Charles Perrow analyzes why catastrophic accidents occur in complex technological systems—nuclear plants, chemical facilities, aircraft, shipping—and argues that safety measures often make matters worse by adding complexity that creates new failure modes. His framework distinguishes between systems with tight coupling (where events follow quickly and processes cannot be stopped once started) and loose coupling (where there is slack and time to intervene), and between systems with linear interactions (where cause and effect are obvious) and complex interactions (where components interact in unexpected ways). Systems that are both tightly coupled and complexly interactive—like nuclear power plants—are prone to 'normal accidents': failures that arise not from abnormal events but from the system's inherent characteristics. Perrow's analysis, first published in 1984, anticipated later disasters including Chernobyl and Fukushima. He examines specific accidents in detail, tracing how seemingly trivial initiating events cascaded through systems in ways operators could not understand or control. His policy conclusions include questioning whether some technologies should be deployed regardless of their potential benefits, since no amount of safety engineering can eliminate normal accident risk in certain system types. Readers interested in technology policy, risk management, or why disasters happen will find a framework that remains influential decades after its introduction. The analysis applies beyond industrial accidents to any complex system—including organizations—where tight coupling and complex interactions create unexpected failures.