On the Concept of Concept in the Context of Autonomous Agents

This paper deals with some fundamental questions regarding the concept of concept in the context of autonomous agents. The most basic of these is defining what it actually means for someone to have a concept. Rather than trying to state a number of conditions that should be satisfied in order to have the concept, it is concluded that having a concept is a matter of degree, which can be defined in terms of the functions the concept can serve. The more functions it can serve and the better it can serve these functions, the higher is the degree to which one has the concept. Moreover, the distinction between entity and dispositional theories of concepts is discussed, and it is concluded that they are complementary in that both perspectives are necessary to get a full picture of the concept of concepts. A conceptualistic entity theory and a dispositional theory based on which functions the concept should be able to serve are then put forward and discussed in a representational framework that supports these functions. Furthermore, we discuss the meaning of concepts, i.e., the problem of interpreting the symbols used to designate concepts, and give some arguments of why an autonomous agent should have the ability to interpret (some of) its own descriptions. We examine the work carried out within the field of logical semantics, and conclude that since traditional truth conditional semantics requires a human who grounds the meaning of elementary symbols, i.e., one who assigns objects and sets of objects to constants and predicates, this approach is not appropriate. Instead, a subjective intensionalistic approach based on the grounding of symbols is suggested, which is more in line with the verificationist and procedural approaches to semantics. Finally, we show that theories of meaning are closely linked with views on universals.