Book Review OF Information and Mlslnformat|on: An Investigation of the Notions of Information, Mtsinforma~lon, Informing, and Misinforming. Fox, Christopher John. Westport Conn.: Greenwood Press; 1983:223 pp. The concept of information is among the most important of our time, yet its meaning is unclear. What is this mysterious thing called information which we use to label systems, mental processes, and even the current age of civilization? This is the thorny and longstanding question that Fox examines. The answer the he reaches has interesting implications for the information sciences, both on a theoretical and a practical level. Fox's basic premise is that an adequate definition of information must encompass the various common uses of the term. He reviews previous defmiLions of information, which have ranged from Fairthorne's view that information is a bogus concept akin to phlogiston, to precise definitions in a narrow domain such as Shannon's. Using the technique of analytic philosophy, Fox argues that none of these definitions captures the meaning of information as we use it in ordinary speech. The author argues persuasively that a definition which does capture our ordinary use of "information" is that information is propositions, i.e. descriptive statements about the world. This result strongly indicates that the voluminous research that has been done in philosophy, mathematics, and other fields, on the nature and expression of propositions has direct relevance to the foundations of computer and information science. Fox's results also have practical consequences for the ways information should be represented and manipulated in systems. If information is propositions, then an adequate formalism for expressing infbrmation in a way useful for computatior~ will most likely be one allowing the expression and manipulation of propositions. This is not the case with current information retrieval systems which are term based. For example, the logical statements and production rules used in used in expert systems should be more useful for representing the information in a document than a listing of the terms in a document. As a n illustration of this, consider a document consisting of the sentence, "John loves Mary". In an information retrieval system, whether boolean or probabalistic, the information will be represented by the words, "John", "loves", and "Mary" losing the semantic links between the words. The Same document could be represented in a logical language, such as PROLOG, as "loves(John,Mary)" with no such semantic loss. This book is an important addition to the literature of the information sciences. The author does a good job of explaining the methodology he uses, and thus this book can and should be read by anyone interested in the foundations of the information sciences, information systems, or artificial intelligence. William B. F~akes AT&T Bell Laboratories Holmdel, NJ. 07733

/lp/association-for-computing-machinery/book-review-information-and-misinformation-an-investigation-of-the-I0cpQaScfO