The main part of this book has explored different ways of thinking about and developing the basic idea that cognition is a form of information processing. As we have discussed, there are different models of information processing, and so different ways of developing this fundamental framework assumption of cognitive science. From the perspective of classical cognitive science, digital computers are the best models we have of how information can be processed. And so, from the perspective of classical cognitive science, we need to think about the mind as a very complex digital computer that processes information in a step-by-step, serial manner. From a more neurally inspired perspective, in contrast, information processing is a parallel rather than serial process. Neural networks can be used to model information processing and problem-solving through the simultaneous activation of large populations of neuron-like units.
Chapter 13 explored alternative ways of analyzing and predicting the behavior of cognitive systems. The dynamical systems approach is one alternative, analyzing cognition in terms of variables evolving through state space, rather than the physical manipulation of symbols carrying information about the environment. Closely related to the dynamical systems approach is the situated cognition movement. This second alternative to information-processing models is inspired by studies of how insects and other low-level organisms solve complex ecological problems, and by illustrations from robotics of how complex behaviors can emerge in individuals and groups from a small repertoire of hard-wired basic behaviors.