In The Minds of Dogs
They were the first domesticated animals and, bred for their ability to serve human needs, have co-evolved with us. Eons of association have turned once-wolves into remarkably social creatures. Just how much do they know about us, the world around them, and themselves?
By September 5, 2017 - last reviewed on September 13, 2017published
Conversing with Canines
Think Cat in the Hat for talking to dogs.
If you want to start an argument among psychologists, behavioral biologists, and next-door-neighbor dog owners, just ask the question: Do dogs understand and use language? The argument tends to focus on whether dogs understand the words and expressions that humans use. A related concern is whether dogs use their various barks, growls, whines, and whimpers, combined with tail wags, body postures, and ear positions, to communicate with people as well as with one another.
Some scientists argue that dogs are more attuned to the emotional aspects of our word sounds than their actual meaning, and that their own signals are just visible expressions of their emotional state. Accordingly, any information such signals communicate about a dog and its intentions is just a byproduct, and those signals provide no more evidence of language ability than does our capacity to understand that other humans are happy because they're smiling or are angry because they're scowling.
With the right tools, it's possible to explore what dogs are capable of cognitively. The study of animal cognition in general, and dog cognition in particular, is now a growth industry.
In the early 1990s, it dawned on me that one of the ways to learn whether dogs actually have language was to deploy tests already developed for assessing human children—and simply modify them for use with dogs. I borrowed the MacArthur Communicative Development Inventory, which assesses language ability in very young children in terms not only of words but also gestures. When someone points a finger and we know that they are trying to communicate the location of something of interest, that's a linguistic gesture. An individual demonstrates an understanding of such an elementary message by looking or moving in that direction.
My data led to the conclusion that the average dog can learn to recognize about 165 words and gestures. "Super dogs"—those in the top 20 percent of canine intelligence—can learn 250 or more.
Other scientists soon tested my predictions. One study showed that a border collie named Rico is able to recognize more than 200 words. Perhaps the most linguistically advanced dog so far is another border collie, named Chaser. She is owned by a retired psychologist, John Pilley, and her vocabulary is around 1,000 words. What's more, Chaser understands some of the basics of grammar involved in simple sentence construction and seems to infer intention.
Evidence from testing dogs suggests that language is not an ability possessed only by humans. The knowledge that dogs have basic language skills offers further insight into the canine mind. The test scores I recorded allowed me to assign each dog a mental age representing the animal's cognitive ability. Dogs have a mental ability roughly equivalent to a human toddler age 2 to 2-and-a-half. Super dogs like Chaser have minds that might be similar to that of a 3-year-old child.
Tests of canine language ability offer a new way of looking at dogs' mental skills. If a problem can't be solved by a 2- to 3-year-old child, then it is not likely that a dog can solve it either. And if a training technique won't work for a toddler, then it likely won't work for a dog. —Stanley Coren
Stanley Coren, PH.D., is a professor emeritus of psychology at the University of British Columbia, Canada, whose research has focused on human cognition as well as dog intelligence. His latest book is Gods, Ghosts, and Black Dogs: The Fascinating Folklore and Mythology of Dogs.
Taking the IQ Test
One of the hardest tricks is coming up with a way to measure dog intelligence.
Humans have language and are mostly willing to follow the sonorous imperative,"You may now turn over your test papers." Still, it took a while to develop reliable IQ exam questions. In species less amenable to turning over—and more given to eating—said papers, the task of creating reliable test items is significantly harder.
Last year, Mark J. Adams and I published a study of 68 border collies to whom a set of six tests had been administered. We wanted to know whether dogs' cognitive abilities "hang together" the way they do in people. Four of the tests were related (from a human perspective). They comprised various barriers around which each dog had to navigate to find food. A fifth test ascertained the dogs' capacity to discriminate between quantities (to choose the bigger or smaller snack). A final test assessed their ability to understand and respond to a human gesture, specifically, a pointing arm directed at one of two inverted beakers, each covering a food reward.
We found a tendency for dogs who were better at one task to be better at others, and dogs who were faster were also more accurate. Three correlated elements—detour time, choice time, and choice accuracy—provided evidence that in dogs, as with people, cognitive abilities are associated with each other at the trait level. And as with humans, there appears to be an underlying factor exerting general influence on cognitive processes—a canine general IQ, or g-factor. The bottom line: Some dogs are smarter than others. This may sound obvious, but it has to be established empirically.
How does training fit in? In our sample, all the dogs were working farm dogs, so they had received similar training exposure. But training does not make all dogs alike. As with humans, brighter dogs learn new tricks faster. Famous dogs like Betsy, who could pick up a new word after two exposures, have had countless hours of training lavished on them, but it seems likely they were all smart dogs to begin with.
As well as being smart, a highly trainable dog must be biddable. Personality and test performance are not easy to decouple with dogs because, like Bartleby the Scrivener, a dog who would prefer not to simply does not. Such recalcitrance is somewhat awkward for psychometricians working with dogs (and with nonhuman animals more generally). It would be nice to be able to cleave cleanly between intelligence and other aspects of canine behavior, such as motivation and obedience. Yet we cannot hang over their heads the threat of tanking on an SAT. We have to go with food bribes instead.
What are the properties of an ideal test item for assessing canine intelligence? All dogs should be able to do it to some extent; it must reflect mental ability, not motor skills or training; it should have a graded outcome that is not simply pass or fail. But we should crack measuring dog intelligence because dogs are a great model for learning how cognitive abilities are associated with constellations of traits such as health, dementia risk, lifespan, and biological fitness. Among humans, for example, intelligence predicts health. Since dogs' outcomes are not subject to influence from the big hitters of epidemiology—smoking, alcohol, and drug abuse—they are terrific animals to partner with. In addition, their propensity to acquire naturally some of the same diseases that we suffer from (including dementia) makes learning about dog cognition a research priority. —Rosalind Arden
Dogs know how to have fun, and encoded in their antics is a deep understanding of fair play.
We've all seen it. When dogs play, they look as if they're going crazy, frenetically wrestling, mouthing, biting, chasing, and rolling over, and doing it again and again until they can hardly stand. They use actions like those seen during fighting or mating in random and unpredictable ways. But play sequences don't reflect the more predictable patterns of behavior seen in real fighting and mating. The random nature of play is one marker that dogs are indeed playing with one another. They know it and so do we.
Despite vastly different shapes, sizes, speeds, and strengths, dogs play together with such reckless abandon—flying around, tumbling, tackling, biting, and running, often with unbelievable rapidity—that it's remarkable there's little conflict or injury. (Dog play escalates into real aggression only around 0.5 percent of the time, studies show, although people think it happens far more often, most likely because it's an attention-getter.) How does play remain playful? It's because dogs' minds are very active, and the animals process information rapidly and accurately, even on the run.
By studying dog play we learn a lot about fairness, empathy, and trust. Based on extensive research, we've discovered that dogs exhibit four basic aspects of fair play: Ask first, be honest, follow the rules, and admit when you're wrong. Dogs keep track of what is happening when they play. They can read what other dogs are doing, and they trust that others want to play rather than fight.
When we carefully study the landscape of play we learn that dogs know very well how to tell other dogs "I want to play with you." They use a number of actions: bowing, face pawing, approaching, and rapidly withdrawing, faking left and going right, mouthing, and running right at a potential playmate. Bows also can be used to tell another dog, "I'm sorry I bit you so hard, let's keep playing."
Bows—crouching on forelimbs, perhaps with barking and tail wagging—essentially are contracts to play, and they change the meaning of the actions that follow, such as biting and mounting. They also serve to reinitiate play after a pause.
Dogs and other animals know they must play fair for play to work at all. Bigger, stronger, and more dominant dogs hold back through role-reversing and self-handicapping. Role-reversing occurs when a dominant animal performs an action during play that would not normally occur during real aggression. A dominant or higher-ranking dog would not roll over on its back during fighting but will when playing.
A hot topic in ethology and animal research today is whether nonhuman animals have a theory of mind—that is, do animals know that other individuals have their own thoughts and feelings, ones that may be the same as or different from their own and that they can anticipate and account for?
For dogs to know that another dog wants to play rather than fight or mate, they need to know what the other is thinking and what its intentions are. Each needs to pay close attention to what the other dog is doing, and each uses this information to predict what the other is likely to do next. Evidence is mounting that dogs likely have a theory of mind, and confirmation is coming from research on play.
There's a good deal of mind reading going on during play, and without empathy and trust, play wouldn't happen. Most dogs are moral mutts: When fairness breaks down, so too does play. —Marc Bekoff
Marc Bekoff, Ph.D., is a professor emeritus of ecology and evolutionary biology at the University of Colorado, Boulder.
How Dog Brains Work
Dogs use the same neural pathways we do to get where they can't go.
At play as at other activities, dogs exert some degree of self-control to inhibit impulses that would take them out of the game or otherwise spoil their social relationships. In this they are much like humans; in fact, social play is a major way young children learn self-regulation. And while the canine brain is a tenth the size of ours, the effortful control of behavior is accomplished much the same way—in the same part of the brain and through a similar biological mechanism.
We know this because psychologists Gregory Berns and Peter Cook, of Emory University's Canine Cognitive Neuroscience Lab, went where no one had gone before. They painstakingly trained a number of dogs to enter an fMRI scanner of their own accord, tolerate earplugs to block out the unsettling noise, sit absolutely still when necessary, and respond to assorted commands in a fully awake state.
Their studies to understand canine brain function pinpoint the neural pathways activated in a variety of behavioral states. The goal, Berns and Cook report in a recent issue of Current Directions in Psychological Science devoted to dog cognition, is, yes, to learn about the dog brain, but it's also to gain comparative insight into human brain function.
Trained on go/no-go hand signals, dogs were scanned to see what happens in their brain when they have to suppress a predominating response to nose-poke a target in front of them. Inhibiting responses is an executive function carried out by the frontal lobes of the cortex in humans.
The dog brain is about the size of a lemon, and the frontal lobes are very small. In humans, the frontal lobes—seat of abstract thought, planning, decision making, and more—take up the front one-third of our much larger brain. In dogs, they take up only about a tenth of the organ.
The bigger the brain of a species, the more modular it gets. Nevertheless, the researchers found, an analogous part of the brain—a small area of the frontal lobe—comes online during active inhibition. What's more, the level of brain activation correlated with the dogs' behavioral performance on the inhibition task and on other tests of self-control—including a canine version of the famed marshmallow test. The researchers were sure they were picking up a generalized behavioral trait of self-control, a facet of dog temperament.
Much as with people, there are individual differences in canine neural response, and they correlate with dog behavior and temperament. Self-control is often hard. One dog barked all the way through the task of actively inhibiting the nose-poke in the scanner—sound like anyone you know?—but still managed to restrain himself until given the release signal.
At a dizzying pace, neuroscience is providing unprecedented information about mental states. One thing studies show is that dog brains are organized similarly to ours in many ways. According to Berns, similarities in physiological processes suggest similarities in internal subjective experiences. At the very least, they imply that dog experience is richer than many people believe.
For Berns, the research also shows more. The knowledge of brain structure and cognitive function holds the key to understanding what it's like to be a dog. "Where structure-function relationships in an animal's brain are similar to those in our brains," he writes in his new book, What It's Like to Be a Dog, "it is likely that the animal is capable of having a similar subjective experience." Everyone knows what it feels like to exert self-control, he notes. "The brain data suggest that a dog's experience [is] very much the same." —Hara Estroff Marano
Leashed To the Here and Now
Do dogs know that we know that they're thinking of us?
For all the neural sophistication of dogs, science also reveals there are categorical differences in the nature of dog experience.
When we think about dogs' minds, we instinctively fall back on anthropomorphism, the idea that animals have thoughts somewhat like our own, just (in some undefined way) less so. Yet even a casual appraisal of the differences between our two species suggests that this can be no more than a crude approximation. Dogs build their picture of the world through their acute sense of smell; we humans are visual creatures first and foremost. Dogs' brains follow the standard carnivore pattern that prioritizes processing sensory information and turning it into precise and rapid action. Ours is dominated by cerebral cortices that give us unparalleled thinking abilities, including a facility for language.
Over a lifetime, we commit thousands of faces to memory; dogs must memorize the characteristic odors of hundreds of butts.
We also differ in how we process this information. Not only do our minds continually review our relationships with others, we also try to imagine how those people relate to one another. With dogs, it's more a case of "out of scent, out of mind."
The sensory and cognitive divide between dogs and their masters suggests that dogs' minute-by-minute experience of the world is significantly different from our own. Dogs seem to live almost entirely in the present, neither ruminating on the past nor planning for the future.
The evolutionary legacy of dogs makes it obvious that their social intelligence originated with their wild ancestor, the wolf. Wolves live in well-coordinated packs in which it's not only important to communicate effectively with one another but also crucial to be able to predict the intentions of other members of the pack by reading their body language. Dogs have inherited these basic building blocks, and the process of domestication has modified them to incorporate an almost uncanny ability to understand our human body language, to the point where it's easy—although probably inaccurate—to credit dogs with considerable emotional intelligence.
Our own behavior in social situations is driven by our conviction that those we interact with are capable of thinking about us, and that they know that we know that they are. Dogs' social intelligence seems to be driven by much simpler, though highly effective, processes, whereby they compare what is happening in the here-and-now with what has happened in similar situations in the past. What it does give them is an almost Zen-like detachment from the baggage of expectation and concern for the future that serves them extremely well as man's best friend. —John Bradshaw
John Bradshaw, Ph.D., is the founding director of the Anthrozoology Institute at the University of Bristol, England. His newest book, to be released this fall, is The Animals Among Us: How Pets Make Us Human.
Submit your response to this story to firstname.lastname@example.org. If you would like us to consider your letter for publication, please include your name, city, and state. Letters may be edited for length and clarity.
Pick up a copy of Psychology Today on newsstands now or subscribe to read the rest of the latest issue.
Facebook image: huesleihoppen/Shutterstock