At best, we tend to view the field of animal psychology as a combination of hypotheses that can never be adequately tested; at worst, we end up viewing the field as a pseudo-science. Animal psychology, or the study of behavior in non-human animals, sits at the juncture of scientific experimentation, psychological inference and, perhaps most troubling, pure imagination. Because we cannot ask other animals what their experience is like, we must instead use our own imaginative skills to look at the animal mind to gain a better understanding of it. The use of this kind of imagination in a scientific field understandably invokes skepticism. However, it is this aspect of animal psychology that potentially allows us to see through the eyes of other animals, making the imagination the most important tool in the field. If we can imagine seeing through the eyes of other animals, we can begin to see more of ourselves. That is, the field of animal psychology has the potential to
uncover more about our own mind by giving us a glimpse of our brain’s evolutionary past and cognitive individuality. At stake within the field of animal psychology, then, is the definition of intelligence and the identification of emotions in ways that we can appreciate without mistakenly projecting human versions of intelligence and emotions on animals. As we search to define another animal’s level of intelligence, we run the risk of imposing our own understanding of intelligence on them.
Taking up this challenge, Alexandra Horowitz’s book Inside of a Dog: What Dogs See, Smell, and Know seeks to help us better understand the field. As she takes us into the mind of our canine companions, she tries to keep human bias out of her tests and conclusions. In itself, it is a laudable goal, but as we discover it is also nearly impossible to achieve. For instance, she writes “the first things to forget are anthropomorphisms” (14); and yet, as we will see, she uses anthropomorphisms to draw her conclusions. We might turn such inconsistencies against her, arguing that animal psychology is laden with an impossible mission. Yet, as I argue in this paper, these problems ultimately help us gain a better understanding of the challenges of animal psychology itself. In other words, there is something greater to these inconsistencies: by reading Inside of a Dog, we come to understand what is at the heart of the problem. In short, the problem Horowitz encounters – and indeed, the problem nearly any animal psychologist encounters – is based squarely on the tests and testing standards we use when we test for animal intelligence. A careful reading of Inside of a Dog shows us that we do not yet know where to draw the line between testing for animal intelligence and putting animals on a human standard of
intelligence. This subtle but important distinction that Horowitz struggles with is central in defining animal intelligence.
Let’s first take a more sustained look at the problem of animal psychology. At heart, the problem can be described as anthropomorphism, or giving human attributes to other animals. The question here is how to evaluate animal intelligence without human influence, or how to understand animal intelligence without using human standards. The instinct to dismiss the field of animal psychology based solely on this conundrum is neither out of line nor misinformed; in fact, it is understandable. We can easily read the problem of anthropomorphism as an impossible barrier. Horowitz herself quickly comes up against this barrier as she defines the possible range of dog intelligence, and while she points out that animal psychology has progressed a long way from using anthropomorphism to describe animal behavior and thought, we can still read her book as falling into it. She sets out to move away from testing animals based on human standards of intellect. For instance, she points out that we must look at what she calls a dog’s “umwelt” (20), or its way of perceiving the world, in order to understand how we should best test the limits of a dog’s intelligence and emotions.
However, this is where Horowitz would encounter a complication with skeptics. In order to define the scope of dog psychology, we have to be able to put the results of our experiments into terms that we understand. Often, this means using tests for animals that were originally designed to examine humans. Thus, we end up judging animals not on an individual species basis but on a human cognitive scale. Because we cannot speak to a dog to understand how it perceives the
cannot come up with a dog definition of dog intelligence or know what it thinks is important in such a test. Instead we have to test animals using measures we can understand and then extrapolate from these results for our conclusions. It is particularly difficult to make distinctions between where the human and animal scale should be separated, and this uncertainty means that animal psychologists can easily come to widely differing definitions of an animal’s intellect. We can easily read these different conclusions as proof that the field is not yet ready for real exploration.
One test that falls into the problem of imposing human standards on nonhuman intelligence is known as the Piaget test and was initially used to study a human infant’s ability to understand object permanence, a steppingstone of human cognitive development. The basic premise of the test is that children are shown an object, such as a toy, and then the toy is hidden under one of three boxes. The child is asked to find the toy. If the child understands that the toy still exists (even though he or she cannot still see it), then the child will lift the box under which they have just seen the toy placed. On the other hand, if the child does not understand that the toy still exists, he or she will randomly pick up one of the other boxes to see if the toy is under it. In the animal version, dogs are asked to do a similar task. They are asked to find a ball after they have seen it placed under one of three buckets.
As Horowitz tries to define the boundaries of dog intelligence for her readers, she cites this dog version of the Piaget test. For many, using a test originally designed to study human intelligence contradicts her earlier statement that we should base our conclusions about canine intelligence on “what dogs actually have the capacity to feel, know and understand” (Horowitz 14).
There is little doubt that a contradiction exists here; however, this inconsistency helps us better appreciate the complications that are inherent in understanding the psychology of another animal. It is hard for humans to conceive of a scale of intelligence different from our own, and thus we have to start testing another animal’s intellect using measures we can relate to. This means we often find ourselves needing to use experiments initially designed for humans. There is no doubt that Horowitz contradicts herself on these grounds. However, her claims do not need to be dismissed outright; instead, we should take them as examples of how these types of contradictions are inherent in the field of animal psychology. In order to understand more about other animals, and know what test to use to define their scale of intelligence, we must start from experiments based on a human scale of intelligence. When reading Inside of a Dog, we need to remember that animal psychology is still in development. It is more productive to read the book as a test for how we might need to develop the field than as a test for the field’s validity.
With this intrinsic problem in mind, Horowitz faces another dilemma: she must either design tests that can accurately tell us more about the limits of animal cognition or find a way to extrapolate unbiased conclusions from tests designed for humans. It is this latter task that opens Horowitz’s book up to criticism. In her conclusions, she points out that most dogs either fail the Piaget test or provide data that is inconclusive. To her, though, these conclusions do not equate to a failure on the dog’s part. This is important: it is not the failure of the dog but a failure of the experiment. At first, when the test was administered and the dogs were asked to find the ball, they
would look to the humans for help, reading their subtle movements or body language to find hints as to where the ball was hidden. When it became apparent that this was what the dogs were doing, the experimenters tried their best to hide the human in the room so the dogs could not look to them for help. At this point, the dog’s success rate began to rapidly decline. As Horowitz points out, what is not taken into account in these conclusions is the fact that dogs excel at reading our social cues. Horowitz defends the dogs’ poor performance, saying “It appears that the very skill at social cognition that is their triumph as a companion to humans contributes to the dog’s failure” (179).
Such a statement suggests what is at stake when we test for intelligence. Perhaps we should administer tests based on the social intelligence of dogs rather than on a human trait like object permanence. This would be a simple conclusion if Horowitz did not further note that “dogs who are less well socialized … set right to the problem, while dogs who live inside the house more often plead quietly with their owners to help” (180). Critics might point out that Horowitz again is contradicting herself while not offering any definitive answers as to the intelligence of a dog. But what is more enlightening about this moment is that it is here where defining dog intelligence on any type of linear scale begins to fall apart. By a scientific definition, the less socialized dog did better on the test. He was less dependent on humans and instead used his own intellect to perform the test. This is exactly what any scientist would want: pure data from the dog, untainted by human meddling. Yet by Horowitz’s definition of dog intellect, based on the ability of a dog to read social cues, the yard dog actually could be said to be less intelligent. So which dog is more intelligent, and which dog
more accurately reflects the scope of dog intelligence? The dog that can better read humans, or the dog that can better pass our intelligence tests because it is not as socially aware? These are questions that could be easily overlooked if we are merely looking to criticize Horowitz’s contradictions. They are, however, some of the most important moments in her book as they provide us with a deeper understanding of the inherent struggles the field of animal psychology faces.
For some readers, Horowitz’s framework for dog intelligence potentially becomes more confusing as she cites a study designed to test a dog’s response to a human emergency. We might read her use of this study and the failure of her conclusions to come up with definitive answers as other examples of the contradictory nature of her book, but again, this is not from any true failing of her own. Instead, it is another reflection of the nature of animal psychology itself. The test Horowitz cites is designed to test the “dog hero” theory. This theory holds that a very high form of dog intelligence might be measured by a dog’s ability to respond to a human emergency simply by reading our body language and other social cues. In this test, a dog owner acted out a particular emergency situation to test the dog’s response. Yet again, the test proved to be a failure for the dog, at least on the surface. Few, if any, dogs responded to their owner’s plight. But Horowitz reconciles the results of the experiment, saying that dogs have to be “taught to understand emergency situations” (240). If this is the case, what of the less-socialized dogs that did not pass the “human interaction” standard of intelligence in the Piaget test? Where does their form of intelligence fall on the canine intellect scale? Yet another problem with animal psychology is reconciling the different experience levels each individual animal has. For
animals that have less contact with humans, and whose intelligence levels would not be based on recognizing human social cues, this task would seem to be much simpler; but by Horowitz’s definition of dog intelligence that makes reading social cues a central feature of that intelligence, this is not the case. Thus, yet again, Horowitz uncovers many more questions about the feasibility of defining a dog’s range of intelligence than she is able to give answers. As it turns out, this ability to uncover the core questions of animal psychology may well be the most important contributions her book makes.
As Horowitz helps us journey into the mind of a dog, she unwittingly takes us on a tour of some of the greater problems of animal psychology. Skeptics would argue that her conclusions about dog intelligence contradict her stated goals, that she is basing her conclusions on unsound assumptions. This may well be true. However, with a closer reading, her seemingly inconclusive and unfounded claims do not necessarily point to a problem with her analysis but to a problem with the concept of animal psychology. As we search for a deeper understanding of the beings around us, we have to question our methods and challenge our assumptions. This may cause some people to question the feasibility of defining animal intelligence, and even the viability of animal psychology as a whole. However, this field has much to offer in helping us understand ourselves. We, too, are animals, and perhaps knowing more about the minds of other creatures will help to illuminate the questions we have about our own psyche.
For more than 20 years, Exposé has annually showcased outstanding essays written by students in Harvard’s first-year course in the Harvard College Writing Program. All students are required to take “Expos,” a tradition since 1872. Expos asks students to work intensively with their teachers on the skill of writing thoughtful arguments that respond to challenging ideas, complex debates, and puzzling social phenomena.
The annotations found throughout this essay were prepared by James Herron, Ph.D., a preceptor in the Harvard Writing Program, the Director of the Harvard Writing Project, and a cultural and linguistic anthropologist.