An Interactivist Perspective of Autonomous Robotic Sign Users

In this paper we present an interactivist perspective on design principles for autonomous sign users such as autonomous robots. Past discussions have centered around symbol grounding or anchoring where concepts are considered prototypes in the perceptual space of a robot or animal. Here we advocate a more complex view starting from the purpose of using signs or pursuing sign acts. We argue that concepts actually serve a number of purposes and revisit techniques for representing and memorizing concepts. Our own proposal focuses on the intentional and anticipatory aspects of sign acts and on anticipatory representations. This research is aimed at designing architectures for growing up autonomous sign users.