Having rejected intersubjective agreement and verifiability as criteria for the acceptance of observation reports, Skinner seems to open the way for first- person reports of private stimuli to qualify as observation reports. However, Skinner does not take this final step. As explained in chapter 11, Skinner's theory implies that the connection between a private event and a verbal response will not usually be a reliable one. He therefore does not trust first- person reports of private events and does not use them as observation reports. His reasons for rejecting introspection are thus based, in part, on an unproven empirical hypothesis and constitute a fifth behaviorist objection to introspection. ( Skinner's other reasons for ignoring introspection are discussed in chapter 11.) In practice, then, Skinner does agree with the behaviorist exclusion of introspection from scientific methods of observation.
Some behaviorists follow Skinner's lead but without his cautious attitude. Given Skinner's theory of how verbal responses to private stimuli can be learned, some behaviorists admit first-person reports of private events as full- fledged and significant sources of scientific data. Although these behaviorists still emphasize the importance of observing behavior, they also accept introspection as a valid method for observing private data. 54 Their position, however, must be located at the periphery of the behaviorist conceptual family. The central behaviorist position is the rejection of introspection as a method of scientific observation for the various reasons reviewed above.
Behaviorism rejects introspection as a method of scientific observation on a number of grounds. Those based on the nature of the contents of introspection (e.g., that they are subjective or private) do not stand up to close analysis. On the other hand, objections based on the nature of the method itself fare better. First, there is the methodological objection that introspection fails to achieve intersubjective agreement. Second is the empirically based objection that introspective reports are necessarily unreliable because of the way they are supposedly learned and maintained. These two objections demonstrate two features of behaviorist conceptions of objectivity: (1) Objectivity is closely associated with intersubjective agreement and communication; and (2) standards of objectivity depend, in part, on the findings and theories of the science of behavior.