Strong Equivalence in Non-monotonic Reasoning

From International Center for Computational Logic

Strong Equivalence in Non-monotonic Reasoning

Talk by Hannes Straß
Abstract: In classical logic, when two theories have the same models, they are

mutually replaceable in arbitrary contexts. This is not the case in various non-monotonic formalisms, e.g. logic programs under the stable model semantics, or abstract argumentation frameworks. There, two theories (programs/frameworks) can have the same models (stable models/extensions), but yet differ in their semantics when both are extended with the same third theory. To obtain mutual replaceability, a stronger notion is needed – this equivalence across all possible extensions is known as strong equivalence.

We address the question on how non-monotonicity is related to this divergence of standard and strong equivalence, show examples of how strong equivalence for concrete formalisms is analysed in the literature, and finally preview an abstract, algebraic approach to

characterising strong equivalence model-theoretically.