The book presents a collection of eight papers that provide detailed descriptions and empirical investigations of natural language systems that have made a substantial contribution to the idea of cooperative interfaces to information systems. The editors call an interface cooperative “if it does not just accept user requests passively or answer them literally, but actively attempts to understand the users’ intentions and to help them solve their application problems.” This notion covers the aspects of the formulation and acceptance of user requests, the presentation of the answer, the reaction to exceptional situations, the navigation through complex dialogs, and the adaptability of the interface to new applications or even languages.

The book is divided into three parts. Part 1 addresses general application-inde- pendent issues in natural language comprehension and (graphical) output presentation. Chapter 1 describes a large and complex augmented phrase-structure grammar for interpreting English language dialogs. In chapter 2, a system is presented that constructs graphic displays according to incomplete declarations of their content, structure, and appearance.

Part 2 introduces three natural language interfaces that were or still are considered for commercial product development. Chapter 3 discusses considerations for the development of natural-language interfaces to database management systems and concludes with a comprehensive checklist of necessary functional capabilities of such systems. In chapter 4, the need for a formal empirical evaluation of information systems interfaces is emphasized, the framework for such studies is designed, and its application to a domain-inde- pendent natural language query system is described. In chapter 5, an interactive customization program is introduced, allowing end users to define their own vocabulary and grammar rules to be used when querying a relational database. These users need not be linguists but should be familiar with the database to be accessed.

Part 3 describes three major projects in knowledge-based natural language processing. Chapter 6 concentrates on a semantics-based approach to the analysis of natural language queries over a relational database. A thorough treatment is given to problems of ambiguity, conjunction, and ellipsis. The contribution demonstrates that an approach ignoring most of the syntactic features of the query can yield nontrivial results. A particular result of language-oriented research in the field of artificial intelligence is described in chapter 7. It is a broadly based cooperative natural language dialog system, HAM-ANS, which can provide access to a wide variety of information systems--a database, an image processing system, and an expert system. Detailed examples of these applications are given. This chapter emphasizes the need for explicit user modeling and presents suggestions for the representation and application of user models. The final chapter 8 describes an expert interface to an information retrieval system. Starting from the natural language query by the user, the rule-based system selects and executes an appropriate strategy to evaluate the query.

The book gives a clear idea of the rapidly evolving field of cooperative interfaces to information systems and introduces the goals of the research, the problems, and the important solutions. Its best feature is that it presents the wealth of problems that still remain to be solved before truly cooperative interfaces are achieved--this area of research is still in its early phase. Some results, however, are rather outdated. Chapter 3 is based on experience with a system whose development ended in 1981, and the examples in chapter 7 quote the system version of 1983.

The book is aimed at AI researchers and students; only a general background in AI is required. References from all the contributions are collected at the end of the book, and a sufficiently detailed subject index is provided.