Attributive Description Formalisms

Research in knowledge representation has led to the development of so-called terminological logics, the purpose of which is to support the representation of the conceptual and terminological part of Artificial Intelligence applications. Independently, in computational linguistics, so-called feature logics have been developed which are aimed at representing the semantic and syntactic information natural language sentences convey. Since both of these logics rely mainly on attributes as the primary notational primitives for representing knowledge, they can be jointly characterized as attributive description formalisms. Although the intended applications for terminological logics and feature logics are not identical, and the computational services of systems based on the respective formalisms are quite different for this reason, the logical foundations turn out to be very similar – as we pointed out elsewhere. In this paper, we will show how attributive description formalisms relate to “the rest of the world.” Recently, a number of formal results in the area of attributive description formalisms have been obtained by exploiting other research fields, such as formal language theory, automata theory, and modal logics. This connection between these different fields of formal research will be highlighted in the sequel.