In lieu of an abstract, here is a brief excerpt of the content:

Current Literature Edited by Elizabeth Crumley I. Book Reviews Book Review Panel: George Agoston, Rudolf Arnheim, John E. Bowlt, Hans Brill, Donald Brook, Patricia Butler, Robert Dixon, Elmer Duncan, Vic Gray, Yusuf Grillo, John G. Hanhardt, Sharon Lebell, Alan Lee, Joy Turner Luke, John Mallinckrodt, Leo Narodny, Sean O’Driscoll, David Pariser, Yehuda Safran, Shao Dazhen, Allan Shields, David Topper, Stephen Wilson. THE COGNITIVE COMPUTER: ON FICIAL INTELLIGENCE by Roger C. Schank with Peter Childers. Addison-Wesley, Reading, MA, U.S.A., LANGUAGE, LEARNING, AND ARTI1984 .268~~. $12.95. ISBN: 0-201-06446-4. Reviewed by Stephen Wilson, Art Department, San Francisco State University, San Francisco, CA 94110, U.S.A. Roger Schank, one of the pioneers in Artificial Intelligence (AI), heads a fruitful research group at Yale. Over the years he has focused on many topics, but he is perhaps best known for his work on natural language understanding and knowledge structures. The Cognitive Computeris a refreshingreflection on this many-year research program by one of its prime movers. Leonardo readers especially will appreciate Schank’s comprehensive approach. He is not interested just in describing the technical details of the research. Rather, he wants readers to understand the whole research context and the implications of the research. The preface lists three questions that the book tries to address: What do we have to know about computers in order to live in a world full of them? What can we learn about what it means to be intelligent? How will intelligent computers affect the world we live in? Clarification of the sophistication of human understanding has been one of Schank’s main contributions. In the early days of A1 research, enthusiasts thought it would be an easy matter to write programs which could read and understand texts such as newspapers. Schank’s research demonstrated the highly elliptical nature of human communication. Even simple children’s stories rely on store- @ 1988 ISAST houses of readers’ ‘common sense’ knowledgeto fill in essential connections. Humans do not spell out every little detail when communicating. For example, Schank suggests that in order for A1programs to understand, the programmer must supply the expectations . He calls these sets of expectations ‘scripts’. For example, a restaurant script allows a program to fill in details about menus, waiters, checks, etc., even if they are not explicitly mentioned. Similarly, he tries to analyze basic human motives and goals in order to allow programs to understand text in areas that are difficult to describe in scripts. Schank’s programs can handle understanding tasks not well served by simple semantic analyzers. Leonardo readers will enjoy Schank’s view that the appreciation of the subtlety of human cognitive functioning is one of the prime reasons for being involved in this research. Furthermore, Schank’s research highlights important opportunities and challenges for the arts. The arts always have focused on the question of what it is to be human. A1 research is a prime contemporary mode of exploring that question. The analysis of human experienceand the attempt to embed that analysis in A1 programs call for artistic sensibilities in addition to technical skills. Inevitably artists will need to become involved. In one place Schank suggests that computer science and A1 research will ultimately disappear as specialties and will be absorbed into other disciplines. The Cognitive Computer is well written and easy to read. I strongly recommend it to readers. I do have a few reservations: Readers should remember that Schank’s approach is only one of many, and other A1 researchers believe other methods will be more fruitful. Also, the section of the book that explores the implications of A1 for the humanities is shallow when compared with other sections. Finally, Schank alternates between respect for the reader and condescension. He suggests that few people will need to program in the future because of the sophisticationof the A1programs. 1feel he plays down the importance of wide dissemination of programming skills. Wide dissemination is necessary as both a technique for guarding against the mythification of this technology and a way of encouraging the widest participation in this quest to understand human understanding. ELECTRONIC AND COMPUTER MUSIC by Peter Manning. Clarendon Press, Oxford, U.K...

pdf

Share