The computer as a metaphor for psychological inquiry: Considerations and recommendations

Abstract

My concern is with the computer as a metaphor for explaining perception and action. A representative sample of arguments for and against the paradigm are presented and evaluated. The conclusion is that the idea of computation is productive for achieving a functionalist description of how we perceive and act. This level of description can contribute to our understanding independently of description achieved at the levels of neurophysiology and phenomenology. Some of the perceived limitations in the computational method rest on the assumption that the symbolic level must be discrete and abstract. In fact, worthwhile explanations within the information processing framework utilize continuous, modality-specific processes and representations as explanatory devices. One suggestion for a movement from the discrete to the continuous mode is advised to bring computational theories in line with the psychological phenomena they describe. Various alternatives to the computational framework are considered and found to be inadequate substitutes. An example of research is used to demonstrate the value of the continuous mode and the computational level of explanation.