A violation that produces no bad outcome gets positively reinforced and repeats more often

What it is

There's an asymmetric feedback relationship between a violation and its consequences: as long as a given deviation doesn't immediately cause trouble, the actual feedback the operator receives is "this works fine, and it's faster" — that feedback is itself a reward, and it makes the same deviation more likely to be repeated next time. The genuine bad outcome only shows up in the low-probability failure case, so most of the time a person committing the violation receives nothing but positive signals.

Why it happens

This is a textbook case of intermittent reinforcement: the benefit (saved time, less effort) pays off almost every time, while the cost only shows up occasionally, at low probability. This "benefit almost always, cost almost never" feedback pattern is exactly the kind that most readily entrenches a behavior — harder to correct with a single warning or a single punishment than a behavior that pays off every time or costs every time, because the overwhelming majority of the person's actual experience keeps telling them this is the right thing to do.

Where it stops holding

This mechanism explains why a violation gets sustained and spreads, not why it first appeared — the original trigger is usually an unworkable rule or situational pressure. Nor does it mean that harsher punishment can reverse this maintaining mechanism: as long as the probability of an actual incident stays low, an individual will still experience "getting punished" far less often than "nothing went wrong, and it was faster." Punishment itself gets swallowed by the same intermittent-reinforcement structure, and simply increasing its severity struggles to outweigh a behavioral tendency built up from a large volume of positive experience.

Applying it

Don't wait for an incident before paying attention to a given violation — as soon as a deviation is observed to be repeating with no visible consequence, treat it as already in the positively-reinforced stage, and target the intervention at making the real cost of deviating perceptible (for instance, visibly surfacing the safety margin the deviation has been quietly eroding, instead of leaving it hidden behind "nothing happened this time"), rather than waiting for an actual incident to supply the "lesson." Verification: track how the frequency of a known deviation changes over time. If the frequency keeps rising with no corresponding negative feedback stepping in, it's on a track of being reinforced and will keep spreading — it needs proactive intervention before an incident occurs, rather than treating "nothing's happened yet" as a reason it can wait.

Related

• Same group: A10.13.1 Routine violations form when a rule is impractical to follow, and drift into the accepted norm · A10.13.2 Exceptional violations occur in rare, high-pressure situations to reach a goal · A10.13.3 An unworkable rule is the root cause of widespread routine violation, and punishing individuals doesn't fix it
• Nearby: A10.07 The Swiss cheese model
• Search terms: intermittent reinforcement · normalization of deviance · near miss