Rechercher des projets européens

Category theory for meaning assembly and the semantics-pragmatics interface (MEANCATS)
Date du début: 1 sept. 2013, Date de fin: 31 août 2015 PROJET  TERMINÉ 

The project investigates the interaction between semantic and pragmatic factors in determining the meaning of an utterance, particularly how the factors can be unified in a single account of meaning. Addressing this question is important for two reasons. First, semantic and pragmatic approaches to natural language meaning have diverged into two separate fields, with independent sets of assumptions. While this has been useful in uncovering a number of phenomena, many scholars have pointed out that the division is artificial and prevents linguistic theory from properly exploring a large number of central phenomena in natural language at the intersection of the two dimensions of meaning. Second, a better integration of semantics and pragmatics will have significant practical consequences both for linguistics and for applications in industry. Currently, any attempt at combining semantic and pragmatic approaches is a daunting task, as the interactions between the two aspects are poorly understood. This also has the effect of limiting computational systems for deep understanding of linguistic input. Given the increasing ubiquity of voice interface systems, improving our understanding of these interactions is a priority.The project will have three goals. The first is to investigate the viability of the mathematical framework of category theory for richer formal models of meaning. I have already obtained promising initial results with this method for some phenomena. I intend to study its applicability to a wider range of phenomena. The second goal will be the implementation of a prototype computational system based on the results of the first goal. The formal system will be implemented as an extension of the Xerox Linguistic Environment and made available to the research and industrial community. Finally, I will test some predictions of the proposal with behavioural experiments. This will provide solid experimental data that will allow wider testing of the theory.