An erudite but meandering amble through the land of memory, from the author of So Much to Say: How to Help Your Child Learn to Talk (1982) and The Penguin Adoption Handbook (1984). Bolles sets out to startle the reader by asserting that memory is an act of imagination--that we create our memories rather than push a mental button that retrieves them from storage for scanning--although the psychologist Frederick Bartlett established as much in the 30's in classic studies of complex American Indian folk tales. Bolles posits a ""ladder of memory"" that evolves--in tune with biological development--from emotional memory (essentially a pain/pleasure axis) to factual memory (all the bloody details) to interpretive memory (giving meaning to memory). By that scale, argues Bolles, John Dean, for all the purported claims of perfect Watergate recall, turns out to have been stuck at an emotional level of memory--his memory was perfect in terms of praise he received from Nixon, and in terms of fears for his own safety, but, elsewhere, notoriously false (Bolles offers an impressive side-by-side comparison between Dean's testimony and White House tapes). The famous Russian ""S"" studied by psychologist Alexander Luria was stuck, on the other hand, in factual memory--all trees, no forest. Proust was the very epitome of interpretive memory. Bolles also embarks on excursions on Piaget and Freud, and on Karl Lashley, Karl Pribram, and others who have pursued the substrate of memory. A problem: Bolles' emphasis on the creative constructive nature of human memory appears to leave no base at all for neural coding. Thus, while he discusses how brain damage can affect particular categories of memory, he has little to say about what is being destroyed. Not fully convincing, but still a provocative--and memorable--read on a difficult subject.