Endo-System View as a Method for Constructive Science

Cognitive Science talks about internal processes and representations of intelligent agents. Since we ourselves are the best examples of such intelligent agents, observation from inside (introspection) is useful. However introspection violates objective criteria required for analytical science. On the other hand, observation from outside has only limited access to internal processes and representations. We propose to use endo-system view, a methodology in which a constituent of a system of interactions observes the whole system from inside. Meta-cognition, cognition of its own cognitive processes, is an example of such endo-system view. This makes contrast to exo-system view implicitly employed for analytical science. Endo-system view, intrinsically due to its definition, interacts with and therefore affects the very interactions being observed. We, rather, take advantage of this interaction as means to analyze the system of interactions scientifically. More generally, we are proposing a new way of science, "constructive science", in contrast to "analytical science". The basic methodology of analytical science is to divide a system into its simpler subsystems and recursively analyze them till we understand every subsystems and the structure of their connection. A precondition of doing this is that the analysis process does not affect the system being analyzed. Constructive science, on the other hand, starts from simple units and tries to construct a complex system. In this paradigm, therefore, the system is to be observed as it changes dynamically. This paper is to discuss a possibility to employ endo-system view as objectively testable methodology in constructive science.