An Alternative Cognitive Architecture
Chapters 3 through 6 examined the bearing of the systematicity and productivity arguments on cognitive architectures that were committed to Representationalism and the Computational Theory of the Attitudes. Chapter 7 extended this examination to two Connectionist models, models that were committed to Representationalism and at least not obviously inconsistent with the Computational Theory of the Attitudes. Chapter 8 extended the argument to two versions of Functionally Combinatorial theories of cognitive representations. Now, in this chapter, we shall consider an alternative cognitive architecture that part ways with both Representationalism and CTA, namely, an architecture based on Robert Cummins’s very interesting work on naturalizing meaning in Representations, Targets, and Attitudes. While Cummins believes that cognition involves computational processes and representations, he does not accept Representationalism or the Computational Theory of the Attitudes found in Classicism. Insofar as the many kinds of systematicity and productivity are bona fide features of human cognition, however, it is relevant to ask how this alternative cognitive architecture, designed to meet other explanatory ends, fares when facing the family of systematicity and productivity arguments.
Unable to display preview. Download preview PDF.
- 1.Cf., e.g., Dretske, (1981, 1988), Fodor, (1990).Google Scholar
- 2.The original formulation of the disjunction problem is to be found in Fodor, (1986). Fodor’s (1990) version for informational semantics suffers from a number of problems discussed in Adams & Aizawa, (1994).Google Scholar
- 3.Cf., Cummins, (1996a), p. 17.Google Scholar