Architectural/Rational Analysis Dichotomy

An important issue in the study of Cognitive Science is the effect of the cognitive architecture on observable behavior. Anderson, in Architectures for Intelligence, Chapter 1, takes the position that the specific architecture underlying an organism is unimportant, and instead behavioral mechanisms can be derived purely from the environment. He describes this technique as "rational analysis." The basis for his position is that, through evolution, behavior is optimized to the external environment and the goals of the organism.

Other cognitive scientists, such as Newell and Simon, view rational analysis as extreme. One criticism of this approach is that one cannot know for certain the goals of an organism. Also, evolution does not arrive at an absolute maximum, only a local maximum. Instead, it is more important to determine which primitive architectural mechanisms are present in human cognition, and how they combine to give rise to intelligent behavior.


Back to List of Dichotomies