The broken-windows theory of crime argues that permitting minor misdemeanors such as loitering and vagrancy to go unpunished only encourages more serious crime. The author argues that this theory has existed for 30 years, but empirical verification was lacking. Instead, existing data suggest that it was false in that no statistically significant relationship exists between disorder and crime in four of five tests. The text’s second part examines whether the order-maintenance approach is theoretically sound. The analysis argues that this approach rests conceptually on unexamined categories of law abiders and disorderly people and of order and disorder that had no intrinsic reality independent of the punishment techniques that society uses. The proposed alternative vision recognizes the possibility of numerous alternatives to policies of emphasizing misdemeanor stops and frisks and arrests, and the need for qualitative, historical, comparative, ethnographic, and quantitative analyses to accompany the implementation and rejection of policy initiatives. Figures, index, and approximately 400 references