They're smart, but not in the ordinary way.
How smart are dogs, really?
If you ask most owners, they'll assure you their dogs are very smart indeed. And they would not be far off. But dogs, like most animals, are specialists. Dogs are terrific examples of how it is more useful to think of intelligence in terms of specific brainpower domains than as “general intelligence,” the stuff IQ tests purport to measure. In their strong domains, dogs are stellar. In others, not so much.
Dog strengths
Dogs can estimate the passage of time with astounding accuracy. Not only do they know what time dinners and walkies are, they know whether or not it's agility-class night. Dogs are also masterful at noticing small cues in their environments to tip them off about what's going on.
Ever noticed that your dog has learned that your turning off the TV at night means it is bedtime and so trundles off unbidden? Turning it off on Saturday afternoon is more ambiguous – could be walkies, car rides or nothing at all. Has your dog doped out that keys plus briefcase equal solitude for a few hours, but that keys plus a Chuckit mean nirvana at the dog park? Does your dog know the turnoff to the dog park? To the groomer? A Border Collie in Germany can discriminate 200 different words (all, amazingly, variations on the theme of “tennis ball”).
People underestimate dogs' ability to finely discriminate. It's a well-worn myth that you shouldn't give a dog an old shoe to chew lest he learns to chew all shoes. It might actually be tricky to teach a dog a category like “shoe,” given that shoes can be leather, fabric, rubber, have laces or not, be light, strappy and open, or weigh several pounds. Humans are very good at classification by function or other high concept. Dogs are less about headings and more about specific case.
It's a no-brainer for a dog to learn that this shoe here is his to dissect at will, but those leather goods on the racks in the closet? Strictly verboten. I know dog trainers who, for kicks, teach their dogs to discriminate U.S. currency by scent. I have difficulty discriminating U.S. currency at all.
Recent research has demonstrated that dogs read human gestures more readily than both chimpanzees and wolves. Even kennel-reared dogs outperformed hand-reared wolves, suggesting the capacity is not a product of experience around people.
Part of dogs' specialization is making the most of their human-rich environment. Another study suggests that dogs gauge their play solicitations to other dogs based on whether the potential playmate is attending or not. Dog-behaviour researcher Marc Bekoff points out that dogs are by no means “dumbed-down wolves.”
Imitative ability
There's no way to – pardon the pun – intelligently discuss cognition without drawing the distinction between what has been scientifically proven and what most people believe based on interpretations of their own day-to-day experience with dogs.
This topic tends to leave people cold, as though science rains on some sort of confabulation parade. But the (very interesting) fact of the matter is, there's enormous discrepancy in key areas between what people think dogs are good at and what dogs are actually good at as soon as other explanations are ruled out. It's why we love science so. Let me give you an example.
I doubt not a day goes by wherein a dog trainer somewhere isn't regaled by an owner about how his dog imitates. This dog imitated that dog. The puppy imitated the resident adult. The dog imitated the owner, the cat or the budgie. Poll dog owners, even dog fanciers, and they will tell you, yes, dogs can imitate. One dog peed. Then the other one did. See? Or the more sophisticated version: one dog opened the latch, broke into the yard and terrorized the ducks. Then the other dog did it. First time ever. See? The sticky thing is that many processes – not just imitation – can account for an animal seeming to do something after witnessing a model. As soon as conditions are controlled so those other means are not available, dogs suddenly can't “imitate.” Imitation is defined as an animal replicating the behaviour of another animal after a single viewing.
The behaviour can't be a natural, pre-programmed behaviour such as barking, urinating, chasing something, etc. Many such behaviours are socially facilitated in dogs – triggered in a non-imitative way, or ratcheted up in intensity. Right off the bat, many imitation anecdotes are out of the running. Does witnessing random but intense and apparently successful activity of a model dog at the latch help as much as witnessing a skilful latch-manipulator dog?
If it does, it's not imitation. Such “two-action tests” are the current gold standard for inferring imitation and so far, dogs are unimpressive at them. The behaviour must be novel to make the criteria for imitation, and learning processes such as operant conditioning must be ruled out. Bottom line: Despite some effort to find evidence of imitation in dogs, researchers have come up empty-handed. So, if dogs can imitate, there is a global canine conspiracy to keep science in the dark about it.
Projecting ability
Now, that said, there are some intriguing whiffs of proto-imitation in the dog-cognition research. For example, if you put a dog behind a fence, put a toy on the other side and demonstrate one of two possible detours around the fence, the dog is more likely to take the detour you showed him.
This brings us back to domain-specific talents: this detour-copying is imitative but limited to route-taking. Dogs don't have the all-purpose imitate-anyone-doing-anything module we humans take for granted.
To me, this is the heart of the divide between what science tells us dogs can (and can't) do and what people continue to insist they can. Imitation comes so incredibly readily to us that we can't imagine a mind that can't do it. We especially can't imagine a familiar mind that can't do it. And, given that humans are capable of ascribing agency to cartoon sponges, it's not surprising our theory-of-mind module runs amok when it comes to those most splendid of surrogate children, our dogs. They are very, very like us. They bond readily and strongly. They feel palpably similar emotions. They are fabulous consequence learners. Their proto-theory-of-mind module seems able to dope out whether another dog is attending or not.
So when they appear to do something that we would have done via imitation, we project imitative ability. Vive la différence!
