Browse Books

Reviewer: James Loftin Snell

The intended audience for this book about human creativity and the use of computational concepts in understanding it—what Boden calls “computational psychology”—is the educated layperson. In her opening chapter, Boden remarks that older explanations of creativity, such as Koestler's <__?__Pub Fmt italic>The act of creation<__?__Pub Fmt /italic> [1], have been quite vague, and she promises to “take up the question from where Koestler left it” and provide a more detailed and scientific explanation. The book succeeds modestly: for the lay reader, it provides some acquaintance with elementary computing concepts, particularly in artificial intelligence (AI), that are relevant to creativity, although for readers technically knowledgeable in these areas, it offers little new insight. The book consists of 11 chapters. The first four describe some classic examples of human creativity (such as Kekule's discovery of the benzene ring); define a few terms and distinctions (for example, between personal and historical creativity); pose various questions about the relationship between computers and creativity; and attempt to “dissolve the mystery” of creativity by citing some speculations on the psychological process (including Hadamard's preparation-incubation-illumination-verification hypothesis). The second four chapters provide nontechnical descriptions of various computational concepts (such as generative grammar, heuristics, search space, semantic network, neural network, and production system) and some AI programs that use them, and show how these relate to <__?__Pub Caret>aspects of human creativity. Chapters 9 and 10 address the issues of randomness (how can a deterministic machine exhibit unpredictability__?__ and the like) and genius (for instance, Was Mozart different qualitatively or merely in “quantity” of creativity__?__) The final chapter provides refutations of several arguments denying the possibility of creativity or “true intelligence” by computers (including a brief treatment of Searle's “Chinese Room” argument) and offers reassurances that computer creativity would not (or should not) threaten the stature of humans. Boden's writing style is conversational, friendly, and smooth, if somewhat verbose in places. She demonstrates an unusual combination of diversity and coherence of knowledge, presenting apt examples from music, graphic arts, mathematics, linguistics, chemistry, physics, astronomy, literature, philosophy, and psychology, and showing persuasively that the phenomena of creativity she describes occur similarly across all these fields. Her explanations of computing concepts, while technically superficial, are accurate and clear and should be reassuringly accessible to the layperson. She provides adequate references and bibliography for readers seeking greater depth, as well as a good index. The book's main purpose, to explain in computational terms how human creativity works, is fulfilled only to the extent of providing a general outline. Yet this partial success is a useful achievement, not least in providing coherent, well-informed arguments to oppose the ignorant writing of authors like Hubert Dreyfus. Moreover, Boden's book does fully succeed in a secondary purpose, namely to dispel the romantic and mystical assumptions about human creativity presumably still held by most educated laypeople (and perhaps even some computing professionals). On these grounds alone, the book is worthwhile reading—it contains just the kind of material that the computing community would like to see broadcast widely (say, as a television documentary) to promote improved understanding of computers by the general population.