By Duane M. Rumbaugh and David A. Washburn
Yale University Press, 2003
Intelligence, creativity, and complex learning in non-human animals have concerned psychologists and others throughout the history of psychology and the evolutionary sciences (e.g., Romanes, 1895; Köhler, 1925; Yerkes, 1916). In spite of the vast accumulation of information that this interest has generated, general theories of intelligence have rarely, if ever, incorporated or attempted to explain what is known about intelligent behavior in animals. From an evolutionary standpoint, the strategy of developing theories of human intelligence and then trying to fit into them the data on intelligent, creative, and complex behavior in other animals seems exactly backward. Rumbaugh and Washburn have addressed these issues by amassing data from animal research and developing concepts that can be useful in understanding intelligent behavior in both non-human animals and human beings.
One interesting aspect of the book is the authors’ tactical departure in confronting the nature of intelligence. Rather than asking questions about the contents of consciousness, or the structural components of intelligence, or even the correlates of individual differences in intelligence, Rumbaugh and Washburn have asked how mental activities that are assumed to require intelligence are performed. The answers they have obtained so far appear to have considerable generality.
By drawing substantially on attempts to promote language acquisition and use and creative, complex behavior in the great apes (gorillas, chimpanzees, bonobos, and orangutans), the authors have taken on a major challenge. Psychologists have traditionally viewed intelligence as heavily loaded (factor analytic pun intended) with language. Anyone who doubts that has only to examine the standard intelligence tests: vocabulary, verbal fluency, verbal comprehension, verbal absurdities, numerical competence, and verbal memory. To the extent that animals, in particular the great apes, can be shown to have the ability to acquire language, to use it to get what they want, to communicate with others of their species and with human beings, and to create novel communication strings, they will have been shown to have access to some quality that can be called “intelligence” without egregious anthropomorphizing.
Much of the book is devoted to research on language acquisition, elaboration, and use in apes, especially chimpanzees and bonobos. The evidence is convincing because the methods were strong: “strong inference” (attempting to disprove alternative hypotheses), adaptation of research instruments to fit the species characteristics of the subjects, and placing control of the experimental situations in the hands of the subjects wherever possible and appropriate. The broad body of evidence on language in animals has been reviewed recently by Hillix and Rumbaugh (2004), whose treatment extends to non-primates as well.
In reviewing work on learning, skill acquisition, and problem solving, the authors have introduced or applied a set of explanatory concepts that can have great generality. While acknowledging our heavy reliance on conditioning models for the last 100 years, these authors have resisted the temptation to abandon those models. Instead, they have built upon the knowledge that has been generated by use of conditioning procedures, and then have extended the explanatory concepts to reach far beyond simple acts and S-R or S-S associations into the understanding of creative, abstract, and original behavior.
Conditioning models can explain only one way in which simple acts and associations may be acquired and sustained – or extinguished. The authors have proposed a system, Rational Behaviorism, that can encompass – and help us understand – that magnificent leap from conditioned responses or the association of stimuli into complex, abstract, and creative behavior, including reasoning, language, numerical proficiency, and some cognitive operations that are distinctly Piagetian, such as comparison, classification, class inclusion, and transitive relations. In doing so, they have reconsidered the concept of reinforcement and its role in strengthening response probabilities and associative learning.
Rational Behaviorism embraces classical and instrumental learning as well as sensory preconditioning and instinctive behavior. The learning systems of organisms are sensitive to certain kinds of stimuli, and when they are reliably paired with other stimuli there is sharing of their response-eliciting properties as some positive function of their relative salience. Reinforcers are viewed as strong stimuli that share their behavior-eliciting properties with other stimuli, including those that are response-produced, as in operant conditioning. What is learned is not responses but predictive relationships between salient and reliable events. The ways the learning systems of diverse organisms then integrate those relationships into creative and efficient behavioral patterns will certainly constitute important questions for future research.
The essential constructs in Rumbaugh and Washburn’s Rational Behaviorism are “salience” and “emergents,” although they have retained a distinct place for the “operants” and “respondents” of an earlier generation of behaviorism. Salience refers to the relative importance to the learner of the conditions of learning: what is to be learned, the personal value of the outcomes of learning. Learning occurs best in situations that have high “personal” value to the learner. That concept is the bridge between mechanistic and rational behaviorism, assuming an importance in complex and abstract learning that is at least as great as that held in the acquisition of simple conditioned responses.
The construct “emergents” is at the center of their system. Emergents (Rumbaugh, 2002) are defined as creative behavior based on previous experience. Emergents benefit from the respondents and operants of classical and operant conditioning procedures, but go beyond those elements to reflect creative adaptation and novel and complex patterns of behavior. Emergents are the frequent but not inevitable products of an accumulation of knowledge followed by a cognitive reorganization of that knowledge that yields new understanding. So, whereas increasing one’s accumulation of knowledge may enable one to get better and better at performing some set of acts, the re-organization of knowledge that emerges from such an accumulation enables one, through new understanding, to perform new and creative behavior that was not possible before. For example, practicing comparisons may make one better at comparing, but using that skill to classify constitutes a new insight and a new ability – in other words, an emergent ability.
The authors’ accounts of language acquisition and use by the great apes are engaging and convincing, partly because they have emphasized the importance of using the right tools and procedures to elicit the behavior that one wishes to understand. Rational Behaviorism represents a breakthrough in concepts of intelligence and its applications. Although the book is a difficult read for those not steeped in the experimental and comparative psychology of learning, it is well worth whatever effort is required. It should have an influence far beyond its own pages on conceptions of the nature of “intelligence” in human beings as well as in non-human primates.
Hillix, W. A. & Rumbaugh, D. M. (2004). Animal bodies, human minds: Ape, dolphin, and parrot language skills. New York: Kluwer/Plenum.
Köhler, W. (1925). The mentality of apes. London: Routledge & Kegan Paul.
Romanes, G. J. (1882). Animal intelligence. New York: Appleton.
Rumbaugh, D. M. (2002). Emergents and rational behaviorism. Eye on Psi Chi, 6_(Winter), 8-14. Reprinted for online availability in Journal of Cognitive Education and Psychology, 2(2), 163-171 (www.iacep.coged.org/journal/).
Yerkes, R. M. (1916). The mental life of monkeys and apes. New York: H. Holt.