The specific objectives of DYANA-2 are four-fold: (i) To determine those aspects of the structure of information states which are relevant to natural language interpretation; (ii) To examine specific semantic and pragmatic phenomena which currently prevent the effective use of natural language in human-computer interaction; (iii) To develop sufficient understanding of the logical and computational properties of sign-based grammar architectures to allow application across all components of the grammar; (iv) To demonstrate the utility of sign-based grammar architecture by providing a principled account of cross-linguistic variation.
The research is concerned with the foundations of natural language interpretation and its consequences for knowledge representation, focusing on the themes of partial information and dynamic interpretation, with attention to the goal of developing a formal theory of natural language interpretation which admits computational implementation.
Results have been produced in the following areas:
grammar architecture (the solution to spurious ambiguity in the Lambek calculus, substructural categorial logic, second order Lambek calculus, categorical polymorphism, constraint based linguistic description);
interpretation in context (integrating dynamic and update semantics, simultaneous abstraction in situation theory, generalized quantifiers and plural anaphora, scope properties of indefinites, presupposition);
implementation and integration (implementation of the constraint based unification formalism, implementation of a DRT fragment using situation theory).
APPROACH AND METHODS
Research is carried out on an interdisciplinary, collaborative basis drawing on concepts and tools from linguistics, logic, artificial intelligence and computer science. The work is divided into three interdependent themes:
- Grammar Architecture, in which the formalisations of the sign-based approach are developed, thus enabling study of the abstract specification and parametric variation of linguistic structure within a computational setting. Here we aim to further the work in DYANA-1 on sign-based grammars, regarding both categorial type systems and unification based theories.
- Interpretation in Context, in which the further development of the outlook on meaning developed in DYANA-1 is addressed, and its application to phenomena such as quantification, anaphora, ellipsis, tense & aspect, questions, presupposition and modal subordination.
- Implementation and Integration, in which the aim is to show the feasibility of using the results from these theoretical investigations in the design of real-life computational applications. This will involve work in the maturing paradigm of computational syntax and the relatively underdeveloped field of computational semantics. It intends to give direction to work on integrating the two.
The research themes that constitute the DYANA project aim to meet some of the challenges that lie ahead on the way towards computer systems that use natural language. Although the major aim of the project is to carry out fundamental research in natural language processing, work is also directed towards the development of prototypes and proof-of-concept implementations. A DYANA Newsletter is published which contains up to date information about recent publications, new deliverables, and DYANA-related workshops.
Topic(s)Data not available
Call for proposalData not available
Funding SchemeData not available
3512 JK Utrecht