01
Run the service
Choose Sunday service. A tight program with its roles filled predicts most of what happens, and steady low-grade repair keeps it there.
communicoupling
The room runs
a program.
A Sunday service, a post office counter, a Monday standup. Each is a bounded piece of the world with a standing pattern of behavior fixed to a time, a place, and a set of objects. Walk in and the pattern is already running; occupants step into its role-slots and are, for its purposes, interchangeable. Roger Barker's field station in a Kansas town logged these units for a quarter century and called them behavior settings.
Barker's finding scandalized his own discipline: to predict what a person is doing, knowing which setting they are in beats knowing everything about their personality. Settings hold that grip by maintaining themselves. Departure from the program draws a response — vetoing, correcting, recruiting, ejecting — until the program runs again.
program strengthhow tightly the setting scripts what happens
staffing ratiooccupants relative to the roles the program needs
boundary permeabilityhow freely people enter and leave mid-session
setting's share
—
acts matching the program; the remainder tracks dispositions
repair events
—
corrections, ejections, and recruitments spent on the session
role load
—
role slots each occupant carries at once
The setting's share of behavior across the session
gold: behavior the program accounts for · blue: repair effort per occupant · red mark: the program's claim on its role holders drops under their dispositions
Set the conditions
what to try
Move the staffing, watch the persons appear.
02
Understaff it
Choose Small school. Roles outnumber bodies, so each occupant carries several and the program's claim on each one rises. Repair nearly stops: a setting this thin cannot afford to correct anyone.
03
Weaken the program
Choose Contested town hall. A loose script, a surplus of occupants, and a door that keeps swinging leave dispositions in charge from the opening minute.
the mechanism
The setting claims the behavior; the person supplies the rest.
Each occupant arrives with a disposition drawn from a spread and a tenure that sets how well they know the program. Every episode the model weighs the setting's claim on that occupant — program strength, times the program's current integrity, times familiarity, times how much of the program the present staff can cover, times whether the occupant holds a role or stands at the margin — against the pull of their own disposition, and draws their behavior from the ratio. What the chart plots is the share of acts the program accounts for.
Deviation above the setting's tolerance draws repair, and repair capacity comes out of the same occupants. Understaffing raises the claim on everyone present and raises tolerance at the same time, so a thin setting works its members harder and corrects them less. Overstaffing turns the surplus into spectators whose behavior the program predicts poorly, and selection ejects them until the roster matches the roles. Correction shrinks a disposition, ejection removes it, recruitment adds a stranger who does not yet know the program. When the program's claim on its own role holders falls under their dispositions for several episodes running, the run is marked collapsed and the room belongs to the people in it.
the social payoff
Coordination that nobody has to perform.
A queue, a service, a ward round, and a standup each do most of their coordinating before anyone speaks. The program assigns positions, sequences turns, and specifies what counts as being here properly, which is why rewriting a meeting's program changes the behavior of everyone who attends more reliably than changing who attends. Barker's practical claim follows: to alter conduct, alter the settings people pass through.
The mechanism sits alongside two nearby accounts. Definition of the situation tracks how participants agree on what is happening; behavior settings hold that the agreement is often preinstalled in the place and its objects. Garbage-can decision shows what surfaces when a setting's program leaves the outcome open.
the mapping
Mechanism ↔ social life.
| In the model | In the world |
|---|---|
| program strength | How firmly time, place, and objects script the standing pattern. |
| staffing ratio | Occupants relative to the role slots the program needs filled. |
| boundary permeability | Turnover during the session; newcomers arrive without the program. |
| role load | Roles carried at once. Barker's small-school students carried several apiece. |
| tolerance | Departure the setting absorbs before it acts. Rises as staffing thins. |
| repair events | Vetoing, correcting, recruiting, ejecting — the setting keeping itself running. |
| spectators | Surplus occupants with no role slot, whose conduct the program barely reaches. |
| collapse | The program stops predicting behavior; the persons govern the room. |
where it tears
Limits.
Settings govern where they are intact.
Contested, thin, and novel situations hand control back to persons, and the theory says little about how a setting is born or dies. The model marks collapse without showing what fills the room afterward.
"The setting did it" launders responsibility.
Somebody wrote the program, set the staffing, and chose the tolerance. Attributing conduct to the setting can shield whoever designed it, which matters most where the program does harm.
Barker's data behaved.
Midwest, Kansas was small, stable, and observable, and its settings had edges a fieldworker could count and time. Findings drawn from so cooperative a habitat travel less well than the method's confidence suggests.
Bounded time and place is under strain.
Feeds, group chats, and drive-throughs stretch the unit until its edges blur. Where a setting has no boundary and no shared clock, the concept may no longer cut anything.