This review of the book, "Conceptual Structures: Information Processing in Mind and Machine," by John F. Sowa, argues that anyone who plans to get involved with issues of knowledge representation should have at least a passing acquaintance with Sowa's conceptual graphs for a database interface. (Used to model the underlying semantics of a database, these graphs describe data from the user's point of view and access data according to the system's point of view. They also represent functional dependencies in a relational database and support inferences and computations that are not explicit in the initial query.) Sowa's ideas on knowledge representation, cognitive science, and expert systems as presented in the book are discussed, and it is argued that: (1) Sowa's ideas were limited by his concern for Codd's relational model of databases; (2) developments in cognitive science during the eight years that elapsed between Sowa's article in the IBM Journal of Research and Development (July 1986) and the publication of the book have raised issues that he does not confront adequately; and (3) perhaps the most valuable legacy of the book will be its impressive set of references to early scholarship. (RP)