The rational behavioristic framework here advanced allows for cognitive operations and emergent phenomena that are, in measure, based on the syntheses of general life experiences and observations and also upon the learning induced by traditional respondent and operant conditioning. The framework is based on the assumption that notably the primate brains, and particularly those of the apes and humans, have evolved so as to be acutely sensitive to the detection and storage of predictive relationships. Memory systems store the knowledge thereby derived in a logical set of "files," which allows the natural operations of the brain to formulate creative behavioral patterns and solutions to novel challenges. Metaphorically, their storage might be thought of in terms of chords which, in their interaction with other memory chords, can be synergistic through the natural physics known otherwise to be characteristic of chords (e.g., fundamentals, resonance, harmonics, overtones, etc.).
Primate research of the 20th century affords strong support for the argument that primates employadvanced cognitive operations (Heyes & Huber, 2000). Neither their behavior nor ours is to be accounted for satisfactorily in terms of basic associative conditioning. For example, learning set formation; speech comprehension by apes; creative use by apes of stone for use as tools; and primates' planning, prediction, and symbolic skills for language and counting present major difficulties for explanation as conditioned responses. Here, the author provides an introduction to a framework called rational behaviorism. In stark contrast to the radical behaviorism of the 20th century, it allows for the emergence of new behaviors generated by cognitive operations of the primate brain. In addition to recognizing behaviors formed by respondent (Pavlovian) and operant (Skinnerian) conditioning procedures (Domjan, 1998), it incorporates a new category of acquired behavior—called emergents (Rumbaugh, Washburn, & Hillix, 1996). Emergents have their defined antecedents and consequences, though they are quite different from those for respondent and operant conditioning. The behavior class of emergents embraces all forms of cognition of animal life: their abilities to acquire concepts, to learn insightfully, to make and use tools (see Figures 2 and 3), to learn the basic dimensions of language, and in many other ways to manifest advanced intelligence. Thus, rational behaviorism is advanced in the hope of providing a comparative behavioral framework, one that rests on the Darwinian principle of continuity, both psychological and biological, particularly from apes to humans. The term rational refers to the inclusion of cognition; the term behaviorism acknowledges the fact that the only data available to our science are behavior. Rational behaviorism acknowledges and embraces all behaviors—respondent, operant, and emergent—and also those that are usually classified as instinctual.
The classic stimulus-response (S-R) model, used to account for the selection and shaping of behaviors contingent upon reinforcement, needs elaboration. That model focuses upon the conditioning of responses through acquisition, change, and execution that are called respondents and operants. Respondents and operants are noted for their specificity, fixedness, and predictability. By contrast, emergents are not conditioned as generally conceived, but rather are generated by integrative processes of relatively complex brains. Emergents can be manifested as new behavior patterns that are noted for being synergistic, integrative, and clever. Emergents also can be manifested as new capabilities, such as speech comprehension, that are not to be accounted for as responses or behaviors altered by basic conditioning procedures (Rumbaugh, Savage-Rumbaugh, & Washburn, 1996; Rumbaugh et al., 1996). The framework here presented proposes that learning instated via basic respondent and operant conditioning affords elements essential to the formation of emergent behaviors and capabilities. They afford predictive relationships and patterns between events and behaviors that can be useful to the integrative processes of complex brains. At the risk of oversimplification, conditioned behaviors are responses based on glands and muscles in relation to antecedent and consequent events; emergents are new behavior patterns based on principles and/or new capabilities generated by the natural integrative processes of brains.
Emergents reflect the natural operations of the brain as being comprised of keen pattern-detection and synthesizing systems. The patterns detected include those from classical and instrumental conditioning. In addition, they include observations obtained through the course of everyday life. Whether the patterns to be detected are provided by conventional conditioning paradigms or through observations of daily events, it is the logic structure of them that affords information to the organism for the generation of emergents.
Although emergents likely incorporate a very broad spectrum of the primate's respondent and operant conditioning history, there are several significant points that serve to differentiate them from respondents and operants (see Rumbaugh et al., 1996, for explication). Some of the main points of contrast are that:
- Emergents' initial appearances come as unanticipated "surprises" to the researcher.
- Emergents provide novel response patterns and solutions to problems.
- Emergents form covertly, hence unobtrusively or silently.
- Emergents afford new behaviors that have no specific reinforcement history.
- The formation of emergents generally cannot be charted.
- The formation of emergents emphasizes classes of experiences.
- Emergents entail the syntheses of individually acquired responses and experiences.
- Emergents are not subject to specific stimulus control as are respondents and operants.
- Emergents frequently reflect rearing conditions and/or early experience.
- Emergents tend to be associated with brain complexity (as per species and maturation).
Examples of emergents are provided in Table 1. In addition, the reader will be reminded of a host of other emergents—new behaviors generated by the subject—ones not otherwise taught or conditioned.
Special attention here is given to advances in ape-language research. Apes don't speak, hence we need a working definition to guide us in the controversy regarding their language skills (see Rumbaugh & Savage-Rumbaugh, 1994, for a review). The definition here provided is: Language is a neurobehavioral system that provides for the construction and use of symbols with which information and novel ideas are exchanged between individuals. The meanings of symbols are defined and changed through social interactions. With only one third the brain size of ours, it would be unreasonable to expect of nonhuman animals that they could master language to a level that would allow them to be a full partner in our culture. Notwithstanding, apes, dolphins, sea lions, and even a parrot have remarkable accomplishments in selected language skills. Without question, they learn the meanings of symbols and can both comprehend and use them as we use words—to represent events and things not necessarily present (e.g., Savage-Rumbaugh, 1986; Savage-Rumbaugh & Lewin, 1994). They initiate use of symbols to bring the attention of others to things and activities (e.g., games; see Figure 5) and occasionally have recounted events not otherwise known to the listener. Their use of word-symbols facilitates their ability to make accurate cross-modal judgments of sameness and difference (Rumbaugh, 1977), and they can comprehend novel commands and requests, conveyed both by gestures in dolphins and by speech in the case of Kanzi, a bonobo (Savage-Rumbaugh et al., 1993). Well beyond any reasonable doubt it is concluded that life-forms other than humans are capable of significant language and skills and that they use their skills to communicate in social contexts. That life-forms can master language skills is a strong data point in support of Darwin's principle of continuity, both biological and psychological, from animal to human (Greenberg & Haraway, 2002; Matsuzawa, 2001; Papini, 2002). The language they are capable of mastering in association with humans neither dimishes their accomplishments nor is grounds to conclude that they have no natural language of their own in the wild. To date, animals have been more facile in learning language systems that we afford to them than we are at learning any language system that might be native to them.
Let us consider how traditional conditioning terms might be modified and how new ones might be defined for the writing of a rational behaviorism intended to unify respondents, operants, and emergents.
Be reminded that here it is assumed that, notably in the case of apes, the evolution of brains includes systems that are highly sensitive to patterns among things and events and to temporal/predictive relationships between events. Brains are inherently sensitive to things and events that tend to occur either together or sequentially. Brains have become highly proficient in these operations in response to selective pressures of survival. Brains have evolved to be sensitive to the logic patterns of environments and experiences associated with them. If an environment has no logic structure, it generates only noise. Brains learn nothing from noise. On the other hand, in environments where there is logic and/or relationships between things and events, brains might take in a great deal of information and organize it so as to achieve "best fits" among both the specifics and principles to be titrated from life experiences. In the extreme, the brain "learns everything," though it places differential emphasis upon some things over others because of their relative salience (see below).
Assume that through the course of development and attendant general exploration and experience in the environment, the organism (especially if primate) will learn both specifically and comprehensively about it. The organism should learn that the environment affords:
The organism also will learn of:
- Events—where some are controllable and others are not
- Resources—where some prove useful, while others are useless or irrelevant
- Risks—the unknown, dangers, or painful/aversive events—all to be avoided
Assume that organisms might learn, through happenstance, curiosity, exploratory behavior, and observation of others, how some events and resources can be managed. Assume also that they might also acquire emergent (i.e., new) ways, more efficient ways, more interesting ways of obtaining and/or using resources than those prescribed by convention.
- Costs—All activity, all behavior entails metabolic costs and, in the extreme, discomfort and pain. Every opportunity entails risks and costs.
Rational behaviorism posits that the organism will, given the opportunity, attend to the consequences of its behavior. It will monitor the environment, its own behavior, and the behavior of others, and will be sensitive to the consequences thereof. The organism will behave and monitor its behavior to the physical and social environs as though it is alert to the detection of possible cause - effect relationships. In sum, even in its elementary expressions, behavior has been selected to control events in the service of sustaining life and maintaining states of being that, insofar as possible, are not aversive and that, in the best of all states, afford relative comfort. Regardless, through their behavior, organisms strive for control over their environments and themselves. Whereas control might only be suggested in the behaviors of relatively simple organisms, control is for real in complex organisms and notably in the larger primates (particularly the great apes and humans). Control can be achieved in measure by what we will call biological smartness, but strategies and tactics in the pursuit of control are perhaps most facilely designed and carried out by psychological intelligence.
Other terms to employ include:
Learning—Units of knowing about (a) what events are associated contiguously in time and (b) what events are sequentially ordered. Learning can be inferred only through behavior; what specifically has been learned can be inferred only by changes in behavior. The scope of learning, though difficult to assess, can be more comprehensive than the first observed changes in behavior might indicate. Simple learning might be limited or adequately accounted for in the traditional S-R model. Notwithstanding, as a positive function of brain and neural system complexity, the structure and scope of complex learning might involve the formulation of principles that are, in their originations, well beyond the bounds that can be accounted for by a basic S-R model.
Salience—Relative to one another, things and events that capture the attention are said to be salient. Things and events might be inherently salient (e.g., thunder and lightning on a clear, hot summer day) or salient due to past experience and at-tendant learning. Inherently weak stimuli can become highly salient because of what other things and/or events have reliably been associated with them, either contiguously or sequentially in time. Salience is a basic parameter of perceptual attention, learning, and performance, and can be very powerful in that function. Highly salient cues can be inherently powerful and used as unconditional stimuli (i.e., shock, bright light, tapping a tendon, loud noise, etc.) in classical conditioning. Relatively subtle (i.e., typically nonsalient) cues can become highly salient if used as conditional stimuli in classical conditioning—if they have histories of reliable association with highly salient things or events (e.g., they can function as Hull's secondary reinforcer and as Skinner's conditioned reinforcer).
Stimulus—A unit of raw energy that impinges upon one or more receptors of the body. A stimulus may have its origin in the external environment or within the body. It may evoke a sensation, depending upon its strength. (Note: A stimulus lacks definition in the perceptual field of the subject in that it lacks interpretation.) To the degree that a stimulus is specific, it can acquire cue property (below); to the degree that a stimulus is strong, it is salient and can induce a drive or motivation for the organism to pursue or avoid it. (This perspective is consistent with Miller and Dollard's.)
Cue—A stimulus that, because of its distinctiveness, has come to serve as a sign or marker to signify possible consequences if responded to or acted upon by the organism.
Response—A rather specific action of a muscle or gland. In a conditioning experiment, the response might be targeted for becoming conditioned to some conditional/discriminative stimulus. It is relatively constrained in its form and function. Examples are as in a reflex arc, pressing a bar, pecking a key, and so on.
Behavior—Actions that generally entail the patterning of neural events (as in thinking) and/or the coordination of muscle groups. Examples include running a maze, riding a bicycle, dancing, singing, climbing, talking, attending conventions, foraging, writing, and so on.
Resource—Foods, liquids, shelters, objects (i.e., tools, rocks, ropes), money, other animates, and so on. (Portions of resources may serve as benefits, defined below, for both responses and behaviors.)
Incentive and Goal—Something relatively specific to be earned or obtained, given the appropriate response or behavior in the appropriate temporal and physical contexts. Incentives are bits of the organism's perceived or known environmental resource bank. An incentive might be a controllable event or a preferred activity (such as play) as well as objects that are quite tangible (such as money). A goal generally is more complex than is an incentive, as here defined. Notwithstanding, a goal has incentive value and functions accordingly.
Benefits—Benefits generally are resources harvested from the environment through behavior. Benefits also can be inherent in certain behavioral changes and states (e.g., resting/sleeping after high energy expenditures) and help to sustain both life and the quality of life (e.g., good foods instead of "just foods" to eat, etc.). Benefits serve to justify the execution of behaviors by their servicing biological needs (e.g., food and water) and quality of life. Benefits are obtained by foraging, running mazes, using manipulanda, social interactions, and so on. Because a benefit can have the consequence of bringing sharp focus to a specific behavior and its perseveration, as in conventional instrumental conditioning procedures, it might erroneously be concluded that a specific behavior or response is all that the organism has learned. The revised framework here advanced would allow for the subject to be learning, all the while, rather comprehensive principles about the context and attendant qualities/relationships thereof for future adaptive behavior patterns. (Thus, the reader should not conclude that what is here called a benefit is the same as what traditionally has been termed a reinforcer or reward.)
Costs—All responses, all behavior entails costs of execution (fatigue, risk of injury, boredom, burning of blood sugars, time, tissue wear, injuries, etc.).
Note: No longer needed is the concept of reinforcement. Reinforcement now is subsumed, viewed, or redefined as an environmental resource or as an incentive obtained as a benefit because of behaviors engaged in by an organism. The subject learns about it, how it might be used or what it can be used for, and how it can be accessed when needed. In its traditional role, then, reinforcement as a term is no longer needed and hence is deleted as a useful term. (It should be noted that the term reinforcement has been difficult to define except in terms of its effect, such as being anything that increases the probability of a specific response.)
Rational behaviorism incorporates basic learning afforded by classical and instrumental conditioning paradigms along with learning induced by logic structure of the environment as perceived by the organism. Rational behaviorism anticipates that what is learned by an organism likely will be far more comprehensive and complex than single responses to single stimuli. Synergistic use of whatever is learned will enable the organism to be creative and clever as some positive function of the complexity of its brain, as particularly exemplified in great apes and humans.
Implications for Research
It is clear that we need to refine our research strategies and tactics so that the detection and measurement of the emergents is more probable and that we then succeed in defining the antecedents that generated them. In turn, we need to be more sensitive to the possible effects that early rear-ing and social contexts might have both upon what is learned and the expression of what has been learned. Generally, we should no longer be content to study how single responses are acquired by the procedures of traditional reinforcement. To do so biases our findings to support the proclamations of Descartes and traditional behaviorism that animals are stupid and that there are no rational bridges between the dynamics of their behaviors and ours.
Question—How might the brain work to accommodate this revised perspective? What natural mechanisms of physics might generate emergent behavioral options from neural systems?
Assume that both field and laboratory data now lend substantial support to the views that:
- The brain, in its evolution and design, detects and stores predictive relationships or patterns between events/things, particularly during infancy.
- Relationships between events/things are stored by codes that metaphorically can be conceived of as having the attributes of musical tones and chords—with fundamentals, timbre, overtones, harmonics, and so on.
- Through neural processes and circuits that afford or service memory, these chords recurrently announce their presence; but whether or not anything generative happens depends upon their selective activation by motivational states, the interactions of two or more chords, and/or the formation of new chords that share certain attributes (e.g., portions of code, notes, chord structure and harmonics, and so on).
- To the degree that new chords induce resonance in other existing chords, synergism (Hebb, 1949; Lyon & Krasnegor, 1996) and generativity commence and provide for what otherwise is called thinking, from which new patterns of behavior might emerge, reflecting the formation of new chords with their own unique timbre, overtones, resonance. Thinking, in turn, might produce creativity, insightfulness, and/or new capabilities.
Question—Behaviorally, how might the validity of this model be evaluated even now?
- We benefit from a quiet environment when we are problem solving; a quiet context constrains the activation of irrelevant chords (i.e., thoughts, ideas). Background sounds should not impede thought through the activation of irrelevant chords (i.e., neural memory systems).
- Cues and/or hints can enhance deliberate efforts to think, to be creative. Herein lies the value of brainstorming ideas with others who have relevant interests and backgrounds.
- Specific contexts, across time, can support specific themes/topics of thought.
- It takes "time to recall the distant past."
- Long-term memory is served by having a period of silence and rest after a learning session. This process of consolidation is served by not having extraneous stimulation or activity following learning.
The complexities of phenomena defined in comparative studies of learning, cognition, and language over the past half century press us to go beyond traditional conditioning frameworks. To that end, a new class of behavior, emergents, and framework, rational behaviorism, have been introduced. It has been posited that in their evolution, brains generally have become more complex and increasingly sensitive to the predictive relationships between temporal and sequential events (both environmental and behavioral) as some positive function of their salience. In their initial appearance, emergents have an element of surprise because they have no specific history of training or conditioning. Their formation has been silent, unobtrusive. It is posited that they are logical extensions of information provided by general life experiences, observational learning, and both respondent and operant conditioning. Early environment plays an important role in the development of the integrative processes by which a brain generates emergent behaviors and competencies. Emergents, once discovered, may be understood in terms of the conditions which give rise to them (i.e., their antecedents), their neurophysiological bases, and how they function to service adaptation through organisms' management of environmental resources and risks.
Beran, M. J., & Rumbaugh, D. M. (2001). "Constructive" enumeration by chimpanzees (Pan troglodytes) on a computerized task. Animal Cognition, 4, 81-89.
Blodgett, H. C. (1929). The effect of the introduction of reward upon the maze performance of rats. University of California Publications in Psychology, 4, 113-134.
Domjan, M. (1998). The principles of learning and behavior (4th ed.). Pacific Grove, CA: Brooks/Cole.
Ellen, P., Soteres, B. J., & Wages, C. (1984). Problem solving in the rat: Piecemeal acquisition of cognitive maps. Animal Learning and Behavior, 12, 232-237.
Gallup, G. G., Jr. (1987). Self-awareness. In G. Mitchell & J. Erwin (Eds.), Comparative primate biology: Vol. 2B. Behavior, cognition and motivation (pp. 3-16). New York: Alan R. Liss.
Gibson, K. R., Rumbaugh, D. M., & Beran, M. J. (2001). Bigger is better: Primate brain size in relationship to cognition. In D. Falk & K. R. Gibson (Eds.), Evolutionary anatomy of the primate cerebral cortex (pp. 79-97). Cambridge: Cambridge University Press.
Greenberg, G., & Haraway, M. M. (2002). Principles of comparative psychology. Boston: Allyn & Bacon.
Harlow, H. F. (1949). The formation of learning sets. Psychological Review, 56, 51-65.
Hebb, D. O. (1949). The organization of behavior: A neuropsychological theory. New York: Wiley.
Heyes, C., & Huber, L. (2000). The evolution of cognition. Cambridge, MA: MIT Press.
Lyon, G. R., & Krasnegor, N. A. (Eds.). (1996). Attention, memory, and executive function. Baltimore, MD: Paul H. Brookes.
Matsuzawa, T. (2001). Chimpanzee mind. Kyoto, Japan: Primate Research Institute.
Menzel, C. R. (1999). Unprompted recall and reporting of hidden objects by a chimpanzee (Pan troglodytes) after extended delays. Journal of Comparative Psychology, 113, 426-434.
Menzel, E. W. (1973). Chimpanzee spatial memory organization. Science, 182, 943-946.
Papini, M. R. (2002). Comparative psychology. Upper Saddle River, NJ: Prentice Hall.
Rumbaugh, D. M. (1977). Language learning by a chimpanzee: The Lana Project. New York: Academic Press.
Rumbaugh, D. M., Hopkins, W. D., Washburn, D. A., & Savage-Rumbaugh, E. S. (1989). Lana chimpanzee learns to count by "Numath": A summary of a videotaped experimental report. The Psychological Record, 39, 459-470.
Rumbaugh, D. M., & Pate, J. L. (1984). The evolution of cognition in primates: A comparative perspective. In H. L. Roitblat, T. G. Bever, & H. S. Terrace (Eds.), Animal cognition (pp. 403-420). Hillsdale, NJ: Erlbaum.
Rumbaugh, D. M., & Savage-Rumbaugh, E. S. (1994). Language in a comparative perspective. In N. J. Mackintosh (Ed.), Animal learning and cognition (pp. 307-333). San Diego, CA: Academic Press.
Rumbaugh, D. M., Savage-Rumbaugh, E. S., & Washburn, D. A. (1996). Toward a new outlook on primate learning and behavior: Complex learning and emergent processes in comparative perspective. Japanese Psychological Research, 38, 113-125.
Rumbaugh, D. M., Washburn, D. A., & Hillix, W. A. (1996). Respondents, operants, and emergents: Toward an integrated perspective on behavior. In K. Pribram & J. King (Eds.), Learning as a self-organizing process (pp. 57-73). Hillsdale, NJ: Erlbaum.
Savage-Rumbaugh, E. S. (1986). Ape language: From conditioned response to symbol. New York: Columbia University Press.
Savage-Rumbaugh, E. S., & Lewin, R. (1994). Kanzi: At the brink of the human mind. New York: John Wiley.
Savage-Rumbaugh, E. S., Murphy, J., Sevcik, R. A., Brakke, K. E., Williams, S., & Rumbaugh, D. M. (1993). Language comprehension in ape and child. Monographs of the Society for Research in Child Development, 58 (3-4, Serial No. 233), 1-242.
Schick, K. D., Toth, N., Garufi, G., Savage-Rumbaugh, E. S., Rumbaugh, D. M., & Sevcik, R. A. (1999). Continuing investigations into the stone tool-making and tool-using capabilties of a bonobo (Pan paniscus). Journal of Archaeological Science, 26, 821-832.
Schrier, A. M., Harlow, H. F., & Stollnitz, F. (Eds.). (1965). Behavior of nonhuman primates: Modern research trends. New York: Academic Press.
Sidman, M. (1994). Equivalence relations and behavior: A research story. Boston: Authors Cooperative.
Tolman, E. C. (1948). Cognitive maps in rats and men. Psychological Review, 55, 189-208.
Treichler, F. R., & van Tilburg, D. (1996). Concurrent conditional discrimination tests of transitive inference by macaque monkeys: List linking. Journal of Experimental Psychology: Animal Behavior Processes, 22, 105-117.
Washburn, D. A., & Rumbaugh, D. M. (1991). Ordinal judgments of numerical symbols by macaques (Macaca mulatta). Psychological Science, 2, 190-193.
At right: Nyota (at age 3 years, soon to be 4 years old) is being reared in a language-structured environment. The emergence of his comprehension and use of lexigrams and speech comprehension is being closely studied by E. Sue Savage-Rumbaugh and William Fields (photograph courtesy of Elizabeth Pugh).
Kanzi learned how to knap flint so as to obtain sharp cutting edges by watching an expert, Dr. Nick Toth (Schick et al., 1999). He generally holds the cobble, from which a large flake is to be made, in his left hand and strikes it with great force with a hammer stone in his right hand.
The larger the rope to be cut, the larger the chip Kanzi makes for the task. Once dull, chips are discarded. Kanzi slices at rope or hide with precision.
Figure 4 - The Emergence of Positive Transfer of Learning in Relation to Evolution of the Large Primate Brain
The horizontal axis presents various primate taxa that include prosimians, New and Old World monkeys, and the lesser and great apes in an order that generally approximates the complexity of their brains. The vertical axis indicates changes in transfer of learning as a result of increasing the amount learned from 67% to 84% correct on groups of two-choice discrimination problems (Rumbaugh, Savage-Rumbaugh, & Washburn, 1996). The emergence of positive transfer is a qualitative shift in effect in relation to brain evolution (r = 0.79) and likely is a reflection both of a shift in learning process from associative to relational as well as an enhancement in transfer of learning of skills. In addition, a positive correlation is posited between cranial capacity and indications of intelligence and culture for the hominids (Gibson, Rumbaugh, & Beran, 2001). [Phaner and Microcebus data provided by W. Cooper; Cebus data provided by E. Visalberghi.]
This article is based on Dr. Rumbaugh's Psi Chi/Frederick Howell Lewis Distinguished Lecture presented on August 24, 2001, during the 72nd Annual Psi Chi National Convention, held in conjunction with the 109th Annual Convention of the American Psychological Association in San Francisco, California.
For her first few years of life, Panzee was co-reared with Panbanisha by E. Sue Savage-Rumbaugh in an environment that was language structured. English was spoken in coordinated use of word-lexigrams. Both apes came to understand English words and, in the case of Panbanisha, simple sentences of request. The comprehension and competent use of lexigrams are excellent examples of emergents. Panzee is now 16 and spontaneously recruits uninformed persons for assistance and directs them by pointing and gesture to food hidden for her in the woods (C. R. Menzel, 1999).
Through skilled use of a joystick, Lana controls the cursor so as to count out sets of items as required by the value of the target number on a given trial. She can construct sets of items quite accurately up to 7 (Beran & Rumbaugh, 2001). Lana was our first subject in language research that commenced in 1971.