|
Abstract : |
We introduce a model and a system architecture for second-order reasoning about statements expressed in a (first-order) terminological representation language. This meta-reasoning approach is based on the reification of first-order propositions and the mediation between second- and first-order expressions via translation rules operating on multiple contexts. An application to a concept learning task in a text understanding environment shows how different degrees of credibility can be assigned to alternative concept hypotheses on the basis of such a schema for terminological meta-reasoning. Appeared in:, |
