|
|
|
|
|
|
|
arguments, it is difficult to eliminate ambiguities and contradictionsas with proverbs, proper application depends upon the intuition of the user. Specific applications of the formalism may arise, but these will probably be in areas of little uncertainty, where theory was not actually required; the formal procedures will be primarily corroborative. Only after considerable effort at this level can we hope for a theory mathematically rigorous and conceptually generala genuinely predictive theory organizing large masses of data at many levels. |
|
|
|
|
|
|
|
|
One of the earliest suggestions (or corroborations) from the formal underpinnings of the  framework is quite fundamental. It can be established that major aspects of the behavior of any very complex system fall outside the explanatory power of simple input-output (S-R or switching) theory. This result is a rigorous version of the observation that ongoing activity in a complex system usually depends upon the past history of that system. This dependence, which both psychologists and computer theorists call "memory," finds its formal counterpart in the notion of state: distinct stimulus-state pairs generally giving rise to different responses. If there are many states (and, by any reasonable definition of state, the CNS has an astronomical number) the same stimulus may give rise to a great many different responses. Thus, observation of stimulus-response pairs will not enable us to discover the mode of operation of any system with a substantial number of states. For a system as complex as the CNS, such a result can be ignored only to the great detriment of the ensuing theory. It is a corollary of this result that complex systems can act in autonomous fashion, producing continuing response sequences in the absence of new stimulus. Thus, a stimulus may serve only to modify ongoing activity rather than to initiate it. In short, the responses of the CNS cannot be explained wholly in terms of concurrent stimuli. |
|
|
|
|
|
|
|
|
The framework also emphasizes a second important point. An adequate theory must include more than a formal counterpart of the internal processes of the system being studied. The environment (or range of possible environments), the information received therefrom, and the ways the system can affect the environment, must also be represented. Moreover, the criterion  emphasizes the importance of performance "along the way." The CNS cannot wait indefinitely for "useful" outcomes; some minimal level of ongoing performance is required. (E.g., if food is not obtained with sufficient frequency, death ensues, totally removing the possibility of further goal-oriented behavior.) Such observations are not new, but the framework does provide a form for fitting and arranging them, and it lends them emphasis. This at least gives us a fresh look at familiar facts, occasionally suggesting new consequences which might otherwise be overwhelmed in the plethora of macro- and micro-data (behavioral and physiological). |
|
|
|
|
|