Stanley Coren's The Intelligence of Dogs: Canine Consciousness and Capabilities has received more media hype than any other dog book in recent memory (with the exception, of course, of The Hidden Life of Dogs). Yet Coren is no Elizabeth Marshall Thomas: he is a dog trainer as well as an experimental psychologist, and he approaches his subject with none of the let's-look-at-the-exotic-natives attitude of artificial wonder that Thomas encourages. Stripped to its barest essence, Coren's book is an extended attempt to answer one simple question: do dogs think? Arguing against the many psychologists who define "thought" as a process which can only occur in human beings, Coren replies resoundingly in the affirmative. The quality of the book is uneven. Unfortunately, the strongest chapters are those that many readers will be tempted to skip, and the weakest are the ones that have received the most attention on the talk-show circuit.
Coren's discussion of the various historical views of the dog's mind is fascinating. Aristotle, for example, argued that dogs and human beings differ only in the degree to which they possess certain mental abilities. Both have emotions, but those of human beings are more complex. Both learn, remember, solve problems, and benefit from experience, but human beings do these things better. In time, the messy ethical and metaphysical complications of this scheme became apparent to philosophers. If animals have intelligence, then they could have souls. If they had souls, then they would be candidates for the afterlife. The prospect of a heaven filled with cattle, pigs, spiders, and flies was a bit overwhelming to many squeamish theologians. The ethical conundrums were equally troubling. If animals had souls, could we justify killing them for food, forcing them into servitude, and refusing to grant them access into the church? Ren‚ Descartes solved these puzzles neatly by positing that animals are merely biological machines, that everything in animal behavior could be reproduced mechanically. No matter how complex, Descartes insisted, the activities going on in animals take place without consciousness, thought, emotion, or feeling.
Descartes' argument is not a relic of the seventeenth century. Many contemporary behavorists, believing dogs to be nothing more than bundles of nerve impulses and raw instinct, are the direct heirs of Descartes. Coren's counterargument is immediately persuasive to those of us who spend our time observing dogs. He asserts that we have no right, in the absence of other data, to deny the possibility of consciousness and intelligence to higher animals such as dogs, animals that have been provided with the same general neurological building blocks and who operate according to the same physiological principles as those of human beings. He writes that, if we observe "a situation where conscious reasoning takes place in humans and we note that dogs respond in much the same way and are affected by the same factors that affect the behaviors of humans, then I would propose that we should accept consciousness and intelligent reasoning in the dog as a plausible hypothesis. If by placing ourselves (figuratively) in the place of the dog, we can accurately predict its behaviors using our own reasoning and consciousness, I would say that this is further evidence consistent with the view that consciousness and intelligent analysis play a role in the dog's behavior" (74). Obvious? To us, most assuredly. But to academics and scientists, the commonsensical often needs to be carefully delineated and described with appropriate jargon before it is taken at all seriously. Our collective hats should be offered to Coren for attempting this translation of what every dog trainer knows into the realm of professional experimental psychology.
Coren jumps off from Howard Gardner's theory of multiple intelligences--the idea that there is no single thing called intelligence but at least several different forms--and extends it into the realm of dogs. He isolates three separate forms of canine intelligence. Adaptive intelligence consists of both learning ability (the number of experience an individual needs to code something into relatively permanent memory) and problem solving (the ability to put together bits of information into a correct answer or to discover new ways to apply previously learned information to novel situations). Adaptive intelligence allows individuals, both human beings and animals, to adapt to their environments or provides them with the skills necessary to modify their environments to suit their needs. Working or obedience intelligence allows dogs to respond appropriately to various commands. The characteristics of working intelligence include a long attention span, the mental flexibility to try another approach when the first responses to a particular command are not rewarded, and the communication skills to realize when its handler is attempting to communicate something. Instinctive intelligence is defined as those genetically determined abilities and behavioral predispositions that allow the dog to do a particular job: for example, terriers to go to ground, bloodhounds to track, spaniels to flush game birds, or border collies to herd sheep. Coren argues that these three types of intelligence all come into play when one discusses the broader category of the "intelligence" of dogs.
Coren's analysis of working or obedience intelligence is by far the weakest link in his book. In attempting to rank the various breeds in terms of working intelligence, Coren found no laboratory research at all. He quickly realized how expensive a scientific study of canine intelligence would be: by his conservative estimate, a grant of at least $14 million would be necessary to acquire, house, train, and test enough individuals of the various breeds to make the study useful. Coren also attempted several sophisticated means of analyzing AKC obedience trial results before abandoning this line entirely. Ultimately, he decided to send out a survey to all obedience judges in the United States and Canada. 208 (about half of those who received the survey) responded, and he followed up about two dozen of these with a more detailed telephone interview.
I wrote to Stanley Coren and asked him to send me his original survey, which is not included in the book. The survey consisted of three parts. Part I asked the judges to score the 74 most popular dog breeds on a scale of 7 to 1 (7: one of the top breeds for obedience competition; 6: can be expected to produce a good qualifying score most of the time; 5: turns in a qualifying performance more than half of the time; 4: expected to qualify about half the time; 3: qualifies less than half the time; 2: can be expected to fail most of the time; and 1: among the worst breeds for obedience competition). Part II required the judges to rate the same dogs on a similar scale for their learning and problem-solving ability. Part III requested that the judges fill in what they consider the ten MOST intelligent and LEAST intelligent breeds of dogs. In his followup telephone interviews, Coren requested the judges' opinions on the length of time it would take selected breeds to learn simple commands, the amount of time various breeds would respond to the first command given, whether a quick performance or noticeable delays could be expected, how well the different breeds remember commands, whether practice is important for a given breed, and whether the distance at which a particular command is given appears to make a difference.
Coren uses these data to arrive at six basic clusters of working intelligence. At the top of the heap, he finds the border collie, poodle, German shepherd, golden retriever, Doberman pinscher, Shetland sheepdog, Labrador retriever, papillon, rottweiler, and Australian cattle dog. These dogs "are the brightest dogs in terms of obedience and working intelligence. Most of these breeds will begin to show an understanding of simple new commands in less than five exposures and will remember these new habits without noticeable need for practice. They obey the first command given by their handler around 95 percent of the time or better. Furthermore, they respond to commands within seconds after they are given, even when the owner is a distance away. These are clearly the top breeds along this intelligence dimension and seem to learn well even with inexperienced or relatively inept trainers" (181). At the bottom, we find the shih tzu, basset hound, mastiff, beagle, pekingese, bloodhound, borzoi, chow chow, bulldog, basenji, and Afghan hound. "During initial training, these breeds may need thirty or forty repetitions before they show the first inkling that they have a clue as to what is expected of them. It is not unusual for these dogs to require over a hundred reiterations of the basic practice activities, often spread over several training sessions, before any reliability is obtained. Even then, their performance may seem slow and unsteady. Once learning is achieved, practice sessions must be repeated a number of times; otherwise, the training seems to evaporate, and these dogs will behave as if they never learned the exercise in the first place" (185). The public and press alike love lists, and it is this analysis of working intelligence, ranked by breed, that has provided the most grist for the media mill since the book's publication earlier this year.
Unfortunately, the methodology underlying Coren's conclusions is extremely faulty. All Coren has managed to do is to obtain a rough list of the success of various breeds in the sport of dog obedience in North America; jumping from that to the number of repetitions it took the various dogs to learn commands is impossible. We can even use Coren himself to challenge his own methodology. In his analysis of adaptive intelligence, Coren includes an interesting canine IQ test. The "CIQ" consists of twelve separate tests, designed to assess the dog's learning and problem-solving ability. I tested two dogs: Connie, my own basset hound (a breed ranked in the bottom tier of intelligence) and Dream, a border collie (a member of the top echelon). The results were interesting. Connie scored in the "brilliant" category, a group that fewer than five percent of the dogs in Coren's standardization group reached (no, I didn't skew the results!). Dream, on the other hand, scored in the low average range of intelligence, where, according to Coren, a dog will need to work rather hard to understand what is required of it. Connie has obedience scores which range from a low of 173 to a high of 186; she currently has two legs on her UD (and plenty of NQ's in our quest for that elusive third leg). Dream is an OTCH who has garnered many high in trials and placed at this year's Gaines Classic. Clearly, an obedience judge seeing the two dogs in the ring would conclude that Dream was by far the easier dog to train. Yet such was not the case. Connie is an extremely quick study who retains what she learns. Dream, according to her handler, always has difficulty learning and retaining new behaviors. Obviously, only erroneous conclusions could be drawn from their respective ring performances as to the amount of time and repetition it took them to learn the commands.
The most striking difference between the two dogs is a personality issue, not a matter of anything that can be labeled "intelligence." Although Coren devotes a full chapter to what he terms the "personality factor," he does not seem to realize how critical a role it plays in the obedience ring. Connie is like many bassets: she's bright and happy to learn if you can convince her that the learning was her idea in the first place (i.e., if you train with food). But she doesn't have a strong sense of duty; if she's under stress or a bit distracted, she'd as soon not obey a command as obey it. Let's indulge in speculation and generalization for a moment, dangerous though it might be. Bassets are perfectly capable of shutting down entirely under stress; more than anything else, their tendency toward negative stress management is the reason why judges see so many slow-moving, tail-drooping, lagging bassets in the ring. Border collies are an entirely different story. Once a behavior is learned, most border collies seem to perform regardless of stress; indeed, many respond to stress by getting sharper and sharper. Dream is not such a successful obedience dog because of her learning ability. She has excelled because, quite simply, she loves to perform in the ring in front of a crowd of spectators. It is this showy sparkle--a je ne sais quoi which would never appear on a personality or intelligence test--that makes Dream unusually good; her learning pattern is all but irrelevant. My basset loves to learn new things and loves to practice but gets a bit overwhelmed in stressful situations, freezing and refusing to work at all. Again, her learning pattern would be impossible to predict in an assessment of her ring performance. In both cases, an obedience judge, based on what she sees at a trial, would be unable to make any meaningful statement about these dogs' trainability. In general, the difference between bassets and border collies is far more a difference of intensity, energy level, and desire to obey commands in the face of adversity than it is a difference of trainability or problem-solving aptitude. Coren would have done much better to follow his initial survey with phone interviews about the temperaments and personalities of particular breeds for obedience, rather than attempting to ask the judges to comment on individual learning patterns.
In corresponding with me, Coren admitted to certain methodological problems in his breed rankings, but he regards what he has produced as an acceptable first approximation. I, however, do not. I think that the book would have been far stronger if this chapter were deleted entirely. In a review in the Wall Street Journal, Manuela Hoelterhoff writes that we can all spare ourselves the trouble of assessing our dogs' intelligence and "just accept Mr. Coren's ranking of breeds in descending order of dimness. After years of observation and interrogating hundreds of vets and trainers, he has fearlessly rated 79 breeds, and the news is not good for proud owners of Afghans, basenjis and bulldogs. They are the Igors of dogdom, occupying the bottom of the list, far away from the genius-level border collies, poodles, German shepherds, and golden retrievers" (Wall Street Journal May 11, 1994: A18). It is such reasoning, however lighthearted, that makes Coren's conclusions extremely dangerous. The inherent problems in the consultation of obedience judges as "experts" are far too deep, and the influence of his conclusions in the minds of the general public is far too profound, to have allowed the rankings to stand as even a rough approximation of reality.
For all that, I still think that The Intelligence of Dogs is a book worth considering. Coren's argument might be uneven in spots, but he has written a work for the intelligent student of dog behavior. Even the weak methodology of his ranking of the breeds by working ability is not sufficient to cripple the work as a whole. While the book will be of little practical use to working dog trainers, the issues that ground it are fascinating. Read it, as much to grapple with its problems as to learn from its strengths.