Formal Aspects of Context

This volume collects the proceedings of the International and Interdisciplinary Conference on Modelling and Using Context, held in Rio de Janeiro in January 1997, which proved to be the first in a series. (The second was held in Trento in 1999, the third in Dundee in 2001.) The theory of context is important and problematic--problematic because the intuitions are confused, because of the disparate disciplines that are involved, and because the chronic problem in cognitive science of how to arrive at a productive relation between formalizations and applications applies with particular force to this area. This uneven but interesting volume illustrates all of these points. "Formal and computational models of context for natural language generation" by Kees van Deemter and Jan Odijk illustrates the point concerning theory and practice. Concentrating on the generation of appropriate pitch accents in monologues, it is essentially a system description, emphasizing constraints on desirable output and functionality of system components. It is impossible to tell from the description how principled and portable the system is, or what the relation is to the relevant theories. "Requirements for dialogue context management" by Harry Bunt is mainly concerned with producing an eclectic assortment of categories for thinking about contexts. "Contextual constraints on thematization in written discourse: An empirical study" by Julia Lavid describes a corpus study of "chaining strategies" in various genres. The author hopes that the corpus will be useful in generation applications. "Context and implicitness: Consequences for traditional and computer-assisted text analysis" by Mark Galliker and Daniel Weimer represents the inverse relation between importance of the application area and strength of the methodology. This application of social psychology to texts involving the expression of prejudice is highly subjective and appears to lack a methodology for analyzing and evaluating the relevant texts. "A context-based mechanization of multi-agent reasoning" by Alessandro Cimatti and Luciano Serafini discusses a formalization and mechanization of the reasoning involved in the "three wise men puzzle" in a belief-contexts framework (Giunchiglia 1993). The paper would be hard to understand without some familiarity with this material. The wise-men puzzle is not a toy problem, but it would be good to know how to use the framework to deal with more realistic examples. It would be good to know what the advantages are of this approach compared with multimodal theorem-