In lieu of an abstract, here is a brief excerpt of the content:

Reviewed by:
  • Exploring time, tense and aspect in natural language database interfaces by Ion Androutsopoulos
  • Maria Filiouchkina
Exploring time, tense and aspect in natural language database interfaces. By Ion Androutsopoulos. (Natural language processing 6.) Amsterdam: John Benjamins, 2002. Pp. ix, 306. ISBN 1588112691. $116 (Hb).

Natural language interfaces to databases (NLITDBs) allow their users to retrieve information from an underlying computer database by submitting natural language queries. This book explores one important aspect of natural language database interfaces, namely the notion of time. It studies how [End Page 334] computer systems may interpret temporal information in natural language queries and generate automatically appropriate responses in natural language.

A revised version of the author’s University of Edinburgh Ph.D. dissertation, the book consists of seven chapters, an index of terms and concepts, and an appendix. In the introduction, Androutsopoulos explains what the book is about and presents a short summary of the volume. He also provides information on natural language interfaces, tense and aspect theories, temporal logics, and temporal databases. Ch. 2 examines how temporal information is conveyed in English, focusing on the following linguistic phenomena: verb forms, temporal verbs, temporal nouns, temporal adjectives, temporal adverbials, subordinate clauses, and anaphora. Following Zeno Vendler’s taxonomy of aspectual classes, A distinguishes between states, activities, culminating activities, and points. Several criteria are proposed for classifying base verb forms. These include the simple present criterion, the point criterion, and the imperfective paradox criterion.

Ch. 3 defines TOP (a temporal meaning representation language), which can be used to represent the formal semantics of the English questions that are submitted to the NLITDB. Several temporal operators are defined in terms of their denotations with respect to speech time (st), event time (et), and localization time (lt). Ch. 4 shows how HPSG grammars (following Carl Pollard and Ivan Sag) can be used to translate English questions involving time to appropriate TOP formulas. In Ch. 5, A introduces TSQL2 (a database language), together with its model of time, and presents a set of provably correct translation rules. These rules can transform every TOP formula into an appropriate TSQL2 query that preserves the semantics of the TOP formula.

To show that the theoretical framework of the book is workable, a prototype NLITDB is suggested (Ch. 6), using Prolog and ALE. This prototype is freely available. In addition, A provides information about other mechanisms that would have to be added if the system were to be used in real-life applications. The final chapter discusses related work on NLITDBs and suggests directions for further research in the field.

In sum, this book develops a unified theoretical framework that can be used to build natural language interfaces to databases. It will interest researchers whose work falls across the areas of natural language interaction, computational linguistics, formal semantics, tense and aspect, HPSG, and temporal logics. Although the research in A’s book is limited to English, it could be used to develop a corresponding theoretical framework for other natural languages as well. As the volume contains numerous semantic and computer-processed formulas, it assumes that the reader has a solid background in symbolic approaches to natural language processing and computational linguistics. Otherwise, the book is written in a clear and readable style and contains examples from English as well as explanations of what the different formulas stand for.

Maria Filiouchkina
University of Tromsø
...

pdf

Share