Getting a computer to understand our everyday language may be a long way off; the computer has to know how to think as well as follow the rules of a particular language. This book makes an important contribution to the study of pragmatics and discourse by presenting an explicit and precise computational approach to the complex problem of the structure of discourse. As Professor Stephen Pinker of MIT notes, "Reichman has laudably chosen to inject some precision into a field much in need of it. The book first looks at extended person-machine communication, beginning with person-person communication and focusing in particular on the conversational flow itself - what makes for coherent discourse? It then describes a computer model that describes this phenomenon as an augmented transition network (ATN). Rachel Reichman is Professor, Department of Electrical Engineering and Computer Science, University of California, San Diego.