The human conceptual system

ABSTRACT

The human conceptual system contains knowledge about the world that supports categorical inference, conceptual combination, and basic cognitive tasks (e.g., perception, memory, language, thought). Standard views of semantic memory typically portray the conceptual system as modular, amodal, static, abstractive, and taxonomic. Recent connectionist accounts are dynamic and contextual but often remain modular, amodal, and taxonomic (although not necessarily so). An alternative view is presented, along with supporting empirical evidence. On this view, the conceptual system shares representational mechanisms with sensory-motor systems, thereby making it non-modular and modal. Within this framework, simulators represent a category dynamically by constructing diverse simulations of category instances across occasions (i.e., static concepts do not represent categories). To represent a category on a given occasion, a simulator reenacts a small subset of modal information for the category. In addition, these simulations include knowledge about settings, actions, and mental perspectives, thereby contextualizing the category's representations. Finally, systems of categories are organized to support situated action. Ad hoc categories arise at the interface between action sequences and the world, with taxonomic categories being subsidiary to this interface. Behavioral and neural evidence is presented to support this view of the conceptual system.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.0 General
          Subjects: Cognitive simulation

General Terms:
Human Factors, Theory