When the Holes Line Up: James Reason and the Swiss Cheese Model
"Reason distinguished between *active failures*—the unsafe..."

When the Holes Line Up: James Reason and the Swiss Cheese Model
How a psychologist’s layered metaphor ended the era of blaming the last pilot at the controls—and gave aviation a vocabulary for system-wide safety.
For most of the twentieth century, accident investigation in aviation followed a familiar script. When something went wrong, investigators traced the final minutes of flight, identified the crew’s last error, and closed the case. The pilot who touched the controls bore the weight of causation. That narrative was simple, legally tidy, and profoundly incomplete.
British psychologist James Reason offered a different picture. In Human Error (1990), he argued that serious accidents rarely stem from a single mistake. They emerge when multiple failures—some obvious, some buried deep in the organization—align at one unlucky moment. Reason illustrated the concept with an image that would outlive nearly everything else he wrote: slices of Swiss cheese, each representing a layer of defense against error. Every layer has holes. Most of the time, the holes do not align. When they do, an accident slips through.
Active Failures and Latent Conditions
Reason drew a crucial distinction between two kinds of failure. Active failures are the unsafe acts of front-line operators—the misread altimeter, the missed checklist item, the continuation below minimums. They are visible, immediate, and easy to blame. Latent conditions are the hidden pathogens introduced long before anyone boards the aircraft: ambiguous procedures written by distant offices, training syllabi that never addressed a known trap, maintenance cultures that reward speed over scrutiny, regulatory gaps that normalized risk. Latent conditions lie quietly in the system until an active failure finds them.
A pilot’s slip at the controls may be the last link in a chain that began months earlier in a conference room nobody on the flight deck ever entered.
Four Tiers of Organizational Risk
Reason refined the architecture in Managing the Risks of Organizational Accidents (1997). Serious organizational accidents, he proposed, propagate through four tiers. At the top sit organizational influences—resource allocation, corporate culture, regulatory climate. Below that comes unsafe supervision—inadequate oversight, planned operations that exceed reasonable limits, failure to correct known deficiencies. The third tier holds preconditions for unsafe acts: fatigue, complacency, poor communication, misleading instrumentation, time pressure. Only at the bottom do unsafe acts themselves appear—the errors and violations investigators have traditionally treated as the whole story.
The model does not excuse unsafe acts. It situates them inside a structure that either contains error or channels it toward catastrophe.
From Metaphor to Method
The international aviation community recognized the framework’s power early. The ICAO Human Factors and Flight Safety Working Group adopted Reason’s approach in the early 1990s, embedding systems thinking into how states and operators analyzed human factors in flight safety. The United States followed with concrete tooling: the U.S. Navy and the Federal Aviation Administration sponsored development of the Human Factors Analysis and Classification System (HFACS) by Scott Shappell and Douglas Wiegmann—a taxonomy derived directly from Reason’s tiered architecture. Investigators could now classify not only what the crew did, but what the organization failed to do.
Reason brought the metaphor to a wider audience in a 2000 British Medical Journal article, “The Swiss Cheese Model of System Accidents.” Medicine, nuclear power, and transportation adopted the same lens. Eurocontrol’s 2004 critical review at its Experimental Centre in Brétigny-sur-Orge acknowledged both the model’s influence and its limits—reminding practitioners that cheese slices are a teaching image, not a substitute for rigorous analysis. Even so, the vocabulary endured.
The End of the Last-Person Doctrine
The Swiss Cheese Model did not eliminate human error. It redefined what error means in a complex system. Blaming the last person who touched the controls had satisfied courts and headlines; it had not prevented the next accident. Reason’s layers made visible what prior investigations had discarded as background: the schedule that encouraged haste, the manual that contradicted training, the report that never reached the cockpit.
Aviation’s shift toward Safety Management Systems—proactive hazard identification, organizational accountability, data-driven risk mitigation—rests on this intellectual foundation. The holes in the cheese are not accidents of manufacturing. Many are cut by the institution itself.
Why it matters to you
When you study NTSB reports, FAA accident briefs, or your own operator’s safety bulletins, you are reading investigations shaped by Reason’s architecture whether the document names him or not. The chain of events in a runway incursion, a fuel exhaustion, or a controlled flight into terrain almost always includes active failures at the pointy end of the airplane and latent conditions upstream—supervisory pressure, ambiguous procedures, normalization of deviance. Threat and Error Management training asks you to trap errors before they propagate; the Swiss Cheese Model explains why trapping one error is sometimes not enough. Your challenge is not merely to fly without mistakes. It is to recognize when organizational holes are aligning around you: the rushed turnaround, the unpublished NOTAM, the company culture that treats a go-around as failure. Reason gave pilots permission to look past the last hand on the controls—including, on your worst day, your own—and ask what in the system allowed the holes to line up. That question is not academic. It is the skill modern safety culture expects every rated pilot to carry into the cockpit.