Archive of events from 2019
An archive of events from the year
-
Research Lindström Lecture: Johan van Benthem (University of Amsterdam, Stanford University, Tsinghua University Beijing)
Dynamic Logics of Model Change
Information update and real-world action suggest a universe of changing models, where finding the valid dynamic laws involves logical languages with modalities for model change. I will discuss two families of such dynamic modal logics, stating some typical results, as well as open problems.
Dynamic-epistemic logics add dynamic superstructure to existing static logics, and tend to not increase complexity of satisfiability and model checking. Logics of graph change, arising e.g. in the study of games or of languages that change their own models under evaluation, tend to jump to higher complexities, becoming undecidable or worse. I will report some recent results that zoom in on the border line between the two kinds of system, and end with some issues about redesign and restoring decidability for more complex logics of action and information dynamics.
-
Public Lindström Lecture: Johan van Benthem (University of Amsterdam, Stanford University, Tsinghua University Beijing)
Logic and Agency: The Promises and The Challenges
One face of logic is turned toward truth and eternal consequence, but another face is dynamic, looking toward activities of reasoning and information handling by agents. In this lecture, I will develop the dynamic perspective, showing how key aspects of rational agency fit in the agenda of logic, such as handling and integrating information from various sources, revising erroneous beliefs, and balancing information with preferences and goals. In this lecture, the vehicle for achieving this will be dynamic-epistemic logics for various sorts of update. These mesh eventually with richer logics of games and strategic interaction. After all, much of reasoning is a social multi-agent process: ‘intelligence seldom comes alone’. I end this part by noting how the two faces of logic share the same methodology, and are in fact complementary.
However, ‘more logic’ is just one way to go in studying agency. I briefly discuss current challenges from the ‘less logic’ camp, where successful behavior is explained by dynamical systems with very simple agents, or from learning systems that may not have a logical formulation at all. I hope to show that logic retains a valuable role even in that stormy setting.