The noted author Stanley Coren argues that one of the reasons dogs and cats have such a rocky relationship is that each species misinterprets the body language of the other. For example, dogs roll on their back to signal submission, whereas cats roll on their back in preparation for disemboweling an attacker or prey. As Coren (2000) puts it, “The opportunities for canine and feline misinterpretation are obvious. An angry cat rolls over on its back, which the dog reads as the doggish message that the feline is giving up and now wants peace. He approaches the cat to give the ritualized sniff, which is the doggish acceptance of such an offer of truce, only to have his face raked by four sets of flashing paws.” In other words, dogs seem to interpret cat behavior as if they were dogs, and this leads them astray. Indeed, numerous authors suggest that dogs interpret human behavior in very doglike terms as well, while we interpret dog behavior in very humanlike terms, all too often with unfortunate consequences for both. For example, the loving hug of a child may be interpreted as a display of dominance by the dog.
Of course, we are not dogs and would not resort to such silly misinterpretations—or would we? In 1998, Clifford Nass and Byron Reeves of Stanford did a number of experiments that suggested that people have an innate tendency to interact with a computer using the same social behaviors and rules they display with other human beings [Reeves and Nass, 1998]. That is, people use what they know about human behavior to predict and explain the behavior of systems that have neither human form nor behavior. The people in the study clearly knew the difference between a computer and a human, and yet when they interacted with the computer they interpreted the behavior of the computer in very humanlike terms. The results of Nass’s and Reeves’ study make it clear that human–computer interface designers would do well to examine human psychology and social behavior since so much of human interaction with the computer occurs within this context.
Online access to SPIE eBooks is limited to subscribing institutions.