One, class A, projects into action language as agent-acts, while the other, class N, represents nonaction language. The behaviorist interpretations of agency, reviewed above, are attempts to accomplish this task.
At least one member of class N is already known, namely behavior brought about by respondent reflex action. A tap to the patellar tendon elicits a knee jerk. This kind of environment-behavior relationship projects into action language as an example of a nonaction, not under the control of the agent who may resist the movement but remain helpless to prevent it. The conceptual error is to assimilate all behavior caused by, or under the functional control of, environmental variables to the model of the patellar reflex. 81 If the only relationship between environment and behavior is that represented by the patellar reflex, then all behavior belongs in class N, and class A is the null set, given the behaviorist assumption of lawfulness in behavior. The behaviorist is then accused of portraying the organism as "passive," "helpless," and "coerced" into acting.
This erroneous assimilation ignores the variety of relationships reviewed in the preceding chapters on S-R psychology. Its roots are to be found in the simple reflexological model which similarly assimilates all behavior to a stimulus-elicitation-response picture. With the rejection of this model arises the possibility of discovering a distinction within response language corresponding to class A. This distinction would then characterize those features of environment-behavior interaction which define agent-acts. That humans can reliably discriminate between actions and nonactions indicates that such features exist. Whether these features also correspond to the properties of behavior singled out as important in the behavioral science is a question yet to be answered. 82
S-R psychology thus leaves organisms as it found them. All behavior is assumed to be lawfully related to the environment, but some of this behavior is related to the environment in such a way as to be categorized in action language as an agent-act, volitional, active, and brought about by a person. S-R psychology can thus be seen as compatible with the concept of agency understood in this way. 83 Whether this concept of agency will prove useful in the behavioral science is as yet unknown.
S-R psychology is not, however, compatible with the notion of a free-willed self-initiating agent. Indeed, any deterministic psychology, behavioristic or not, must reject this type of agency. Determinism is a critical working assumption not readily abandoned by any scientific psychology, even if the former is restricted to the macro-level and to statistical laws.
Behaviorism's externalism, on the other hand, with its concentration on external environmental causes of behavior, does not follow necessarily from