This book is intended to introduce the reader to the developing technology of natural language processing. It is organized primarily around case studies of implemented natural language processors. For clarity, the various aspects of language analysis, syntax, semantics, discourse, and knowledge representation are considered separately, but they are, in fact, interdependent. The case study approach illustrates the mutual dependence of the various aspects through examples. In addition, case studies give the reader indirect exposure to the areas of application to which natural language systems have been applied. To some degree, case studies can also give an indication of the extent of coverage that the systems have over their domains of discourse. Finally, the case studies illustrate some of the open questions of natural language processing.
In this book, a number of concepts will be defined. This specialized vocabulary facilitates later reference. It should also familiarize readers with the nomenclature of the field so that they can follow later developments in the technical literature if they wish to. Although new vocabulary is defined, no prior familiarity with natural language processing or other branches of artificial intelligence is presupposed. The book is written only with the expectation that the reader is familiar with concepts of general computer science. As for linguistics, the level of sophistication required of readers is about the level one attains in a high school grammar class.
This is a relatively short book. It is intended as an introduction to the field, not as a comprehensive survey covering every facet. It could be read with one of two goals: to familiarize the reader with the directions and status of ongoing research, or to serve as a starting point leading to deeper investigation. This introductory book concentrates on the issues central to natural language processing technology. Many related topics are not covered, topics the interested reader may wish to consider. There is a great deal of natural language research currently being conducted in linguistics and psychology. In artificial intelligence, there is related work in language generation, speech understanding, problem solving and knowledge representation. Knowledge representation is discussed to some extent, but thorough coverage would require a book in itself. The emerging discipline of cognitive science is meant to study cognition through the methodologies of linguistics, psychology, artificial intelligence, and other disciplines. In computer science, related work is going on in database technology, interactive system design, and human factors design; these are not covered here.
The material presented here is based primarily on reports of research in the open literature. To a lesser extent, some of the material is based on my personal experience of research in natural language processing. Available sources are not always as complete or detailed as one might wish, but I have made an attempt to discuss each system and its relation to others as fairly and accurately as possible. There are bound to be some errors, and for these, I apologize in advance.