In previous chapters, the elements of classical, of connectionist, and of embodied cognitive science have been presented. We have proceeded in a fashion that accentuated potential differences between these three schools of thought. However, now that the elements of all three approaches have been presented, we are in a position to explore how real and extensive these differences are. Is there one cognitive science, or many? One approach to answering this question is to consider whether the distinctions between the elements of the cognitive sciences are truly differences in kind.
The position of the current chapter is that there are strong relations amongst the three schools of thought in cognitive science; differences between these schools are more matters of degree than qualitative differences of kind. Let us set a context for this discussion by providing an argument similar in structure to the one framed by Adams and Aizawa (2008) against the notion of the extended mind.
One important critique of embodied cognitive science’s proposal of the extended mind is based on an analysis of the mark of the cognitive (Adams & Aizawa, 2008). The mark of the cognitive is a set of necessary and sufficient features that distinguish cognitive phenomena from other phenomena. Adams and Aizawa’s central argument against the extended mind is that it fails to provide the required features.
If one thinks that cognitive processing is simply any sort of dynamical system process, then—so understood—cognitive processing is again likely to be found spanning the brain, body and environment. But, so understood, cognitive processing will also be found in the swinging of a pendulum of a grandfather clock or the oscillations of the atoms of a hydrogen molecule. Being a dynamical system is pretty clearly insufficient for cognition or even a cognitive system. (Adams & Aizawa, 2008, p. 23)
Connectionist and embodied approaches can easily be characterized as explicit reactions against the classical viewpoint. That is, they view certain characteristics of classical cognitive science as being incorrect, and they propose theories in which these characteristics have been removed. For instance, consider Rodney Brooks’ reaction against classical AI and robotics:
During my earlier years as a postdoc at MIT, and as a junior faculty member at Stanford, I had developed a heuristic in carrying out research. I would look at how everyone else was tackling a certain problem and find the core central thing that they all agreed on so much that they never even talked about it. I would negate the central implicit belief and see where it led. This often turned out to be quite useful. (Brooks, 2002, p. 37)
This reactive approach suggests a context for the current chapter: that there should be a mark of the classical, a set of necessary and sufficient features that distinguish the theories of classical cognitive science from the theories of either connectionist or of embodied cognitive science. Given the material presented in earlier chapters, a candidate set of such features can easily be produced: central control, serial processing, internal representations, explicit rules, the disembodied mind, and so on. Alternative approaches to cognitive science can be characterized as taking a subset of these features and inverting them in accordance with Brooks’ heuristic.
In the sections that follow we examine candidate features that define the mark of the classical. It is shown that none of these features provide a necessary and sufficient distinction between classical and non-classical theories. For instance, central control is not a required property of a classical system, but was incorporated as an engineering convenience. Furthermore, central control is easily found in non-classical systems such as connectionist networks.
If there is no mark of the classical, then this indicates that there are not many cognitive sciences, but only one. Later chapters support this position by illustrating theories of cognitive science that incorporate elements of all three approaches.