One of These Things Is Not Like the Others
Chaser the Border Collie with Her Owner
A recent study at Northwestern University informs us that “babies can think before they can speak.” In the study, 7-month-old infants were able to understand what researchers describe as the simplest and most basic of abstract relations: the levels of sameness or difference between two objects, a key cognitive ability underlying human intelligence, and one that differentiates us from all other animals including great apes.
The test was quite simple.
“Infants were shown pairs of items, two Elmo dolls—or two different—an Elmo doll and a toy camel—until their ‘looking-time’ declined.”
This decline in the amount of time the infants spent looking at different pairs of objects indicated to the researchers that the relation between the objects—the fact that they were alike—had quickly become encoded in the babies’ brains. Meanwhile, when the infants spent more time looking, it coincided with the fact that objects were not alike.
Dedre Gentner, co-author of the study said the infants “were able to form an abstract, same-or-different relation after seeing only 6 - 9 examples.” She goes on to say, “It appears that relational learning is something that humans, even very young humans, are much better at than other primates.” As an example Gentner cites a recent study using baboons whose success at this required over 15,000 trials.
Co-author Susan Hespos sums things up: “Infants can form abstract relations before they learn the words that describe [those] relations, meaning that relational learning in humans does not require language and is a fundamental human skill of its own.”
What does this have to do with dogs?
Similar studies have been done where dogs were tested to see if they were capable of comparing the differences between smaller and larger amounts of food or the differences between carrots and doggie treats or between the amount of treats they got to what another dog got. The conclusions drawn were that dogs seem able to make comparisons in a way similar to a young child’s abilities in this area.
Of course one might offer the simple idea that there’s a difference between making comparisons and having preferences. For example, before she learned sign language Helen Keller was able to show preferences for certain foods. But it wasn’t until she learned to sign that she had the ability to make comparisons between the things she liked and the things she disliked. This shows that ability to understand and use human language is what enabled her to make such comparisons, an ability dogs don’t have but that the children in this study do, even though they’re still too young to talk.
Let me make this very clear: it’s not possible to make comparisons without having the ability to use and understand language. No other animal—including chimps and great apes—has demonstrated this ability. And it’s not something that comes in gradients. You either have the full ability to use and understand language or you don’t.
Logic & Language vs. Pattern Recognition
Remember, the researchers say that 7-month olds can form abstract relations between objects before they learn the words that describe those relations. Yet we know now that children learn and understand some aspects of human language while still in the womb. So while the infants in this study can’t yet use language, they do understand that words have meanings.
Do dogs understand that words have meanings? No. Dogs respond to all kinds of cues, some verbal, some not. It’s entirely possible to teach a dog to sit, for example, without ever using the word sit, all you need are hand signals, and in some cases, just your body language. It’s true that, generally speaking, the family dog is far more aware of its environment than the toddlers in the house are; yet toddlers are still able to understand things like language, logic, and object permanence while dogs have no such abilities.
Do children understand abstract relations? To a certain degree, yes. Do dogs? No. Dogs are experts at an unconscious process called pattern recognition, something we (and young children) also excel at. But unlike dogs and children, adult humans often think too much and rarely pay as much attention to changing patterns the way our dogs and children do. And, if we’re not careful, it can be easy to mistake pattern recognition for the ability to think, reason, and make inferences, abilities dogs don’t have.
Of course, some people would disagree.
Learning Through Inference
In his book, If Dogs Could Talk, famed Hungarian ethologist Vilmos Csányi writes a great deal about how he thought his dog Flip was capable of making inferences. But in an interview for The BARk, Csányi says, “A family dog constantly observes human behavior and always tries to predict interesting actions in which he could participate. Dogs can learn any tiny signal for the important actions and is always ready to contribute.”
This is absolutely true. And I would argue that all the examples of “inferences” Csányi observed in his dog Flip can be boiled down to this simple ability.
In their book The Genius of Dogs primatologists Brian Hare and Vanessa Woods write, “What is special about children is that if you show a child a red block and a green block, then ask for the chromium block, not the red block, most children will give you the green block, despite not knowing that the word 'chromium' can refer to a shade of green. The child inferred the name of the object.”
Actually, we don’t know that the child inferred the name of the object and didn’t, instead, read some unconscious signal given by the person who asked for it. Like dogs, children are very good at reading our patterns of behavior. In fact, there was a study done where detection dogs were asked to find hidden items. But their handlers had been told where the objects were hidden, though in some cases the dogs were lied to. The study found that most of the dogs went to the false locations rather than the spots where the items were actually hidden. Why? Most probably because they picked up information from their handlers through pattern recognition; reading their eye contact and body language.
In his book Chaser: Unlocking the Genius of the Dog Who Knows a Thousand Words, John Pilley discusses how he taught his dog Chaser over 1,000 words for various toys. Pilley would place a new object Chaser had never seen before in a different room with several toys she already knew by name. Then he’d tell Chaser to go fetch the toy—let’s say it was called Fuzzbee—a word Chaser had never heard before. And, Pilley says, that just like children, Chaser inferred that the new word referred to a new toy.
But did she?
We know that dogs are descended from wolves. And wolves are different from other mammalian predators, most of whom go through the fairly predictable predatory sequence : search, stalk, chase, grab-bite and kill-bite. But wolves are unique in that they routinely hunt animals that are 10 times their size. Plus they exhibit a unique behavior that other predators don’t, called culling where they target the weakest member of the herd. Culling is also a necessary aspect of a herding dog’s job, and in dogdom, border collies are the Einsteins of culling.
So Chaser’s abilities can be explained via high-octane pattern recognition, which includes a border collie’s innate ability to cull sheep, which is inherited from the way wolves cull the weakest elk or buffalo from the rest of the herd. It does not in any way require an ability to make inferences.
One of the tests given to Chaser is where she’s been told to find “Fuzzbee.” She doesn’t know who or what Fuzzbee is, but she goes into a room full of toys. She looks around. She sees a toy she’s never seen before. And since she’s operating through an already established pattern of being told to find toys, and has been bred for her ability cull sheep (or toys), bringing a new toy back is not an example of making inferences but is, rather, the simple ability to respond to verbal cues and pre-determined patterns of behavior.
When Brian Hare did his initial research on a dog’s ability to follow where a human points (an ability chimpanzees don’t have), the headline was “Dogs Are Smarter Than Chimps,” with a sub-title “…Plus They’re Smarter Than Wolves!” That’s because in Hare’s studies none of the chimps and wolves were able to follow where Hare and his research team pointed while most of the dogs were. Hare thought that the ability to follow where humans point must have been a by-product of domestication.
But then Monique Udell did a more carefully-designed study with a) dogs living in human households, b) hand-reared wolves, and c) stray dogs, recently re-located to animal shelters. And guess what? The wolves’ ability to follow where humans point was no different from that of the pet dogs, while the stray dogs weren’t able to follow where humans point.
It seems to me that if we’re to be scientific about it—i.e., apply Ockham’s razor and Morgan’s canon—and since canine behavior can be explained most parsimoniously via the process of pattern recognition, it’s a no-brainer, that dogs and wolves can both follow where humans point.
And Speaking of Brains…
However, when it comes to brain power, here’s how dogs really stack up against both chimps and toddlers, based on the number of neurons in the cerebral cortex of each.
Dogs: 450 million neurons. (Cats have 300 million, pigs 450 million.)
Chimps: 5.5 billion neurons.
Humans: 86 billion neurons.
To give you an idea of the scale involved, a chimpanzee has over 34 million neurons to every single neuron in a dog’s brain. And a toddler has a ratio of more than 537 million neurons to every neuron in the dog’s brain. So if dogs are really smarter than chimps and are on a mental par with toddlers, where does all this supposed brain power come from?
It’s very simple, and it goes back to Hare’s initial research, and a dog’s ability to follow human cues. (In the real world dogs don’t always follow where we point except as they did in Hare’s laboratory studies; they generally just look around in a general way).
So the real reason behind the idea that dogs seem so “smart” comes from 5 simple things.
1) Dogs share a common ancestor with wolves,
2) Our dogs’ ancestors became domesticated by our ancestors because of their ability to chase, harass, cull, and corner large prey animals, the way wolves still do today.
3) This ability made it easier for our ancestors to kill and consume the flesh of mastodons, wooly mammoths, etc., giving our ancestors an advantage over Neanderthals.
4) Humans have been using certain aspects of the wolf’s prey drive to get our dogs to work with us and for us for tens of thousands of years.
5) The reason dogs in Hare’s laboratory studies seem so smart is that they live with human beings, and we exhibit the most complex patterns of behavior of any animal on earth.
In other words, dogs are “smarter” than cats, pigs, chimps and (in some ways) toddlers* not because of their own innate IQs, but because of their daily interactions with us. It has nothing to do with making inferences or comparisons as young children are able to do.
Lee Charles Kelley
“Life Is an Adventure—Where Will Your Dog Take You?”