Music has long occupied a position on the periphery of artificial intelligence research.' Nevertheless, Artificial Intelligence has yet to run an article whose central topic concerns the relationship between music and artificial intelligence. Is this because music is "just an application ?" Certainly, the applications-oriented literature has yielded some interesting reports, particularly in automating music composition. (Most of the July 1991 issue of Computer was devoted to such reports.) Or is it because the relationship between music and mind is still beyond the scientific paradigms of artificial intelligence and there are major conceptual barriers which must be overcome before music "finds its place" in the world of artificial intelligence research?
The books we are reviewing in this collection will not answer these thorny questions, but they may help us better appreciate why they are so thorny. Three are reports of major research efforts. The books by David Cope and Robert Rowe both take on significant questions of what it means to have machines make and listen to music. Because much of this work may rest more on an understanding of the nature of music itself than on an appreciation of the supporting technology, both of these books are being reviewed by academically credentialled music scholars who are also fluent in the capabilities of computer technology. Jonathan Berger, who is at The Center for Studies in Music Technology at Yale University, is an accomplished music theorist who has recently applied his theoretical background to the publication of a CD-ROM. Eliot Handelman recently received his Ph.D. from the Music Department at Princeton University for a thesis with the provocative title, "Music as Secondary Consciousness", whose text was supplemented by an original composition. The third report is a collection of papers by Peter Desain and Henkjan Honing which are more