Too simple a story: A response to Rollin

Abstract

Rollin’s ambitious paper is provocative, but, in trying to do too much, it oversimplifies complex issues. Specifically, Rollin gives an account of the origin of behaviorism, a set of criticisms of behaviorism, and ethical arguments and exhortations toward the better treatment of animals in research, but he does each of these things sketchily, weakening his conclusion about the use and abuse of animals in research. In general, I agree with Rollin’s ethical qualms concerning animal studies, but I am afraid I have little more to actually say on the subject. Instead, I will discuss Rollin’s oversimplified and contradictory treatment of the rise and philosophy of behaviorism, and his misplaced hopes for cognitive psychology. Rollin says that the early comparative psychologists happily imputed mental states to animals, and that Lloyd Morgan’s famous canon was meant not to outlaw inference to mind but only reasonably to restrict it. So much is true, but there is an important complication. In a paper in Milld, Morgan (1886) distinguished two kinds of inferences from animal behavior to animal mind. First, there are objective inferences from behavior to objectively characterizable cognitive powers. So, if I see a dog pick its master out from a crowd of other little boys, I may reasonably infer that the dog possesses recognition memory and some level of discriminative ability. Objective inferences are not made on the basis of analogy to our own minds, and posit powers and processes that may be further characterized by research. The other kind of inference Morgan called “ejective inference”, using a now unfamiliar term that we may change to “subjective” inference. Subjective inferences are inferences of emotions and feelings made on analogy with our own experience and do not posit objective, positively characterizable mental powers or processes. Saying the dog is /~U~PJ to see its master is making a subjective inference. Morgan then argued that objective inferences are acceptable to the science of animal intelligence, because the powers and processes imputed to animals are objectively characterizable. Subjective inferences, however, involving no objectively characterizable power or process, therefore have no place in the science of animal intelligence. Morgan was careful not to deny that animals have emotions and feelings - they do - but he did maintain that subjective states of awareness can find no place in science. Morgan’s point is worth making, for it anticipates