Tuesday, 10:40 AM. A child pushes a chair. A teacher sees the chair fall, the child's face, the moment of contact. From this vantage point, the sequence is simple: the child pushed the chair. Something must have happened to cause that. What happened? Nothing — not in the last sixty seconds. The room is the same room it was a minute ago. Whatever produced this moment did not begin in this moment. It became visible in this moment.
This is easy to state and difficult to hold onto, because behavior is the only part of the process that arrives with a timestamp. The push happens at 10:40. The observer's attention arrives at 10:40. It is natural — almost unavoidable — to treat 10:40 as the beginning, because 10:40 is when the observer's own process of noticing began.
But the observer's attention beginning at 10:40 is not the same as the child's process beginning at 10:40. These are two different timelines that happen to intersect at one visible point.
The behavior is where two timelines cross. It is not the origin of either one.What was the child's timeline before 10:40? A morning of transitions — arrival, a change in seating, an unexpected substitute, a fire drill at 9:15 that was loud and unscheduled, a worksheet that required sustained attention immediately afterward. None of these, individually, would be remarkable. Together, across two and a half hours, they describe a system that has been organizing around accumulating demand since the day began — with very little recovery in between.
By 10:40, that system meets an ordinary request — put your chair back — and the chair is pushed instead. The request did not cause this. The request arrived at a point in the morning where very little capacity remained to meet any request gently.
This is the same architecture as regulatory margin, read from a different angle. Margin describes the distance that narrows across time. This reading describes what happens at the moment that distance reaches zero: a visible event, arriving into view, that appears — to anyone who starts watching at 10:40 — to have come from nowhere.
It did not come from nowhere. It came from everywhere that came before it.This has a direct implication for how behavior is approached, and it is not the implication that behavior should be ignored or excused. It is that behavior cannot be accurately read by examining only the interval in which it became visible. The behavior is data about the moment of visibility. It is not data about the moment of origin — and the moment of origin is where the explanation lives.
Asking "what happened right before this" is asking the right question about the wrong interval. The interval that matters did not begin a minute ago. It began whenever the current cycle of accumulation began — that morning, that hour, sometimes that week.
This reframes what "intervention" means. If behavior is the visible endpoint of a process, then intervening only at the endpoint — responding to the push, addressing the push, consequating the push — intervenes at the one point in the process that is already finished. The process that produced the push is no longer occurring by the time the push is visible. It has already run its course.
What remains available is the next cycle. Reading what the morning required, what recovery occurred and where it did not, and what conditions are about to repeat — this is where intervention still has something to act on. Not the behavior that already happened. The conditions that are about to happen again.
The behavior was legible. The organization generating it was not.