Logic in Action Themes
Surely it is preposterous to think that present day logic is able to change, say, the habits of tax-payers in the foreseeable future. However, it is equally inconceivable that we were to neglect the logical inclinations of the players in the emerging information society. Logic, understood in a broad sense, eventually ought to provide us with the tools and concepts to approach and analyze this realm of information interchange.
The Spinoza project Logic in Action explicitly aims at enhancing and furthering the scope and role of logic in an upcoming information science. In the project logic constitutes the common approach to information, information flow and information exchange, and by locating information, interpretation and reasoning in the context of rational, decision making agents, a focus of common interest is created for various disciplines.
Of course, for such an enterprise to be feasible, deliberate choices have to be made, and themes have to be selected to focus upon. We mention three themes which illustrate the interactions between logic, linguistics, mathematics, and computer science characteristic for the ILLC research environment. These themes reflect and enhance the long-standing tradition of information-oriented logic in Amsterdam with such highlights as intuitionistic and modal logic and dynamic semantics.
Dynamic models of information and communication
A central aim of the Spinoza project is the design and study of formal models of the patterns of information and information flow. Even in the simplest forms of communication diverse notions such as knowledge, physical action and information change are intertwined, and a multi-agent perspective is called for. Many interesting research problems arise from finding out how such features interact, in rich epistemic action logics that combine individual information states with collective ones.
More on the empirical side, we are after a formal characterization of the linguistic `presuppositions' for successful information processing. A unifying perspective is looked for in the area of game theory, whose (modal) logical properties are investigated.
Correspondences between computation and information processing
Modern information technology has blurred the borderline between natural and artificial languages. A similar blurring of boundaries reveals itself at the level of foundational research.
The `dynamic turn' in the semantics of natural language was partly inspired by the theoretical study of the semantics of programming languages: a command like `increase the value of register X by 1' relates an `old' memory state to a `new' one. Similarly, mention of a new topic of conversation in natural language relates an `old' context of discourse items to a `new' one. Interestingly, one of the tools designed for the analysis of this context change phenomenon in natural language, dynamic predicate logic, gives rise in turn to a computational interpretation: dynamic predicate logic can be turned into a programming language. By the looks of it, programming with natural language is just around the corner.
Modular reasoning with light-weight representations
Informatics has become a common name for the new science of information, together with its associated applications and human dimensions. One of the most pressing issues facing informatics is content finding, accessing, structuring, and presenting the information we need. Content can be represented in many ways, ranging from simple keywords to light-weight semantic analyses to deep ones. The key challenges is to understand the balance between the richness of representations and the computational efficiciency of constructing representations and reasoning with them.
The strategy we have adopted is a mixture of foundational and experimental work with an emphasis on developing small, dedicated logical techniques and lean natural language processing tools. Novel in its avoidance of baroque supersystems, this project analyzes semantic complexity, makes it explicit, and harnesses it.