CEPA eprint 1341 (EVG-051)

Another minor revision, or the disregard for control theory and the analysis of inductive feedback systems

Glasersfeld E. von (1978) Another minor revision, or the disregard for control theory and the analysis of inductive feedback systems. The Behavioral and Brain Sciences 1(1): 79–80. Available at http://cepa.info/1341
[Commentary on Bindra D. (1978) How adaptive behavior is produced: A perceptual-motivational alternative to response reinforcements. The Behavioral and Brain Sciences 1(1): 41–52.]
Many who have in the past taken issue with the dogmatic simplifications and distortions perpetrated in the name of Behaviorism will tend to agree with Bindra’s introductory declaration that “the basic tenets of the response-reinforcement framework are invalid” and that “the difficulties cannot be simply patched up with minor revisions.” Reading on, however, it soon becomes clear that different critics of that framework may have quite different ideas as to which basic tenets are invalid and what major changes would have to be made in our view of behavior in order to arrive at a more plausible explanatory framework or model
Bindra, clearly, is struggling to get away from the primitive assumption that behavior is determined in a direct, linear cause-effect way by stimuli Yet he undermines his efforts and minimizes his chances of success by perpetuating a terminology that inevitably brings with it much of the traditional behaviorist conceptualization Learning, he postulates, consists in the forming of an association between a perceptual stimulus Sn and the hedonic stimulus SH, and this learned association then manifests itself because Sn comes to generate “the same motivation (central motive state cms) as is normally generated by the hedonic stimulus (SH) itself.” And, he adds, “this central motive state, in combination with the detailed sensory-spatial features of the situation, then determines what response will emerge” Contrary to S-R theory (according to which a response is “emitted” under certain stimulus conditions because it was linked to one or more of the stimuli by means of reinforcement), Bindra proposes that the organism does not have a repertoire of response behaviors, but that each individual response “is a fresh construction dependent on the motivational state and the pexgo (i.e., presently excited gnostic organization) generated by the prevailing stimulus situation.”
I would suggest that the word “motivation” is being used here in two rather different senses and that this is misleading because, in conjunction with the term “hedonic stimulus,” it makes it all but impossible to focus on the crucial connection between motivation and satisfaction On the one hand, a hedonic stimulus is said to generate motivation (and to transmit this capacity to other stimuli by association), and this motivation No 1 leads to the construction of a response behavior that. on the basis of “structural gnostic elements” (structured, presumably, by prior experience), is expected to lead to some kind of consumption (or avoidance) and hence to satisfaction On the other hand, there is the organism’s general motivational state, that is to say, motivation No 2, which refers to the kind of consuming the organism wants or is driven to look for at the moment (e g, hunger) The indiscriminate use of “motivation” for both these items hides the fact that when Bindra talks about “detailed sensory-spatial features of the situation” determining “what response will emerge,” or “an environmental stimulus complex that is critical for the production of a certain act,” he is making the tacit assumption that the organism, in its prior experience, has established that, given certain “sensory-spatial features” or “environmental stimulus complexes,” certain ways of acting are likely to lead to particular forms of satisfaction In other words, if we are discussing “learning” (and not primary reflexes), the central representation of the visual stimulus I call “apple” becomes a hedonic stimulus only after I have at least once (a) bitten into an apple, (b) liked the result, and (c) in some way recorded that connection If from this we form the hypothesis that the satisfaction derived from biting into an apple generates a specific kind of association between the apple-stimulus and the act of biting, we get a rather widely applicable principle or model that, of course, is really no more or no less than the well-known principle of inductive inference As Maturana (1970) put it: “The living system, due to its circular organization, is an inductive system and functions always in a predictive manner what occurred once will occur again.” For the inductivist, the association between stimulus and response is, of course, not causal but instrumental in that it becomes an operative link if, and only if, the organism wants or needs the kind of satisfaction that was experienced in conjunction with that stimulus and that act (i.e., when there is motivation No 2)
Though Bindra says “there is little doubt that animals learn the correlations between their actions and the consequences of those actions,” he maintains that these correlations in no way explain how specific actions get produced This difficulty, I would suggest, arises only if one assumes that the sensory signals out of which an organism builds up its “central representation” of a perceptual stimulus or stimulus configuration (e.g, an object-concept of apple) are data that are altogether different in kind or dimension from the proprioceptive signals out of which the organism builds up central representations of its motor acts [see Roland, this issue] From a cyberneticist’s point of view, they are all input (Powers, 1973) That is to say, they are all proximal, and the problem of making combinations with distal stimuli simply cannot arise for any organism, because there is no such thing as a distal stimulus in an organism’s experience (Only an observer can meaningfully distinguish “distal” and “proximal” relative to an observed organism ) Also, there is a considerable body of evidence from work with infants that indicates that the representation of objects, as items independent of the infant’s motor activity involved in touching or grasping them, is only gradually and laboriously developed and externalized (Piaget, 1937) [see also Brainerd, this volume, next issue, forthcoming]
The main problem Bindra sees in the formulation of “specific principles of response production” springs from two sources First, his rigid and rather arbitrary dichotomy between fixed “motor programs” that last half a second or less, and “long sequences of acts” for which, he says, “there is no lasting ‘memory’ basis.” Since he considers the short acts to be rigidly fixed and stimulus dependent, he needs an altogether different mechanism for the construction of “act-assemblies.” Second (and this, I believe, is due also to his profound involvement in the traditional behaviorist conceptualization), he completely disregards the principles of control theory and hierarchical feedback systems (Craik, 1966; MacKay, 1966; McFarland, 1971), which, ever since the pioneering work of Craik and Wiener in the 1940s, have given us not only theoretical but also functioning models of goal directed motor programs and act-assemblies that operate on the basis of continuous, situation- specific adaptation
In short, from this reader’s point of view, Bindra’s proposal is still an attempt at “minor revision” that does not get away from the traditional tenet that living organisms are passive receivers of stimuli to which they react according to pseudo-mechanical principles A detailed analysis of what is covered by the term “gnostic organization” would, I believe, reveal the indispensability of induction and of inductively built-up representations of desirable and undesirable states and events that can then serve as goals in the assembly of adaptively modifiable behaviors
Craik K. J. W. (1966) The nature of psychology. Cambridge, Cambridge University Press.
MacKay D. M. (1966) Cerebral organization and the conscious control of action. In: J. C. Eccles (ed.) Brain and Conscious Experience. New York, Springer: 422–45.
Maturana H. R. (1970) Neurophysiology of cognition. In: P. L. Garvin (ed.) Cognition: A Multiple View. New York, Spartan Books, pp. 3–23. http://cepa.info/536
McFarland D. J. (1971) Feedback mechanisms in animal behavior. New York, Academic Press.
Piaget J. (1937) La construction du réel chez l’enfant. Neuchâtel, Delachaux et Niestlé.
Powers W. T. (1973) Behavior: The Control of Perception. Chicago, Aldine.
Found a mistake? Contact corrections/at/cepa.infoDownloaded from http://cepa.info/1341 on 2016-12-31 · Publication curated by Alexander Riegler