Nellie’s fine being left alone at home for several hours at a time. She never bites. She’s welcoming to humans and other dogs alike. She’ll even share her food, unless french fries are involved. And she soothed Alonso’s pain after her two sons left for college and her two pre-Nellie dogs died. The pair have shaped each other, exactly as companions would be expected to do.