A rambling, commonsensical treatise with a host of mini-insights but no big picture. Ekman is a professor of psychology (U. Cal., San Francisco) who has done research on nonverbal communication for the last 20 years, and now reviews some of his findings for a popular audience. He discusses the ""Brokaw hazard"": Tom Brokaw once said that ""convoluted answers or sophisticated evasions"" were the best clue that an interviewee was lying; but good liars don't make that mistake and some honest people talk that way naturally. He takes up the ""Othello error"": charging someone with a crime may stir the emotions--as measurable on a polygraph, say--of an innocent person. He analyzes the ways deceit may be revealed by ""emblems"" (unequivocal gestures such as giving someone the finger), ""illustrators"" (body movements that illustrate or accentuate thought), and ""manipulators"" (one part of the body doing something with another part). Emblematic slips, if the liar makes one, are good evidence, but illustrators and manipulators are less reliable. The forehead is an ideal place to look for telltale muscle twitches because few people (only 15 percent) can deliberately fake them. False smiles don't reach the muscles around the eyes, etc. By the end of the book Ekman has assembled a long ""lying checklist"" (38 questions), part of which he applies, with dubious results, to two notable instances of diplomatic lying: Hitler to Chamberlain at Berchtesgaden, and Kennedy and Gromyko to each other during the Cuban missile crisis. The upshot of it all is that a) some lies and liars just can't be caught; and b) though the FBI and the CIA wish they had foolproof lie-detectors, most of us couldn't live in an absolutely truthful world. Fitfully intriguing, too often prolix and obvious.