Does your dog love you?

What if the "love and affection" that your dog seems to shower upon you isn't really that at all? Imagine, deep in dog and human evolutionary history, that dogs were attracted to food scraps and other leavings around human camp sites. To the humans, dogs probably appeared as just another potential food item. Under these circumstances, natural selection would favor any sort of behavior, on the dog's part, which would avert being eaten. What behaviors did the dog already have in its repertoire? Wild canid social behavior includes affiliative and submissive behaviors which plug very nicely into human expectations for social bonding. In this scenario, natural selection should favor the expression of affiliation and submission by the dog, whether the dog feels any "emotional" attachment or not.

But, really, the question is whether we can ascribe emotion--joy, sorrow, love, sadness--to a dog (or any animal) as a feeling that underlies behavioral actions. Beyond our own emotional response--yes, of course my dog loves me--can the hypothesis that animals feel emotion be tested? A direct test is probably imposssible; we'll simply never know what your dog is thinking.

However, a line of evidence can be considered if you're looking for support of the hypothesis that animals have emotions of the human sort.

First, the neural architecture that regulates behavior, and their underlying emotions, in humans didn't arise from nowhere when humans evolved. We share brain structure, neurotransmitters, and hormones with the entire verebrate lineage. Our commonality with frogs, lizards, birds, and mammals extends beyond having four limbs and similar bone structures. We know that human emotions are greatly affected by neurotransmitters such as serotonin, and that, in fact, treatment of animals such as dogs with drugs such as Prozac (fluoxetine) has behavioral effects on dogs which are in line with the behavioral effects that Prozac has on humans. Even mice exhibit behavior that reflects "anxiety" (Nyberg et al. 2003).

Then consider the logic of how emotions compel human behavior. In "normal" behavior, our emotions serve as motivators to express evolutionarily appropriate behavioral responses. (See the discussion of drive theory for more on this.) An excellent example of this is the role of emotion in stimulating parental care of infants. If emotion is the great motivator of human behavior, would it be surprising that animals with similar brains are motivated in the same way? If emotion does not drive the behavior of animals with similar physiology to ours, what does?

Ultimately, emotionality in animals is probably an untestable hypothesis. Descriptors of emotion, such as fear, sadness, loneliness, and affection, are subjective representations of a human state. They are difficult enough to quantify in humans and to compare among humans, much less to document in animals.

For the last century biologists have been trained to avoid anthropomorphism, probably rightly so, as ascribing human emotions and motivations to animals interferes with objective hypothesis testing. It is certainly unscientific to uncritically apply subjective descriptors to the behavior we observe. It is equally unscientific to ignore reasonable hypotheses because of a well-established taboo among scientists. We know enough now about the neurochemical bases of behavior that the brain is no longer a black box; rather than relying on imagination when trying to understand the neural bases of behavior, we can look to an extensive fund of knowledge concerning the roles of hormones and neurotransmitters in behavioral modulation.

What should students and researchers of animal behavior do? When appropriate, emotion should be considered as a hypothetical shaper of behavior. In manipulative experiments and clinical treatment of animal behavior, matching subjective descriptions of human emotion with an animal's state gives excellent clues about how pharmacological treatment might affect the animal's behavior. This approach, though, needs to be tempered with the recognition that subjective representations can lead to gross misinterpretations or oversimplications of animal behavior and social systems. As with any hypothesis, emotion should be regarded with skepticism; as a potentially testable statement, rather than as an assumption.

Bahlig-Pieren Z, Turner DC 1999. Anthropomorphic interpretations and ethological descriptions of dog and cat behavior by lay people. ANTHROZOOS 12 (4): 205-210
Hare B, Brown M, Williamson C, Tomasello M 2002 The domestication of social cognition in dogs SCIENCE 298 (5598): 1634-1636
Overall KL, Dunham AE 2002 Clinical features and outcome in dogs and cats with obsessive-compulsive disorder: 126 cases (1989-2000). JOURNAL OF THE AMERICAN VETERINARY MEDICAL ASSOCIATION. 221 (10): 1445-1452
Virga V, Houpt KA, Scarlett JM 2001 Efficacy of amitriptyline as a pharmacological adjunct to behavioral modification in the management of aggressive behaviors in dogs J AM ANIM HOSP ASSOC 37 (4): 325-330

Nyberg, J. M. O.Vekovischeva, N. K. Sandnabba 2003 Anxiety Profiles of Mice Selectively Bred for Intermale Aggression Behavior Genetics 33:503-511