Photo: Michael Marsland/Yale University

The accident at the Three Mile Island nuclear plant in Pennsylvania on 28 March 1979 eventually prompted quite a few people to alter their careers, oftentimes involuntarily. But for Yale University sociologist Charles Perrow, author of this issue’s “Shrink the Targets,” the change in direction was very much his own doing. Asked by a U.S. government commission to participate in its investigation of the mishap, he readily agreed, even though he had never given industrial accidents a thought.

Accidents in those days were routinely attributed to faults in machines or to lapses in their human operators, problems outside the purview of sociology. “I looked into the reports from Three Mile Island and began to see that the problem went much deeper,” recalls Perrow, now professor emeritus at Yale. He went on to apply what he’d learned to other industrial accidents, in the process developing a theory that traced big failures to the very structure of the organizations that suffered them. For the mishaps he coined a name that became the title of his 1984 book: Normal Accidents.

Despite the acclaim for that book, Perrow never hit it big as an industrial consultant, perhaps because he didn’t bother to disguise his disapproval of the nuclear power industry. “I’ve changed my views on that,” he says. “I believe nuclear power plants could be made safer than I had thought, and I’ve become more afraid of the alternative to them: global warming.”

Now, at age 81, he finds the consulting gigs are starting to roll in, but not from the nuclear industry. The U.S. government has roped him into studies on software reliability and, in the wake of the 9/11 attacks, on preparedness for terrorism. Terrorists hit large, complex organizations much as natural disasters and industrial accidents do, Perrow says, but with a twist: the terrorists as well as the victims can study the results and use the knowledge for the next round.