The Rules Were Written by Other People’s Mistakes

Overconfidence likes to treat instructions as redundant: they are written for the careless, and an experienced person can get by on common sense. In practice many rules are a short record of accidents that have already happened to someone else.

James Reason, a researcher of human error, said his interest in slips began with a domestic scene: distracted by a demanding cat, he put a spoonful of cat food into the teapot. The mistake looked ridiculous, but it was produced by a perfectly normal sequence of actions in a badly arranged environment.

Reason later proposed his famous model of layers of defence, which came to be drawn as slices of Swiss cheese. Every barrier has holes; an accident passes through when weaknesses at several levels line up for a moment. The model matters because it moves the question from “who was not paying enough attention?” to “why was one human error able to travel through the entire system?”

You cannot count on each of them working perfectly. A person will forget a step. A device will turn out to be faulty. Circumstances will differ from what was expected.

That is why safety is built in layers:

• the right tool;
• a guard;
• a safe position;
• restricted access;
• a check;
• a way to stop the process quickly.

One barrier can fail. Several independent barriers lower the chance that all the errors will meet in one place at one time.

A safety device should not be ignored. That is exactly what it is installed for. But it must not become permission to break the other rules.

A seat belt does not make any speed reasonable. A circuit breaker does not turn a bare wire into a normal part of the interior.

Before departing from an established method, you need to understand:

• which hazard it controls;
• why it was designed this way;
• what will happen after the change;
• what will replace the protection you are removing;
• who takes the consequences.

Innovation without an understanding of the old system often turns out to be the reinvention of a long-familiar accident.

In the kitchen there is a rule: you do not catch a falling knife. It looks cowardly and superfluous right up to the first time your hand goes after the blade on its own, with no involvement from consciousness at all. The point of the rule is not to explain the physics but to have the decision made in advance — because at the moment of falling there is no time to decide. Almost any line in an instruction that looks as though it was written for an idiot works the same way: someone already did exactly that, and the result was expensive enough to end up in the text.

Hence an unpleasant consequence for personal experience. “Nothing has ever happened to me” is a sample of one lucky person over a small number of attempts. The instruction is a larger sample, collected by people who were less lucky.