Understanding Natural Language

This book describes a computer system for understanding English: a program designed to achieve three goals: to discover what intelligence is and how it can be formalized; to clarify what language is and how it is put together; to develop a practical language understanding system. The experimental system—a simulated robot with a hand and an eye and the ability to manipulate toy blocks upon a table—is capable of entering into a dialogue with a person, responding to English sentences with actions and English replies, accepting and remembering information, and asking for clarification when its heuristic programs cannot understand a sentence. Thus, it deals in an integrated way with all the aspects of language understanding—syntax, semantics, inference, and, perhaps most important, knowledge of the subject matter.
The book first describes the system in general and compares it with other work on semantics, inference, and syntactic parsing. The author includes an introduction to LISP, the computer language in which the system is written, for readers who are not familiar with it, and then describes the system in detail:
• The English grammar (based largely on Halliday's theory of systemic grammar);
• PROGRAMMAR, the language created for expressing the details of grammar within the system, used primarily for parsing sentences;
• PLANNER, a language for representing problem solving procedures, and used in the system for answering questions, creating plans, and making inferences;
• The detailed model of the world of toy blocks used as the system's problem domain, and the way this world is represented in the formalism;
• The semantic analysis of English which was developed for the system, including a basic analysis of the way meaning is conveyed by each syntactic unit, and a discussion of the different types of context and their importance.
The book promises to become a classic. It is necessary reading for everyone interested in artificial intelligence, semantics, linguistic…