pmd@cbsck.UUCP (Paul Dubuc) (07/30/85)
A response to Tim Maroney: >... Behaviorism does not explain behavior in terms of a >deterministic stimulus-response model. That is characteristic of the >Pavlovian models that preceded behaviorism. Stimuli are dealt with as a >minor sub-topic, being a factor that correlates with the frequency of >emission of operants, but there is really no such thing as a "response" in >behaviorist models. Instead, a probabilistic model based on frequency of >emission of observable behaviors called "operants" is used. > >Behaviorist models deal in a completely scientific way with behavior by >dealing only with behaviors that can be objectively measured in the >laboratory, and by reporting observations rather than speculating on >unprovable causative factors. It NEVER says anything like "Organism O >emitted behavior B at time T because of factors F". It does, however, say >things like "The frequency with which organism O emitted behavior B was >higher when reinforcement schedule S was applied than when S-prime was >applied." Does this explanation really make much difference in how the behaviorist model is applied? What is the extent of its usefulness if it deals "only with behaviours that can be objectively measured in the laboratory" (What are these?) and produces only a statistical model of behaviour? How is the model completely scientific? Is it predictive? I can think of applications for the model in the areas of economic theory, politics, and criminal justice. When it is applied it seems to take the emphasis off of individual responsibility and on to the individual's environment. The method of changing the behaviour of people emphasises changing their environment. But what do statistics say about how I behave when a particular "reinforcement schedule" is applied to me? Not much. The probability is obtained from an experimental population and really can't predict very well how any given individual will act. If the behaviorist model is to be applied at all, it seems that it must be applied in a deterministic way. Its application to society is an application to me even though it may say nothing about how I will behave. Also, how does the limitation to "objectively measureable" behaviour impact on the usefulness of the model for any application. Again, if the behaviorist model is going to be applied to anything, it must be with the assumption that this limitation is insignificant. This seems to make it more of a deterministic application. Paul Dubuc