A common strategy for dealing with semantic interoperability is to create large, shared conceptual schemas (often called ontologies) that are then used as a common reference for agents using autonomous (and typically heterogeneous) conceptualizations. Shared conceptualizations are supposed to play the role that the reference to a common world of objects and facts plays in human communication. However, if the reference to a common world does not always guarantee the success of communication (i.e., communication free of ambiguities and misunderstandings), the problem gets even worse when we try to replace the common world with an abstract structure (e.g., a shared ontology), which - being a linguistic structure - is by definition compatible with many different interpretations.
The aim of the workshop is to investigate an alternative approach to semantic interoperability based on Meaning Negotiation (MN). The idea of MN is that any real-world approach to semantic interoperability between autonomous entities (namely entities that cannot assess semantic problems by "looking into each other's head", like humans or software agents) should involve (among other things) a social process of negotiating an agreement on the content (semantics) and the speaker's intention (pragmatics) of a communication. As a simple example, imagine that you ask a travel agent in Miami information about hotels in Venice. In the first place, the agent will try to establish which Venice you are talking about (e.g., Venice in Florida or Venice in Italy?), which is the semantics, and then the purpose of your trip (e.g., leisure or business), the dates (e.g., summer or winter), the price range (e.g., cheap or expensive), which is the intention. After this negotiation phase (in this case, based on questions and answers), the travel agent will be able to provide meaningful suggestions. Other forms of MN may involve: existence (do black holes exist?); appropriateness of a classification (is the attack to the Twin Towers to be classified as a "terrorist attack" or as a "war declaration"?); relationships between concepts (is "artificial intelligence" a sub topic of "computer science" or "psychology"?); ambiguity (under what conditions "bug" and "error" can be treated as synonyms?); and so on. The problem of meaning negotiation can be addressed from many different perspectives, using different conceptual and technological tools, and with different motivations in mind. We invite contributions from a variety of areas, including knowledge representation, multi-agent systems, databases, natural language processing, machine learning, game theory, epistemology, philosophy of language, cognitive science, psychology, sociology, organization and management sciences.
March 15, 2002: Deadline for WS submissions April 26, 2002: Notification of acceptance to authors May 12, 2002: Deadline for camera-ready workshop notes and other information July 28, 2002: workshop on Meaning Negotiation
University of TrentoFrank van Harmelen
Department of Information and Communication Technologies / Cognitive Science Laboratory
Via Sommarive, Povo - I-38050 Trento (Italy)
E-mail: bouquet@cs.unitn.it
Phone: +39-0461-882135 Fax: +39-0461-882124
Vrije UniversiteitFausto Giunchiglia
Faculty of Sciences
Division of Mathematics and Computer Science
E-mail: Frank.van.Harmelen@cs.vu.nl
University of TrentoDeborah McGuinness
Department of Information and Communication Technologies
E-mail: fausto@cs.unitn.it
Stanford UniversityJohn Mylopoulos
Knowledge Systems Laboratory
E-mail: dlm@ksl.stanford.edu
Department of Computer ScienceMassimo Warglien
University of Toronto (Canada)
E-mail: jm@cs.toronto.edu
University of Venice
Dipartimento di Economia e Direzione Aziendale and
University of Trento
Cognitive Science Laboratory
E-mail: warglien@unive.it