Even though well-known frameworks, such as Shannon’s information theory and first-order logic, exist that strive to define what information actually is, these theories are hardly adequate means of tackling the kind of information processing that is the basis of human-to-human communication. Addressing this issue by introducing a brand new logic, specifically based on information and inspired by situation theory, is the purpose of Devlin’s mind-provoking book.

A general overview of the whole approach is given in the first chapter. Infons, the basis information quanta, are the subject of the second chapter, together with situations, the models of infons, which are the agents’ views of the world; more formal definitions of situations and situation types are given in chapter 3. The relationship between human language and this new logic is studied in chapter 4, where the semantics of sentences is seen as constraints between situation types. The general mathematical properties of this “infon logic” are developed in chapter 5. Even though this book does not pretend to introduce a new theory of mind, it looks into the effect of this new logic on our views of what cognition and belief (chapter 6), perception and action (chapter 7), and language as action (chapters 8 and 9) are in some detail. The bibliography is small, but seems to include most of the basic papers and books. An index completes the book.

We have in this book an almost complete description of a new approach to information, with formal definitions and practical consequences for human understanding and language. Most of the important aspects are covered, and both their theoretical and their practical implications are mentioned. The pace is leisurely, and Devlin gives good explanations for the introduction of the more advanced concepts. Even though the book is described as mathematical, the prerequisites are light, a background in formal logic being enough to grasp the abstract definitions. As recognized by the author himself, few theorems exist as yet to flesh out this new framework, but this paucity is the price to pay for pioneering a new approach.

Mathematically oriented readers interested in artificial intelligence and human cognition should be able to get new ideas from this well-written book.