Cognitive science: From computers to anthills as models of human thought

Peter Gärdenfors


With the development of computers in the 1940s and 1950s, a new model for human thinking became available. The initial period of cognitive science was driven by the analogy that the brain functions like a computer. Consequently, thinking was viewed as the processing of symbols. This was also the methodology of classical artificial intelligence. As a result of criticism of the symbol manipulation paradigm, there have recently been two main kinds of reaction to it. The first one is connectionism, where thinking is modelled as associations in artificial neuron networks. Some connectionist models are tightly connected to developments in the neurosciences, while others are more general models of cognitive processes such as concept formation. The second reaction consists of theories of embodied and situated cognition, where cognition is seen as taking place not only in the brain, but in interaction with the body and the surrounding world. In line with this, modern studies of robotics are based on so called reactive systems, the actions of which depend directly on the world instead of a symbolic model of it. The situated view on cognition will also be central for future developments of man-machine interaction, in particular in educational tools that exploit information technology.

Full Text: