Auteur: Marti Hooijmans
Vakgroep: Theoretische Filosofie
Robotics and representation
My thesis tries to find an answer to the question why many researchers working in situated robotics are holding on to representational terminology when describing information processing mechanisms. For various reasons the notion of representation is extremely difficult to pin down in any precise way and often criticized within this field.
My answer is that on the one hand it is easiest to use the readily available terminology. Workers in situated robotics still have the habit of approaching the concepts of representation and meaning in a common-sense way, which originates from the fact situated robotics has its roots in cognitive science and artificial intelligence. Cognitive science in turn stands in the broader background of ideas I call the Fregean tradition: a body of ideas that is common to analytical philosophy, logic, computer science, linguistics and cognitive science.
On the other hand, there is the unavailability of an alternative terminology, despite various attempts in this direction. These alternatives are either not very coherent, or are unsuccesful in their attempt to replace representation-like concepts with usable descriptions of information processing mechanisms.
Pointing to their incompleteness is the most common counterargument employed by proponents of representational terminology; there are representation-hungry problems that need to be coped with. Recent situated robotics tries to deal with this counterargument. Some alternatives try to develop new ways of looking at information processing mechanisms; other alternatives try to reformulate the classical representational terminology in a way to make it fit in with the situated, embodied approach. Only very few develop new concepts and support those with an empirical theory.
|Laatst gewijzigd:||01 november 2013 14:56|