In this paper we present the main features of SHARDS - a Semantically-based HPSG Approach to the Resolution of Dialogue fragments. This implemented system interprets short questions (sluices) and short answers.

We present an approach to VP ellipsis that allows the direct derivation of source and target sentences (the former need not be unique) during semantic construction. Specific syntactic constituent structures are associated with ellipsis potential, which can then be discharged by pro-verbs like did (too). The determination of source and target sentence, which is done with semantic features in an HPSG framework, is coupled with a comprehensive analysis of ellipsis, which also handles its interaction with scope and anaphora.

The Constraint Language for Lambda Structures (CLLS) is an expressive tree description language. It provides a uniform framework for underspecified semantics, covering scope, ellipsis, and anaphora. Efficient algorithms exist for the sublanguage that models scope. But so far no terminating algorithm exists for sublanguages that model ellipsis. We introduce well-nested parallelism constraints and show that they solve this problem.

We propose a procedure for the syntactic reconstruction of ellipsis that covers (i) antecedent contained ellipsis, (ii) intersentential VP ellipsis, (iii) bare constituent ellipsis, and (iv) gapping. The procedure is head-driven, and we implement it in a typed feature structure grammar which is a version of HPSG. It differs from most of the current approaches to ellipsis in that it does not require an abstract level of logical form as the input to the reconstruction procedure, and it reconstructs all elided elements in situ, rather than after extraction from an antecedent VP or clause. The procedure generates an AVM which specifies the reconstructed feature structure of the ellipsis site. This representation captures the primary syntactic and semantic properties of the elided elements. 1. INTRODUCTION Ellipsis structures pose an important problem for computational systems designed to model sentence interpretation and discourse. They contain information which is not overtly expressed, ...

Introduction A major task for a dialogue system is the resolution of fragments. Basic examples include bare NP answers (1), where the bare NP John is resolved as the assertion John saw Mary, and sluicing (2), where the WhP who is interpreted as the question Which student saw John. (1) (2) (3) A: Who saw Mary? B: John A: A student saw John. B. Who? Either the antecedent or the fragment (or both) may be embedded (3). 1. A: Bill wonders who saw Mary. B: John. (John saw Mary) 2. A: Bill thinks a student saw John. B: Who? (Which student does Bill think saw John?) 3. A: Who saw Mary? B: John thinks Bill. (John thinks Bill saw Mary) 4. A: A student saw John. B: Bill wonders who. (Bill wonders which student saw John) Constructing a computational framework for interpreting such fragments in a principled and systematic way is important for the development of viable grammar-driven text understanding and dialogue management systems. 2 Theoretical Background 2.1 Ellipsis Resolution:

Abstract. This paper lays the foundations for a processing model of relative acceptability levels in verb phrase ellipsis (VPE). In the proposed model, mismatching VPE examples are grammatical but less acceptable because they violate heuristic parsing strategies. This analysis is presented in a Minimalist Grammar formalism that is compatible with standard parsing techniques. The overall proposal integrates computational assumptions about parsing with a psycholinguistic linking hypothesis. These parts work together with the syntactic analysis to derive novel predictions that are confirmed in a controlled experiment. 1