So far we have looked at the nature of intelligence and discussed the cognitivistic paradigm, which still by far dominates scientific and everyday thinking about intelligence. In a number of places, however, we have alluded to potential problems with this paradigm. In this chapter, we inspect more closely what these problems really are and why they have arisen in the first place. As we argue, the cognitivistic paradigm’s neglect of the fact that intelligent agents, humans, animals, and robots are embodied agents thatlive in a real physical world leads to significant shortcomings in explaining intelligence.
Outlining the cognitivistic paradigm’s problems and understanding their origins helps us, on the one hand, to avoid making the same mistakes again; on the other hand, it provides us with inspiration about what needs to be done differently. The chapter is relatively short. Most of the issues it raises have been discussed at length in the literature, and an overview of those issues is sufficient here without repeating the details of the arguments. The goal is to outline the main problems that historically have led researchers to reconsider their approach to the study of intelligence.
We proceed as follows in the chapter: first, we work out the main distinctive characteristics of real and virtual worlds. We then present an overview of some of the well-known problems of traditional systems, followed by an inspection of some of the fundamental issues involved. We conclude with a number of suggestions as to what might be to be done in order to overcome these problems.